Merge pull request #24 from vzhomeexperiments/dev_0.4.1

optimized code

R/aml_make_model.R

@@ -21,6 +21,8 @@
 #' @param path_data           Path where the aggregated historical data is stored, if exists in rds format
 #' @param force_update        Boolean, by setting this to TRUE function will generate new model
 #'                            (useful after h2o engine update)
+#' @param num_nn_options      Integer, value from 1 to 20 or more. Used to change number of variants
+#'                            of the random neural network structures
 #'
 #' @return Function is writing a file object with the best Deep Learning Regression model
 #' @export
@@ -62,7 +64,8 @@
 #'                timeframe = 60,
 #'                path_model = path_model,
 #'                path_data = path_data,
-#'                force_update=FALSE)
+#'                force_update=FALSE,
+#'                num_nn_options = 2)
 #'
 #' # stop h2o engine
 #' h2o.shutdown(prompt = FALSE)
@@ -75,7 +78,8 @@
 #'
 #'
 aml_make_model <- function(symbol, timeframe, path_model, path_data,
-                           force_update=FALSE){
+                           force_update=FALSE,
+                           num_nn_options = 24){
 
   requireNamespace("dplyr", quietly = TRUE)
   requireNamespace("readr", quietly = TRUE)
@@ -129,7 +133,7 @@ aml_make_model <- function(symbol, timeframe, path_model, path_data,
   #h2o.init()
 
   ### random network structure
-  nn_sets <- sample.int(n = 100, 24) %>% matrix(ncol = 3)
+  nn_sets <- sample.int(n = 100, num_nn_options) %>% matrix(ncol = 3)
 
   ###
 

R/aml_score_data.R

@@ -55,7 +55,9 @@
 #' aml_make_model(symbol = 'USDJPY',
 #'                timeframe = 60,
 #'                path_model = path_model,
-#'                path_data = path_data)
+#'                path_data = path_data,
+#'                force_update=FALSE,
+#'                num_nn_options = 2)
 #'
 #'
 #' path_sbxm <- normalizePath(tempdir(),winslash = "/")

R/aml_test_model.R

@@ -63,7 +63,8 @@
 #'                timeframe = 60,
 #'                path_model = path_model,
 #'                path_data = path_data,
-#'                force_update=FALSE)
+#'                force_update=FALSE,
+#'                num_nn_options = 2)
 #'
 #' path_sbxm <- normalizePath(tempdir(),winslash = "/")
 #' path_sbxs <- normalizePath(tempdir(),winslash = "/")

R/mt_make_model.R

@@ -25,6 +25,8 @@
 #' @param path_model          String, Path where the models are be stored
 #' @param path_data           String, Path where the aggregated historical data is stored, if exists in rds format
 #' @param activate_balance    Boolean, option to choose if to balance market type classes or not, default TRUE
+#' @param num_nn_options      Integer, value from 1 to 20 or more. Used to change number of variants
+#'                            of the random neural network structures
 #'
 #' @return Function is writing file object with the model
 #' @export
@@ -56,7 +58,8 @@
 #'               timeframe = 60,
 #'               path_model = path_model,
 #'               path_data = path_data,
-#'               activate_balance = TRUE)
+#'               activate_balance = TRUE,
+#'               num_nn_options = 2)
 #'
 #' # stop h2o engine
 #' h2o.shutdown(prompt = FALSE)
@@ -72,7 +75,8 @@ mt_make_model <- function(indicator_dataset,
                           num_bars,
                           timeframe = 60,
                           path_model, path_data,
-                          activate_balance = TRUE){
+                          activate_balance = TRUE,
+                          num_nn_options = 24){
 
   requireNamespace("dplyr", quietly = TRUE)
   requireNamespace("readr", quietly = TRUE)
@@ -111,7 +115,7 @@ mt_make_model <- function(indicator_dataset,
 
   # try different models and choose the best one...
   ### random network structure
-  nn_sets <- sample.int(n = 100, 24) %>% matrix(ncol = 3)
+  nn_sets <- sample.int(n = 100, num_nn_options) %>% matrix(ncol = 3)
 
   for (i in 1:dim(nn_sets)[1]) {
 

cran-comments.md

@@ -12,5 +12,8 @@ There were no ERRORs or WARNINGs or NOTEs
 
 As of email from 07.10.2020 package was re-checked again
 
+- removed tidyr from imports
+- optimized code to allow faster and less CPU intensive calculations
+
 ## Downstream dependencies
 R version 3.6.3 (2020-02-29)