#include<bits/stdc++.h>
using namespace std;

int main() {
string filename = "data.txt";
ifstream infile(filename);
if (!infile) {
cerr << "Error opening file: " << filename << endl;
return 1;
}

int numFeatures = 0;
vector<vector<double>> X;
vector<double> T;

string line;
while (getline(infile, line)) {
stringstream ss(line);
vector<double> features;
double val;
while (ss >> val) {
features.push_back(val);
}
// Last element is target
double target = features.back();
features.pop_back();
X.push_back(features);
T.push_back(target);

numFeatures = features.size();
}

int nSamples = X.size();
double learningRate = 0.01;
int epochs = 100;

// Initialize weights and bias
vector<double> weights(numFeatures, 0.0);
double bias = 0.0;

// Training
for (int epoch = 0; epoch < epochs; epoch++) {
double totalError = 0;
for (int i = 0; i < nSamples; i++) {
double net = bias;
for (int j = 0; j < numFeatures; j++) {
net += weights[j] * X[i][j];
}
double output = net; // Adaline uses linear output
double error = T[i] - output;
totalError += error * error;

for (int j = 0; j < numFeatures; j++) {
weights[j] += learningRate * error * X[i][j];
}
bias += learningRate * error;
}
cout << "Epoch " << epoch + 1 << " - MSE: " << totalError / nSamples << endl;
}

cout << "\nFinal weights: ";
for (auto w : weights) cout << w << " ";
cout << "\nFinal bias: " << bias << endl;

// Prediction
cout << "\nPredictions:\n";
for (int i = 0; i < nSamples; i++) {
double net = bias;
for (int j = 0; j < numFeatures; j++) {
net += weights[j] * X[i][j];
}
cout << "Sample " << i + 1 << ": " << net << " (target: " << T[i] << ")" << endl;
}

return 0;
}