rmontanana/BayesNet
rmontanana/BayesNet
1
0
Fork 0
Code Issues 2 Pull Requests Packages Projects Releases 3 Wiki Activity

block_update and install in local folder

Browse Source
This commit is contained in:
Ricardo Montañana Gómez 2024-04-10 00:55:36 +02:00
parent 1326891d6a
commit cf9b5716ac
Signed by: rmontanana
GPG Key ID: 46064262FD9A7ADE
7 changed files with 90 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
.vscode/launch.json vendored
View File

@ -8,7 +8,7 @@
"program": "${workspaceFolder}/build_release/sample/bayesnet_sample",
"args": [
"${workspaceFolder}/tests/data/glass.arff"
],
]
},
{
"type": "lldb",
@ -16,11 +16,33 @@
"name": "test",
"program": "${workspaceFolder}/build_debug/tests/TestBayesNet",
"args": [
"[Network]"
//"-c=\"Metrics Test\"",
// "-s",
"Block Update"
],
"cwd": "${workspaceFolder}/build_debug/tests",
"cwd": "${workspaceFolder}/build_debug/tests"
},
{
"name": "(gdb) Launch",
"type": "cppdbg",
"request": "launch",
"program": "enter program name, for example ${workspaceFolder}/a.out",
"args": [],
"stopAtEntry": false,
"cwd": "${fileDirname}",
"environment": [],
"externalConsole": false,
"MIMode": "gdb",
"setupCommands": [
{
"description": "Enable pretty-printing for gdb",
"text": "-enable-pretty-printing",
"ignoreFailures": true
},
{
"description": "Set Disassembly Flavor to Intel",
"text": "-gdb-set disassembly-flavor intel",
"ignoreFailures": true
}
]
}
]
}

2
CMakeLists.txt
View File

@ -1,7 +1,7 @@
cmake_minimum_required(VERSION 3.20)
project(BayesNet
VERSION 1.0.4
VERSION 1.0.4.1
DESCRIPTION "Bayesian Network and basic classifiers Library."
HOMEPAGE_URL "https://github.com/rmontanana/bayesnet"
LANGUAGES CXX

3
Makefile
View File

@ -53,9 +53,10 @@ uninstall: ## Uninstall library
xargs rm < $(f_release)/install_manifest.txt
@echo ">>> Done";
prefix = "/usr/local"
install: ## Install library
@echo ">>> Installing BayesNet...";
@cmake --install $(f_release)
@cmake --install $(f_release) --prefix $(prefix)
@echo ">>> Done";
debug: ## Build a debug version of the project

65
bayesnet/ensembles/BoostAODE.cc
View File

@ -127,6 +127,50 @@ namespace bayesnet {
}
return { weights, alpha_t, terminate };
}
std::tuple<torch::Tensor&, double, bool> BoostAODE::update_weights_block(int k, torch::Tensor& ytrain, torch::Tensor& weights)
{
//
// Make predict with only the last k models
//
std::unique_ptr<Classifier> model;
std::vector<std::unique_ptr<Classifier>> models_bak;
auto significance_bak = significanceModels;
auto n_models_bak = n_models;
// Remove the first n_models - k models
for (int i = 0; i < n_models - k; ++i) {
model = std::move(models[0]);
models.erase(models.begin());
models_bak.push_back(std::move(model));
}
assert(models.size() == k);
significanceModels = std::vector<double>(k, 1.0);
n_models = k;
auto ypred = predict(X_train);
//
// Update weights
//
double alpha_t;
bool terminate;
std::tie(weights, alpha_t, terminate) = update_weights(y_train, ypred, weights);
//
// Restore the models if needed
//
if (k != n_models_bak) {
for (int i = k - 1; i >= 0; --i) {
model = std::move(models_bak[i]);
models.insert(models.begin(), std::move(model));
}
}
significanceModels = significance_bak;
n_models = n_models_bak;
//
// Update the significance of the last k models
//
for (int i = 0; i < k; ++i) {
significanceModels[n_models - k + i] = alpha_t;
}
return { weights, alpha_t, terminate };
}
std::vector<int> BoostAODE::initializeModels()
{
std::vector<int> featuresUsed;
@ -156,7 +200,7 @@ namespace bayesnet {
std::unique_ptr<Classifier> model = std::make_unique<SPODE>(feature);
model->fit(dataset, features, className, states, weights_);
models.push_back(std::move(model));
significanceModels.push_back(1.0);
significanceModels.push_back(1.0); // They will be updated later in trainModel
n_models++;
}
notes.push_back("Used features in initialization: " + std::to_string(featuresUsed.size()) + " of " + std::to_string(features.size()) + " with " + select_features_algorithm);
@ -229,13 +273,15 @@ namespace bayesnet {
std::unique_ptr<Classifier> model;
model = std::make_unique<SPODE>(feature);
model->fit(dataset, features, className, states, weights_);
torch::Tensor ypred;
ypred = model->predict(X_train);
// Step 3.1: Compute the classifier amout of say
std::tie(weights_, alpha_t, finished) = update_weights(y_train, ypred, weights_);
if (finished) {
VLOG_SCOPE_F(2, "** epsilon_t > 0.5 **");
break;
alpha_t = 0.0;
if (!block_update) {
auto ypred = model->predict(X_train);
// Step 3.1: Compute the classifier amout of say
std::tie(weights_, alpha_t, finished) = update_weights(y_train, ypred, weights_);
if (finished) {
VLOG_SCOPE_F(2, "** epsilon_t > 0.5 **");
break;
}
}
// Step 3.4: Store classifier and its accuracy to weigh its future vote
numItemsPack++;
@ -245,6 +291,9 @@ namespace bayesnet {
n_models++;
VLOG_SCOPE_F(2, "numItemsPack: %d n_models: %d featuresUsed: %zu", numItemsPack, n_models, featuresUsed.size());
}
if (block_update) {
std::tie(weights_, alpha_t, finished) = update_weights_block(k, y_train, weights_);
}
if (convergence && !finished) {
auto y_val_predict = predict(X_test);
double accuracy = (y_val_predict == y_test).sum().item<double>() / (double)y_test.size(0);

3
bayesnet/ensembles/BoostAODE.h
View File

@ -25,6 +25,7 @@ namespace bayesnet {
void buildModel(const torch::Tensor& weights) override;
void trainModel(const torch::Tensor& weights) override;
private:
std::tuple<torch::Tensor&, double, bool> update_weights_block(int k, torch::Tensor& ytrain, torch::Tensor& weights);
std::vector<int> initializeModels();
torch::Tensor X_train, y_train, X_test, y_test;
// Hyperparameters
@ -36,7 +37,7 @@ namespace bayesnet {
std::string select_features_algorithm = Orders.DESC; // Selected feature selection algorithm
FeatureSelect* featureSelector = nullptr;
double threshold = -1;
bool block_update = true;
bool block_update = false;
};
}
#endif

2
docs/BoostAODE.md
View File

@ -8,7 +8,7 @@ The hyperparameters defined in the algorithm are:
- ***order*** (*{"asc", "desc", "rand"}*): Sets the order (ascending/descending/random) in which dataset variables will be processed to choose the parents of the *SPODEs*. Default value: *"desc"*.
- ***block_update*** (*boolean*): Sets whether the algorithm will update the weights of the models in blocks. If set to false, the algorithm will update the weights of the models one by one. Default value: *true*.
- ***block_update*** (*boolean*): Sets whether the algorithm will update the weights of the models in blocks. If set to false, the algorithm will update the weights of the models one by one. Default value: *false*.
- ***convergence*** (*boolean*): Sets whether the convergence of the result will be used as a termination condition. If this hyperparameter is set to true, the training dataset passed to the model is divided into two sets, one serving as training data and the other as a test set (so the original test partition will become a validation partition in this case). The partition is made by taking the first partition generated by a process of generating a 5 fold partition with stratification using a predetermined seed. The exit condition used in this *convergence* is that the difference between the accuracy obtained by the current model and that obtained by the previous model is greater than *1e-4*; otherwise, one will be added to the number of models that worsen the result (see next hyperparameter). Default value: *true*.

1
update_coverage.py
View File

@ -1,5 +1,4 @@
import subprocess
import os
import sys
readme_file = "README.md"