These results are based on our legacy PHP analysis, consider migrating to our new PHP analysis engine instead. Learn more
1 | <?php |
||
2 | |||
3 | declare(strict_types=1); |
||
4 | |||
5 | namespace Phpml\Helper; |
||
6 | |||
7 | use Phpml\Classification\Classifier; |
||
8 | |||
9 | trait OneVsRest |
||
10 | { |
||
11 | /** |
||
12 | * @var array |
||
13 | */ |
||
14 | protected $classifiers = []; |
||
15 | |||
16 | /** |
||
17 | * All provided training targets' labels. |
||
18 | * |
||
19 | * @var array |
||
20 | */ |
||
21 | protected $allLabels = []; |
||
22 | |||
23 | /** |
||
24 | * @var array |
||
25 | */ |
||
26 | protected $costValues = []; |
||
27 | |||
28 | /** |
||
29 | * Train a binary classifier in the OvR style |
||
30 | */ |
||
31 | public function train(array $samples, array $targets): void |
||
32 | { |
||
33 | // Clears previous stuff. |
||
34 | $this->reset(); |
||
35 | |||
36 | $this->trainByLabel($samples, $targets); |
||
37 | } |
||
38 | |||
39 | /** |
||
40 | * Resets the classifier and the vars internally used by OneVsRest to create multiple classifiers. |
||
41 | */ |
||
42 | public function reset(): void |
||
43 | { |
||
44 | $this->classifiers = []; |
||
45 | $this->allLabels = []; |
||
46 | $this->costValues = []; |
||
47 | |||
48 | $this->resetBinary(); |
||
49 | } |
||
50 | |||
51 | protected function trainByLabel(array $samples, array $targets, array $allLabels = []): void |
||
52 | { |
||
53 | // Overwrites the current value if it exist. $allLabels must be provided for each partialTrain run. |
||
54 | if (!empty($allLabels)) { |
||
55 | $this->allLabels = $allLabels; |
||
56 | } else { |
||
57 | $this->allLabels = array_keys(array_count_values($targets)); |
||
58 | } |
||
59 | |||
60 | sort($this->allLabels, SORT_STRING); |
||
61 | |||
62 | // If there are only two targets, then there is no need to perform OvR |
||
63 | if (count($this->allLabels) == 2) { |
||
64 | // Init classifier if required. |
||
65 | if (empty($this->classifiers)) { |
||
66 | $this->classifiers[0] = $this->getClassifierCopy(); |
||
67 | } |
||
68 | |||
69 | $this->classifiers[0]->trainBinary($samples, $targets, $this->allLabels); |
||
70 | } else { |
||
71 | // Train a separate classifier for each label and memorize them |
||
72 | |||
73 | foreach ($this->allLabels as $label) { |
||
74 | // Init classifier if required. |
||
75 | if (empty($this->classifiers[$label])) { |
||
76 | $this->classifiers[$label] = $this->getClassifierCopy(); |
||
77 | } |
||
78 | |||
79 | [$binarizedTargets, $classifierLabels] = $this->binarizeTargets($targets, $label); |
||
80 | $this->classifiers[$label]->trainBinary($samples, $binarizedTargets, $classifierLabels); |
||
0 ignored issues
–
show
|
|||
81 | } |
||
82 | } |
||
83 | |||
84 | // If the underlying classifier is capable of giving the cost values |
||
85 | // during the training, then assign it to the relevant variable |
||
86 | // Adding just the first classifier cost values to avoid complex average calculations. |
||
87 | $classifierref = reset($this->classifiers); |
||
88 | if (method_exists($classifierref, 'getCostValues')) { |
||
89 | $this->costValues = $classifierref->getCostValues(); |
||
90 | } |
||
91 | } |
||
92 | |||
93 | /** |
||
94 | * Returns an instance of the current class after cleaning up OneVsRest stuff. |
||
95 | * |
||
96 | * @return Classifier|OneVsRest |
||
97 | */ |
||
98 | protected function getClassifierCopy() |
||
99 | { |
||
100 | // Clone the current classifier, so that |
||
101 | // we don't mess up its variables while training |
||
102 | // multiple instances of this classifier |
||
103 | $classifier = clone $this; |
||
104 | $classifier->reset(); |
||
105 | |||
106 | return $classifier; |
||
107 | } |
||
108 | |||
109 | /** |
||
110 | * @return mixed |
||
111 | */ |
||
112 | protected function predictSample(array $sample) |
||
113 | { |
||
114 | if (count($this->allLabels) == 2) { |
||
115 | return $this->classifiers[0]->predictSampleBinary($sample); |
||
116 | } |
||
117 | |||
118 | $probs = []; |
||
119 | |||
120 | foreach ($this->classifiers as $label => $predictor) { |
||
121 | $probs[$label] = $predictor->predictProbability($sample, $label); |
||
122 | } |
||
123 | |||
124 | arsort($probs, SORT_NUMERIC); |
||
125 | |||
126 | return key($probs); |
||
127 | } |
||
128 | |||
129 | /** |
||
130 | * Each classifier should implement this method instead of train(samples, targets) |
||
131 | */ |
||
132 | abstract protected function trainBinary(array $samples, array $targets, array $labels); |
||
133 | |||
134 | /** |
||
135 | * To be overwritten by OneVsRest classifiers. |
||
136 | */ |
||
137 | abstract protected function resetBinary(): void; |
||
138 | |||
139 | /** |
||
140 | * Each classifier that make use of OvR approach should be able to |
||
141 | * return a probability for a sample to belong to the given label. |
||
142 | * |
||
143 | * @return mixed |
||
144 | */ |
||
145 | abstract protected function predictProbability(array $sample, string $label); |
||
146 | |||
147 | /** |
||
148 | * Each classifier should implement this method instead of predictSample() |
||
149 | * |
||
150 | * @return mixed |
||
151 | */ |
||
152 | abstract protected function predictSampleBinary(array $sample); |
||
153 | |||
154 | /** |
||
155 | * Groups all targets into two groups: Targets equal to |
||
156 | * the given label and the others |
||
157 | * |
||
158 | * $targets is not passed by reference nor contains objects so this method |
||
159 | * changes will not affect the caller $targets array. |
||
160 | * |
||
161 | * @param mixed $label |
||
162 | * |
||
163 | * @return array Binarized targets and target's labels |
||
164 | */ |
||
165 | private function binarizeTargets(array $targets, $label): array |
||
166 | { |
||
167 | $notLabel = "not_${label}"; |
||
168 | foreach ($targets as $key => $target) { |
||
169 | $targets[$key] = $target == $label ? $label : $notLabel; |
||
170 | } |
||
171 | |||
172 | $labels = [$label, $notLabel]; |
||
173 | |||
174 | return [$targets, $labels]; |
||
175 | } |
||
176 | } |
||
177 |
This check looks for variables that are accessed but have not been defined. It raises an issue if it finds another variable that has a similar name.
The variable may have been renamed without also renaming all references.