In a previous post, we tried to model the daily and the weekly candle with an Elman Neural Network. The results were not encouraging. The problem that we face with neural networks is that most of the time these networks get stuck up with a local minima. So everytime you run the model, it will give you a different result. Another major problem that we faced was the time being taken by R to make calculation. Time was around 20 minutes which is excessively if you want to trade on the intraday timeframe.
R Is A Powerful Machine Learning And Data Analysis Software
You must be familiar with R software by now. R is a powerful machine learning and data analysis software that is open source and can be downloaded freely. MT4 MQL4 language just lacks the power of R. MQL4 is just good for making a few indicators and EAs. What we can do is connect R with MT4 and then use the power of R to make the predictions. Python is another powerful machine learning and data analysis language that you should learn.
VAR Model For Predicting Next Few Candles Infographic
In this post we are going to discuss how we can use Vector Autoregression Models in predicting the next candle. R package that we will use for this is known as VaR package. This package calculates 3 models: Vector Autoregression Model (VAR), Structured Vector Autoregression Model (SVAR) and Vector Error Correction Model (VECM). In this post we are not going to discuss SVAR and VECM model. We will discuss these models when I write the post on how to model cointegration. Infographic below gives a brief overview of how to make these predictions.
If you don’t know what Vector Autoregression is, you should watch this video below that explains the VAR model:
The problem that we face with most time series models is that most of the time we are only using the close price in making the predictions. In a VAR model we can use Open, High, Low and Close in one autoregressive model that then is used to predict Open, High, Low and Close for the next few candles. So let’s start and see how good are the predictions. If we calculate correlation of Open, High, Low and Close with each other we get values above 0.9 which is an indication that these variables are highly correlated.
> # Import the csv file > quotes <- read.csv("E:/MarketData/GBPUSD10080.csv", header=FALSE) > > > x <-nrow(quotes) > > #correlation of High and Close price > > cor(quotes[,4], quotes[,6]) [1] 0.9958719 > > #correlation of Low and Close price > > cor(quotes[,5], quotes[,6]) [1] 0.9968037
You can see the high correlation between Open, High, Low and Close. We already know this from our trading experience that Open, High, Low and Close predict each other. So we will develop a VAR model in this post uses Open, High, Low, Close and the Volume as the endogenous variables.
Weekly Candle Prediction With VAR Model
> # Import the csv file > quotes <- read.csv("E:/MarketData/GBPUSD10080.csv", header=FALSE) > > > x <-nrow(quotes) > > #convert the data frame into an xts object > quotes1 <- as.ts(quotes) > > tail(quotes1) V1 V2 V3 V4 V5 V6 V7 [1021,] 1021 1 1.29514 1.34811 1.28507 1.31937 767484 [1022,] 1022 1 1.31973 1.33151 1.30651 1.30971 613903 [1023,] 1023 1 1.31467 1.33015 1.30570 1.32313 565129 [1024,] 1024 1 1.32262 1.33719 1.30216 1.30762 513608 [1025,] 1025 1 1.30809 1.30969 1.29084 1.29100 355199 [1026,] 1026 1 1.29274 1.29400 1.29007 1.29318 6854 > > lr <- quotes1[(x-1000):(x-2),3:7] > > > #install vars package > > library(vars) Loading required package: MASS Loading required package: strucchange Loading required package: zoo Attaching package: ‘zoo’ The following objects are masked from ‘package:base’: as.Date, as.Date.numeric Loading required package: sandwich Loading required package: urca Loading required package: lmtest > > summary(VAR(lr, p=10, type="both")) VAR Estimation Results: ========================= Endogenous variables: V3, V4, V5, V6, V7 Deterministic variables: both Sample size: 989 Log Likelihood: 1362.321 Roots of the characteristic polynomial: 0.9932 0.9694 0.956 0.8815 0.8815 0.8686 0.8686 0.8622 0.8622 0.8616 0.8616 0.8526 0.8526 0.8138 0.8138 0.8131 0.8131 0.8122 0.8122 0.7956 0.7956 0.7944 0.7944 0.789 0.7872 0.7872 0.7772 0.7772 0.7746 0.7746 0.7659 0.7659 0.7577 0.7577 0.7575 0.7575 0.7493 0.7493 0.7477 0.7477 0.7297 0.7297 0.7205 0.6671 0.6671 0.6554 0.6554 0.6389 0.6389 0.3242 Call: VAR(y = lr, p = 10, type = "both") Estimation results for equation V3: =================================== V3 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 1.246e-02 3.475e-02 0.359 0.72003 V4.l1 -7.164e-02 1.733e-02 -4.133 3.9e-05 *** V5.l1 4.368e-02 1.588e-02 2.751 0.00606 ** V6.l1 1.008e+00 1.469e-02 68.654 < 2e-16 *** V7.l1 1.151e-09 2.400e-09 0.479 0.63184 V3.l2 -1.031e-01 3.476e-02 -2.967 0.00309 ** V4.l2 3.383e-02 1.760e-02 1.922 0.05485 . V5.l2 -1.707e-02 1.620e-02 -1.054 0.29232 V6.l2 -1.327e-02 3.625e-02 -0.366 0.71427 V7.l2 -2.974e-09 3.184e-09 -0.934 0.35046 V3.l3 -3.677e-02 3.491e-02 -1.053 0.29243 V4.l3 -1.234e-02 1.762e-02 -0.700 0.48379 V5.l3 2.445e-02 1.638e-02 1.493 0.13578 V6.l3 9.166e-02 3.611e-02 2.538 0.01130 * V7.l3 1.728e-09 3.179e-09 0.544 0.58689 V3.l4 9.987e-02 3.480e-02 2.870 0.00420 ** V4.l4 9.725e-05 1.764e-02 0.006 0.99560 V5.l4 2.562e-02 1.651e-02 1.552 0.12110 V6.l4 3.591e-02 3.627e-02 0.990 0.32240 V7.l4 2.094e-10 3.191e-09 0.066 0.94769 V3.l5 -2.444e-02 3.485e-02 -0.701 0.48332 V4.l5 1.417e-02 1.767e-02 0.802 0.42267 V5.l5 -4.530e-02 1.654e-02 -2.738 0.00630 ** V6.l5 -1.115e-01 3.637e-02 -3.065 0.00224 ** V7.l5 -1.419e-09 3.206e-09 -0.442 0.65828 V3.l6 -5.473e-02 3.491e-02 -1.567 0.11735 V4.l6 -1.870e-02 1.774e-02 -1.054 0.29220 V5.l6 -1.664e-02 1.651e-02 -1.008 0.31382 V6.l6 6.187e-02 3.657e-02 1.692 0.09096 . V7.l6 1.650e-10 3.213e-09 0.051 0.95905 V3.l7 4.949e-02 3.508e-02 1.411 0.15860 V4.l7 -2.370e-02 1.821e-02 -1.301 0.19359 V5.l7 -4.293e-03 1.667e-02 -0.258 0.79680 V6.l7 8.131e-02 3.644e-02 2.232 0.02588 * V7.l7 1.186e-09 3.255e-09 0.364 0.71573 V3.l8 -1.002e-01 3.488e-02 -2.872 0.00418 ** V4.l8 1.292e-02 1.823e-02 0.709 0.47853 V5.l8 9.711e-03 1.646e-02 0.590 0.55544 V6.l8 -3.283e-02 3.659e-02 -0.897 0.36984 V7.l8 -8.295e-11 3.258e-09 -0.025 0.97969 V3.l9 -8.366e-03 3.490e-02 -0.240 0.81059 V4.l9 1.550e-04 1.798e-02 0.009 0.99312 V5.l9 -1.959e-02 1.642e-02 -1.193 0.23305 V6.l9 9.683e-02 3.642e-02 2.659 0.00798 ** V7.l9 2.957e-10 3.269e-09 0.090 0.92793 V3.l10 2.671e-02 1.430e-02 1.867 0.06214 . V4.l10 4.090e-03 1.731e-02 0.236 0.81324 V5.l10 -2.607e-02 1.629e-02 -1.601 0.10979 V6.l10 9.962e-03 3.621e-02 0.275 0.78331 V7.l10 4.051e-10 2.460e-09 0.165 0.86923 const -4.424e-03 1.449e-03 -3.054 0.00232 ** trend 5.016e-07 6.392e-07 0.785 0.43288 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.00385 on 937 degrees of freedom Multiple R-Squared: 0.9995, Adjusted R-squared: 0.9995 F-statistic: 3.523e+04 on 51 and 937 DF, p-value: < 2.2e-16 Estimation results for equation V4: =================================== V4 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 8.893e-02 1.102e-01 0.807 0.4198 V4.l1 7.385e-02 5.496e-02 1.344 0.1794 V5.l1 -1.203e-01 5.035e-02 -2.390 0.0170 * V6.l1 9.804e-01 4.656e-02 21.056 <2e-16 *** V7.l1 2.215e-09 7.611e-09 0.291 0.7711 V3.l2 -1.110e-02 1.102e-01 -0.101 0.9198 V4.l2 4.702e-02 5.579e-02 0.843 0.3996 V5.l2 -3.964e-02 5.137e-02 -0.772 0.4405 V6.l2 -7.074e-02 1.149e-01 -0.616 0.5384 V7.l2 -3.774e-09 1.010e-08 -0.374 0.7086 V3.l3 -1.826e-01 1.107e-01 -1.650 0.0993 . V4.l3 1.202e-01 5.586e-02 2.152 0.0316 * V5.l3 5.866e-02 5.193e-02 1.130 0.2589 V6.l3 -2.245e-02 1.145e-01 -0.196 0.8446 V7.l3 9.167e-09 1.008e-08 0.909 0.3634 V3.l4 1.296e-01 1.103e-01 1.175 0.2404 V4.l4 -1.065e-01 5.594e-02 -1.904 0.0573 . V5.l4 -5.299e-03 5.235e-02 -0.101 0.9194 V6.l4 1.376e-01 1.150e-01 1.197 0.2318 V7.l4 2.304e-10 1.012e-08 0.023 0.9818 V3.l5 6.076e-02 1.105e-01 0.550 0.5826 V4.l5 1.267e-01 5.603e-02 2.261 0.0240 * V5.l5 -1.201e-01 5.246e-02 -2.289 0.0223 * V6.l5 -1.057e-01 1.153e-01 -0.916 0.3597 V7.l5 -1.476e-08 1.017e-08 -1.452 0.1469 V3.l6 -4.499e-02 1.107e-01 -0.406 0.6845 V4.l6 8.071e-03 5.625e-02 0.143 0.8859 V5.l6 -1.419e-02 5.234e-02 -0.271 0.7864 V6.l6 -1.836e-02 1.159e-01 -0.158 0.8742 V7.l6 -5.189e-09 1.019e-08 -0.509 0.6106 V3.l7 -3.695e-02 1.112e-01 -0.332 0.7398 V4.l7 -1.799e-02 5.775e-02 -0.311 0.7555 V5.l7 3.676e-02 5.284e-02 0.696 0.4869 V6.l7 5.590e-02 1.155e-01 0.484 0.6286 V7.l7 8.583e-09 1.032e-08 0.832 0.4058 V3.l8 -5.701e-02 1.106e-01 -0.515 0.6064 V4.l8 3.337e-02 5.779e-02 0.577 0.5638 V5.l8 -8.257e-02 5.220e-02 -1.582 0.1141 V6.l8 -1.854e-02 1.160e-01 -0.160 0.8731 V7.l8 5.132e-10 1.033e-08 0.050 0.9604 V3.l9 3.938e-02 1.107e-01 0.356 0.7220 V4.l9 3.747e-02 5.702e-02 0.657 0.5112 V5.l9 -2.008e-02 5.206e-02 -0.386 0.6998 V6.l9 1.212e-01 1.155e-01 1.049 0.2944 V7.l9 -3.628e-09 1.036e-08 -0.350 0.7264 V3.l10 3.882e-02 4.535e-02 0.856 0.3922 V4.l10 2.298e-02 5.488e-02 0.419 0.6756 V5.l10 -7.085e-02 5.164e-02 -1.372 0.1704 V6.l10 -5.189e-02 1.148e-01 -0.452 0.6515 V7.l10 6.544e-09 7.799e-09 0.839 0.4017 const 4.096e-03 4.593e-03 0.892 0.3727 trend -7.971e-08 2.027e-06 -0.039 0.9686 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.01221 on 937 degrees of freedom Multiple R-Squared: 0.9948, Adjusted R-squared: 0.9945 F-statistic: 3493 on 51 and 937 DF, p-value: < 2.2e-16 Estimation results for equation V5: =================================== V5 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 -7.499e-02 1.272e-01 -0.590 0.555636 V4.l1 -8.007e-02 6.346e-02 -1.262 0.207318 V5.l1 1.849e-01 5.813e-02 3.181 0.001518 ** V6.l1 9.423e-01 5.376e-02 17.528 < 2e-16 *** V7.l1 -3.280e-09 8.788e-09 -0.373 0.709082 V3.l2 8.457e-02 1.272e-01 0.665 0.506420 V4.l2 -1.338e-01 6.442e-02 -2.077 0.038030 * V5.l2 2.096e-01 5.931e-02 3.534 0.000429 *** V6.l2 -3.433e-02 1.327e-01 -0.259 0.795899 V7.l2 2.196e-09 1.166e-08 0.188 0.850627 V3.l3 -2.865e-01 1.278e-01 -2.242 0.025206 * V4.l3 -3.530e-02 6.449e-02 -0.547 0.584232 V5.l3 2.402e-01 5.995e-02 4.007 6.65e-05 *** V6.l3 -1.725e-01 1.322e-01 -1.305 0.192218 V7.l3 7.677e-09 1.164e-08 0.660 0.509700 V3.l4 1.077e-01 1.274e-01 0.846 0.397930 V4.l4 -1.462e-01 6.458e-02 -2.264 0.023795 * V5.l4 -1.739e-01 6.044e-02 -2.878 0.004096 ** V6.l4 2.954e-01 1.328e-01 2.225 0.026302 * V7.l4 -1.246e-08 1.168e-08 -1.066 0.286560 V3.l5 2.131e-01 1.276e-01 1.670 0.095300 . V4.l5 -2.800e-02 6.469e-02 -0.433 0.665205 V5.l5 -5.179e-02 6.057e-02 -0.855 0.392735 V6.l5 4.773e-02 1.331e-01 0.358 0.720099 V7.l5 9.485e-11 1.174e-08 0.008 0.993555 V3.l6 -2.762e-01 1.278e-01 -2.161 0.030955 * V4.l6 2.070e-02 6.495e-02 0.319 0.750030 V5.l6 7.136e-02 6.044e-02 1.181 0.238017 V6.l6 -1.489e-01 1.339e-01 -1.113 0.266200 V7.l6 -1.263e-11 1.176e-08 -0.001 0.999144 V3.l7 8.459e-02 1.284e-01 0.659 0.510259 V4.l7 -7.361e-02 6.668e-02 -1.104 0.269881 V5.l7 5.939e-02 6.101e-02 0.973 0.330577 V6.l7 2.549e-01 1.334e-01 1.911 0.056359 . V7.l7 5.319e-09 1.192e-08 0.446 0.655386 V3.l8 -1.573e-01 1.277e-01 -1.232 0.218429 V4.l8 7.412e-02 6.672e-02 1.111 0.266908 V5.l8 -3.500e-02 6.027e-02 -0.581 0.561589 V6.l8 -1.293e-01 1.339e-01 -0.966 0.334467 V7.l8 -7.355e-09 1.193e-08 -0.617 0.537630 V3.l9 2.760e-02 1.278e-01 0.216 0.828985 V4.l9 6.122e-02 6.583e-02 0.930 0.352622 V5.l9 -2.453e-02 6.011e-02 -0.408 0.683241 V6.l9 1.342e-01 1.333e-01 1.006 0.314494 V7.l9 5.821e-09 1.197e-08 0.486 0.626768 V3.l10 -1.628e-02 5.236e-02 -0.311 0.755849 V4.l10 -8.096e-02 6.337e-02 -1.278 0.201703 V5.l10 7.221e-03 5.962e-02 0.121 0.903618 V6.l10 3.439e-02 1.326e-01 0.259 0.795387 V7.l10 8.116e-10 9.005e-09 0.090 0.928208 const 5.373e-03 5.303e-03 1.013 0.311189 trend -1.758e-07 2.340e-06 -0.075 0.940128 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.01409 on 937 degrees of freedom Multiple R-Squared: 0.993, Adjusted R-squared: 0.9926 F-statistic: 2600 on 51 and 937 DF, p-value: < 2.2e-16 Estimation results for equation V6: =================================== V6 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 1.003e-01 1.832e-01 0.547 0.58418 V4.l1 2.105e-02 9.137e-02 0.230 0.81786 V5.l1 -2.414e-04 8.371e-02 -0.003 0.99770 V6.l1 9.950e-01 7.741e-02 12.853 < 2e-16 *** V7.l1 2.109e-09 1.265e-08 0.167 0.86767 V3.l2 1.494e-01 1.832e-01 0.815 0.41511 V4.l2 -1.232e-01 9.276e-02 -1.329 0.18433 V5.l2 1.898e-01 8.541e-02 2.222 0.02650 * V6.l2 -2.334e-01 1.911e-01 -1.222 0.22218 V7.l2 -3.473e-09 1.678e-08 -0.207 0.83612 V3.l3 -4.602e-01 1.840e-01 -2.501 0.01255 * V4.l3 3.126e-02 9.287e-02 0.337 0.73648 V5.l3 2.181e-01 8.633e-02 2.526 0.01169 * V6.l3 -1.945e-01 1.903e-01 -1.022 0.30700 V7.l3 1.025e-08 1.676e-08 0.612 0.54082 V3.l4 1.643e-01 1.834e-01 0.896 0.37052 V4.l4 -2.561e-01 9.300e-02 -2.754 0.00601 ** V5.l4 -1.499e-01 8.703e-02 -1.722 0.08534 . V6.l4 5.116e-01 1.912e-01 2.676 0.00758 ** V7.l4 -9.885e-09 1.682e-08 -0.588 0.55690 V3.l5 7.083e-02 1.837e-01 0.386 0.69993 V4.l5 8.401e-02 9.315e-02 0.902 0.36735 V5.l5 -1.521e-01 8.721e-02 -1.743 0.08159 . V6.l5 1.482e-02 1.917e-01 0.077 0.93840 V7.l5 -1.063e-08 1.690e-08 -0.629 0.52958 V3.l6 -2.494e-01 1.840e-01 -1.355 0.17565 V4.l6 4.918e-02 9.353e-02 0.526 0.59910 V5.l6 5.305e-02 8.702e-02 0.610 0.54229 V6.l6 -2.493e-02 1.928e-01 -0.129 0.89712 V7.l6 -4.569e-09 1.694e-08 -0.270 0.78741 V3.l7 1.223e-01 1.849e-01 0.661 0.50855 V4.l7 2.154e-02 9.601e-02 0.224 0.82250 V5.l7 5.986e-02 8.786e-02 0.681 0.49586 V6.l7 1.890e-01 1.921e-01 0.984 0.32528 V7.l7 5.562e-09 1.716e-08 0.324 0.74585 V3.l8 -1.093e-01 1.839e-01 -0.595 0.55225 V4.l8 -1.184e-02 9.608e-02 -0.123 0.90196 V5.l8 -8.629e-02 8.679e-02 -0.994 0.32038 V6.l8 -1.994e-01 1.929e-01 -1.034 0.30155 V7.l8 7.015e-09 1.717e-08 0.408 0.68304 V3.l9 -3.546e-02 1.840e-01 -0.193 0.84719 V4.l9 1.151e-01 9.479e-02 1.215 0.22477 V5.l9 3.148e-02 8.655e-02 0.364 0.71615 V6.l9 1.233e-01 1.920e-01 0.642 0.52090 V7.l9 2.161e-09 1.723e-08 0.125 0.90022 V3.l10 -3.399e-02 7.539e-02 -0.451 0.65218 V4.l10 6.509e-03 9.124e-02 0.071 0.94314 V5.l10 -5.308e-02 8.585e-02 -0.618 0.53651 V6.l10 4.494e-02 1.909e-01 0.235 0.81395 V7.l10 -3.758e-09 1.297e-08 -0.290 0.77204 const 1.383e-02 7.636e-03 1.811 0.07043 . trend 5.040e-07 3.370e-06 0.150 0.88115 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.02029 on 937 degrees of freedom Multiple R-Squared: 0.9855, Adjusted R-squared: 0.9847 F-statistic: 1251 on 51 and 937 DF, p-value: < 2.2e-16 Estimation results for equation V7: =================================== V7 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 7.716e+05 4.718e+05 1.635 0.10233 V4.l1 -1.081e+05 2.354e+05 -0.459 0.64627 V5.l1 -6.181e+04 2.156e+05 -0.287 0.77447 V6.l1 5.567e+04 1.994e+05 0.279 0.78019 V7.l1 8.553e-01 3.260e-02 26.237 < 2e-16 *** V3.l2 -7.817e+05 4.720e+05 -1.656 0.09801 . V4.l2 -3.345e+04 2.390e+05 -0.140 0.88870 V5.l2 1.447e+05 2.200e+05 0.658 0.51097 V6.l2 -8.731e+05 4.922e+05 -1.774 0.07641 . V7.l2 7.572e-03 4.324e-02 0.175 0.86101 V3.l3 4.075e+05 4.740e+05 0.860 0.39018 V4.l3 6.469e+04 2.392e+05 0.270 0.78691 V5.l3 2.367e+04 2.224e+05 0.106 0.91528 V6.l3 8.217e+05 4.903e+05 1.676 0.09410 . V7.l3 7.772e-02 4.318e-02 1.800 0.07217 . V3.l4 -3.742e+05 4.725e+05 -0.792 0.42859 V4.l4 9.746e+03 2.396e+05 0.041 0.96756 V5.l4 -7.457e+04 2.242e+05 -0.333 0.73950 V6.l4 -4.846e+05 4.925e+05 -0.984 0.32537 V7.l4 -6.512e-02 4.333e-02 -1.503 0.13319 V3.l5 -1.097e+05 4.733e+05 -0.232 0.81679 V4.l5 -8.191e+03 2.400e+05 -0.034 0.97278 V5.l5 1.029e+04 2.247e+05 0.046 0.96350 V6.l5 3.926e+05 4.939e+05 0.795 0.42685 V7.l5 -2.886e-02 4.354e-02 -0.663 0.50767 V3.l6 2.079e+05 4.741e+05 0.439 0.66110 V4.l6 -3.581e+05 2.409e+05 -1.486 0.13752 V5.l6 1.481e+05 2.242e+05 0.661 0.50896 V6.l6 2.031e+05 4.966e+05 0.409 0.68262 V7.l6 8.119e-02 4.363e-02 1.861 0.06308 . V3.l7 3.814e+04 4.764e+05 0.080 0.93620 V4.l7 3.997e+05 2.473e+05 1.616 0.10644 V5.l7 -2.280e+04 2.263e+05 -0.101 0.91978 V6.l7 -1.949e+05 4.948e+05 -0.394 0.69381 V7.l7 -1.734e-02 4.420e-02 -0.392 0.69485 V3.l8 1.724e+04 4.737e+05 0.036 0.97098 V4.l8 -5.745e+04 2.475e+05 -0.232 0.81650 V5.l8 7.599e+04 2.236e+05 0.340 0.73404 V6.l8 -2.405e+05 4.969e+05 -0.484 0.62848 V7.l8 -1.071e-01 4.424e-02 -2.420 0.01569 * V3.l9 3.979e+05 4.739e+05 0.840 0.40134 V4.l9 1.173e+05 2.442e+05 0.480 0.63108 V5.l9 1.859e+05 2.230e+05 0.834 0.40462 V6.l9 -1.805e+05 4.946e+05 -0.365 0.71521 V7.l9 -7.430e-02 4.439e-02 -1.674 0.09451 . V3.l10 -1.265e+05 1.942e+05 -0.652 0.51486 V4.l10 3.112e+04 2.351e+05 0.132 0.89469 V5.l10 3.205e+05 2.211e+05 1.449 0.14757 V6.l10 -7.664e+05 4.918e+05 -1.558 0.11948 V7.l10 2.249e-01 3.340e-02 6.733 2.89e-11 *** const 2.655e+04 1.967e+04 1.350 0.17745 trend 2.326e+01 8.681e+00 2.679 0.00751 ** --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 52280 on 937 degrees of freedom Multiple R-Squared: 0.936, Adjusted R-squared: 0.9325 F-statistic: 268.6 on 51 and 937 DF, p-value: < 2.2e-16 Covariance matrix of residuals: V3 V4 V5 V6 V7 V3 1.482e-05 9.228e-06 7.926e-06 3.905e-06 -1.093e+01 V4 9.228e-06 1.490e-04 9.214e-05 1.898e-04 4.843e+01 V5 7.926e-06 9.214e-05 1.986e-04 2.306e-04 -7.977e+01 V6 3.905e-06 1.898e-04 2.306e-04 4.119e-04 -3.903e+01 V7 -1.093e+01 4.843e+01 -7.977e+01 -3.903e+01 2.733e+09 Correlation matrix of residuals: V3 V4 V5 V6 V7 V3 1.00000 0.19637 0.1461 0.04999 -0.05433 V4 0.19637 1.00000 0.5356 0.76621 0.07588 V5 0.14608 0.53558 1.0000 0.80621 -0.10827 V6 0.04999 0.76621 0.8062 1.00000 -0.03678 V7 -0.05433 0.07588 -0.1083 -0.03678 1.00000 > > fitvar1 = VAR(lr, p=10, type="both") > > # auto correlation function for residuals for close price > > acf(residuals(fitvar1)[,4]) > > # auto correlation function for residuals for open, high and low price > > acf(residuals(fitvar1)[,1]) > > acf(residuals(fitvar1)[,2]) > > acf(residuals(fitvar1)[,3]) > > #vector auto regression of Open, High, Low and Close > > #predict the next 10 values > > var.pred <- predict(fitvar1, n.ahead=3, ci=0.95) > > var.pred $V3 fcst lower upper CI [1,] 1.309286 1.301741 1.316832 0.007545514 [2,] 1.310732 1.270236 1.351228 0.040496033 [3,] 1.312006 1.255404 1.368609 0.056602615 $V4 fcst lower upper CI [1,] 1.320224 1.296299 1.344149 0.02392477 [2,] 1.338595 1.293872 1.383318 0.04472325 [3,] 1.333909 1.275277 1.392541 0.05863241 $V5 fcst lower upper CI [1,] 1.287867 1.260244 1.315490 0.02762322 [2,] 1.290540 1.241703 1.339377 0.04883692 [3,] 1.303158 1.239837 1.366480 0.06332178 $V6 fcst lower upper CI [1,] 1.310808 1.271032 1.350584 0.03977584 [2,] 1.310701 1.254293 1.367109 0.05640817 [3,] 1.325592 1.257826 1.393357 0.06776577 $V7 fcst lower upper CI [1,] 538530.5 436063.6 640997.4 102466.9 [2,] 467243.1 332563.5 601922.7 134679.6 [3,] 434972.6 280427.9 589517.3 154544.7 > > quotes[(x-1):x,] V1 V2 V3 V4 V5 V6 V7 1025 2016.08.07 00:00 1.30809 1.30969 1.29084 1.29100 355199 1026 2016.08.14 00:00 1.29274 1.29400 1.29007 1.29318 6854
You can see the predicted Open, High, Low and Close for next week is 1.30928, 1.320224, 1.28786 and 1.31080 while the actual Open, High, Low and Close was 1.30809, 1.30969, 1.29084 and 1.29100. So you can see this model did not make a good prediction of the weekly candle.
Daily Candle Prediction Using VAR Model
Now we run this model for the daily candle and see how good prediction we get for the daily candle.
> # Import the csv file > quotes <- read.csv("E:/MarketData/GBPUSD1440.csv", header=FALSE) > > > x <-nrow(quotes) > > lr <- quotes[(x-1000):(x-5),3:7] > > > #install vars package > > library(vars) > > summary(VAR(lr, p=10, type="both")) VAR Estimation Results: ========================= Endogenous variables: V3, V4, V5, V6, V7 Deterministic variables: both Sample size: 986 Log Likelihood: 5634.754 Roots of the characteristic polynomial: 0.9936 0.9662 0.8731 0.8731 0.8717 0.8717 0.8629 0.8498 0.8498 0.8237 0.8237 0.8191 0.8169 0.8169 0.8147 0.8147 0.8112 0.8112 0.7894 0.7894 0.7882 0.7882 0.7756 0.7756 0.757 0.757 0.7566 0.7522 0.7522 0.7465 0.7465 0.7376 0.7376 0.7369 0.7369 0.7111 0.7111 0.7051 0.6933 0.6933 0.6797 0.6797 0.6461 0.6305 0.6284 0.6284 0.4777 0.4777 0.2771 0.2771 Call: VAR(y = lr, p = 10, type = "both") Estimation results for equation V3: =================================== V3 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 -9.319e-02 3.536e-02 -2.635 0.00855 ** V4.l1 -2.637e-02 1.576e-02 -1.673 0.09460 . V5.l1 1.094e-01 1.420e-02 7.706 3.32e-14 *** V6.l1 9.860e-01 1.511e-02 65.272 < 2e-16 *** V7.l1 1.855e-09 1.477e-09 1.256 0.20950 V3.l2 -3.533e-02 3.557e-02 -0.993 0.32082 V4.l2 6.974e-03 1.588e-02 0.439 0.66071 V5.l2 -1.143e-02 1.460e-02 -0.783 0.43411 V6.l2 3.161e-02 3.565e-02 0.887 0.37542 V7.l2 -3.795e-10 1.571e-09 -0.242 0.80918 V3.l3 -2.396e-02 3.560e-02 -0.673 0.50116 V4.l3 4.055e-03 1.581e-02 0.257 0.79758 V5.l3 8.547e-03 1.457e-02 0.587 0.55759 V6.l3 3.223e-02 3.557e-02 0.906 0.36504 V7.l3 5.576e-10 1.583e-09 0.352 0.72475 V3.l4 3.055e-02 3.559e-02 0.858 0.39090 V4.l4 -4.362e-02 1.588e-02 -2.748 0.00612 ** V5.l4 -3.874e-02 1.459e-02 -2.655 0.00807 ** V6.l4 6.579e-02 3.567e-02 1.844 0.06543 . V7.l4 -1.230e-09 1.591e-09 -0.773 0.43978 V3.l5 -9.809e-02 3.615e-02 -2.713 0.00678 ** V4.l5 1.357e-02 1.595e-02 0.851 0.39493 V5.l5 -7.406e-03 1.470e-02 -0.504 0.61457 V6.l5 -4.511e-03 3.561e-02 -0.127 0.89924 V7.l5 5.364e-11 1.586e-09 0.034 0.97303 V3.l6 -4.473e-03 3.633e-02 -0.123 0.90206 V4.l6 -1.902e-03 1.590e-02 -0.120 0.90482 V5.l6 1.518e-02 1.473e-02 1.030 0.30311 V6.l6 8.784e-02 3.612e-02 2.432 0.01521 * V7.l6 3.086e-10 1.585e-09 0.195 0.84571 V3.l7 -6.115e-02 3.609e-02 -1.694 0.09053 . V4.l7 4.979e-02 1.591e-02 3.129 0.00181 ** V5.l7 3.044e-02 1.474e-02 2.065 0.03917 * V6.l7 -4.487e-02 3.624e-02 -1.238 0.21591 V7.l7 2.041e-11 1.585e-09 0.013 0.98973 V3.l8 2.435e-02 3.634e-02 0.670 0.50298 V4.l8 -1.709e-02 1.591e-02 -1.074 0.28296 V5.l8 -1.685e-02 1.474e-02 -1.143 0.25343 V6.l8 3.794e-02 3.611e-02 1.051 0.29368 V7.l8 -1.661e-09 1.577e-09 -1.053 0.29240 V3.l9 -6.258e-02 3.629e-02 -1.725 0.08491 . V4.l9 3.590e-02 1.593e-02 2.253 0.02446 * V5.l9 8.078e-03 1.479e-02 0.546 0.58501 V6.l9 -1.806e-02 3.643e-02 -0.496 0.62010 V7.l9 1.041e-09 1.563e-09 0.666 0.50556 V3.l10 -1.596e-02 1.388e-02 -1.150 0.25046 V4.l10 4.676e-03 1.597e-02 0.293 0.76982 V5.l10 9.758e-03 1.478e-02 0.660 0.50929 V6.l10 3.242e-02 3.631e-02 0.893 0.37206 V7.l10 1.579e-09 1.466e-09 1.076 0.28199 const 1.038e-03 1.400e-03 0.741 0.45873 trend -1.283e-07 2.258e-07 -0.568 0.57002 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.001302 on 934 degrees of freedom Multiple R-Squared: 0.9998, Adjusted R-squared: 0.9998 F-statistic: 8.245e+04 on 51 and 934 DF, p-value: < 2.2e-16 Estimation results for equation V4: =================================== V4 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 -2.137e-02 1.355e-01 -0.158 0.8747 V4.l1 -9.351e-02 6.037e-02 -1.549 0.1218 V5.l1 6.453e-02 5.439e-02 1.186 0.2358 V6.l1 1.042e+00 5.787e-02 18.007 < 2e-16 *** V7.l1 8.568e-09 5.660e-09 1.514 0.1305 V3.l2 9.777e-02 1.363e-01 0.718 0.4732 V4.l2 4.072e-02 6.085e-02 0.669 0.5035 V5.l2 -6.018e-02 5.594e-02 -1.076 0.2823 V6.l2 2.610e-02 1.366e-01 0.191 0.8485 V7.l2 6.496e-09 6.020e-09 1.079 0.2808 V3.l3 2.076e-01 1.364e-01 1.522 0.1284 V4.l3 -6.666e-02 6.055e-02 -1.101 0.2713 V5.l3 -9.251e-02 5.582e-02 -1.657 0.0978 . V6.l3 -5.625e-02 1.363e-01 -0.413 0.6798 V7.l3 -6.398e-09 6.065e-09 -1.055 0.2917 V3.l4 3.484e-01 1.364e-01 2.555 0.0108 * V4.l4 -1.202e-01 6.083e-02 -1.976 0.0485 * V5.l4 -1.007e-01 5.590e-02 -1.802 0.0719 . V6.l4 1.772e-02 1.367e-01 0.130 0.8969 V7.l4 -4.718e-10 6.096e-09 -0.077 0.9383 V3.l5 3.264e-02 1.385e-01 0.236 0.8137 V4.l5 8.265e-03 6.110e-02 0.135 0.8924 V5.l5 -5.888e-02 5.632e-02 -1.045 0.2961 V6.l5 -2.329e-01 1.364e-01 -1.707 0.0882 . V7.l5 -2.278e-09 6.076e-09 -0.375 0.7078 V3.l6 4.075e-03 1.392e-01 0.029 0.9767 V4.l6 -1.283e-02 6.093e-02 -0.211 0.8332 V5.l6 -5.057e-02 5.642e-02 -0.896 0.3703 V6.l6 5.970e-02 1.384e-01 0.431 0.6663 V7.l6 -8.713e-10 6.074e-09 -0.143 0.8860 V3.l7 -2.582e-02 1.383e-01 -0.187 0.8519 V4.l7 4.276e-02 6.097e-02 0.701 0.4832 V5.l7 9.950e-02 5.647e-02 1.762 0.0784 . V6.l7 -8.214e-02 1.388e-01 -0.592 0.5542 V7.l7 1.060e-09 6.073e-09 0.175 0.8615 V3.l8 -4.635e-02 1.392e-01 -0.333 0.7392 V4.l8 1.038e-02 6.094e-02 0.170 0.8647 V5.l8 4.205e-02 5.648e-02 0.745 0.4567 V6.l8 -8.289e-02 1.383e-01 -0.599 0.5492 V7.l8 4.292e-09 6.041e-09 0.710 0.4776 V3.l9 4.436e-02 1.390e-01 0.319 0.7497 V4.l9 -3.187e-02 6.103e-02 -0.522 0.6017 V5.l9 -4.553e-02 5.665e-02 -0.804 0.4217 V6.l9 7.967e-02 1.396e-01 0.571 0.5682 V7.l9 -7.693e-11 5.987e-09 -0.013 0.9898 V3.l10 -2.719e-03 5.317e-02 -0.051 0.9592 V4.l10 1.819e-02 6.120e-02 0.297 0.7663 V5.l10 -1.833e-02 5.663e-02 -0.324 0.7462 V6.l10 3.164e-04 1.391e-01 0.002 0.9982 V7.l10 -2.223e-09 5.618e-09 -0.396 0.6924 const 2.908e-02 5.364e-03 5.422 7.5e-08 *** trend -9.603e-07 8.652e-07 -1.110 0.2674 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.004989 on 934 degrees of freedom Multiple R-Squared: 0.9966, Adjusted R-squared: 0.9965 F-statistic: 5431 on 51 and 934 DF, p-value: < 2.2e-16 Estimation results for equation V5: =================================== V5 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 -1.829e-02 1.848e-01 -0.099 0.921159 V4.l1 -3.141e-01 8.235e-02 -3.814 0.000146 *** V5.l1 7.342e-02 7.419e-02 0.990 0.322618 V6.l1 1.108e+00 7.893e-02 14.041 < 2e-16 *** V7.l1 -1.244e-08 7.720e-09 -1.611 0.107462 V3.l2 1.848e-01 1.858e-01 0.995 0.320178 V4.l2 -8.525e-02 8.300e-02 -1.027 0.304609 V5.l2 -1.704e-01 7.630e-02 -2.234 0.025733 * V6.l2 3.169e-01 1.863e-01 1.701 0.089231 . V7.l2 6.312e-09 8.210e-09 0.769 0.442200 V3.l3 -6.741e-02 1.860e-01 -0.362 0.717163 V4.l3 -7.400e-02 8.259e-02 -0.896 0.370533 V5.l3 -8.990e-02 7.613e-02 -1.181 0.237969 V6.l3 -2.240e-02 1.858e-01 -0.121 0.904100 V7.l3 -1.540e-11 8.272e-09 -0.002 0.998515 V3.l4 -7.687e-01 1.860e-01 -4.133 3.90e-05 *** V4.l4 -1.738e-01 8.296e-02 -2.095 0.036407 * V5.l4 1.159e-01 7.625e-02 1.521 0.128693 V6.l4 1.470e-01 1.864e-01 0.789 0.430351 V7.l4 -2.929e-09 8.315e-09 -0.352 0.724718 V3.l5 -1.853e-01 1.889e-01 -0.981 0.326722 V4.l5 -9.323e-02 8.333e-02 -1.119 0.263548 V5.l5 8.284e-02 7.682e-02 1.078 0.281116 V6.l5 8.068e-01 1.861e-01 4.335 1.61e-05 *** V7.l5 5.350e-09 8.288e-09 0.646 0.518699 V3.l6 -6.964e-02 1.899e-01 -0.367 0.713830 V4.l6 1.697e-01 8.310e-02 2.042 0.041456 * V5.l6 2.516e-01 7.696e-02 3.270 0.001117 ** V6.l6 -6.916e-04 1.887e-01 -0.004 0.997077 V7.l6 2.371e-09 8.284e-09 0.286 0.774752 V3.l7 -9.441e-02 1.886e-01 -0.501 0.616740 V4.l7 5.609e-02 8.315e-02 0.675 0.500160 V5.l7 1.198e-01 7.703e-02 1.556 0.120161 V6.l7 -1.815e-01 1.893e-01 -0.958 0.338059 V7.l7 7.488e-09 8.284e-09 0.904 0.366239 V3.l8 -4.161e-02 1.899e-01 -0.219 0.826589 V4.l8 2.061e-02 8.312e-02 0.248 0.804253 V5.l8 2.639e-03 7.704e-02 0.034 0.972678 V6.l8 1.763e-02 1.887e-01 0.093 0.925598 V7.l8 -3.987e-09 8.240e-09 -0.484 0.628572 V3.l9 -4.596e-02 1.896e-01 -0.242 0.808522 V4.l9 1.326e-01 8.325e-02 1.593 0.111572 V5.l9 1.477e-01 7.727e-02 1.912 0.056184 . V6.l9 -2.284e-01 1.903e-01 -1.200 0.230370 V7.l9 1.755e-09 8.166e-09 0.215 0.829873 V3.l10 -1.327e-01 7.252e-02 -1.830 0.067562 . V4.l10 7.385e-02 8.347e-02 0.885 0.376505 V5.l10 6.001e-02 7.723e-02 0.777 0.437346 V6.l10 -3.097e-02 1.897e-01 -0.163 0.870372 V7.l10 9.871e-10 7.662e-09 0.129 0.897529 const 2.898e-04 7.316e-03 0.040 0.968414 trend -1.780e-06 1.180e-06 -1.509 0.131738 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.006805 on 934 degrees of freedom Multiple R-Squared: 0.9941, Adjusted R-squared: 0.9938 F-statistic: 3111 on 51 and 934 DF, p-value: < 2.2e-16 Estimation results for equation V6: =================================== V6 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 3.367e-01 2.253e-01 1.494 0.135387 V4.l1 -3.881e-01 1.004e-01 -3.866 0.000118 *** V5.l1 -2.200e-02 9.044e-02 -0.243 0.807873 V6.l1 1.179e+00 9.623e-02 12.255 < 2e-16 *** V7.l1 -2.882e-09 9.412e-09 -0.306 0.759468 V3.l2 2.041e-01 2.266e-01 0.901 0.367788 V4.l2 -7.237e-02 1.012e-01 -0.715 0.474661 V5.l2 -2.325e-01 9.301e-02 -2.500 0.012591 * V6.l2 7.984e-02 2.271e-01 0.352 0.725235 V7.l2 7.645e-09 1.001e-08 0.764 0.445163 V3.l3 2.409e-01 2.268e-01 1.062 0.288462 V4.l3 -1.585e-01 1.007e-01 -1.574 0.115798 V5.l3 -1.465e-01 9.281e-02 -1.578 0.114903 V6.l3 2.439e-02 2.266e-01 0.108 0.914284 V7.l3 3.626e-09 1.008e-08 0.360 0.719269 V3.l4 -3.998e-01 2.267e-01 -1.763 0.078145 . V4.l4 -1.811e-01 1.011e-01 -1.790 0.073703 . V5.l4 -1.719e-03 9.295e-02 -0.018 0.985251 V6.l4 2.633e-03 2.272e-01 0.012 0.990758 V7.l4 -3.595e-09 1.014e-08 -0.355 0.722889 V3.l5 -1.582e-01 2.303e-01 -0.687 0.492234 V4.l5 -9.912e-02 1.016e-01 -0.976 0.329491 V5.l5 6.789e-02 9.365e-02 0.725 0.468687 V6.l5 4.993e-01 2.269e-01 2.201 0.027989 * V7.l5 3.021e-09 1.010e-08 0.299 0.765016 V3.l6 -1.322e-02 2.314e-01 -0.057 0.954458 V4.l6 1.370e-01 1.013e-01 1.353 0.176508 V5.l6 1.814e-01 9.382e-02 1.934 0.053474 . V6.l6 6.317e-02 2.301e-01 0.275 0.783727 V7.l6 2.521e-09 1.010e-08 0.250 0.802954 V3.l7 4.250e-02 2.299e-01 0.185 0.853386 V4.l7 4.719e-02 1.014e-01 0.465 0.641693 V5.l7 1.682e-01 9.390e-02 1.791 0.073546 . V6.l7 -2.684e-01 2.308e-01 -1.163 0.245204 V7.l7 2.276e-09 1.010e-08 0.225 0.821770 V3.l8 9.332e-04 2.315e-01 0.004 0.996784 V4.l8 1.361e-02 1.013e-01 0.134 0.893166 V5.l8 -6.082e-02 9.392e-02 -0.648 0.517421 V6.l8 -1.062e-01 2.300e-01 -0.462 0.644388 V7.l8 -4.049e-09 1.005e-08 -0.403 0.686981 V3.l9 1.532e-01 2.311e-01 0.663 0.507662 V4.l9 2.968e-02 1.015e-01 0.292 0.770000 V5.l9 7.221e-02 9.420e-02 0.767 0.443524 V6.l9 -8.450e-02 2.320e-01 -0.364 0.715843 V7.l9 4.529e-09 9.955e-09 0.455 0.649265 V3.l10 -1.155e-03 8.841e-02 -0.013 0.989576 V4.l10 -2.097e-02 1.018e-01 -0.206 0.836787 V5.l10 -2.282e-02 9.416e-02 -0.242 0.808558 V6.l10 -1.200e-01 2.313e-01 -0.519 0.603910 V7.l10 3.200e-10 9.341e-09 0.034 0.972677 const 2.554e-02 8.919e-03 2.863 0.004289 ** trend -3.021e-06 1.439e-06 -2.100 0.035987 * --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.008295 on 934 degrees of freedom Multiple R-Squared: 0.991, Adjusted R-squared: 0.9905 F-statistic: 2024 on 51 and 934 DF, p-value: < 2.2e-16 Estimation results for equation V7: =================================== V7 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 -9.316e+05 8.352e+05 -1.116 0.2649 V4.l1 -1.262e+05 3.722e+05 -0.339 0.7345 V5.l1 -1.918e+05 3.353e+05 -0.572 0.5674 V6.l1 4.801e+05 3.567e+05 1.346 0.1787 V7.l1 3.832e-01 3.489e-02 10.981 < 2e-16 *** V3.l2 -1.699e+06 8.400e+05 -2.023 0.0433 * V4.l2 1.251e+05 3.751e+05 0.334 0.7388 V5.l2 2.759e+05 3.448e+05 0.800 0.4239 V6.l2 7.475e+05 8.419e+05 0.888 0.3748 V7.l2 1.455e-01 3.711e-02 3.922 9.44e-05 *** V3.l3 1.653e+06 8.408e+05 1.965 0.0497 * V4.l3 -9.239e+05 3.733e+05 -2.475 0.0135 * V5.l3 2.028e+05 3.441e+05 0.589 0.5558 V6.l3 1.570e+06 8.399e+05 1.869 0.0619 . V7.l3 5.747e-02 3.739e-02 1.537 0.1246 V3.l4 1.445e+06 8.406e+05 1.719 0.0860 . V4.l4 7.055e+05 3.750e+05 1.882 0.0602 . V5.l4 -5.551e+05 3.446e+05 -1.611 0.1076 V6.l4 -1.127e+06 8.424e+05 -1.338 0.1813 V7.l4 -1.332e-02 3.758e-02 -0.355 0.7230 V3.l5 1.013e+06 8.537e+05 1.187 0.2356 V4.l5 -2.501e+05 3.766e+05 -0.664 0.5067 V5.l5 5.286e+05 3.472e+05 1.522 0.1282 V6.l5 -1.949e+06 8.411e+05 -2.317 0.0207 * V7.l5 3.748e-01 3.746e-02 10.006 < 2e-16 *** V3.l6 -1.229e+05 8.581e+05 -0.143 0.8861 V4.l6 -1.658e+05 3.756e+05 -0.442 0.6589 V5.l6 -3.440e+05 3.478e+05 -0.989 0.3229 V6.l6 -6.986e+05 8.531e+05 -0.819 0.4130 V7.l6 -7.677e-02 3.744e-02 -2.050 0.0406 * V3.l7 -9.764e+05 8.523e+05 -1.146 0.2523 V4.l7 4.369e+05 3.758e+05 1.163 0.2453 V5.l7 1.690e+05 3.481e+05 0.485 0.6276 V6.l7 -5.993e+04 8.558e+05 -0.070 0.9442 V7.l7 -6.295e-02 3.744e-02 -1.681 0.0930 . V3.l8 -4.692e+05 8.582e+05 -0.547 0.5846 V4.l8 4.179e+05 3.757e+05 1.112 0.2662 V5.l8 5.177e+05 3.482e+05 1.487 0.1374 V6.l8 3.444e+04 8.528e+05 0.040 0.9678 V7.l8 1.138e-02 3.724e-02 0.305 0.7601 V3.l9 -1.097e+06 8.569e+05 -1.280 0.2008 V4.l9 -3.034e+05 3.762e+05 -0.806 0.4203 V5.l9 -2.661e+05 3.492e+05 -0.762 0.4463 V6.l9 8.136e+05 8.603e+05 0.946 0.3445 V7.l9 7.410e-03 3.691e-02 0.201 0.8409 V3.l10 -7.367e+04 3.278e+05 -0.225 0.8222 V4.l10 -1.570e+05 3.773e+05 -0.416 0.6774 V5.l10 4.374e+05 3.491e+05 1.253 0.2105 V6.l10 8.825e+05 8.574e+05 1.029 0.3036 V7.l10 7.415e-02 3.463e-02 2.141 0.0325 * const 6.062e+04 3.307e+04 1.833 0.0671 . trend 1.057e+01 5.334e+00 1.981 0.0479 * --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 30750 on 934 degrees of freedom Multiple R-Squared: 0.7602, Adjusted R-squared: 0.7471 F-statistic: 58.06 on 51 and 934 DF, p-value: < 2.2e-16 Covariance matrix of residuals: V3 V4 V5 V6 V7 V3 1.696e-06 1.916e-06 1.985e-06 2.317e-06 -1.072e+00 V4 1.916e-06 2.489e-05 1.241e-05 2.951e-05 2.352e+01 V5 1.985e-06 1.241e-05 4.630e-05 4.676e-05 -5.108e+01 V6 2.317e-06 2.951e-05 4.676e-05 6.881e-05 -1.904e+01 V7 -1.072e+00 2.352e+01 -5.108e+01 -1.904e+01 9.458e+08 Correlation matrix of residuals: V3 V4 V5 V6 V7 V3 1.00000 0.2949 0.2240 0.21444 -0.02676 V4 0.29492 1.0000 0.3655 0.71301 0.15328 V5 0.22403 0.3655 1.0000 0.82829 -0.24409 V6 0.21444 0.7130 0.8283 1.00000 -0.07464 V7 -0.02676 0.1533 -0.2441 -0.07464 1.00000 > > fitvar1 = VAR(lr, p=10, type="both") > > # auto correlation function for residuals for close price > > acf(residuals(fitvar1)[,4]) > > # auto correlation function for residuals for open, high and low price > > acf(residuals(fitvar1)[,1]) > > acf(residuals(fitvar1)[,2]) > > acf(residuals(fitvar1)[,3]) > > #vector auto regression of Open, High, Low and Close > > #predict the next 10 values > > var.pred <- predict(fitvar1, n.ahead=3, ci=0.95) > > var.pred $V3 fcst lower upper CI [1,] 1.332669 1.330116 1.335221 0.002552364 [2,] 1.333993 1.316781 1.351204 0.017211745 [3,] 1.333547 1.309619 1.357475 0.023928297 $V4 fcst lower upper CI [1,] 1.341456 1.331678 1.351234 0.009778128 [2,] 1.341383 1.321789 1.360978 0.019594577 [3,] 1.342218 1.316616 1.367819 0.025601667 $V5 fcst lower upper CI [1,] 1.326643 1.313306 1.339979 0.01333679 [2,] 1.325935 1.304385 1.347485 0.02155004 [3,] 1.325274 1.297976 1.352572 0.02729799 $V6 fcst lower upper CI [1,] 1.334641 1.318382 1.350900 0.01625882 [2,] 1.333610 1.310357 1.356863 0.02325269 [3,] 1.335006 1.306402 1.363610 0.02860379 $V7 fcst lower upper CI [1,] 112702.0 52424.20 172979.7 60277.75 [2,] 107960.4 43233.35 172687.4 64727.03 [3,] 119106.5 51915.36 186297.6 67191.10 > > quotes[(x-4):x,] V1 V2 V3 V4 V5 V6 V7 2052 2016.08.04 00:00 1.33255 1.33450 1.31027 1.31072 120701 2053 2016.08.05 00:00 1.31037 1.31751 1.30216 1.30762 96414 2054 2016.08.08 00:00 1.30809 1.30969 1.30280 1.30401 71881 2055 2016.08.09 00:00 1.30407 1.30486 1.29560 1.30021 84611 2056 2016.08.10 00:00 1.30051 1.30941 1.29945 1.30280 74582
Now you can check above the predicted closing price is 1.33464 while the actual closing price is 1.31072. So the prediction is way off the mark. VAR model is not good for the daily timeframe as well as the weekly timeframe. Why? Keep this in mind that price relationship is highly non linear while VAR is a linear model. So most of the time this model will not be able to capture the non linearity in price.
H4 Candle Prediction Using VAR Model
> # Import the csv file > quotes <- read.csv("E:/MarketData/GBPUSD240.csv", header=FALSE) > > > x <-nrow(quotes) > > lr <- quotes[(x-1000):(x-5),3:7] > > > #install vars package > > library(vars) > > summary(VAR(lr, p=10, type="both")) VAR Estimation Results: ========================= Endogenous variables: V3, V4, V5, V6, V7 Deterministic variables: both Sample size: 986 Log Likelihood: 9436.685 Roots of the characteristic polynomial: 0.9896 0.9896 0.988 0.9771 0.9028 0.8803 0.8502 0.8502 0.8499 0.8499 0.8335 0.8335 0.8311 0.8311 0.8173 0.8173 0.8098 0.8098 0.8072 0.8072 0.7951 0.7951 0.7832 0.7832 0.7826 0.7826 0.7796 0.7796 0.7778 0.7778 0.7774 0.7774 0.7714 0.7701 0.7701 0.7547 0.7547 0.7461 0.7461 0.6952 0.6952 0.6634 0.6634 0.6049 0.6049 0.5589 0.5589 0.5488 0.5488 0.1861 Call: VAR(y = lr, p = 10, type = "both") Estimation results for equation V3: =================================== V3 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 -1.037e-01 3.453e-02 -3.003 0.002748 ** V4.l1 1.417e-02 1.355e-02 1.046 0.296059 V5.l1 7.148e-03 1.378e-02 0.519 0.604168 V6.l1 9.963e-01 1.406e-02 70.877 < 2e-16 *** V7.l1 6.820e-10 5.010e-09 0.136 0.891763 V3.l2 -2.832e-04 3.464e-02 -0.008 0.993479 V4.l2 -4.638e-03 1.394e-02 -0.333 0.739342 V5.l2 1.439e-02 1.439e-02 1.000 0.317613 V6.l2 6.864e-02 3.597e-02 1.908 0.056674 . V7.l2 -4.326e-09 5.517e-09 -0.784 0.433168 V3.l3 -8.623e-03 3.449e-02 -0.250 0.802651 V4.l3 -1.653e-04 1.388e-02 -0.012 0.990498 V5.l3 -1.455e-02 1.440e-02 -1.010 0.312834 V6.l3 2.691e-02 3.591e-02 0.749 0.453842 V7.l3 3.468e-09 5.509e-09 0.630 0.529131 V3.l4 7.907e-02 3.444e-02 2.296 0.021896 * V4.l4 2.202e-03 1.381e-02 0.159 0.873365 V5.l4 -1.102e-02 1.443e-02 -0.764 0.445163 V6.l4 -2.170e-02 3.567e-02 -0.608 0.543219 V7.l4 4.447e-09 5.321e-09 0.836 0.403504 V3.l5 -1.896e-02 3.430e-02 -0.553 0.580555 V4.l5 -9.700e-02 1.379e-02 -7.033 3.91e-12 *** V5.l5 9.335e-02 1.436e-02 6.500 1.30e-10 *** V6.l5 5.174e-03 3.558e-02 0.145 0.884400 V7.l5 1.681e-08 4.422e-09 3.802 0.000153 *** V3.l6 6.702e-03 2.851e-02 0.235 0.814221 V4.l6 -8.106e-03 1.418e-02 -0.572 0.567714 V5.l6 1.757e-02 1.465e-02 1.199 0.230715 V6.l6 -1.948e-02 3.546e-02 -0.549 0.582846 V7.l6 -2.532e-09 4.434e-09 -0.571 0.568130 V3.l7 -7.122e-03 2.615e-02 -0.272 0.785413 V4.l7 1.663e-02 1.422e-02 1.170 0.242278 V5.l7 -4.681e-03 1.479e-02 -0.316 0.751718 V6.l7 -3.376e-02 2.958e-02 -1.141 0.254041 V7.l7 2.606e-09 5.390e-09 0.484 0.628813 V3.l8 -5.688e-02 2.621e-02 -2.170 0.030268 * V4.l8 2.543e-02 1.418e-02 1.794 0.073195 . V5.l8 -1.678e-02 1.471e-02 -1.141 0.254291 V6.l8 1.446e-02 2.737e-02 0.528 0.597325 V7.l8 -5.162e-09 5.525e-09 -0.934 0.350409 V3.l9 3.163e-03 2.591e-02 0.122 0.902853 V4.l9 -7.183e-03 1.421e-02 -0.505 0.613355 V5.l9 5.744e-04 1.472e-02 0.039 0.968874 V6.l9 3.546e-02 2.740e-02 1.294 0.195974 V7.l9 -6.583e-09 5.527e-09 -1.191 0.233992 V3.l10 -6.957e-03 1.158e-02 -0.601 0.548160 V4.l10 2.902e-02 1.377e-02 2.108 0.035300 * V5.l10 -3.057e-04 1.408e-02 -0.022 0.982678 V6.l10 -1.507e-02 2.719e-02 -0.554 0.579540 V7.l10 -5.012e-09 5.016e-09 -0.999 0.317964 const 9.394e-04 8.722e-04 1.077 0.281736 trend 1.526e-07 1.082e-07 1.411 0.158608 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.0006991 on 934 degrees of freedom Multiple R-Squared: 0.9998, Adjusted R-squared: 0.9998 F-statistic: 1.035e+05 on 51 and 934 DF, p-value: < 2.2e-16 Estimation results for equation V4: =================================== V4 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 1.867e-01 1.498e-01 1.246 0.21305 V4.l1 2.615e-01 5.878e-02 4.449 9.66e-06 *** V5.l1 -1.529e-01 5.980e-02 -2.556 0.01073 * V6.l1 8.369e-01 6.099e-02 13.723 < 2e-16 *** V7.l1 -1.827e-08 2.174e-08 -0.841 0.40077 V3.l2 1.987e-01 1.503e-01 1.322 0.18640 V4.l2 -5.715e-02 6.046e-02 -0.945 0.34478 V5.l2 -1.485e-01 6.242e-02 -2.379 0.01756 * V6.l2 6.441e-03 1.561e-01 0.041 0.96709 V7.l2 -1.395e-08 2.394e-08 -0.583 0.56012 V3.l3 1.987e-01 1.497e-01 1.328 0.18451 V4.l3 -1.652e-02 6.021e-02 -0.274 0.78390 V5.l3 -1.091e-02 6.249e-02 -0.175 0.86145 V6.l3 -8.968e-02 1.558e-01 -0.576 0.56500 V7.l3 -3.601e-09 2.390e-08 -0.151 0.88029 V3.l4 8.586e-02 1.494e-01 0.575 0.56564 V4.l4 -1.257e-01 5.993e-02 -2.097 0.03622 * V5.l4 -1.263e-01 6.259e-02 -2.018 0.04384 * V6.l4 -9.618e-02 1.548e-01 -0.621 0.53447 V7.l4 1.666e-08 2.309e-08 0.722 0.47067 V3.l5 -2.921e-02 1.488e-01 -0.196 0.84442 V4.l5 -6.779e-02 5.984e-02 -1.133 0.25755 V5.l5 1.024e-01 6.231e-02 1.643 0.10070 V6.l5 1.214e-01 1.544e-01 0.786 0.43192 V7.l5 2.520e-08 1.918e-08 1.314 0.18929 V3.l6 7.593e-02 1.237e-01 0.614 0.53949 V4.l6 -2.472e-02 6.152e-02 -0.402 0.68797 V5.l6 -4.074e-02 6.356e-02 -0.641 0.52168 V6.l6 -1.102e-02 1.538e-01 -0.072 0.94293 V7.l6 3.820e-08 1.924e-08 1.986 0.04737 * V3.l7 -2.973e-01 1.135e-01 -2.620 0.00893 ** V4.l7 -2.097e-02 6.168e-02 -0.340 0.73396 V5.l7 -3.466e-02 6.417e-02 -0.540 0.58928 V6.l7 -7.907e-02 1.283e-01 -0.616 0.53795 V7.l7 1.849e-09 2.338e-08 0.079 0.93699 V3.l8 2.285e-02 1.137e-01 0.201 0.84082 V4.l8 2.888e-02 6.151e-02 0.469 0.63884 V5.l8 -3.901e-02 6.382e-02 -0.611 0.54118 V6.l8 3.719e-01 1.187e-01 3.133 0.00179 ** V7.l8 -9.722e-09 2.397e-08 -0.406 0.68515 V3.l9 -1.176e-01 1.124e-01 -1.046 0.29566 V4.l9 -4.954e-03 6.165e-02 -0.080 0.93597 V5.l9 -2.734e-03 6.385e-02 -0.043 0.96586 V6.l9 -1.201e-02 1.189e-01 -0.101 0.91953 V7.l9 -4.428e-08 2.398e-08 -1.847 0.06512 . V3.l10 1.182e-02 5.024e-02 0.235 0.81400 V4.l10 -1.088e-02 5.974e-02 -0.182 0.85551 V5.l10 -7.017e-02 6.107e-02 -1.149 0.25085 V6.l10 1.743e-01 1.180e-01 1.478 0.13987 V7.l10 4.819e-09 2.176e-08 0.221 0.82480 const 4.270e-03 3.784e-03 1.128 0.25943 trend 7.916e-07 4.693e-07 1.687 0.09200 . --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.003033 on 934 degrees of freedom Multiple R-Squared: 0.9966, Adjusted R-squared: 0.9964 F-statistic: 5395 on 51 and 934 DF, p-value: < 2.2e-16 Estimation results for equation V5: =================================== V5 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 -5.862e-02 2.342e-01 -0.250 0.80242 V4.l1 -2.930e-01 9.191e-02 -3.188 0.00148 ** V5.l1 6.265e-01 9.349e-02 6.701 3.58e-11 *** V6.l1 7.544e-01 9.535e-02 7.912 7.16e-15 *** V7.l1 1.955e-08 3.399e-08 0.575 0.56527 V3.l2 7.000e-02 2.350e-01 0.298 0.76583 V4.l2 1.559e-01 9.453e-02 1.649 0.09949 . V5.l2 -2.304e-01 9.759e-02 -2.361 0.01842 * V6.l2 -2.050e-02 2.440e-01 -0.084 0.93305 V7.l2 -2.204e-08 3.742e-08 -0.589 0.55608 V3.l3 3.358e-01 2.340e-01 1.435 0.15157 V4.l3 -3.799e-02 9.413e-02 -0.404 0.68660 V5.l3 2.207e-01 9.770e-02 2.259 0.02411 * V6.l3 -1.128e-01 2.436e-01 -0.463 0.64325 V7.l3 1.193e-08 3.737e-08 0.319 0.74966 V3.l4 9.250e-02 2.336e-01 0.396 0.69220 V4.l4 -1.414e-01 9.369e-02 -1.509 0.13172 V5.l4 -2.187e-01 9.786e-02 -2.234 0.02569 * V6.l4 -3.167e-01 2.420e-01 -1.309 0.19094 V7.l4 -8.423e-08 3.610e-08 -2.334 0.01983 * V3.l5 -3.966e-01 2.327e-01 -1.704 0.08863 . V4.l5 6.291e-02 9.356e-02 0.672 0.50149 V5.l5 1.890e-01 9.742e-02 1.941 0.05262 . V6.l5 1.466e-01 2.413e-01 0.608 0.54360 V7.l5 3.859e-08 2.999e-08 1.287 0.19851 V3.l6 -6.720e-02 1.934e-01 -0.347 0.72831 V4.l6 -1.019e-01 9.619e-02 -1.060 0.28947 V5.l6 -2.023e-01 9.937e-02 -2.036 0.04202 * V6.l6 3.183e-01 2.405e-01 1.323 0.18603 V7.l6 -5.574e-08 3.008e-08 -1.853 0.06420 . V3.l7 -1.416e-01 1.774e-01 -0.798 0.42489 V4.l7 -1.737e-01 9.644e-02 -1.801 0.07201 . V5.l7 1.918e-01 1.003e-01 1.911 0.05627 . V6.l7 1.216e-01 2.006e-01 0.606 0.54460 V7.l7 6.025e-08 3.656e-08 1.648 0.09970 . V3.l8 -1.668e-01 1.778e-01 -0.938 0.34833 V4.l8 2.354e-01 9.617e-02 2.448 0.01455 * V5.l8 -6.005e-02 9.978e-02 -0.602 0.54744 V6.l8 1.347e-01 1.856e-01 0.726 0.46815 V7.l8 -4.564e-08 3.748e-08 -1.218 0.22356 V3.l9 3.909e-02 1.757e-01 0.222 0.82405 V4.l9 -1.251e-01 9.639e-02 -1.298 0.19469 V5.l9 1.235e-01 9.983e-02 1.237 0.21630 V6.l9 5.619e-02 1.859e-01 0.302 0.76249 V7.l9 7.502e-08 3.749e-08 2.001 0.04568 * V3.l10 7.762e-02 7.855e-02 0.988 0.32331 V4.l10 -1.006e-01 9.340e-02 -1.077 0.28162 V5.l10 -9.781e-02 9.548e-02 -1.024 0.30595 V6.l10 1.029e-01 1.845e-01 0.558 0.57724 V7.l10 3.770e-09 3.403e-08 0.111 0.91179 const 1.237e-02 5.916e-03 2.090 0.03688 * trend -7.838e-07 7.337e-07 -1.068 0.28573 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.004742 on 934 degrees of freedom Multiple R-Squared: 0.992, Adjusted R-squared: 0.9916 F-statistic: 2273 on 51 and 934 DF, p-value: < 2.2e-16 Estimation results for equation V6: =================================== V6 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 6.143e-02 2.705e-01 0.227 0.82042 V4.l1 3.120e-03 1.062e-01 0.029 0.97656 V5.l1 4.992e-01 1.080e-01 4.622 4.34e-06 *** V6.l1 5.393e-01 1.101e-01 4.896 1.15e-06 *** V7.l1 1.022e-08 3.926e-08 0.260 0.79468 V3.l2 3.753e-01 2.714e-01 1.383 0.16712 V4.l2 7.276e-02 1.092e-01 0.666 0.50536 V5.l2 -2.912e-01 1.127e-01 -2.583 0.00994 ** V6.l2 -3.699e-03 2.819e-01 -0.013 0.98953 V7.l2 -1.620e-08 4.323e-08 -0.375 0.70796 V3.l3 5.037e-01 2.703e-01 1.863 0.06271 . V4.l3 5.418e-03 1.087e-01 0.050 0.96027 V5.l3 1.728e-01 1.129e-01 1.531 0.12614 V6.l3 -3.499e-01 2.814e-01 -1.243 0.21403 V7.l3 -6.099e-09 4.317e-08 -0.141 0.88768 V3.l4 1.073e-01 2.698e-01 0.398 0.69087 V4.l4 -2.038e-01 1.082e-01 -1.883 0.05996 . V5.l4 -2.368e-01 1.130e-01 -2.094 0.03649 * V6.l4 -4.544e-01 2.795e-01 -1.626 0.10439 V7.l4 -4.130e-08 4.170e-08 -0.990 0.32220 V3.l5 -2.926e-01 2.688e-01 -1.088 0.27667 V4.l5 1.135e-01 1.081e-01 1.050 0.29379 V5.l5 1.755e-01 1.125e-01 1.559 0.11923 V6.l5 2.073e-01 2.788e-01 0.744 0.45731 V7.l5 3.454e-08 3.465e-08 0.997 0.31910 V3.l6 -4.680e-02 2.234e-01 -0.209 0.83411 V4.l6 -7.567e-02 1.111e-01 -0.681 0.49604 V5.l6 -2.154e-01 1.148e-01 -1.876 0.06094 . V6.l6 1.749e-01 2.779e-01 0.630 0.52916 V7.l6 -2.357e-08 3.475e-08 -0.678 0.49779 V3.l7 -2.877e-01 2.049e-01 -1.404 0.16068 V4.l7 -1.925e-01 1.114e-01 -1.728 0.08425 . V5.l7 2.428e-01 1.159e-01 2.095 0.03644 * V6.l7 2.674e-02 2.318e-01 0.115 0.90816 V7.l7 7.484e-08 4.223e-08 1.772 0.07671 . V3.l8 3.902e-02 2.054e-01 0.190 0.84936 V4.l8 2.676e-01 1.111e-01 2.409 0.01618 * V5.l8 -2.263e-02 1.153e-01 -0.196 0.84438 V6.l8 2.070e-01 2.144e-01 0.965 0.33459 V7.l8 -4.861e-08 4.329e-08 -1.123 0.26181 V3.l9 -2.785e-02 2.030e-01 -0.137 0.89093 V4.l9 -8.262e-02 1.114e-01 -0.742 0.45827 V5.l9 1.045e-01 1.153e-01 0.906 0.36530 V6.l9 -1.622e-01 2.147e-01 -0.755 0.45031 V7.l9 1.587e-08 4.331e-08 0.366 0.71413 V3.l10 1.076e-01 9.074e-02 1.186 0.23583 V4.l10 -8.320e-02 1.079e-01 -0.771 0.44081 V5.l10 -1.501e-01 1.103e-01 -1.361 0.17395 V6.l10 1.609e-01 2.131e-01 0.755 0.45027 V7.l10 1.089e-08 3.931e-08 0.277 0.78176 const 1.696e-02 6.835e-03 2.481 0.01327 * trend -3.600e-07 8.476e-07 -0.425 0.67116 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.005478 on 934 degrees of freedom Multiple R-Squared: 0.9891, Adjusted R-squared: 0.9885 F-statistic: 1665 on 51 and 934 DF, p-value: < 2.2e-16 Estimation results for equation V7: =================================== V7 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 -1.274e+03 2.769e+05 -0.005 0.99633 V4.l1 2.046e+05 1.087e+05 1.883 0.06004 . V5.l1 -1.854e+05 1.105e+05 -1.677 0.09390 . V6.l1 -3.199e+04 1.127e+05 -0.284 0.77668 V7.l1 4.511e-01 4.019e-02 11.227 < 2e-16 *** V3.l2 -2.431e+04 2.778e+05 -0.087 0.93030 V4.l2 -2.200e+05 1.118e+05 -1.969 0.04930 * V5.l2 2.693e+04 1.154e+05 0.233 0.81550 V6.l2 1.571e+05 2.885e+05 0.545 0.58618 V7.l2 2.700e-02 4.425e-02 0.610 0.54190 V3.l3 -2.750e+05 2.766e+05 -0.994 0.32041 V4.l3 -3.928e+04 1.113e+05 -0.353 0.72423 V5.l3 -1.132e+05 1.155e+05 -0.980 0.32746 V6.l3 1.190e+05 2.880e+05 0.413 0.67954 V7.l3 -1.076e-01 4.419e-02 -2.435 0.01509 * V3.l4 1.284e+05 2.762e+05 0.465 0.64216 V4.l4 5.291e+04 1.108e+05 0.478 0.63306 V5.l4 2.115e+04 1.157e+05 0.183 0.85497 V6.l4 4.433e+05 2.861e+05 1.549 0.12160 V7.l4 1.215e-01 4.268e-02 2.847 0.00451 ** V3.l5 1.319e+05 2.751e+05 0.479 0.63183 V4.l5 5.867e+04 1.106e+05 0.530 0.59601 V5.l5 3.046e+04 1.152e+05 0.264 0.79152 V6.l5 -2.834e+05 2.854e+05 -0.993 0.32085 V7.l5 5.727e-02 3.546e-02 1.615 0.10664 V3.l6 4.160e+05 2.287e+05 1.819 0.06919 . V4.l6 -2.258e+05 1.137e+05 -1.985 0.04738 * V5.l6 5.037e+05 1.175e+05 4.287 2.00e-05 *** V6.l6 -3.381e+05 2.844e+05 -1.189 0.23478 V7.l6 6.820e-01 3.556e-02 19.177 < 2e-16 *** V3.l7 -2.509e+05 2.097e+05 -1.196 0.23193 V4.l7 3.214e+05 1.140e+05 2.818 0.00493 ** V5.l7 -1.147e+05 1.186e+05 -0.967 0.33397 V6.l7 -6.141e+05 2.372e+05 -2.589 0.00979 ** V7.l7 -3.123e-01 4.323e-02 -7.224 1.05e-12 *** V3.l8 1.497e+05 2.102e+05 0.712 0.47670 V4.l8 -7.843e+04 1.137e+05 -0.690 0.49053 V5.l8 5.199e+04 1.180e+05 0.441 0.65954 V6.l8 1.576e+05 2.195e+05 0.718 0.47288 V7.l8 4.220e-02 4.431e-02 0.952 0.34116 V3.l9 6.332e+04 2.078e+05 0.305 0.76066 V4.l9 2.956e+05 1.140e+05 2.594 0.00965 ** V5.l9 8.095e+03 1.180e+05 0.069 0.94534 V6.l9 -3.422e+05 2.198e+05 -1.557 0.11981 V7.l9 -2.162e-01 4.433e-02 -4.878 1.26e-06 *** V3.l10 -2.747e+04 9.288e+04 -0.296 0.76748 V4.l10 8.143e+04 1.104e+05 0.737 0.46106 V5.l10 7.613e+04 1.129e+05 0.674 0.50025 V6.l10 -3.288e+05 2.181e+05 -1.507 0.13203 V7.l10 5.448e-02 4.023e-02 1.354 0.17606 const -3.542e+03 6.995e+03 -0.506 0.61276 trend 1.028e+00 8.676e-01 1.185 0.23634 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 5607 on 934 degrees of freedom Multiple R-Squared: 0.7962, Adjusted R-squared: 0.785 F-statistic: 71.53 on 51 and 934 DF, p-value: < 2.2e-16 Covariance matrix of residuals: V3 V4 V5 V6 V7 V3 4.888e-07 5.852e-07 5.053e-07 6.699e-07 -2.234e-01 V4 5.852e-07 9.201e-06 5.158e-06 1.082e-05 4.208e+00 V5 5.053e-07 5.158e-06 2.249e-05 2.298e-05 -1.058e+01 V6 6.699e-07 1.082e-05 2.298e-05 3.001e-05 -5.724e+00 V7 -2.234e-01 4.208e+00 -1.058e+01 -5.724e+00 3.144e+07 Correlation matrix of residuals: V3 V4 V5 V6 V7 V3 1.00000 0.2760 0.1524 0.1749 -0.05699 V4 0.27596 1.0000 0.3586 0.6513 0.24741 V5 0.15240 0.3586 1.0000 0.8845 -0.39769 V6 0.17491 0.6513 0.8845 1.0000 -0.18635 V7 -0.05699 0.2474 -0.3977 -0.1863 1.00000 > > fitvar1 = VAR(lr, p=10, type="both") > > # auto correlation function for residuals for close price > > acf(residuals(fitvar1)[,4]) > > # auto correlation function for residuals for open, high and low price > > acf(residuals(fitvar1)[,1]) > > acf(residuals(fitvar1)[,2]) > > acf(residuals(fitvar1)[,3]) > > #vector auto regression of Open, High, Low and Close > > #predict the next 10 values > > var.pred <- predict(fitvar1, n.ahead=3, ci=0.95) > > var.pred $V3 fcst lower upper CI [1,] 1.332577 1.331207 1.333947 0.001370293 [2,] 1.334250 1.323378 1.345123 0.010872588 [3,] 1.334271 1.319469 1.349074 0.014802333 $V4 fcst lower upper CI [1,] 1.335382 1.329437 1.341327 0.005945184 [2,] 1.336633 1.325868 1.347398 0.010764925 [3,] 1.339282 1.324903 1.353661 0.014378940 $V5 fcst lower upper CI [1,] 1.332261 1.322966 1.341556 0.009294972 [2,] 1.332014 1.316372 1.347656 0.015641808 [3,] 1.332158 1.313221 1.351096 0.018937874 $V6 fcst lower upper CI [1,] 1.333983 1.323246 1.344721 0.01073738 [2,] 1.334070 1.319314 1.348826 0.01475552 [3,] 1.336536 1.318648 1.354424 0.01788796 $V7 fcst lower upper CI [1,] 8565.783 -2424.409 19555.97 10990.19 [2,] 9851.751 -2771.602 22475.10 12623.35 [3,] 24147.332 11166.106 37128.56 12981.23 > > quotes[(x-4):x,] V1 V2 V3 V4 V5 V6 V7 10249 2016.08.04 00:00 1.33255 1.33444 1.33178 1.33364 6610 10250 2016.08.04 04:00 1.33363 1.33450 1.33234 1.33282 6478 10251 2016.08.04 08:00 1.33281 1.33290 1.32788 1.32944 21980 10252 2016.08.04 12:00 1.32946 1.33412 1.31124 1.31369 45662 10253 2016.08.04 16:00 1.31367 1.31460 1.31165 1.31273 5221
Now let’s see what are the predictions and what are the actuals. Actual Open, High, Low and Close is 1.33255, 1.33444. 1.33178 and 1.33364. The predicted Open, High, Low and Close is 1.33257, 1.33538, 1.33261 and 1.33398. The predicted values are pretty close to the actual values. So we have a good predictive model for H4 timeframe.
H1 Candle Prediction Using VAR Model
Below we check how good are the predictions for H1 candle!
> # Import the csv file > quotes <- read.csv("E:/MarketData/GBPUSD60.csv", header=FALSE) > > > x <-nrow(quotes) > > lr <- quotes[(x-1000):(x-5),3:7] > > > #install vars package > > library(vars) > > summary(VAR(lr, p=10, type="both")) VAR Estimation Results: ========================= Endogenous variables: V3, V4, V5, V6, V7 Deterministic variables: both Sample size: 986 Log Likelihood: 11142.161 Roots of the characteristic polynomial: 0.993 0.9659 0.9659 0.9014 0.8919 0.8919 0.8822 0.8822 0.8602 0.8602 0.8429 0.8429 0.842 0.842 0.8405 0.8405 0.8403 0.8403 0.8292 0.8292 0.8221 0.8221 0.8158 0.8158 0.8096 0.8096 0.7852 0.7852 0.7845 0.7845 0.7641 0.7641 0.7591 0.7591 0.7591 0.7591 0.7524 0.7524 0.7495 0.7495 0.7479 0.7479 0.7449 0.7449 0.725 0.725 0.5638 0.5586 0.5586 0.552 Call: VAR(y = lr, p = 10, type = "both") Estimation results for equation V3: =================================== V3 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 4.438e-03 3.728e-02 0.119 0.905256 V4.l1 -6.359e-03 2.282e-02 -0.279 0.780580 V5.l1 8.814e-03 1.788e-02 0.493 0.622079 V6.l1 9.960e-01 1.998e-02 49.850 < 2e-16 *** V7.l1 8.948e-09 1.746e-08 0.512 0.608439 V3.l2 3.239e-02 3.712e-02 0.872 0.383220 V4.l2 -2.060e-02 2.276e-02 -0.905 0.365647 V5.l2 -3.047e-02 1.928e-02 -1.581 0.114269 V6.l2 3.435e-02 3.800e-02 0.904 0.366360 V7.l2 -1.885e-08 2.115e-08 -0.891 0.373010 V3.l3 -1.946e-02 3.697e-02 -0.526 0.598782 V4.l3 -2.882e-05 2.262e-02 -0.001 0.998984 V5.l3 -4.139e-02 2.097e-02 -1.974 0.048703 * V6.l3 1.257e-02 3.809e-02 0.330 0.741537 V7.l3 8.201e-09 2.133e-08 0.384 0.700713 V3.l4 -6.095e-02 3.697e-02 -1.649 0.099570 . V4.l4 5.569e-02 2.234e-02 2.493 0.012833 * V5.l4 2.554e-02 2.189e-02 1.167 0.243566 V6.l4 -1.451e-02 3.841e-02 -0.378 0.705613 V7.l4 -4.947e-09 2.140e-08 -0.231 0.817260 V3.l5 5.157e-02 3.697e-02 1.395 0.163454 V4.l5 -2.074e-02 2.237e-02 -0.927 0.354166 V5.l5 9.107e-02 2.169e-02 4.199 2.94e-05 *** V6.l5 -5.097e-02 3.799e-02 -1.342 0.180012 V7.l5 1.357e-08 2.135e-08 0.636 0.525017 V3.l6 5.659e-03 3.626e-02 0.156 0.876018 V4.l6 -1.431e-02 2.232e-02 -0.641 0.521771 V5.l6 -2.967e-02 2.176e-02 -1.363 0.173149 V6.l6 -2.392e-02 3.806e-02 -0.628 0.529832 V7.l6 -1.296e-08 2.139e-08 -0.606 0.544732 V3.l7 2.646e-02 3.625e-02 0.730 0.465588 V4.l7 -8.082e-02 2.228e-02 -3.627 0.000303 *** V5.l7 -3.216e-02 2.236e-02 -1.438 0.150679 V6.l7 8.674e-02 3.741e-02 2.318 0.020646 * V7.l7 4.074e-08 2.142e-08 1.902 0.057455 . V3.l8 1.983e-02 3.587e-02 0.553 0.580438 V4.l8 -2.340e-02 2.170e-02 -1.079 0.281088 V5.l8 4.672e-02 2.196e-02 2.127 0.033682 * V6.l8 -2.632e-02 3.740e-02 -0.704 0.481757 V7.l8 -3.263e-09 2.119e-08 -0.154 0.877636 V3.l9 7.981e-02 3.570e-02 2.236 0.025620 * V4.l9 -1.102e-01 2.156e-02 -5.110 3.91e-07 *** V5.l9 -5.239e-02 2.112e-02 -2.481 0.013288 * V6.l9 8.600e-02 3.680e-02 2.337 0.019665 * V7.l9 1.277e-08 2.115e-08 0.604 0.546185 V3.l10 3.119e-02 1.902e-02 1.640 0.101345 V4.l10 1.293e-02 2.064e-02 0.627 0.531081 V5.l10 -1.069e-01 2.112e-02 -5.060 5.04e-07 *** V6.l10 5.874e-02 3.629e-02 1.619 0.105882 V7.l10 -3.487e-08 1.671e-08 -2.086 0.037227 * const -1.288e-03 1.057e-03 -1.219 0.223260 trend 1.162e-07 1.471e-07 0.790 0.429900 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.0008625 on 934 degrees of freedom Multiple R-Squared: 0.9998, Adjusted R-squared: 0.9998 F-statistic: 8.027e+04 on 51 and 934 DF, p-value: < 2.2e-16 Estimation results for equation V4: =================================== V4 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 4.588e-02 9.153e-02 0.501 0.616294 V4.l1 7.162e-03 5.604e-02 0.128 0.898325 V5.l1 -7.112e-04 4.389e-02 -0.016 0.987075 V6.l1 9.936e-01 4.906e-02 20.252 < 2e-16 *** V7.l1 2.360e-07 4.287e-08 5.505 4.76e-08 *** V3.l2 2.572e-01 9.115e-02 2.821 0.004882 ** V4.l2 -7.152e-03 5.589e-02 -0.128 0.898208 V5.l2 1.404e-02 4.733e-02 0.297 0.766898 V6.l2 -6.490e-02 9.331e-02 -0.696 0.486908 V7.l2 -7.591e-08 5.193e-08 -1.462 0.144180 V3.l3 1.613e-01 9.077e-02 1.777 0.075867 . V4.l3 4.700e-02 5.553e-02 0.846 0.397567 V5.l3 -1.812e-01 5.150e-02 -3.520 0.000453 *** V6.l3 -1.891e-01 9.353e-02 -2.022 0.043441 * V7.l3 -9.163e-08 5.237e-08 -1.750 0.080530 . V3.l4 -1.145e-01 9.078e-02 -1.261 0.207517 V4.l4 1.224e-01 5.485e-02 2.231 0.025938 * V5.l4 4.819e-03 5.375e-02 0.090 0.928568 V6.l4 -1.639e-01 9.431e-02 -1.738 0.082538 . V7.l4 -2.065e-08 5.255e-08 -0.393 0.694376 V3.l5 8.916e-02 9.079e-02 0.982 0.326319 V4.l5 6.277e-02 5.493e-02 1.143 0.253492 V5.l5 1.631e-01 5.325e-02 3.063 0.002256 ** V6.l5 -1.408e-01 9.328e-02 -1.510 0.131462 V7.l5 3.189e-08 5.241e-08 0.609 0.542960 V3.l6 1.483e-01 8.904e-02 1.666 0.096125 . V4.l6 -1.863e-02 5.481e-02 -0.340 0.734029 V5.l6 -3.543e-01 5.343e-02 -6.631 5.64e-11 *** V6.l6 8.712e-02 9.344e-02 0.932 0.351397 V7.l6 -8.830e-08 5.253e-08 -1.681 0.093071 . V3.l7 3.542e-02 8.902e-02 0.398 0.690815 V4.l7 -1.384e-01 5.472e-02 -2.530 0.011563 * V5.l7 4.650e-03 5.489e-02 0.085 0.932515 V6.l7 1.473e-01 9.186e-02 1.603 0.109175 V7.l7 8.635e-08 5.258e-08 1.642 0.100916 V3.l8 1.048e-01 8.808e-02 1.190 0.234223 V4.l8 7.103e-03 5.328e-02 0.133 0.893965 V5.l8 3.034e-02 5.393e-02 0.563 0.573900 V6.l8 -4.870e-02 9.183e-02 -0.530 0.596040 V7.l8 -7.383e-09 5.202e-08 -0.142 0.887159 V3.l9 -1.012e-01 8.766e-02 -1.154 0.248674 V4.l9 -5.861e-02 5.295e-02 -1.107 0.268619 V5.l9 7.986e-03 5.186e-02 0.154 0.877647 V6.l9 -2.641e-02 9.037e-02 -0.292 0.770147 V7.l9 3.976e-08 5.193e-08 0.766 0.444022 V3.l10 -6.842e-02 4.670e-02 -1.465 0.143212 V4.l10 1.687e-01 5.068e-02 3.329 0.000907 *** V5.l10 -3.109e-02 5.186e-02 -0.599 0.549012 V6.l10 5.191e-05 8.911e-02 0.001 0.999535 V7.l10 -1.038e-07 4.104e-08 -2.529 0.011592 * const -1.719e-03 2.596e-03 -0.662 0.508088 trend -2.279e-07 3.612e-07 -0.631 0.528195 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.002118 on 934 degrees of freedom Multiple R-Squared: 0.9986, Adjusted R-squared: 0.9986 F-statistic: 1.346e+04 on 51 and 934 DF, p-value: < 2.2e-16 Estimation results for equation V5: =================================== V5 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 -2.807e-01 1.443e-01 -1.945 0.052078 . V4.l1 1.546e-01 8.835e-02 1.749 0.080548 . V5.l1 6.578e-01 6.920e-02 9.506 < 2e-16 *** V6.l1 4.983e-01 7.735e-02 6.442 1.88e-10 *** V7.l1 -1.405e-07 6.759e-08 -2.079 0.037880 * V3.l2 2.739e-01 1.437e-01 1.906 0.056953 . V4.l2 -3.151e-02 8.812e-02 -0.358 0.720729 V5.l2 6.902e-01 7.463e-02 9.248 < 2e-16 *** V6.l2 -5.786e-01 1.471e-01 -3.933 9.02e-05 *** V7.l2 2.790e-07 8.188e-08 3.408 0.000683 *** V3.l3 -2.295e-01 1.431e-01 -1.603 0.109235 V4.l3 1.747e-01 8.756e-02 1.995 0.046349 * V5.l3 3.498e-01 8.119e-02 4.308 1.82e-05 *** V6.l3 -5.572e-01 1.475e-01 -3.778 0.000168 *** V7.l3 -1.034e-07 8.258e-08 -1.252 0.210708 V3.l4 -6.276e-02 1.431e-01 -0.438 0.661141 V4.l4 4.878e-02 8.648e-02 0.564 0.572843 V5.l4 -1.861e-01 8.474e-02 -2.196 0.028317 * V6.l4 -8.661e-02 1.487e-01 -0.582 0.560412 V7.l4 -9.678e-08 8.286e-08 -1.168 0.243108 V3.l5 2.084e-01 1.431e-01 1.456 0.145727 V4.l5 -1.693e-01 8.661e-02 -1.954 0.050960 . V5.l5 -9.031e-03 8.397e-02 -0.108 0.914375 V6.l5 7.572e-02 1.471e-01 0.515 0.606803 V7.l5 1.812e-07 8.263e-08 2.193 0.028575 * V3.l6 2.408e-01 1.404e-01 1.715 0.086665 . V4.l6 -2.154e-01 8.643e-02 -2.493 0.012851 * V5.l6 -3.403e-01 8.425e-02 -4.039 5.81e-05 *** V6.l6 2.949e-01 1.473e-01 2.001 0.045639 * V7.l6 -6.397e-08 8.282e-08 -0.772 0.440074 V3.l7 7.131e-02 1.404e-01 0.508 0.611515 V4.l7 5.537e-02 8.627e-02 0.642 0.521118 V5.l7 -1.114e-02 8.655e-02 -0.129 0.897650 V6.l7 -1.766e-02 1.448e-01 -0.122 0.902961 V7.l7 -2.865e-08 8.291e-08 -0.346 0.729789 V3.l8 -9.743e-02 1.389e-01 -0.702 0.483116 V4.l8 9.449e-02 8.400e-02 1.125 0.260939 V5.l8 1.554e-01 8.503e-02 1.828 0.067845 . V6.l8 -1.149e-01 1.448e-01 -0.793 0.427764 V7.l8 -1.020e-08 8.201e-08 -0.124 0.901057 V3.l9 8.082e-02 1.382e-01 0.585 0.558850 V4.l9 -3.180e-02 8.348e-02 -0.381 0.703340 V5.l9 1.428e-01 8.176e-02 1.746 0.081121 . V6.l9 -1.633e-01 1.425e-01 -1.146 0.252178 V7.l9 1.050e-07 8.187e-08 1.283 0.199835 V3.l10 -3.240e-02 7.363e-02 -0.440 0.659979 V4.l10 1.352e-01 7.991e-02 1.692 0.090905 . V5.l10 -1.629e-01 8.176e-02 -1.992 0.046680 * V6.l10 -2.998e-02 1.405e-01 -0.213 0.831091 V7.l10 -1.476e-07 6.470e-08 -2.281 0.022752 * const 7.157e-03 4.093e-03 1.749 0.080659 . trend -6.186e-07 5.695e-07 -1.086 0.277605 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.003339 on 934 degrees of freedom Multiple R-Squared: 0.9965, Adjusted R-squared: 0.9963 F-statistic: 5262 on 51 and 934 DF, p-value: < 2.2e-16 Estimation results for equation V6: =================================== V6 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 -1.664e-01 1.502e-01 -1.108 0.268104 V4.l1 1.311e-01 9.195e-02 1.426 0.154319 V5.l1 7.151e-01 7.202e-02 9.930 < 2e-16 *** V6.l1 4.937e-01 8.050e-02 6.133 1.27e-09 *** V7.l1 9.048e-08 7.034e-08 1.286 0.198646 V3.l2 3.502e-01 1.496e-01 2.341 0.019432 * V4.l2 1.079e-02 9.171e-02 0.118 0.906387 V5.l2 2.867e-01 7.767e-02 3.692 0.000236 *** V6.l2 -5.274e-01 1.531e-01 -3.444 0.000598 *** V7.l2 8.716e-08 8.521e-08 1.023 0.306654 V3.l3 -3.517e-02 1.489e-01 -0.236 0.813393 V4.l3 1.870e-01 9.112e-02 2.052 0.040418 * V5.l3 3.191e-01 8.450e-02 3.777 0.000169 *** V6.l3 -5.392e-01 1.535e-01 -3.513 0.000464 *** V7.l3 -5.424e-08 8.594e-08 -0.631 0.528081 V3.l4 -1.454e-01 1.489e-01 -0.976 0.329251 V4.l4 9.038e-02 9.000e-02 1.004 0.315518 V5.l4 -1.731e-01 8.819e-02 -1.963 0.049924 * V6.l4 -1.513e-01 1.548e-01 -0.978 0.328466 V7.l4 -6.207e-08 8.623e-08 -0.720 0.471804 V3.l5 1.807e-01 1.490e-01 1.213 0.225387 V4.l5 -7.290e-02 9.014e-02 -0.809 0.418866 V5.l5 9.812e-02 8.738e-02 1.123 0.261794 V6.l5 -8.835e-02 1.531e-01 -0.577 0.563937 V7.l5 7.315e-08 8.599e-08 0.851 0.395222 V3.l6 2.917e-01 1.461e-01 1.997 0.046135 * V4.l6 -1.359e-01 8.994e-02 -1.511 0.131117 V5.l6 -6.472e-01 8.767e-02 -7.381 3.46e-13 *** V6.l6 4.206e-01 1.533e-01 2.743 0.006205 ** V7.l6 -1.393e-07 8.619e-08 -1.616 0.106522 V3.l7 8.693e-02 1.461e-01 0.595 0.551876 V4.l7 -7.605e-02 8.978e-02 -0.847 0.397205 V5.l7 3.043e-02 9.007e-02 0.338 0.735574 V6.l7 5.473e-02 1.507e-01 0.363 0.716623 V7.l7 1.021e-07 8.628e-08 1.183 0.237213 V3.l8 -6.197e-02 1.445e-01 -0.429 0.668172 V4.l8 1.019e-01 8.742e-02 1.166 0.243888 V5.l8 1.065e-01 8.849e-02 1.204 0.229040 V6.l8 -2.322e-02 1.507e-01 -0.154 0.877548 V7.l8 -7.937e-08 8.535e-08 -0.930 0.352668 V3.l9 -1.162e-01 1.438e-01 -0.808 0.419217 V4.l9 -3.072e-02 8.688e-02 -0.354 0.723682 V5.l9 9.215e-02 8.509e-02 1.083 0.279092 V6.l9 -1.240e-01 1.483e-01 -0.836 0.403351 V7.l9 1.039e-07 8.521e-08 1.219 0.223036 V3.l10 -8.067e-02 7.662e-02 -1.053 0.292681 V4.l10 1.467e-01 8.316e-02 1.764 0.077983 . V5.l10 -4.688e-02 8.509e-02 -0.551 0.581789 V6.l10 4.479e-02 1.462e-01 0.306 0.759392 V7.l10 -8.263e-08 6.734e-08 -1.227 0.220112 const 4.242e-03 4.259e-03 0.996 0.319568 trend -6.731e-07 5.926e-07 -1.136 0.256290 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 0.003475 on 934 degrees of freedom Multiple R-Squared: 0.9963, Adjusted R-squared: 0.9961 F-statistic: 4913 on 51 and 934 DF, p-value: < 2.2e-16 Estimation results for equation V7: =================================== V7 = V3.l1 + V4.l1 + V5.l1 + V6.l1 + V7.l1 + V3.l2 + V4.l2 + V5.l2 + V6.l2 + V7.l2 + V3.l3 + V4.l3 + V5.l3 + V6.l3 + V7.l3 + V3.l4 + V4.l4 + V5.l4 + V6.l4 + V7.l4 + V3.l5 + V4.l5 + V5.l5 + V6.l5 + V7.l5 + V3.l6 + V4.l6 + V5.l6 + V6.l6 + V7.l6 + V3.l7 + V4.l7 + V5.l7 + V6.l7 + V7.l7 + V3.l8 + V4.l8 + V5.l8 + V6.l8 + V7.l8 + V3.l9 + V4.l9 + V5.l9 + V6.l9 + V7.l9 + V3.l10 + V4.l10 + V5.l10 + V6.l10 + V7.l10 + const + trend Estimate Std. Error t value Pr(>|t|) V3.l1 -5.531e+04 8.556e+04 -0.646 0.518196 V4.l1 -5.090e+04 5.238e+04 -0.972 0.331471 V5.l1 -1.131e+05 4.103e+04 -2.756 0.005966 ** V6.l1 1.151e+05 4.586e+04 2.509 0.012292 * V7.l1 7.999e-01 4.008e-02 19.960 < 2e-16 *** V3.l2 2.077e+04 8.521e+04 0.244 0.807530 V4.l2 8.262e+03 5.225e+04 0.158 0.874384 V5.l2 -9.809e+04 4.425e+04 -2.217 0.026884 * V6.l2 1.681e+05 8.724e+04 1.927 0.054260 . V7.l2 -1.127e-01 4.855e-02 -2.321 0.020506 * V3.l3 1.033e+04 8.486e+04 0.122 0.903155 V4.l3 1.402e+04 5.192e+04 0.270 0.787214 V5.l3 1.750e+03 4.814e+04 0.036 0.971002 V6.l3 -8.681e+03 8.744e+04 -0.099 0.920936 V7.l3 -4.844e-02 4.896e-02 -0.989 0.322751 V3.l4 -4.279e+04 8.486e+04 -0.504 0.614253 V4.l4 2.059e+04 5.128e+04 0.401 0.688162 V5.l4 5.766e+04 5.024e+04 1.148 0.251399 V6.l4 -5.507e+04 8.817e+04 -0.625 0.532395 V7.l4 2.976e-02 4.913e-02 0.606 0.544809 V3.l5 3.251e+04 8.487e+04 0.383 0.701748 V4.l5 -1.811e+04 5.135e+04 -0.353 0.724441 V5.l5 -5.346e+03 4.978e+04 -0.107 0.914515 V6.l5 3.998e+04 8.720e+04 0.458 0.646707 V7.l5 7.042e-02 4.899e-02 1.437 0.150985 V3.l6 -3.894e+04 8.324e+04 -0.468 0.640003 V4.l6 1.325e+04 5.124e+04 0.259 0.796071 V5.l6 -7.529e+04 4.995e+04 -1.507 0.132087 V6.l6 3.656e+04 8.736e+04 0.419 0.675666 V7.l6 -5.088e-02 4.910e-02 -1.036 0.300422 V3.l7 -1.150e+04 8.322e+04 -0.138 0.890149 V4.l7 -5.589e+04 5.115e+04 -1.093 0.274857 V5.l7 1.979e+04 5.132e+04 0.386 0.699820 V6.l7 4.654e+04 8.588e+04 0.542 0.587981 V7.l7 9.150e-02 4.916e-02 1.861 0.062996 . V3.l8 9.021e+04 8.234e+04 1.096 0.273526 V4.l8 -1.029e+04 4.981e+04 -0.207 0.836308 V5.l8 -1.031e+05 5.042e+04 -2.046 0.041057 * V6.l8 9.273e+04 8.585e+04 1.080 0.280344 V7.l8 -1.021e-01 4.863e-02 -2.100 0.036031 * V3.l9 -1.244e+05 8.195e+04 -1.518 0.129358 V4.l9 3.253e+04 4.950e+04 0.657 0.511279 V5.l9 -5.139e+04 4.848e+04 -1.060 0.289372 V6.l9 -9.584e+03 8.448e+04 -0.113 0.909697 V7.l9 -9.003e-02 4.854e-02 -1.855 0.063967 . V3.l10 -5.409e+04 4.366e+04 -1.239 0.215692 V4.l10 1.000e+05 4.738e+04 2.111 0.035063 * V5.l10 1.467e+04 4.848e+04 0.303 0.762223 V6.l10 4.759e+04 8.330e+04 0.571 0.567971 V7.l10 -1.390e-01 3.836e-02 -3.622 0.000308 *** const 1.546e+03 2.427e+03 0.637 0.524256 trend -9.065e-01 3.376e-01 -2.685 0.007385 ** --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 1980 on 934 degrees of freedom Multiple R-Squared: 0.7347, Adjusted R-squared: 0.7202 F-statistic: 50.71 on 51 and 934 DF, p-value: < 2.2e-16 Covariance matrix of residuals: V3 V4 V5 V6 V7 V3 7.440e-07 7.393e-07 7.004e-07 6.898e-07 8.406e-03 V4 7.393e-07 4.485e-06 2.198e-06 4.633e-06 1.386e+00 V5 7.004e-07 2.198e-06 1.115e-05 9.487e-06 -2.354e+00 V6 6.898e-07 4.633e-06 9.487e-06 1.208e-05 -6.411e-01 V7 8.406e-03 1.386e+00 -2.354e+00 -6.411e-01 3.920e+06 Correlation matrix of residuals: V3 V4 V5 V6 V7 V3 1.000000 0.4047 0.2432 0.23013 0.004922 V4 0.404717 1.0000 0.3108 0.62945 0.330465 V5 0.243168 0.3108 1.0000 0.81756 -0.356087 V6 0.230130 0.6294 0.8176 1.00000 -0.093179 V7 0.004922 0.3305 -0.3561 -0.09318 1.000000 > > fitvar1 = VAR(lr, p=10, type="both") > > # auto correlation function for residuals for close price > > acf(residuals(fitvar1)[,4]) > > # auto correlation function for residuals for open, high and low price > > acf(residuals(fitvar1)[,1]) > > acf(residuals(fitvar1)[,2]) > > acf(residuals(fitvar1)[,3]) > > #vector auto regression of Open, High, Low and Close > > #predict the next 10 values > > var.pred <- predict(fitvar1, n.ahead=3, ci=0.95) > > var.pred $V3 fcst lower upper CI [1,] 1.308923 1.307233 1.310614 0.001690542 [2,] 1.309417 1.302398 1.316437 0.007019716 [3,] 1.309859 1.299338 1.320380 0.010520951 $V4 fcst lower upper CI [1,] 1.309923 1.305772 1.314074 0.004150918 [2,] 1.310630 1.302681 1.318578 0.007948780 [3,] 1.310900 1.299789 1.322010 0.011110588 $V5 fcst lower upper CI [1,] 1.308566 1.302021 1.315111 0.006544701 [2,] 1.308913 1.298814 1.319012 0.010098650 [3,] 1.308848 1.295342 1.322355 0.013506637 $V6 fcst lower upper CI [1,] 1.309337 1.302526 1.316148 0.006810988 [2,] 1.309749 1.299393 1.320106 0.010356437 [3,] 1.309778 1.296913 1.322642 0.012864717 $V7 fcst lower upper CI [1,] 2955.973 -924.4980 6836.443 3880.471 [2,] 3783.907 -1276.8589 8844.672 5060.766 [3,] 4644.148 -976.3765 10264.673 5620.525 > > quotes[(x-4):x,] V1 V2 V3 V4 V5 V6 V7 5567 2016.08.08 06:00 1.30875 1.30897 1.30858 1.30871 976 5568 2016.08.08 07:00 1.30870 1.30881 1.30815 1.30852 1145 5569 2016.08.08 08:00 1.30851 1.30867 1.30805 1.30815 1279 5570 2016.08.08 09:00 1.30814 1.30862 1.30457 1.30512 3715 5571 2016.08.08 10:00 1.30516 1.30630 1.30456 1.30576 1748
Now once again we got very good predictions for H1 candle which are pretty close the actual values. I am omitting the M30 timeframe but the calculations will show you that this model can make good predictions on M30 as well.
Conclusion
The predictions on Weekly and Daily timeframe were not good. The most probable reason is the high non linear price patterns on weekly and daily timeframes. The predictions on H4, H1, M30, M15, M5 and M1 are pretty good. The most probable reason is that the non linear structure is not that strong on these timeframes. The time taken by R for these calculations is less 15 seconds. So you can use these predictions in day trading. You can also use these predictions in trading binary options.