From bf1adece8aeb48e136085233d2f5ff2f9600eaf5 Mon Sep 17 00:00:00 2001
From: Mike Vink <mike1994vink@gmail.com>
Date: Sun, 2 May 2021 17:33:26 +0200
Subject: update

---
 scripts/modelling.R | 39 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 39 insertions(+)
 create mode 100644 scripts/modelling.R

(limited to 'scripts/modelling.R')

diff --git a/scripts/modelling.R b/scripts/modelling.R
new file mode 100644
index 0000000..b1fc746
--- /dev/null
+++ b/scripts/modelling.R
@@ -0,0 +1,39 @@
+library(caret)
+library(tidyverse)
+library(MLeval)
+
+source("./data_prep.R")
+
+data_list <- sets_partitions
+results <- list()
+models <- c("rrlda", "naive_bayes", "rf", "regLogistic")
+fitControl <- trainControl( ## 10-fold CV
+    method = "repeatedcv",
+    number = 10,
+    classProbs = TRUE,
+    savePredictions = TRUE,
+    repeats = 2
+)
+for (model in models) {
+    dataset = 1
+    # for (data in data_list) {
+    for (data in data_list[c(14, 16, 19)]) {
+        print(paste("Training", model, "on dataset", dataset))
+        train <- data[["train"]]
+        X_train <- as.data.frame(train[-c(1, 2)])
+        Y_train <- train[c(2)][[1]]
+        levels(Y_train) <- c("Low", "High")
+        set.seed(13121994)
+        model_trained <- train(
+            X_train,
+            y = Y_train,
+            method = model,
+            trControl = fitControl
+        )
+        results[[model]][[dataset]] <- model_trained
+        dataset = dataset + 1
+    }
+}
+save(results, file="./modelling_results_withrrlda.RData")
+# save(results, file="./modelling_results.RData")
+
-- 
cgit v1.2.3