481 lines
59 KiB
HTML
481 lines
59 KiB
HTML
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
|
||
|
||
<html lang="en">
|
||
|
||
<head>
|
||
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
|
||
<title>LCOV - BayesNet Coverage Report - bayesnet/ensembles/BoostAODE.cc</title>
|
||
<link rel="stylesheet" type="text/css" href="../../gcov.css">
|
||
</head>
|
||
|
||
<body>
|
||
|
||
<table width="100%" border=0 cellspacing=0 cellpadding=0>
|
||
<tr><td class="title">LCOV - code coverage report</td></tr>
|
||
<tr><td class="ruler"><img src="../../glass.png" width=3 height=3 alt=""></td></tr>
|
||
|
||
<tr>
|
||
<td width="100%">
|
||
<table cellpadding=1 border=0 width="100%">
|
||
<tr>
|
||
<td width="10%" class="headerItem">Current view:</td>
|
||
<td width="10%" class="headerValue"><a href="../../index.html" target="_parent">top level</a> - <a href="index.html" target="_parent">bayesnet/ensembles</a> - BoostAODE.cc<span style="font-size: 80%;"> (source / <a href="BoostAODE.cc.func-c.html">functions</a>)</span></td>
|
||
<td width="5%"></td>
|
||
<td width="5%"></td>
|
||
<td width="5%" class="headerCovTableHead">Coverage</td>
|
||
<td width="5%" class="headerCovTableHead" title="Covered + Uncovered code">Total</td>
|
||
<td width="5%" class="headerCovTableHead" title="Exercised code only">Hit</td>
|
||
</tr>
|
||
<tr>
|
||
<td class="headerItem">Test:</td>
|
||
<td class="headerValue">BayesNet Coverage Report</td>
|
||
<td></td>
|
||
<td class="headerItem">Lines:</td>
|
||
<td class="headerCovTableEntryHi">98.3 %</td>
|
||
<td class="headerCovTableEntry">237</td>
|
||
<td class="headerCovTableEntry">233</td>
|
||
</tr>
|
||
<tr>
|
||
<td class="headerItem">Test Date:</td>
|
||
<td class="headerValue">2024-05-06 17:54:04</td>
|
||
<td></td>
|
||
<td class="headerItem">Functions:</td>
|
||
<td class="headerCovTableEntryHi">100.0 %</td>
|
||
<td class="headerCovTableEntry">9</td>
|
||
<td class="headerCovTableEntry">9</td>
|
||
</tr>
|
||
<tr>
|
||
<td class="headerItem">Legend:</td>
|
||
<td class="headerValueLeg"> Lines:
|
||
<span class="coverLegendCov">hit</span>
|
||
<span class="coverLegendNoCov">not hit</span>
|
||
</td>
|
||
<td></td>
|
||
</tr>
|
||
<tr><td><img src="../../glass.png" width=3 height=3 alt=""></td></tr>
|
||
</table>
|
||
</td>
|
||
</tr>
|
||
|
||
<tr><td class="ruler"><img src="../../glass.png" width=3 height=3 alt=""></td></tr>
|
||
</table>
|
||
|
||
<table cellpadding=0 cellspacing=0 border=0>
|
||
<tr>
|
||
<td><br></td>
|
||
</tr>
|
||
<tr>
|
||
<td>
|
||
<pre class="sourceHeading"> Line data Source code</pre>
|
||
<pre class="source">
|
||
<span id="L1"><span class="lineNum"> 1</span> : // ***************************************************************</span>
|
||
<span id="L2"><span class="lineNum"> 2</span> : // SPDX-FileCopyrightText: Copyright 2024 Ricardo Montañana Gómez</span>
|
||
<span id="L3"><span class="lineNum"> 3</span> : // SPDX-FileType: SOURCE</span>
|
||
<span id="L4"><span class="lineNum"> 4</span> : // SPDX-License-Identifier: MIT</span>
|
||
<span id="L5"><span class="lineNum"> 5</span> : // ***************************************************************</span>
|
||
<span id="L6"><span class="lineNum"> 6</span> : </span>
|
||
<span id="L7"><span class="lineNum"> 7</span> : #include <set></span>
|
||
<span id="L8"><span class="lineNum"> 8</span> : #include <functional></span>
|
||
<span id="L9"><span class="lineNum"> 9</span> : #include <limits.h></span>
|
||
<span id="L10"><span class="lineNum"> 10</span> : #include <tuple></span>
|
||
<span id="L11"><span class="lineNum"> 11</span> : #include <folding.hpp></span>
|
||
<span id="L12"><span class="lineNum"> 12</span> : #include "bayesnet/feature_selection/CFS.h"</span>
|
||
<span id="L13"><span class="lineNum"> 13</span> : #include "bayesnet/feature_selection/FCBF.h"</span>
|
||
<span id="L14"><span class="lineNum"> 14</span> : #include "bayesnet/feature_selection/IWSS.h"</span>
|
||
<span id="L15"><span class="lineNum"> 15</span> : #include "BoostAODE.h"</span>
|
||
<span id="L16"><span class="lineNum"> 16</span> : #include "lib/log/loguru.cpp"</span>
|
||
<span id="L17"><span class="lineNum"> 17</span> : </span>
|
||
<span id="L18"><span class="lineNum"> 18</span> : namespace bayesnet {</span>
|
||
<span id="L19"><span class="lineNum"> 19</span> : </span>
|
||
<span id="L20"><span class="lineNum"> 20</span> <span class="tlaGNC tlaBgGNC"> 168 : BoostAODE::BoostAODE(bool predict_voting) : Ensemble(predict_voting)</span></span>
|
||
<span id="L21"><span class="lineNum"> 21</span> : {</span>
|
||
<span id="L22"><span class="lineNum"> 22</span> <span class="tlaGNC"> 1848 : validHyperparameters = {</span></span>
|
||
<span id="L23"><span class="lineNum"> 23</span> : "maxModels", "bisection", "order", "convergence", "convergence_best", "threshold",</span>
|
||
<span id="L24"><span class="lineNum"> 24</span> : "select_features", "maxTolerance", "predict_voting", "block_update"</span>
|
||
<span id="L25"><span class="lineNum"> 25</span> <span class="tlaGNC"> 1848 : };</span></span>
|
||
<span id="L26"><span class="lineNum"> 26</span> : </span>
|
||
<span id="L27"><span class="lineNum"> 27</span> <span class="tlaGNC"> 504 : }</span></span>
|
||
<span id="L28"><span class="lineNum"> 28</span> <span class="tlaGNC"> 92 : void BoostAODE::buildModel(const torch::Tensor& weights)</span></span>
|
||
<span id="L29"><span class="lineNum"> 29</span> : {</span>
|
||
<span id="L30"><span class="lineNum"> 30</span> : // Models shall be built in trainModel</span>
|
||
<span id="L31"><span class="lineNum"> 31</span> <span class="tlaGNC"> 92 : models.clear();</span></span>
|
||
<span id="L32"><span class="lineNum"> 32</span> <span class="tlaGNC"> 92 : significanceModels.clear();</span></span>
|
||
<span id="L33"><span class="lineNum"> 33</span> <span class="tlaGNC"> 92 : n_models = 0;</span></span>
|
||
<span id="L34"><span class="lineNum"> 34</span> : // Prepare the validation dataset</span>
|
||
<span id="L35"><span class="lineNum"> 35</span> <span class="tlaGNC"> 276 : auto y_ = dataset.index({ -1, "..." });</span></span>
|
||
<span id="L36"><span class="lineNum"> 36</span> <span class="tlaGNC"> 92 : if (convergence) {</span></span>
|
||
<span id="L37"><span class="lineNum"> 37</span> : // Prepare train & validation sets from train data</span>
|
||
<span id="L38"><span class="lineNum"> 38</span> <span class="tlaGNC"> 76 : auto fold = folding::StratifiedKFold(5, y_, 271);</span></span>
|
||
<span id="L39"><span class="lineNum"> 39</span> <span class="tlaGNC"> 76 : auto [train, test] = fold.getFold(0);</span></span>
|
||
<span id="L40"><span class="lineNum"> 40</span> <span class="tlaGNC"> 76 : auto train_t = torch::tensor(train);</span></span>
|
||
<span id="L41"><span class="lineNum"> 41</span> <span class="tlaGNC"> 76 : auto test_t = torch::tensor(test);</span></span>
|
||
<span id="L42"><span class="lineNum"> 42</span> : // Get train and validation sets</span>
|
||
<span id="L43"><span class="lineNum"> 43</span> <span class="tlaGNC"> 380 : X_train = dataset.index({ torch::indexing::Slice(0, dataset.size(0) - 1), train_t });</span></span>
|
||
<span id="L44"><span class="lineNum"> 44</span> <span class="tlaGNC"> 228 : y_train = dataset.index({ -1, train_t });</span></span>
|
||
<span id="L45"><span class="lineNum"> 45</span> <span class="tlaGNC"> 380 : X_test = dataset.index({ torch::indexing::Slice(0, dataset.size(0) - 1), test_t });</span></span>
|
||
<span id="L46"><span class="lineNum"> 46</span> <span class="tlaGNC"> 228 : y_test = dataset.index({ -1, test_t });</span></span>
|
||
<span id="L47"><span class="lineNum"> 47</span> <span class="tlaGNC"> 76 : dataset = X_train;</span></span>
|
||
<span id="L48"><span class="lineNum"> 48</span> <span class="tlaGNC"> 76 : m = X_train.size(1);</span></span>
|
||
<span id="L49"><span class="lineNum"> 49</span> <span class="tlaGNC"> 76 : auto n_classes = states.at(className).size();</span></span>
|
||
<span id="L50"><span class="lineNum"> 50</span> : // Build dataset with train data</span>
|
||
<span id="L51"><span class="lineNum"> 51</span> <span class="tlaGNC"> 76 : buildDataset(y_train);</span></span>
|
||
<span id="L52"><span class="lineNum"> 52</span> <span class="tlaGNC"> 76 : metrics = Metrics(dataset, features, className, n_classes);</span></span>
|
||
<span id="L53"><span class="lineNum"> 53</span> <span class="tlaGNC"> 76 : } else {</span></span>
|
||
<span id="L54"><span class="lineNum"> 54</span> : // Use all data to train</span>
|
||
<span id="L55"><span class="lineNum"> 55</span> <span class="tlaGNC"> 64 : X_train = dataset.index({ torch::indexing::Slice(0, dataset.size(0) - 1), "..." });</span></span>
|
||
<span id="L56"><span class="lineNum"> 56</span> <span class="tlaGNC"> 16 : y_train = y_;</span></span>
|
||
<span id="L57"><span class="lineNum"> 57</span> : }</span>
|
||
<span id="L58"><span class="lineNum"> 58</span> <span class="tlaGNC"> 900 : }</span></span>
|
||
<span id="L59"><span class="lineNum"> 59</span> <span class="tlaGNC"> 88 : void BoostAODE::setHyperparameters(const nlohmann::json& hyperparameters_)</span></span>
|
||
<span id="L60"><span class="lineNum"> 60</span> : {</span>
|
||
<span id="L61"><span class="lineNum"> 61</span> <span class="tlaGNC"> 88 : auto hyperparameters = hyperparameters_;</span></span>
|
||
<span id="L62"><span class="lineNum"> 62</span> <span class="tlaGNC"> 88 : if (hyperparameters.contains("order")) {</span></span>
|
||
<span id="L63"><span class="lineNum"> 63</span> <span class="tlaGNC"> 100 : std::vector<std::string> algos = { Orders.ASC, Orders.DESC, Orders.RAND };</span></span>
|
||
<span id="L64"><span class="lineNum"> 64</span> <span class="tlaGNC"> 20 : order_algorithm = hyperparameters["order"];</span></span>
|
||
<span id="L65"><span class="lineNum"> 65</span> <span class="tlaGNC"> 20 : if (std::find(algos.begin(), algos.end(), order_algorithm) == algos.end()) {</span></span>
|
||
<span id="L66"><span class="lineNum"> 66</span> <span class="tlaGNC"> 4 : throw std::invalid_argument("Invalid order algorithm, valid values [" + Orders.ASC + ", " + Orders.DESC + ", " + Orders.RAND + "]");</span></span>
|
||
<span id="L67"><span class="lineNum"> 67</span> : }</span>
|
||
<span id="L68"><span class="lineNum"> 68</span> <span class="tlaGNC"> 16 : hyperparameters.erase("order");</span></span>
|
||
<span id="L69"><span class="lineNum"> 69</span> <span class="tlaGNC"> 20 : }</span></span>
|
||
<span id="L70"><span class="lineNum"> 70</span> <span class="tlaGNC"> 84 : if (hyperparameters.contains("convergence")) {</span></span>
|
||
<span id="L71"><span class="lineNum"> 71</span> <span class="tlaGNC"> 36 : convergence = hyperparameters["convergence"];</span></span>
|
||
<span id="L72"><span class="lineNum"> 72</span> <span class="tlaGNC"> 36 : hyperparameters.erase("convergence");</span></span>
|
||
<span id="L73"><span class="lineNum"> 73</span> : }</span>
|
||
<span id="L74"><span class="lineNum"> 74</span> <span class="tlaGNC"> 84 : if (hyperparameters.contains("convergence_best")) {</span></span>
|
||
<span id="L75"><span class="lineNum"> 75</span> <span class="tlaGNC"> 12 : convergence_best = hyperparameters["convergence_best"];</span></span>
|
||
<span id="L76"><span class="lineNum"> 76</span> <span class="tlaGNC"> 12 : hyperparameters.erase("convergence_best");</span></span>
|
||
<span id="L77"><span class="lineNum"> 77</span> : }</span>
|
||
<span id="L78"><span class="lineNum"> 78</span> <span class="tlaGNC"> 84 : if (hyperparameters.contains("bisection")) {</span></span>
|
||
<span id="L79"><span class="lineNum"> 79</span> <span class="tlaGNC"> 32 : bisection = hyperparameters["bisection"];</span></span>
|
||
<span id="L80"><span class="lineNum"> 80</span> <span class="tlaGNC"> 32 : hyperparameters.erase("bisection");</span></span>
|
||
<span id="L81"><span class="lineNum"> 81</span> : }</span>
|
||
<span id="L82"><span class="lineNum"> 82</span> <span class="tlaGNC"> 84 : if (hyperparameters.contains("threshold")) {</span></span>
|
||
<span id="L83"><span class="lineNum"> 83</span> <span class="tlaGNC"> 24 : threshold = hyperparameters["threshold"];</span></span>
|
||
<span id="L84"><span class="lineNum"> 84</span> <span class="tlaGNC"> 24 : hyperparameters.erase("threshold");</span></span>
|
||
<span id="L85"><span class="lineNum"> 85</span> : }</span>
|
||
<span id="L86"><span class="lineNum"> 86</span> <span class="tlaGNC"> 84 : if (hyperparameters.contains("maxTolerance")) {</span></span>
|
||
<span id="L87"><span class="lineNum"> 87</span> <span class="tlaGNC"> 44 : maxTolerance = hyperparameters["maxTolerance"];</span></span>
|
||
<span id="L88"><span class="lineNum"> 88</span> <span class="tlaGNC"> 44 : if (maxTolerance < 1 || maxTolerance > 4)</span></span>
|
||
<span id="L89"><span class="lineNum"> 89</span> <span class="tlaGNC"> 12 : throw std::invalid_argument("Invalid maxTolerance value, must be greater in [1, 4]");</span></span>
|
||
<span id="L90"><span class="lineNum"> 90</span> <span class="tlaGNC"> 32 : hyperparameters.erase("maxTolerance");</span></span>
|
||
<span id="L91"><span class="lineNum"> 91</span> : }</span>
|
||
<span id="L92"><span class="lineNum"> 92</span> <span class="tlaGNC"> 72 : if (hyperparameters.contains("predict_voting")) {</span></span>
|
||
<span id="L93"><span class="lineNum"> 93</span> <span class="tlaGNC"> 4 : predict_voting = hyperparameters["predict_voting"];</span></span>
|
||
<span id="L94"><span class="lineNum"> 94</span> <span class="tlaGNC"> 4 : hyperparameters.erase("predict_voting");</span></span>
|
||
<span id="L95"><span class="lineNum"> 95</span> : }</span>
|
||
<span id="L96"><span class="lineNum"> 96</span> <span class="tlaGNC"> 72 : if (hyperparameters.contains("select_features")) {</span></span>
|
||
<span id="L97"><span class="lineNum"> 97</span> <span class="tlaGNC"> 36 : auto selectedAlgorithm = hyperparameters["select_features"];</span></span>
|
||
<span id="L98"><span class="lineNum"> 98</span> <span class="tlaGNC"> 180 : std::vector<std::string> algos = { SelectFeatures.IWSS, SelectFeatures.CFS, SelectFeatures.FCBF };</span></span>
|
||
<span id="L99"><span class="lineNum"> 99</span> <span class="tlaGNC"> 36 : selectFeatures = true;</span></span>
|
||
<span id="L100"><span class="lineNum"> 100</span> <span class="tlaGNC"> 36 : select_features_algorithm = selectedAlgorithm;</span></span>
|
||
<span id="L101"><span class="lineNum"> 101</span> <span class="tlaGNC"> 36 : if (std::find(algos.begin(), algos.end(), selectedAlgorithm) == algos.end()) {</span></span>
|
||
<span id="L102"><span class="lineNum"> 102</span> <span class="tlaGNC"> 4 : throw std::invalid_argument("Invalid selectFeatures value, valid values [" + SelectFeatures.IWSS + ", " + SelectFeatures.CFS + ", " + SelectFeatures.FCBF + "]");</span></span>
|
||
<span id="L103"><span class="lineNum"> 103</span> : }</span>
|
||
<span id="L104"><span class="lineNum"> 104</span> <span class="tlaGNC"> 32 : hyperparameters.erase("select_features");</span></span>
|
||
<span id="L105"><span class="lineNum"> 105</span> <span class="tlaGNC"> 40 : }</span></span>
|
||
<span id="L106"><span class="lineNum"> 106</span> <span class="tlaGNC"> 68 : if (hyperparameters.contains("block_update")) {</span></span>
|
||
<span id="L107"><span class="lineNum"> 107</span> <span class="tlaGNC"> 8 : block_update = hyperparameters["block_update"];</span></span>
|
||
<span id="L108"><span class="lineNum"> 108</span> <span class="tlaGNC"> 8 : hyperparameters.erase("block_update");</span></span>
|
||
<span id="L109"><span class="lineNum"> 109</span> : }</span>
|
||
<span id="L110"><span class="lineNum"> 110</span> <span class="tlaGNC"> 68 : Classifier::setHyperparameters(hyperparameters);</span></span>
|
||
<span id="L111"><span class="lineNum"> 111</span> <span class="tlaGNC"> 144 : }</span></span>
|
||
<span id="L112"><span class="lineNum"> 112</span> <span class="tlaGNC"> 544 : std::tuple<torch::Tensor&, double, bool> update_weights(torch::Tensor& ytrain, torch::Tensor& ypred, torch::Tensor& weights)</span></span>
|
||
<span id="L113"><span class="lineNum"> 113</span> : {</span>
|
||
<span id="L114"><span class="lineNum"> 114</span> <span class="tlaGNC"> 544 : bool terminate = false;</span></span>
|
||
<span id="L115"><span class="lineNum"> 115</span> <span class="tlaGNC"> 544 : double alpha_t = 0;</span></span>
|
||
<span id="L116"><span class="lineNum"> 116</span> <span class="tlaGNC"> 544 : auto mask_wrong = ypred != ytrain;</span></span>
|
||
<span id="L117"><span class="lineNum"> 117</span> <span class="tlaGNC"> 544 : auto mask_right = ypred == ytrain;</span></span>
|
||
<span id="L118"><span class="lineNum"> 118</span> <span class="tlaGNC"> 544 : auto masked_weights = weights * mask_wrong.to(weights.dtype());</span></span>
|
||
<span id="L119"><span class="lineNum"> 119</span> <span class="tlaGNC"> 544 : double epsilon_t = masked_weights.sum().item<double>();</span></span>
|
||
<span id="L120"><span class="lineNum"> 120</span> <span class="tlaGNC"> 544 : if (epsilon_t > 0.5) {</span></span>
|
||
<span id="L121"><span class="lineNum"> 121</span> : // Inverse the weights policy (plot ln(wt))</span>
|
||
<span id="L122"><span class="lineNum"> 122</span> : // "In each round of AdaBoost, there is a sanity check to ensure that the current base </span>
|
||
<span id="L123"><span class="lineNum"> 123</span> : // learner is better than random guess" (Zhi-Hua Zhou, 2012)</span>
|
||
<span id="L124"><span class="lineNum"> 124</span> <span class="tlaGNC"> 16 : terminate = true;</span></span>
|
||
<span id="L125"><span class="lineNum"> 125</span> : } else {</span>
|
||
<span id="L126"><span class="lineNum"> 126</span> <span class="tlaGNC"> 528 : double wt = (1 - epsilon_t) / epsilon_t;</span></span>
|
||
<span id="L127"><span class="lineNum"> 127</span> <span class="tlaGNC"> 528 : alpha_t = epsilon_t == 0 ? 1 : 0.5 * log(wt);</span></span>
|
||
<span id="L128"><span class="lineNum"> 128</span> : // Step 3.2: Update weights for next classifier</span>
|
||
<span id="L129"><span class="lineNum"> 129</span> : // Step 3.2.1: Update weights of wrong samples</span>
|
||
<span id="L130"><span class="lineNum"> 130</span> <span class="tlaGNC"> 528 : weights += mask_wrong.to(weights.dtype()) * exp(alpha_t) * weights;</span></span>
|
||
<span id="L131"><span class="lineNum"> 131</span> : // Step 3.2.2: Update weights of right samples</span>
|
||
<span id="L132"><span class="lineNum"> 132</span> <span class="tlaGNC"> 528 : weights += mask_right.to(weights.dtype()) * exp(-alpha_t) * weights;</span></span>
|
||
<span id="L133"><span class="lineNum"> 133</span> : // Step 3.3: Normalise the weights</span>
|
||
<span id="L134"><span class="lineNum"> 134</span> <span class="tlaGNC"> 528 : double totalWeights = torch::sum(weights).item<double>();</span></span>
|
||
<span id="L135"><span class="lineNum"> 135</span> <span class="tlaGNC"> 528 : weights = weights / totalWeights;</span></span>
|
||
<span id="L136"><span class="lineNum"> 136</span> : }</span>
|
||
<span id="L137"><span class="lineNum"> 137</span> <span class="tlaGNC"> 1088 : return { weights, alpha_t, terminate };</span></span>
|
||
<span id="L138"><span class="lineNum"> 138</span> <span class="tlaGNC"> 544 : }</span></span>
|
||
<span id="L139"><span class="lineNum"> 139</span> <span class="tlaGNC"> 28 : std::tuple<torch::Tensor&, double, bool> BoostAODE::update_weights_block(int k, torch::Tensor& ytrain, torch::Tensor& weights)</span></span>
|
||
<span id="L140"><span class="lineNum"> 140</span> : {</span>
|
||
<span id="L141"><span class="lineNum"> 141</span> : /* Update Block algorithm</span>
|
||
<span id="L142"><span class="lineNum"> 142</span> : k = # of models in block</span>
|
||
<span id="L143"><span class="lineNum"> 143</span> : n_models = # of models in ensemble to make predictions</span>
|
||
<span id="L144"><span class="lineNum"> 144</span> : n_models_bak = # models saved</span>
|
||
<span id="L145"><span class="lineNum"> 145</span> : models = vector of models to make predictions</span>
|
||
<span id="L146"><span class="lineNum"> 146</span> : models_bak = models not used to make predictions</span>
|
||
<span id="L147"><span class="lineNum"> 147</span> : significances_bak = backup of significances vector</span>
|
||
<span id="L148"><span class="lineNum"> 148</span> : </span>
|
||
<span id="L149"><span class="lineNum"> 149</span> : Case list</span>
|
||
<span id="L150"><span class="lineNum"> 150</span> : A) k = 1, n_models = 1 => n = 0 , n_models = n + k</span>
|
||
<span id="L151"><span class="lineNum"> 151</span> : B) k = 1, n_models = n + 1 => n_models = n + k</span>
|
||
<span id="L152"><span class="lineNum"> 152</span> : C) k > 1, n_models = k + 1 => n= 1, n_models = n + k</span>
|
||
<span id="L153"><span class="lineNum"> 153</span> : D) k > 1, n_models = k => n = 0, n_models = n + k</span>
|
||
<span id="L154"><span class="lineNum"> 154</span> : E) k > 1, n_models = k + n => n_models = n + k</span>
|
||
<span id="L155"><span class="lineNum"> 155</span> : </span>
|
||
<span id="L156"><span class="lineNum"> 156</span> : A, D) n=0, k > 0, n_models == k</span>
|
||
<span id="L157"><span class="lineNum"> 157</span> : 1. n_models_bak <- n_models</span>
|
||
<span id="L158"><span class="lineNum"> 158</span> : 2. significances_bak <- significances</span>
|
||
<span id="L159"><span class="lineNum"> 159</span> : 3. significances = vector(k, 1)</span>
|
||
<span id="L160"><span class="lineNum"> 160</span> : 4. Don’t move any classifiers out of models</span>
|
||
<span id="L161"><span class="lineNum"> 161</span> : 5. n_models <- k</span>
|
||
<span id="L162"><span class="lineNum"> 162</span> : 6. Make prediction, compute alpha, update weights</span>
|
||
<span id="L163"><span class="lineNum"> 163</span> : 7. Don’t restore any classifiers to models</span>
|
||
<span id="L164"><span class="lineNum"> 164</span> : 8. significances <- significances_bak</span>
|
||
<span id="L165"><span class="lineNum"> 165</span> : 9. Update last k significances</span>
|
||
<span id="L166"><span class="lineNum"> 166</span> : 10. n_models <- n_models_bak</span>
|
||
<span id="L167"><span class="lineNum"> 167</span> : </span>
|
||
<span id="L168"><span class="lineNum"> 168</span> : B, C, E) n > 0, k > 0, n_models == n + k</span>
|
||
<span id="L169"><span class="lineNum"> 169</span> : 1. n_models_bak <- n_models</span>
|
||
<span id="L170"><span class="lineNum"> 170</span> : 2. significances_bak <- significances</span>
|
||
<span id="L171"><span class="lineNum"> 171</span> : 3. significances = vector(k, 1)</span>
|
||
<span id="L172"><span class="lineNum"> 172</span> : 4. Move first n classifiers to models_bak</span>
|
||
<span id="L173"><span class="lineNum"> 173</span> : 5. n_models <- k</span>
|
||
<span id="L174"><span class="lineNum"> 174</span> : 6. Make prediction, compute alpha, update weights</span>
|
||
<span id="L175"><span class="lineNum"> 175</span> : 7. Insert classifiers in models_bak to be the first n models</span>
|
||
<span id="L176"><span class="lineNum"> 176</span> : 8. significances <- significances_bak</span>
|
||
<span id="L177"><span class="lineNum"> 177</span> : 9. Update last k significances</span>
|
||
<span id="L178"><span class="lineNum"> 178</span> : 10. n_models <- n_models_bak</span>
|
||
<span id="L179"><span class="lineNum"> 179</span> : */</span>
|
||
<span id="L180"><span class="lineNum"> 180</span> : //</span>
|
||
<span id="L181"><span class="lineNum"> 181</span> : // Make predict with only the last k models</span>
|
||
<span id="L182"><span class="lineNum"> 182</span> : //</span>
|
||
<span id="L183"><span class="lineNum"> 183</span> <span class="tlaGNC"> 28 : std::unique_ptr<Classifier> model;</span></span>
|
||
<span id="L184"><span class="lineNum"> 184</span> <span class="tlaGNC"> 28 : std::vector<std::unique_ptr<Classifier>> models_bak;</span></span>
|
||
<span id="L185"><span class="lineNum"> 185</span> : // 1. n_models_bak <- n_models 2. significances_bak <- significances</span>
|
||
<span id="L186"><span class="lineNum"> 186</span> <span class="tlaGNC"> 28 : auto significance_bak = significanceModels;</span></span>
|
||
<span id="L187"><span class="lineNum"> 187</span> <span class="tlaGNC"> 28 : auto n_models_bak = n_models;</span></span>
|
||
<span id="L188"><span class="lineNum"> 188</span> : // 3. significances = vector(k, 1)</span>
|
||
<span id="L189"><span class="lineNum"> 189</span> <span class="tlaGNC"> 28 : significanceModels = std::vector<double>(k, 1.0);</span></span>
|
||
<span id="L190"><span class="lineNum"> 190</span> : // 4. Move first n classifiers to models_bak</span>
|
||
<span id="L191"><span class="lineNum"> 191</span> : // backup the first n_models - k models (if n_models == k, don't backup any)</span>
|
||
<span id="L192"><span class="lineNum"> 192</span> <span class="tlaGNC"> 148 : for (int i = 0; i < n_models - k; ++i) {</span></span>
|
||
<span id="L193"><span class="lineNum"> 193</span> <span class="tlaGNC"> 120 : model = std::move(models[0]);</span></span>
|
||
<span id="L194"><span class="lineNum"> 194</span> <span class="tlaGNC"> 120 : models.erase(models.begin());</span></span>
|
||
<span id="L195"><span class="lineNum"> 195</span> <span class="tlaGNC"> 120 : models_bak.push_back(std::move(model));</span></span>
|
||
<span id="L196"><span class="lineNum"> 196</span> : }</span>
|
||
<span id="L197"><span class="lineNum"> 197</span> <span class="tlaGNC"> 28 : assert(models.size() == k);</span></span>
|
||
<span id="L198"><span class="lineNum"> 198</span> : // 5. n_models <- k</span>
|
||
<span id="L199"><span class="lineNum"> 199</span> <span class="tlaGNC"> 28 : n_models = k;</span></span>
|
||
<span id="L200"><span class="lineNum"> 200</span> : // 6. Make prediction, compute alpha, update weights</span>
|
||
<span id="L201"><span class="lineNum"> 201</span> <span class="tlaGNC"> 28 : auto ypred = predict(X_train);</span></span>
|
||
<span id="L202"><span class="lineNum"> 202</span> : //</span>
|
||
<span id="L203"><span class="lineNum"> 203</span> : // Update weights</span>
|
||
<span id="L204"><span class="lineNum"> 204</span> : //</span>
|
||
<span id="L205"><span class="lineNum"> 205</span> : double alpha_t;</span>
|
||
<span id="L206"><span class="lineNum"> 206</span> : bool terminate;</span>
|
||
<span id="L207"><span class="lineNum"> 207</span> <span class="tlaGNC"> 28 : std::tie(weights, alpha_t, terminate) = update_weights(y_train, ypred, weights);</span></span>
|
||
<span id="L208"><span class="lineNum"> 208</span> : //</span>
|
||
<span id="L209"><span class="lineNum"> 209</span> : // Restore the models if needed</span>
|
||
<span id="L210"><span class="lineNum"> 210</span> : //</span>
|
||
<span id="L211"><span class="lineNum"> 211</span> : // 7. Insert classifiers in models_bak to be the first n models</span>
|
||
<span id="L212"><span class="lineNum"> 212</span> : // if n_models_bak == k, don't restore any, because none of them were moved</span>
|
||
<span id="L213"><span class="lineNum"> 213</span> <span class="tlaGNC"> 28 : if (k != n_models_bak) {</span></span>
|
||
<span id="L214"><span class="lineNum"> 214</span> : // Insert in the same order as they were extracted</span>
|
||
<span id="L215"><span class="lineNum"> 215</span> <span class="tlaGNC"> 24 : int bak_size = models_bak.size();</span></span>
|
||
<span id="L216"><span class="lineNum"> 216</span> <span class="tlaGNC"> 144 : for (int i = 0; i < bak_size; ++i) {</span></span>
|
||
<span id="L217"><span class="lineNum"> 217</span> <span class="tlaGNC"> 120 : model = std::move(models_bak[bak_size - 1 - i]);</span></span>
|
||
<span id="L218"><span class="lineNum"> 218</span> <span class="tlaGNC"> 120 : models_bak.erase(models_bak.end() - 1);</span></span>
|
||
<span id="L219"><span class="lineNum"> 219</span> <span class="tlaGNC"> 120 : models.insert(models.begin(), std::move(model));</span></span>
|
||
<span id="L220"><span class="lineNum"> 220</span> : }</span>
|
||
<span id="L221"><span class="lineNum"> 221</span> : }</span>
|
||
<span id="L222"><span class="lineNum"> 222</span> : // 8. significances <- significances_bak</span>
|
||
<span id="L223"><span class="lineNum"> 223</span> <span class="tlaGNC"> 28 : significanceModels = significance_bak;</span></span>
|
||
<span id="L224"><span class="lineNum"> 224</span> : //</span>
|
||
<span id="L225"><span class="lineNum"> 225</span> : // Update the significance of the last k models</span>
|
||
<span id="L226"><span class="lineNum"> 226</span> : //</span>
|
||
<span id="L227"><span class="lineNum"> 227</span> : // 9. Update last k significances</span>
|
||
<span id="L228"><span class="lineNum"> 228</span> <span class="tlaGNC"> 104 : for (int i = 0; i < k; ++i) {</span></span>
|
||
<span id="L229"><span class="lineNum"> 229</span> <span class="tlaGNC"> 76 : significanceModels[n_models_bak - k + i] = alpha_t;</span></span>
|
||
<span id="L230"><span class="lineNum"> 230</span> : }</span>
|
||
<span id="L231"><span class="lineNum"> 231</span> : // 10. n_models <- n_models_bak</span>
|
||
<span id="L232"><span class="lineNum"> 232</span> <span class="tlaGNC"> 28 : n_models = n_models_bak;</span></span>
|
||
<span id="L233"><span class="lineNum"> 233</span> <span class="tlaGNC"> 56 : return { weights, alpha_t, terminate };</span></span>
|
||
<span id="L234"><span class="lineNum"> 234</span> <span class="tlaGNC"> 28 : }</span></span>
|
||
<span id="L235"><span class="lineNum"> 235</span> <span class="tlaGNC"> 32 : std::vector<int> BoostAODE::initializeModels()</span></span>
|
||
<span id="L236"><span class="lineNum"> 236</span> : {</span>
|
||
<span id="L237"><span class="lineNum"> 237</span> <span class="tlaGNC"> 32 : std::vector<int> featuresUsed;</span></span>
|
||
<span id="L238"><span class="lineNum"> 238</span> <span class="tlaGNC"> 32 : torch::Tensor weights_ = torch::full({ m }, 1.0 / m, torch::kFloat64);</span></span>
|
||
<span id="L239"><span class="lineNum"> 239</span> <span class="tlaGNC"> 32 : int maxFeatures = 0;</span></span>
|
||
<span id="L240"><span class="lineNum"> 240</span> <span class="tlaGNC"> 32 : if (select_features_algorithm == SelectFeatures.CFS) {</span></span>
|
||
<span id="L241"><span class="lineNum"> 241</span> <span class="tlaGNC"> 8 : featureSelector = new CFS(dataset, features, className, maxFeatures, states.at(className).size(), weights_);</span></span>
|
||
<span id="L242"><span class="lineNum"> 242</span> <span class="tlaGNC"> 24 : } else if (select_features_algorithm == SelectFeatures.IWSS) {</span></span>
|
||
<span id="L243"><span class="lineNum"> 243</span> <span class="tlaGNC"> 12 : if (threshold < 0 || threshold >0.5) {</span></span>
|
||
<span id="L244"><span class="lineNum"> 244</span> <span class="tlaGNC"> 8 : throw std::invalid_argument("Invalid threshold value for " + SelectFeatures.IWSS + " [0, 0.5]");</span></span>
|
||
<span id="L245"><span class="lineNum"> 245</span> : }</span>
|
||
<span id="L246"><span class="lineNum"> 246</span> <span class="tlaGNC"> 4 : featureSelector = new IWSS(dataset, features, className, maxFeatures, states.at(className).size(), weights_, threshold);</span></span>
|
||
<span id="L247"><span class="lineNum"> 247</span> <span class="tlaGNC"> 12 : } else if (select_features_algorithm == SelectFeatures.FCBF) {</span></span>
|
||
<span id="L248"><span class="lineNum"> 248</span> <span class="tlaGNC"> 12 : if (threshold < 1e-7 || threshold > 1) {</span></span>
|
||
<span id="L249"><span class="lineNum"> 249</span> <span class="tlaGNC"> 8 : throw std::invalid_argument("Invalid threshold value for " + SelectFeatures.FCBF + " [1e-7, 1]");</span></span>
|
||
<span id="L250"><span class="lineNum"> 250</span> : }</span>
|
||
<span id="L251"><span class="lineNum"> 251</span> <span class="tlaGNC"> 4 : featureSelector = new FCBF(dataset, features, className, maxFeatures, states.at(className).size(), weights_, threshold);</span></span>
|
||
<span id="L252"><span class="lineNum"> 252</span> : }</span>
|
||
<span id="L253"><span class="lineNum"> 253</span> <span class="tlaGNC"> 16 : featureSelector->fit();</span></span>
|
||
<span id="L254"><span class="lineNum"> 254</span> <span class="tlaGNC"> 16 : auto cfsFeatures = featureSelector->getFeatures();</span></span>
|
||
<span id="L255"><span class="lineNum"> 255</span> <span class="tlaGNC"> 16 : auto scores = featureSelector->getScores();</span></span>
|
||
<span id="L256"><span class="lineNum"> 256</span> <span class="tlaGNC"> 100 : for (const int& feature : cfsFeatures) {</span></span>
|
||
<span id="L257"><span class="lineNum"> 257</span> <span class="tlaGNC"> 84 : featuresUsed.push_back(feature);</span></span>
|
||
<span id="L258"><span class="lineNum"> 258</span> <span class="tlaGNC"> 84 : std::unique_ptr<Classifier> model = std::make_unique<SPODE>(feature);</span></span>
|
||
<span id="L259"><span class="lineNum"> 259</span> <span class="tlaGNC"> 84 : model->fit(dataset, features, className, states, weights_);</span></span>
|
||
<span id="L260"><span class="lineNum"> 260</span> <span class="tlaGNC"> 84 : models.push_back(std::move(model));</span></span>
|
||
<span id="L261"><span class="lineNum"> 261</span> <span class="tlaGNC"> 84 : significanceModels.push_back(1.0); // They will be updated later in trainModel</span></span>
|
||
<span id="L262"><span class="lineNum"> 262</span> <span class="tlaGNC"> 84 : n_models++;</span></span>
|
||
<span id="L263"><span class="lineNum"> 263</span> <span class="tlaGNC"> 84 : }</span></span>
|
||
<span id="L264"><span class="lineNum"> 264</span> <span class="tlaGNC"> 16 : notes.push_back("Used features in initialization: " + std::to_string(featuresUsed.size()) + " of " + std::to_string(features.size()) + " with " + select_features_algorithm);</span></span>
|
||
<span id="L265"><span class="lineNum"> 265</span> <span class="tlaGNC"> 16 : delete featureSelector;</span></span>
|
||
<span id="L266"><span class="lineNum"> 266</span> <span class="tlaGNC"> 32 : return featuresUsed;</span></span>
|
||
<span id="L267"><span class="lineNum"> 267</span> <span class="tlaGNC"> 48 : }</span></span>
|
||
<span id="L268"><span class="lineNum"> 268</span> <span class="tlaGNC"> 92 : void BoostAODE::trainModel(const torch::Tensor& weights)</span></span>
|
||
<span id="L269"><span class="lineNum"> 269</span> : {</span>
|
||
<span id="L270"><span class="lineNum"> 270</span> : //</span>
|
||
<span id="L271"><span class="lineNum"> 271</span> : // Logging setup</span>
|
||
<span id="L272"><span class="lineNum"> 272</span> : //</span>
|
||
<span id="L273"><span class="lineNum"> 273</span> <span class="tlaGNC"> 92 : loguru::set_thread_name("BoostAODE");</span></span>
|
||
<span id="L274"><span class="lineNum"> 274</span> <span class="tlaGNC"> 92 : loguru::g_stderr_verbosity = loguru::Verbosity_OFF;</span></span>
|
||
<span id="L275"><span class="lineNum"> 275</span> <span class="tlaGNC"> 92 : loguru::add_file("boostAODE.log", loguru::Truncate, loguru::Verbosity_MAX);</span></span>
|
||
<span id="L276"><span class="lineNum"> 276</span> : </span>
|
||
<span id="L277"><span class="lineNum"> 277</span> : // Algorithm based on the adaboost algorithm for classification</span>
|
||
<span id="L278"><span class="lineNum"> 278</span> : // as explained in Ensemble methods (Zhi-Hua Zhou, 2012)</span>
|
||
<span id="L279"><span class="lineNum"> 279</span> <span class="tlaGNC"> 92 : fitted = true;</span></span>
|
||
<span id="L280"><span class="lineNum"> 280</span> <span class="tlaGNC"> 92 : double alpha_t = 0;</span></span>
|
||
<span id="L281"><span class="lineNum"> 281</span> <span class="tlaGNC"> 92 : torch::Tensor weights_ = torch::full({ m }, 1.0 / m, torch::kFloat64);</span></span>
|
||
<span id="L282"><span class="lineNum"> 282</span> <span class="tlaGNC"> 92 : bool finished = false;</span></span>
|
||
<span id="L283"><span class="lineNum"> 283</span> <span class="tlaGNC"> 92 : std::vector<int> featuresUsed;</span></span>
|
||
<span id="L284"><span class="lineNum"> 284</span> <span class="tlaGNC"> 92 : if (selectFeatures) {</span></span>
|
||
<span id="L285"><span class="lineNum"> 285</span> <span class="tlaGNC"> 32 : featuresUsed = initializeModels();</span></span>
|
||
<span id="L286"><span class="lineNum"> 286</span> <span class="tlaGNC"> 16 : auto ypred = predict(X_train);</span></span>
|
||
<span id="L287"><span class="lineNum"> 287</span> <span class="tlaGNC"> 16 : std::tie(weights_, alpha_t, finished) = update_weights(y_train, ypred, weights_);</span></span>
|
||
<span id="L288"><span class="lineNum"> 288</span> : // Update significance of the models</span>
|
||
<span id="L289"><span class="lineNum"> 289</span> <span class="tlaGNC"> 100 : for (int i = 0; i < n_models; ++i) {</span></span>
|
||
<span id="L290"><span class="lineNum"> 290</span> <span class="tlaGNC"> 84 : significanceModels[i] = alpha_t;</span></span>
|
||
<span id="L291"><span class="lineNum"> 291</span> : }</span>
|
||
<span id="L292"><span class="lineNum"> 292</span> <span class="tlaGNC"> 16 : if (finished) {</span></span>
|
||
<span id="L293"><span class="lineNum"> 293</span> <span class="tlaUNC tlaBgUNC"> 0 : return;</span></span>
|
||
<span id="L294"><span class="lineNum"> 294</span> : }</span>
|
||
<span id="L295"><span class="lineNum"> 295</span> <span class="tlaGNC tlaBgGNC"> 16 : }</span></span>
|
||
<span id="L296"><span class="lineNum"> 296</span> <span class="tlaGNC"> 76 : int numItemsPack = 0; // The counter of the models inserted in the current pack</span></span>
|
||
<span id="L297"><span class="lineNum"> 297</span> : // Variables to control the accuracy finish condition</span>
|
||
<span id="L298"><span class="lineNum"> 298</span> <span class="tlaGNC"> 76 : double priorAccuracy = 0.0;</span></span>
|
||
<span id="L299"><span class="lineNum"> 299</span> <span class="tlaGNC"> 76 : double improvement = 1.0;</span></span>
|
||
<span id="L300"><span class="lineNum"> 300</span> <span class="tlaGNC"> 76 : double convergence_threshold = 1e-4;</span></span>
|
||
<span id="L301"><span class="lineNum"> 301</span> <span class="tlaGNC"> 76 : int tolerance = 0; // number of times the accuracy is lower than the convergence_threshold</span></span>
|
||
<span id="L302"><span class="lineNum"> 302</span> : // Step 0: Set the finish condition</span>
|
||
<span id="L303"><span class="lineNum"> 303</span> : // epsilon sub t > 0.5 => inverse the weights policy</span>
|
||
<span id="L304"><span class="lineNum"> 304</span> : // validation error is not decreasing</span>
|
||
<span id="L305"><span class="lineNum"> 305</span> : // run out of features</span>
|
||
<span id="L306"><span class="lineNum"> 306</span> <span class="tlaGNC"> 76 : bool ascending = order_algorithm == Orders.ASC;</span></span>
|
||
<span id="L307"><span class="lineNum"> 307</span> <span class="tlaGNC"> 76 : std::mt19937 g{ 173 };</span></span>
|
||
<span id="L308"><span class="lineNum"> 308</span> <span class="tlaGNC"> 504 : while (!finished) {</span></span>
|
||
<span id="L309"><span class="lineNum"> 309</span> : // Step 1: Build ranking with mutual information</span>
|
||
<span id="L310"><span class="lineNum"> 310</span> <span class="tlaGNC"> 428 : auto featureSelection = metrics.SelectKBestWeighted(weights_, ascending, n); // Get all the features sorted</span></span>
|
||
<span id="L311"><span class="lineNum"> 311</span> <span class="tlaGNC"> 428 : if (order_algorithm == Orders.RAND) {</span></span>
|
||
<span id="L312"><span class="lineNum"> 312</span> <span class="tlaGNC"> 36 : std::shuffle(featureSelection.begin(), featureSelection.end(), g);</span></span>
|
||
<span id="L313"><span class="lineNum"> 313</span> : }</span>
|
||
<span id="L314"><span class="lineNum"> 314</span> : // Remove used features</span>
|
||
<span id="L315"><span class="lineNum"> 315</span> <span class="tlaGNC"> 856 : featureSelection.erase(remove_if(begin(featureSelection), end(featureSelection), [&](auto x)</span></span>
|
||
<span id="L316"><span class="lineNum"> 316</span> <span class="tlaGNC"> 38800 : { return std::find(begin(featuresUsed), end(featuresUsed), x) != end(featuresUsed);}),</span></span>
|
||
<span id="L317"><span class="lineNum"> 317</span> <span class="tlaGNC"> 428 : end(featureSelection)</span></span>
|
||
<span id="L318"><span class="lineNum"> 318</span> : );</span>
|
||
<span id="L319"><span class="lineNum"> 319</span> <span class="tlaGNC"> 428 : int k = bisection ? pow(2, tolerance) : 1;</span></span>
|
||
<span id="L320"><span class="lineNum"> 320</span> <span class="tlaGNC"> 428 : int counter = 0; // The model counter of the current pack</span></span>
|
||
<span id="L321"><span class="lineNum"> 321</span> <span class="tlaGNC"> 428 : VLOG_SCOPE_F(1, "counter=%d k=%d featureSelection.size: %zu", counter, k, featureSelection.size());</span></span>
|
||
<span id="L322"><span class="lineNum"> 322</span> <span class="tlaGNC"> 1004 : while (counter++ < k && featureSelection.size() > 0) {</span></span>
|
||
<span id="L323"><span class="lineNum"> 323</span> <span class="tlaGNC"> 576 : auto feature = featureSelection[0];</span></span>
|
||
<span id="L324"><span class="lineNum"> 324</span> <span class="tlaGNC"> 576 : featureSelection.erase(featureSelection.begin());</span></span>
|
||
<span id="L325"><span class="lineNum"> 325</span> <span class="tlaGNC"> 576 : std::unique_ptr<Classifier> model;</span></span>
|
||
<span id="L326"><span class="lineNum"> 326</span> <span class="tlaGNC"> 576 : model = std::make_unique<SPODE>(feature);</span></span>
|
||
<span id="L327"><span class="lineNum"> 327</span> <span class="tlaGNC"> 576 : model->fit(dataset, features, className, states, weights_);</span></span>
|
||
<span id="L328"><span class="lineNum"> 328</span> <span class="tlaGNC"> 576 : alpha_t = 0.0;</span></span>
|
||
<span id="L329"><span class="lineNum"> 329</span> <span class="tlaGNC"> 576 : if (!block_update) {</span></span>
|
||
<span id="L330"><span class="lineNum"> 330</span> <span class="tlaGNC"> 500 : auto ypred = model->predict(X_train);</span></span>
|
||
<span id="L331"><span class="lineNum"> 331</span> : // Step 3.1: Compute the classifier amout of say</span>
|
||
<span id="L332"><span class="lineNum"> 332</span> <span class="tlaGNC"> 500 : std::tie(weights_, alpha_t, finished) = update_weights(y_train, ypred, weights_);</span></span>
|
||
<span id="L333"><span class="lineNum"> 333</span> <span class="tlaGNC"> 500 : }</span></span>
|
||
<span id="L334"><span class="lineNum"> 334</span> : // Step 3.4: Store classifier and its accuracy to weigh its future vote</span>
|
||
<span id="L335"><span class="lineNum"> 335</span> <span class="tlaGNC"> 576 : numItemsPack++;</span></span>
|
||
<span id="L336"><span class="lineNum"> 336</span> <span class="tlaGNC"> 576 : featuresUsed.push_back(feature);</span></span>
|
||
<span id="L337"><span class="lineNum"> 337</span> <span class="tlaGNC"> 576 : models.push_back(std::move(model));</span></span>
|
||
<span id="L338"><span class="lineNum"> 338</span> <span class="tlaGNC"> 576 : significanceModels.push_back(alpha_t);</span></span>
|
||
<span id="L339"><span class="lineNum"> 339</span> <span class="tlaGNC"> 576 : n_models++;</span></span>
|
||
<span id="L340"><span class="lineNum"> 340</span> <span class="tlaGNC"> 576 : VLOG_SCOPE_F(2, "numItemsPack: %d n_models: %d featuresUsed: %zu", numItemsPack, n_models, featuresUsed.size());</span></span>
|
||
<span id="L341"><span class="lineNum"> 341</span> <span class="tlaGNC"> 576 : }</span></span>
|
||
<span id="L342"><span class="lineNum"> 342</span> <span class="tlaGNC"> 428 : if (block_update) {</span></span>
|
||
<span id="L343"><span class="lineNum"> 343</span> <span class="tlaGNC"> 28 : std::tie(weights_, alpha_t, finished) = update_weights_block(k, y_train, weights_);</span></span>
|
||
<span id="L344"><span class="lineNum"> 344</span> : }</span>
|
||
<span id="L345"><span class="lineNum"> 345</span> <span class="tlaGNC"> 428 : if (convergence && !finished) {</span></span>
|
||
<span id="L346"><span class="lineNum"> 346</span> <span class="tlaGNC"> 296 : auto y_val_predict = predict(X_test);</span></span>
|
||
<span id="L347"><span class="lineNum"> 347</span> <span class="tlaGNC"> 296 : double accuracy = (y_val_predict == y_test).sum().item<double>() / (double)y_test.size(0);</span></span>
|
||
<span id="L348"><span class="lineNum"> 348</span> <span class="tlaGNC"> 296 : if (priorAccuracy == 0) {</span></span>
|
||
<span id="L349"><span class="lineNum"> 349</span> <span class="tlaGNC"> 60 : priorAccuracy = accuracy;</span></span>
|
||
<span id="L350"><span class="lineNum"> 350</span> : } else {</span>
|
||
<span id="L351"><span class="lineNum"> 351</span> <span class="tlaGNC"> 236 : improvement = accuracy - priorAccuracy;</span></span>
|
||
<span id="L352"><span class="lineNum"> 352</span> : }</span>
|
||
<span id="L353"><span class="lineNum"> 353</span> <span class="tlaGNC"> 296 : if (improvement < convergence_threshold) {</span></span>
|
||
<span id="L354"><span class="lineNum"> 354</span> <span class="tlaGNC"> 176 : VLOG_SCOPE_F(3, " (improvement<threshold) tolerance: %d numItemsPack: %d improvement: %f prior: %f current: %f", tolerance, numItemsPack, improvement, priorAccuracy, accuracy);</span></span>
|
||
<span id="L355"><span class="lineNum"> 355</span> <span class="tlaGNC"> 176 : tolerance++;</span></span>
|
||
<span id="L356"><span class="lineNum"> 356</span> <span class="tlaGNC"> 176 : } else {</span></span>
|
||
<span id="L357"><span class="lineNum"> 357</span> <span class="tlaGNC"> 120 : VLOG_SCOPE_F(3, "* (improvement>=threshold) Reset. tolerance: %d numItemsPack: %d improvement: %f prior: %f current: %f", tolerance, numItemsPack, improvement, priorAccuracy, accuracy);</span></span>
|
||
<span id="L358"><span class="lineNum"> 358</span> <span class="tlaGNC"> 120 : tolerance = 0; // Reset the counter if the model performs better</span></span>
|
||
<span id="L359"><span class="lineNum"> 359</span> <span class="tlaGNC"> 120 : numItemsPack = 0;</span></span>
|
||
<span id="L360"><span class="lineNum"> 360</span> <span class="tlaGNC"> 120 : }</span></span>
|
||
<span id="L361"><span class="lineNum"> 361</span> <span class="tlaGNC"> 296 : if (convergence_best) {</span></span>
|
||
<span id="L362"><span class="lineNum"> 362</span> : // Keep the best accuracy until now as the prior accuracy</span>
|
||
<span id="L363"><span class="lineNum"> 363</span> <span class="tlaGNC"> 32 : priorAccuracy = std::max(accuracy, priorAccuracy);</span></span>
|
||
<span id="L364"><span class="lineNum"> 364</span> : } else {</span>
|
||
<span id="L365"><span class="lineNum"> 365</span> : // Keep the last accuray obtained as the prior accuracy</span>
|
||
<span id="L366"><span class="lineNum"> 366</span> <span class="tlaGNC"> 264 : priorAccuracy = accuracy;</span></span>
|
||
<span id="L367"><span class="lineNum"> 367</span> : }</span>
|
||
<span id="L368"><span class="lineNum"> 368</span> <span class="tlaGNC"> 296 : }</span></span>
|
||
<span id="L369"><span class="lineNum"> 369</span> <span class="tlaGNC"> 428 : VLOG_SCOPE_F(1, "tolerance: %d featuresUsed.size: %zu features.size: %zu", tolerance, featuresUsed.size(), features.size());</span></span>
|
||
<span id="L370"><span class="lineNum"> 370</span> <span class="tlaGNC"> 428 : finished = finished || tolerance > maxTolerance || featuresUsed.size() == features.size();</span></span>
|
||
<span id="L371"><span class="lineNum"> 371</span> <span class="tlaGNC"> 428 : }</span></span>
|
||
<span id="L372"><span class="lineNum"> 372</span> <span class="tlaGNC"> 76 : if (tolerance > maxTolerance) {</span></span>
|
||
<span id="L373"><span class="lineNum"> 373</span> <span class="tlaGNC"> 8 : if (numItemsPack < n_models) {</span></span>
|
||
<span id="L374"><span class="lineNum"> 374</span> <span class="tlaGNC"> 8 : notes.push_back("Convergence threshold reached & " + std::to_string(numItemsPack) + " models eliminated");</span></span>
|
||
<span id="L375"><span class="lineNum"> 375</span> <span class="tlaGNC"> 8 : VLOG_SCOPE_F(4, "Convergence threshold reached & %d models eliminated of %d", numItemsPack, n_models);</span></span>
|
||
<span id="L376"><span class="lineNum"> 376</span> <span class="tlaGNC"> 104 : for (int i = 0; i < numItemsPack; ++i) {</span></span>
|
||
<span id="L377"><span class="lineNum"> 377</span> <span class="tlaGNC"> 96 : significanceModels.pop_back();</span></span>
|
||
<span id="L378"><span class="lineNum"> 378</span> <span class="tlaGNC"> 96 : models.pop_back();</span></span>
|
||
<span id="L379"><span class="lineNum"> 379</span> <span class="tlaGNC"> 96 : n_models--;</span></span>
|
||
<span id="L380"><span class="lineNum"> 380</span> : }</span>
|
||
<span id="L381"><span class="lineNum"> 381</span> <span class="tlaGNC"> 8 : } else {</span></span>
|
||
<span id="L382"><span class="lineNum"> 382</span> <span class="tlaUNC tlaBgUNC"> 0 : notes.push_back("Convergence threshold reached & 0 models eliminated");</span></span>
|
||
<span id="L383"><span class="lineNum"> 383</span> <span class="tlaUNC"> 0 : VLOG_SCOPE_F(4, "Convergence threshold reached & 0 models eliminated n_models=%d numItemsPack=%d", n_models, numItemsPack);</span></span>
|
||
<span id="L384"><span class="lineNum"> 384</span> <span class="tlaUNC"> 0 : }</span></span>
|
||
<span id="L385"><span class="lineNum"> 385</span> : }</span>
|
||
<span id="L386"><span class="lineNum"> 386</span> <span class="tlaGNC tlaBgGNC"> 76 : if (featuresUsed.size() != features.size()) {</span></span>
|
||
<span id="L387"><span class="lineNum"> 387</span> <span class="tlaGNC"> 4 : notes.push_back("Used features in train: " + std::to_string(featuresUsed.size()) + " of " + std::to_string(features.size()));</span></span>
|
||
<span id="L388"><span class="lineNum"> 388</span> <span class="tlaGNC"> 4 : status = WARNING;</span></span>
|
||
<span id="L389"><span class="lineNum"> 389</span> : }</span>
|
||
<span id="L390"><span class="lineNum"> 390</span> <span class="tlaGNC"> 76 : notes.push_back("Number of models: " + std::to_string(n_models));</span></span>
|
||
<span id="L391"><span class="lineNum"> 391</span> <span class="tlaGNC"> 108 : }</span></span>
|
||
<span id="L392"><span class="lineNum"> 392</span> <span class="tlaGNC"> 4 : std::vector<std::string> BoostAODE::graph(const std::string& title) const</span></span>
|
||
<span id="L393"><span class="lineNum"> 393</span> : {</span>
|
||
<span id="L394"><span class="lineNum"> 394</span> <span class="tlaGNC"> 4 : return Ensemble::graph(title);</span></span>
|
||
<span id="L395"><span class="lineNum"> 395</span> : }</span>
|
||
<span id="L396"><span class="lineNum"> 396</span> : }</span>
|
||
</pre>
|
||
</td>
|
||
</tr>
|
||
</table>
|
||
<br>
|
||
|
||
<table width="100%" border=0 cellspacing=0 cellpadding=0>
|
||
<tr><td class="ruler"><img src="../../glass.png" width=3 height=3 alt=""></td></tr>
|
||
<tr><td class="versionInfo">Generated by: <a href="https://github.com//linux-test-project/lcov" target="_parent">LCOV version 2.0-1</a></td></tr>
|
||
</table>
|
||
<br>
|
||
|
||
</body>
|
||
</html>
|