rmontanana/BayesNet
1
0
Fork 0
Code Issues 2 Pull Requests 1 Actions Packages Projects Releases 6 Wiki Activity

Complete predict & predict_proba with voting & probabilities

Browse Source
This commit is contained in:
Ricardo Montañana Gómez
2024-02-23 23:11:14 +01:00
parent 52abd2d670
commit 8477698d8d
5 changed files with 161 additions and 283 deletions
Download Patch File Download Diff File Expand all files Collapse all files

49
src/Ensemble.cc
View File

@@ -51,32 +51,6 @@ namespace bayesnet {
result /= sum;
return result;
}
std::vector<std::vector<double>> Ensemble::voting(std::vector<std::vector<int>>& votes)
{
// Convert n_models x m matrix to a m x n_class_states matrix
std::vector<std::vector<double>> y_pred_final;
int numClasses = states.at(className).size();
auto sum = std::reduce(significanceModels.begin(), significanceModels.end());
// y_pred is m x n_models with the prediction of every model for each sample
std::cout << std::string(80, '*') << std::endl;
for (int i = 0; i < votes.size(); ++i) {
// n_votes store in each index (value of class) the significance added by each model
// i.e. n_votes[0] contains how much value has the value 0 of class. That value is generated by the models predictions
std::vector<double> n_votes(numClasses, 0.0);
for (int j = 0; j < n_models; ++j) {
n_votes[votes[i][j]] += significanceModels.at(j);
}
for (auto& x : n_votes) {
std::cout << x << " ";
}
std::cout << std::endl;
// To only do one division per result and gain precision
std::transform(n_votes.begin(), n_votes.end(), n_votes.begin(), [sum](double x) { return x / sum; });
y_pred_final.push_back(n_votes);
}
std::cout << std::string(80, '*') << std::endl;
return y_pred_final;
}
std::vector<std::vector<double>> Ensemble::predict_proba(std::vector<std::vector<int>>& X)
{
if (!fitted) {
@@ -94,7 +68,6 @@ namespace bayesnet {
std::vector<int> Ensemble::predict(std::vector<std::vector<int>>& X)
{
auto res = predict_proba(X);
std::cout << "res: " << res.size() << ", " << res[0].size() << std::endl;
return compute_arg_max(res);
}
torch::Tensor Ensemble::predict(torch::Tensor& X)
@@ -151,6 +124,13 @@ namespace bayesnet {
}
return y_pred;
}
std::vector<std::vector<double>> Ensemble::predict_average_voting(std::vector<std::vector<int>>& X)
{
torch::Tensor Xt = bayesnet::vectorToTensor(X, false);
auto y_pred = predict_average_voting(Xt);
std::vector<std::vector<double>> result = tensorToVectorDouble(y_pred);
return result;
}
torch::Tensor Ensemble::predict_average_voting(torch::Tensor& X)
{
// Build a m x n_models tensor with the predictions of each model
@@ -169,21 +149,6 @@ namespace bayesnet {
}
return voting(y_pred);
}
std::vector<std::vector<double>> Ensemble::predict_average_voting(std::vector<std::vector<int>>& X)
{
auto Xt = vectorToTensor(X);
auto y_pred = predict_average_voting(Xt);
auto res = voting(y_pred);
std::vector<std::vector<double>> result;
// Iterate over cols
for (int i = 0; i < res.size(1); ++i) {
auto col_tensor = res.index({ "...", i });
auto col = std::vector<double>(col_tensor.data_ptr<double>(), col_tensor.data_ptr<double>() + res.size(0));
result.push_back(col);
}
return result;
//return tensorToVector<double>(res);
}
float Ensemble::score(torch::Tensor& X, torch::Tensor& y)
{
auto y_pred = predict(X);

1
src/Ensemble.h
View File

@@ -36,7 +36,6 @@ namespace bayesnet {
torch::Tensor compute_arg_max(torch::Tensor& X);
std::vector<int> compute_arg_max(std::vector<std::vector<double>>& X);
torch::Tensor voting(torch::Tensor& votes);
std::vector<std::vector<double>> voting(std::vector<std::vector<int>>& votes);
unsigned n_models;
std::vector<std::unique_ptr<Classifier>> models;
std::vector<double> significanceModels;

29
src/bayesnetUtils.cc
View File

@@ -10,28 +10,39 @@ namespace bayesnet {
sort(indices.begin(), indices.end(), [&nums](int i, int j) {return nums[i] > nums[j];});
return indices;
}
template<typename T>
std::vector<std::vector<T>> tensorToVector(torch::Tensor& dtensor)
std::vector<std::vector<int>> tensorToVector(torch::Tensor& dtensor)
{
// convert mxn tensor to nxm std::vector
std::vector<std::vector<T>> result;
std::vector<std::vector<int>> result;
// Iterate over cols
for (int i = 0; i < dtensor.size(1); ++i) {
auto col_tensor = dtensor.index({ "...", i });
auto col = std::vector<T>(col_tensor.data_ptr<T>(), col_tensor.data_ptr<T>() + dtensor.size(0));
auto col = std::vector<int>(col_tensor.data_ptr<int>(), col_tensor.data_ptr<int>() + dtensor.size(0));
result.push_back(col);
}
return result;
}
torch::Tensor vectorToTensor(std::vector<std::vector<int>>& vector)
std::vector<std::vector<double>> tensorToVectorDouble(torch::Tensor& dtensor)
{
// convert nxm std::vector to mxn tensor
long int m = vector[0].size();
long int n = vector.size();
// convert mxn tensor to mxn std::vector
std::vector<std::vector<double>> result;
// Iterate over cols
for (int i = 0; i < dtensor.size(0); ++i) {
auto col_tensor = dtensor.index({ i, "..." });
auto col = std::vector<double>(col_tensor.data_ptr<float>(), col_tensor.data_ptr<float>() + dtensor.size(1));
result.push_back(col);
}
return result;
}
torch::Tensor vectorToTensor(std::vector<std::vector<int>>& vector, bool transpose)
{
// convert nxm std::vector to mxn tensor if transpose
long int m = transpose ? vector[0].size() : vector.size();
long int n = transpose ? vector.size() : vector[0].size();
auto tensor = torch::zeros({ m, n }, torch::kInt32);
for (int i = 0; i < m; ++i) {
for (int j = 0; j < n; ++j) {
tensor[i][j] = vector[j][i];
tensor[i][j] = transpose ? vector[j][i] : vector[i][j];
}
}
return tensor;

6
src/bayesnetUtils.h
View File

@@ -4,8 +4,8 @@
#include <vector>
namespace bayesnet {
std::vector<int> argsort(std::vector<double>& nums);
template<typename T>
std::vector<std::vector<T>> tensorToVector(torch::Tensor& dtensor);
torch::Tensor vectorToTensor(std::vector<std::vector<int>>& vector);
std::vector<std::vector<int>> tensorToVector(torch::Tensor& dtensor);
std::vector<std::vector<double>> tensorToVectorDouble(torch::Tensor& dtensor);
torch::Tensor vectorToTensor(std::vector<std::vector<int>>& vector, bool transpose = true);
}
#endif //BAYESNET_UTILS_H