Logistic regression eval

Author: AlexanderJernstrom

examples/data/titanic_train.csv

@@ -0,0 +1,68 @@
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+#include <math.h>
+#include "logistic_regression_eval.h"
+#include "../regression/logistic_regression.h"
+
+TitanicSurvivor *loadTrainDataFromCsv()
+{
+    int n_training = 890;
+    FILE *stream = fopen("examples/data/titanic_train.csv", "r");
+    if (stream == NULL)
+    {
+        perror("Unable to open the file.");
+        exit(1);
+    }
+    char line[1024];
+    TitanicSurvivor *train_set = malloc(sizeof(TitanicSurvivor) * n_training);
+    int line_number = 0;
+
+    while (fgets(line, 1024, stream))
+    {
+        if (line_number > 0)
+        {
+            char *token = strtok(line, ",");
+            int column = 1; // Start from the first column
+            float firstColumn, thirdColumn;
+
+            while (token != NULL)
+            {
+                if (column == 2)
+                {
+                    firstColumn = atof(token);
+                }
+                else if (column == 3)
+                {
+                    thirdColumn = atof(token);
+                    break; // No need to tokenize further
+                }
+                token = strtok(NULL, ",");
+                column++;
+            }
+            TitanicSurvivor survivor = {.survived = firstColumn, .pClass = thirdColumn};
+            train_set[line_number] = survivor;
+        }
+        line_number++;
+    }
+    return train_set;
+}
+
+void calculateLogisticRegression()
+{
+    TitanicSurvivor *train_data = loadTrainDataFromCsv();
+    int n_training = 890;
+
+    double x[n_training];
+    double y[n_training];
+
+    for (int i = 0; i < n_training; i++)
+    {
+        x[i] = train_data[i].pClass;
+        y[i] = train_data[i].survived;
+    }
+
+    double *parameters = logisticRegression(x, y, n_training);
+
+    printf("\n %f * x + %f", parameters[0], parameters[1]);
+}

examples/logistic_regression_eval.c

@@ -0,0 +1,9 @@
+#ifndef LOGISTIC_REGRESSION_EVAL
+#define LOGISTIC_REGRESSION_EVAL
+
+typedef struct {
+    int pClass;
+    int survived;
+} TitanicSurvivor;
+void calculateLogisticRegression();
+#endif

examples/logistic_regression_eval.h

@@ -10,6 +10,7 @@
 #include <stdlib.h> // pulls in declaration of malloc, free
 #include "nn/tests/nn_test.h"
 #include "examples/linear_regression_eval.h"
+#include "examples/logistic_regression_eval.h"
 // 1 2 3
 // 4 5 6
 
@@ -148,7 +149,8 @@ int main()
     // layers: {denseLayer, reluLayer, denseLayer, sigmoidLayer}
 
     // test_nn();
-    calculateRegression();
+    // calculateRegression();
+    calculateLogisticRegression();
     return 0;
 }
 

main.c

@@ -16,7 +16,7 @@ double *logisticRegression(double xValues[], double yValues[], int n)
     double dA = 0;
     double dB = 0;
     // Can be increased here
-    while (iter < 1000)
+    while (iter < 5000)
     {
         double *predicted = malloc(n * sizeof(double *));
         for (int i = 0; i < n; i++)