NeuralNetwork.h
3.39 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
#ifndef NEURALNETWORK_H
#define NEURALNETWORK_H
#include <vector>
#include <string>
class NeuralNetwork {
public:
// Single neuron calculation
double singleNeuron(double input, double weight);
// Multiple inputs, single output
double multipleInputSingleOutput(std::vector<double> inputs, std::vector<double> weights, double bias);
// Single input, multiple outputs
void singleInputMultipleOutput(double input, std::vector<double> weights, double bias, std::vector<double>& outputs);
// Multiple inputs, multiple outputs
void multipleInputMultipleOutput(std::vector<double>& inputs, std::vector<double>& weights, std::vector<double>& biases, std::vector<double>& outputs, int inputSize, int outputSize);
// Hidden layer function
void hiddenLayer(std::vector<double>& inputs, std::vector<double>& hiddenWeights, std::vector<double>& hiddenBiases, std::vector<double>& hiddenOutputs, int inputSize, int hiddenSize);
// Error calculation
void calculateError(std::vector<double>& predictedOutput, std::vector<double>& groundTruth, std::vector<double>& error);
// Mean Squared Error (MSE)
double calculateMSE(std::vector<double>& error);
// Root Mean Squared Error (RMSE)
double calculateRMSE(double mse);
// Brute-force learning to find the best weight
void bruteForceLearning(double input, double& weight, double expectedValue, double learningRate, int maxEpochs);
// Backpropagation learning function
void backpropagation(const std::vector<double>& input, const std::vector<double>& expectedOutput,
std::vector<std::vector<double>>& inputToHiddenWeights, std::vector<double>& hiddenBiases,
std::vector<std::vector<double>>& hiddenToOutputWeights, std::vector<double>& outputBiases,
double learningRate, int epochs);
// Activation functions (ReLU and Sigmoid)
double relu(double x);
double sigmoid(double x);
// Vectorized activation functions
void vectorReLU(std::vector<double>& inputVector, std::vector<double>& outputVector);
void vectorSigmoid(std::vector<double>& inputVector, std::vector<double>& outputVector);
// Print a 2D matrix
void printMatrix(int rows, int cols, const std::vector<std::vector<double>>& matrix);
// Compute cost for logistic regression
double computeCost(int m, const std::vector<std::vector<double>>& yhat, const std::vector<std::vector<double>>& y);
// Normalize a 2D matrix
int normalizeData2D(const std::vector<std::vector<double>>& inputMatrix, std::vector<std::vector<double>>& outputMatrix);
// Save network
void saveNetwork(const std::string& filename, int numOfFeatures, int numOfHiddenNodes, int numOfOutputNodes,
std::vector<std::vector<double>>& inputToHiddenWeights, std::vector<double>& hiddenLayerBias,
std::vector<std::vector<double>>& hiddenToOutputWeights, std::vector<double>& outputLayerBias);
// Load network
void loadNetwork(const std::string& filename, int numOfFeatures, int numOfHiddenNodes, int numOfOutputNodes,
std::vector<std::vector<double>>& inputToHiddenWeights, std::vector<double>& hiddenLayerBias,
std::vector<std::vector<double>>& hiddenToOutputWeights, std::vector<double>& outputLayerBias);
};
#endif // NEURALNETWORK_H