From 13b0d76c5ca4f35968e33a0828ffbcc0c27127ae Mon Sep 17 00:00:00 2001
From: novak_99 <cooljohn312@gmail.com>
Date: Fri, 28 May 2021 19:37:34 -0700
Subject: [PATCH] "Vectorized" implementation of SGD for ProbitReg

---
 MLPP/ProbitReg/ProbitReg.cpp | 20 ++++++--------------
 main.cpp                     | 18 +++++++++---------
 2 files changed, 15 insertions(+), 23 deletions(-)

diff --git a/MLPP/ProbitReg/ProbitReg.cpp b/MLPP/ProbitReg/ProbitReg.cpp
index 20dfd5b..49817e7 100644
--- a/MLPP/ProbitReg/ProbitReg.cpp
+++ b/MLPP/ProbitReg/ProbitReg.cpp
@@ -94,6 +94,7 @@ namespace MLPP{
     }
 
     void ProbitReg::SGD(double learning_rate, int max_epoch, bool UI){
+        // NOTE: ∂y_hat/∂z is sparse
         LinAlg alg;
         Activation avn;
         Reg regularization;
@@ -111,24 +112,15 @@ namespace MLPP{
             double z = propagate(inputSet[outputIndex]);
             cost_prev = Cost({y_hat}, {outputSet[outputIndex]});
 
+            double error = y_hat - outputSet[outputIndex];
 
-            for(int i = 0; i < k; i++){
-                    
-                // Calculating the weight gradients
-                
-                double w_gradient = (y_hat - outputSet[outputIndex]) * ((1 / sqrt(2 * M_PI)) * exp(-z * z / 2)) * inputSet[outputIndex][i];
-
-                std::cout << exp(-z * z / 2) << std::endl;
-                // Weight updation
-                weights[i] -= learning_rate * w_gradient;
-            }
+            // Weight Updation
+            weights = alg.subtraction(weights, alg.scalarMultiply(learning_rate * error * ((1 / sqrt(2 * M_PI)) * exp(-z * z / 2)), inputSet[outputIndex]));
             weights = regularization.regWeights(weights, lambda, alpha, reg);
             
-            // Calculating the bias gradients
-            double b_gradient = (y_hat - outputSet[outputIndex]);
-            
             // Bias updation
-            bias -= learning_rate * b_gradient * ((1 / sqrt(2 * M_PI)) * exp(-z * z / 2));
+            bias -= learning_rate * error * ((1 / sqrt(2 * M_PI)) * exp(-z * z / 2));
+
             y_hat = Evaluate({inputSet[outputIndex]});
                 
             if(UI) { 
diff --git a/main.cpp b/main.cpp
index 6198bd6..4d58068 100644
--- a/main.cpp
+++ b/main.cpp
@@ -142,22 +142,22 @@ int main() {
     // alg.printVector(model.modelSetTest((alg.transpose(inputSet))));
     // std::cout << "ACCURACY: " << 100 * model.score() << "%" << std::endl;
 
-    // LOGISTIC REGRESSION
-    std::vector<std::vector<double>> inputSet; 
-    std::vector<double> outputSet; 
-    data.setData(30, "/Users/marcmelikyan/Desktop/Data/BreastCancer.csv", inputSet, outputSet);
-    LogReg model(inputSet, outputSet); 
-    model.SGD(0.001, 100000, 0);
+    // // LOGISTIC REGRESSION
+    // std::vector<std::vector<double>> inputSet; 
+    // std::vector<double> outputSet; 
+    // data.setData(30, "/Users/marcmelikyan/Desktop/Data/BreastCancer.csv", inputSet, outputSet);
+    // LogReg model(inputSet, outputSet); 
+    // model.SGD(0.001, 100000, 0);
     // model.MLE(0.1, 10000, 0);
-    alg.printVector(model.modelSetTest(inputSet));
-    std::cout << "ACCURACY: " << 100 * model.score() << "%" << std::endl;
+    // alg.printVector(model.modelSetTest(inputSet));
+    // std::cout << "ACCURACY: " << 100 * model.score() << "%" << std::endl;
 
     // // PROBIT REGRESSION
     // std::vector<std::vector<double>> inputSet;
     // std::vector<double> outputSet;
     // data.setData(30, "/Users/marcmelikyan/Desktop/Data/BreastCancer.csv", inputSet, outputSet);
     // ProbitReg model(inputSet, outputSet); 
-    // model.gradientDescent(0.0001, 10000, 1);
+    // model.SGD(0.001, 10000, 1);
     // alg.printVector(model.modelSetTest(inputSet));
     // std::cout << "ACCURACY: " << 100 * model.score() << "%" << std::endl;