Merge pull request #28 from vzhomeexperiments/dev_0.4.3

Dev 0.4.3

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: lazytrade
 Type: Package
 Title: Learn Computer and Data Science using Algorithmic Trading
-Version: 0.4.2
+Version: 0.4.3
 Author: Vladimir Zhbanko
 Maintainer: Vladimir Zhbanko <[email protected]>
 Description: Provide sets of functions and methods to learn and practice data science using idea of algorithmic trading.

NEWS.md

@@ -3,6 +3,18 @@
 ## Planned Changes
 
 * Add folder generation function
+* Put option predict ahead as a parameter (now 34 bars)
+* aml_make_model use parameter to split data with specific n of rows e.g. n_rows
+
+# lazytrade 0.4.3
+
+# Version 0.4.3
+
+## Changes
+
+* solving warnings by changing example in *mt_make_model* function
+* change function *mt_make_model* to read manually checked dataset
+* update documentation
 
 # lazytrade 0.4.2
 

R/aml_make_model.R

@@ -105,8 +105,8 @@ aml_make_model <- function(symbol, timeframe, path_model, path_data,
     model_status <- readr::read_csv(dec_file_path) %>% select(MaxPerf) %$% MaxPerf
   } else if(force_update == TRUE) {
     # delete the model and previous test results
-    remove(dec_file_path)
-    remove(m_path)
+    #try(remove(dec_file_path), silent = TRUE)
+    #try(remove(m_path),silent = TRUE)
     model_status <- -1
   } else { model_status <- 0 }
 
@@ -129,7 +129,7 @@ aml_make_model <- function(symbol, timeframe, path_model, path_data,
 
 
   # split data to train and test blocks
-  # note: model will be tested on the PAST data and trained on the NEWEST data
+  # note: model will be trained on the OLDEST data
   test_ind  <- 1:round(0.3*(nrow(dat12))) #train indices 1:xxx
   dat21 <- dat12[test_ind, ]    #dataset to test the model using 30% of data
   dat22 <- dat12[-test_ind, ]   #dataset to train the model

R/aml_test_model.R

@@ -16,14 +16,14 @@
 #' @author (C) 2020 Vladimir Zhbanko
 #'
 #' @param symbol              Character symbol of the asset for which to train the model
-#' @param num_bars            Number of bars used to test the model
-#' @param timeframe           Data timeframe e.g. 60 min
-#' @param path_model          Path where the models are be stored
-#' @param path_data           Path where the aggregated historical data is stored, if exists in rds format
-#' @param path_sbxm           Path to the sandbox where file with strategy test results should be written (master terminal)
-#' @param path_sbxs           Path to the sandbox where file with strategy test results should be written (slave terminal)
+#' @param num_bars            Integer, Number of (rows) bars used to test the model
+#' @param timeframe           Integer, Data timeframe e.g. 60 min. This will be equal to 1 bar
+#' @param path_model          String, User path where the models are be stored
+#' @param path_data           String, User path where the aggregated historical data is stored, if exists in rds format
+#' @param path_sbxm           String, User path to the sandbox where file with strategy test results should be written (master terminal)
+#' @param path_sbxs           String, User path to the sandbox where file with strategy test results should be written (slave terminal)
 #'
-#' @return Function is writing file into Decision Support System folder
+#' @return Function is writing file into Decision Support System folders
 #' @export
 #'
 #' @examples
@@ -116,10 +116,6 @@ aml_test_model <- function(symbol, num_bars, timeframe, path_model, path_data,
   # file name with the tick data
   path_tick <- file.path(path_data, "TickSize_AI_RSIADX.csv")
 
-  ## !!!!!! TDL
-  ## only select the latest 30% of data... or only last 2 month
-
-
   #dataset with date column X1
   y <- readr::read_rds(full_path) %>%
     # remove empty rows
@@ -139,6 +135,8 @@ aml_test_model <- function(symbol, num_bars, timeframe, path_model, path_data,
     #value z will contain tick value for this symbol
     X2
 
+  #TDL add fail safe mechanism for 'z'
+
   # generate a file name for model
   m_name <- paste0("DL_Regression", "-", symbol,"-", timeframe)
   m_path <- file.path(path_model, m_name)
@@ -158,9 +156,6 @@ aml_test_model <- function(symbol, num_bars, timeframe, path_model, path_data,
     #NB <- 3
 
 
-
-
-
   # To DO: Add ATR multiplier to simulate TP/SL
   # add for loop for ATR Multiplier
 

R/mt_evaluate.R

@@ -35,15 +35,17 @@
 #' data(macd_ML60M)
 #'
 #' # start h2o engine (using all CPU's by default)
-#' h2o.init()
+#' h2o.init(nthreads = 2)
 #'
 #' # performing Deep Learning Regression using the custom function
 #' # this function stores model to the temp location
 #' mt_make_model(indicator_dataset = macd_ML60M,
 #'               num_bars = 64,
 #'               timeframe = 60,
 #'               path_model = path_model,
-#'               path_data = path_data)
+#'               path_data = path_data,
+#'               activate_balance = TRUE,
+#'               num_nn_options = 3)
 #'
 #'
 #' # Use sample data
@@ -77,30 +79,6 @@ mt_evaluate <- function(x, path_model, num_cols, timeframe){
   # load models
   m1 <- h2o::h2o.loadModel(m_path)
 
-
-  if(class(m1)[1] == "H2ORegressionModel") {
-  # Convert to matrix
-  X_m <- lazytrade::to_m(x, num_cols) %>% as.data.frame()
-  colnames(X_m) <- c(paste("X",1:num_cols,sep=""))
-  # load the dataset to h2o
-  test  <- h2o::as.h2o(x = X_m, destination_frame = "test")
-
-  # retrieve the predicted market type value
-  e1 <- h2o::h2o.predict(m1, test)
-
-  # round the number to achieve class
-  result <- round(e1) %>% as.vector()
-
-  # manage negatives and/or bizzare numbers
-  if(result <= 0 || result > 6) {element <- -1} else {element <- result}
-
-  # output result of prediction from the function
-  return(element)
-
-  }
-
-  if(class(m1)[1] == "H2OMultinomialModel") {
-
     # Convert to matrix
     X_m <- lazytrade::to_m(x, num_cols) %>% as.data.frame()
     colnames(X_m) <- c(paste("X",1:num_cols,sep=""))
@@ -117,6 +95,6 @@ mt_evaluate <- function(x, path_model, num_cols, timeframe){
     # output result of prediction from the function
     return(e1)
 
-  }
+
 
 }

R/mt_make_model.R

@@ -25,8 +25,10 @@
 #' @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
+#' @param num_nn_options      Integer, value from 3 to 24 or more. Used to change number of variants
+#'                            of the random neural network structures. Value 3 will mean that only one
+#'                            random structure will be used. To avoid warnings make sure to set this value
+#'                            multiple of 3. Higher values will increase computation time.
 #'
 #' @return Function is writing file object with the model
 #' @export
@@ -59,7 +61,7 @@
 #'               path_model = path_model,
 #'               path_data = path_data,
 #'               activate_balance = TRUE,
-#'               num_nn_options = 2)
+#'               num_nn_options = 3)
 #'
 #' # stop h2o engine
 #' h2o.shutdown(prompt = FALSE)
@@ -86,18 +88,18 @@ mt_make_model <- function(indicator_dataset,
 
   ## check if the latest data is available
   # construct path to the new data
-  path_file_name <- paste0("macd_ai_classified_", timeframe, "M.rds")
+  path_file_name <- paste0("macd_checked_", timeframe, "M.rds")
   path_newdata <- file.path(path_data, path_file_name)
   if(file.exists(path_newdata)){
     # use new data...
-    macd_ML2 <- read_rds(path_newdata) %>%
+    macd_ML2 <- readr::read_rds(path_newdata) %>%
       # and add new data
-      bind_rows(indicator_dataset) %>%
+      dplyr::bind_rows(indicator_dataset) %>%
       # convert one column to factor
-      mutate_at("M_T", as.factor)
+      dplyr::mutate_at("M_T", as.factor)
   } else {
     # use input data and transform dataset column to factor
-    macd_ML2 <- indicator_dataset %>% mutate_at("M_T", as.factor)
+    macd_ML2 <- indicator_dataset %>% dplyr::mutate_at("M_T", as.factor)
   }
 
   # check if we don't have too much data

cran-comments.md

@@ -2,7 +2,7 @@
 * windows, R 4.0.3
 * R-hub builder
 * R version 4.0.3 (2020-10-10)
-* R Under development (unstable) (2020-10-11 r79327)
+* R Under development (unstable) (2020-12-13 r79623)
 
 ## R CMD check results
 Status: OK