ThePandas_solution.R

solution.R

@@ -1,14 +1,60 @@
 # Load necessary libraries
+rm(list=ls())
 library(dplyr)
 library(lubridate)
 library(ggplot2)
-library(randomForest)
 library(tidyr)
-library(pacman)
-# pacman::p_load(randomForest)
+library(h2o)
+
+
+# Initialize the H2O cluster
+h2o.init(nthreads = -1)
+
+# Define parameters
+start_train <- as.Date("2017-01-01")
+end_train <- as.Date("2023-11-30")
+start_test <- as.Date("2024-01-01")
+end_test <- as.Date("2024-06-30")
+
+n_buys <- 10
+verbose <- FALSE
+
+
+
+# Load sector data
+df_sectors <- read.csv('data/data0.csv')
+
+# Load price and financial data
+df_data <- read.csv('data/data1.csv')
+df_data$date <- as.Date(df_data$date)
+
+df_x <- df_data %>% select(date, security, price, return30, ratio_pe, ratio_pcf, ratio_de, ratio_roe, ratio_roa)
+df_y <- df_data %>% select(date, security, label)
+
+list_vars1 <- c('price', 'return30', 'ratio_pe', 'ratio_pcf', 'ratio_de', 'ratio_roe', 'ratio_roa')
+
+# Create signals DataFrame
+df_signals <- data.frame(date = unique(df_x$date[df_x$date >= start_test & df_x$date <= end_test]))
+df_signals <- df_signals %>% arrange(date)
+
+# Initialize an empty list for storing accuracy results
+df_signals$acc_total <- NA
+df_signals$acc_current <- NA
+
+
+
+# Load necessary libraries
+library(dplyr)
+library(lubridate)
+library(ggplot2)
+library(tidyr)
+library(h2o)
 
 print('---> R Script Start')
 
+# Initialize the H2O cluster
+h2o.init(nthreads = -1)
+
 # Define parameters
 start_train <- as.Date("2017-01-01")
 end_train <- as.Date("2023-11-30")
@@ -55,7 +101,7 @@ for (i in seq_len(nrow(df_signals))) {
   df_testx <- df_x %>% filter(date >= df_signals$date[i])
   df_testy <- df_y %>% filter(date >= df_signals$date[i])
 
-  # Scale data
+  # Scale data (optional, h2o handles scaling internally)
   for (col in list_vars1) {
     scaler <- function(x) (x - min(x)) / (max(x) - min(x)) * 2 - 1
     df_trainx[[col]] <- scaler(df_trainx[[col]])
@@ -64,15 +110,33 @@ for (i in seq_len(nrow(df_signals))) {
 
   df_trainy <- df_trainy %>% mutate(label = as.factor(label)) 
 
-  # Fit the Random Forest classifier
-  if (i == 1) {
-    clf <- randomForest(x = df_trainx[list_vars1], y = df_trainy$label, ntree = 10, mtry = sqrt(length(list_vars1)), nodesize = 1000)
+  # Convert to h2o frames
+  train_h2o <- as.h2o(cbind(df_trainx[list_vars1], label = df_trainy$label))
+  test_h2o <- as.h2o(df_testx[list_vars1])
+
+  # Build and train the H2O deep learning model trains every 10 days 
+  if (i %in% seq(1, nrow(df_signals), 10)) {
+    clf <- h2o.deeplearning(
+      x = list_vars1,
+      y = "label",
+      training_frame = train_h2o,
+      activation = "RectifierWithDropout",
+      hidden = c(64, 32),
+      epochs = 20,
+      input_dropout_ratio = 0.2,
+      hidden_dropout_ratios = c(0.5, 0.5),
+      balance_classes = TRUE,
+      stopping_metric = "misclassification",
+      stopping_rounds = 5,
+      verbose = verbose
+    )
   }
 
   # Predictions and accuracy
-  pred_probs <- predict(clf, newdata = df_testx[list_vars1], type = "prob")
-  df_testy$signal <- pred_probs[, 2]
-  df_testy$pred <- ifelse(pred_probs[, 2] > 0.5, 1, 0)
+  pred_probs <- h2o.predict(clf, newdata = test_h2o)
+  pred_probs <- as.data.frame(pred_probs)
+  df_testy$signal <- pred_probs$p1
+  df_testy$pred <- ifelse(pred_probs$p1 > 0.5, 1, 0)
   df_testy$count <- 1
 
   df_current <- df_testy %>% filter(date == df_signals$date[i])
@@ -89,6 +153,19 @@ for (i in seq_len(nrow(df_signals))) {
   df_signals[i, df_current$security] <- df_current$signal
 }
 
+# Continue with your signal plotting, buy matrix, and payoff calculations
+
+# Shutdown H2O cluster after completion
+h2o.shutdown(prompt = FALSE)
+
+print('---> R Script End')
+
+# Continue with your signal plotting, buy matrix, and payoff calculations
+
+# Shutdown H2O cluster after completion
+h2o.shutdown(prompt = FALSE)
+
+
 # Create buy matrix for payoff plot
 df_signals$`10th` <- apply(df_signals[df_sectors$security], 1, function(x) sort(x, decreasing = TRUE)[n_buys])
 df_index <- df_signals[df_sectors$security] >= df_signals$`10th`
@@ -137,7 +214,7 @@ plot_payoff <- function(df_buys) {
 
   # df_payoff <- df[, 1, drop = FALSE]
   df <- df * (1 / n_buys)  # equally weighted
-
+  
   df_payoff <- NULL
   arr_ret <- as.matrix(df_returns[-1] + 1)
   df_payoff$payoff <- diag(df %*% t(arr_ret-1))
@@ -157,4 +234,3 @@ df_payoff <- plot_payoff(df_buys)
 ggplot(df_payoff, aes(x = date, y = tri)) + geom_line() + ggtitle("Payoff Over Time")
 
 
-print('---> R Script End')