All Versions
15
Latest Version
Avg Release Cycle
101 days
Latest Release
525 days ago

Changelog History
Page 1

  • v0.9.0 Changes

    May 15, 2020

    βž• Added

    • [Preprocessing] Implement LabelEncoder (#369)
    • [Preprocessing] Implement ColumnFilter (#378)
    • [Preprocessing] Implement LambdaTransformer (#381)
    • [Preprocessing] Implement NumberConverter (#377)
    • [Preprocessing] Implement OneHotEncoder (#384)
    • [Workflow] Implement FeatureUnion (#382)
    • [Metric] Add Regression metrics: meanSquaredError, meanSquaredLogarithmicError, meanAbsoluteError, medianAbsoluteError, r2Score, maxError (#373)
    • [Regression] Implement DecisionTreeRegressor (#375)
    • [Tokenization] Added NGramWordTokenizer (#370)
    • πŸ— [Internal] Add php 7.4 for travis build (#393)
    • [FeatureExtraction] FeatureExtraction (#425)

    πŸ›  Fixed

    • [Math] Fix division by zero in ANOVA for small size dataset (#391)
    • πŸ“„ [Internal] Fix grammatical errors in code and docs (#410, #411, #413, #414)

    πŸ”„ Changed

    • ⚑️ [Internal] update phpunit to 8 (#367, #372)
    • πŸ’… [Internal] code style fixes (#366)
    • ⚑️ [Internal] update dependencies (#383, #415, #419, #426)
  • v0.8.0 Changes

    March 20, 2019

    βž• Added

    • [Tokenization] Added NGramTokenizer (#350)
    • editorconfig file (#355)

    πŸ›  Fixed

    • [Dataset] FilesDataset read samples without additional array (#363)
    • πŸ›  [Tokenization] fixed error with numeric token values (#363)

    πŸ”„ Changed

    • 🐎 [Math] improved performance with pow and sqrt replacement (#350)
    • [Math] reduce duplicated code in distance metrics (#348)
    • ⚑️ update phpunit to 7.5.1 (#335)
    • πŸ’… code style fixes (#334)
  • v0.7.0 Changes

    November 07, 2018

    βž• Added

    • [Clustering] added KMeans associative clustering (#262)
    • 🚚 [Dataset] added removeColumns function to ArrayDataset (#249)
    • [Dataset] added a SvmDataset class for SVM-Light (or LibSVM) format files (#237)
    • [Dataset] added Mnist Dataset for MNIST file format (#326)
    • 🐎 [Internal] Add performance test for LeastSquares (#263)

    πŸ”„ Changed

    • [Internal] implement Keep a Changelog format
    • 0️⃣ [Classification] changed the default kernel type in SVC to Kernel::RBF (#267)
    • ⚑️ [Optimizer] removed $initialTheta property and renamed setInitialTheta method to setTheta (#252)
    • πŸ‘» [Imputer] Throw exception when trying to transform without train data (#314)
    • [Math] Micro optimization for matrix multiplication (#255)
    • πŸ‘» [Internal] Throw proper exception (#259, #251)
    • [MLPClassifier] return labels in output (#315)
    • ⚑️ [Internal] Update phpstan to 0.10.5 (#320)

    πŸ›  Fixed

    • βœ… [SVM] ensure DataTransformer::testSet samples array is not empty (#204)
    • ⚑️ [Optimizer] optimizer initial theta randomization (#239)
    • πŸ— [Internal] travis build on osx (#281)
    • [SVM] SVM locale (non-locale aware) (#288)
    • πŸ“š [Internal] typo, tests, code styles and documentation fixes (#265, #261, #254, #253, #251, #250, #248, #245, #243, #317, #328)
    • [Classification] Check if feature exist when predict target in NaiveBayes (#327)
  • v0.6.2 Changes

    February 22, 2018
    • πŸ›  Fix Apriori array keys (#238)
  • v0.6.1 Changes

    February 17, 2018
    • πŸ›  Fix KMeans and EigenvalueDecomposition #235
  • v0.6.0 Changes

    February 16, 2018
    • πŸ”‹ feature [FeatureSelection] implement SelectKBest with scoring functions (#232)
    • πŸ”‹ feature [FeatureSelection] implement VarianceThreshold - simple baseline approach to feature selection. (#228)
    • πŸ”‹ feature [Classification] support probability estimation in SVC (#218)
    • πŸ”‹ feature [NeuralNetwork] configure an Activation Function per hidden layer (#208)
    • πŸ”‹ feature [NeuralNetwork] Ability to update learningRate in MLP (#160)
    • πŸ”‹ feature [Metric] Choose averaging method in classification report (#205)
    • ✨ enhancement Add phpstan strict rules (#233)
    • ✨ enhancement Flatten directory structure (#220)
    • ✨ enhancement Update phpunit/phpunit (#219)
    • ✨ enhancement Cache dependencies installed with composer on Travis (#215)
    • ✨ enhancement Add support for coveralls.io (#153)
    • ✨ enhancement Add phpstan and easy coding standards (#156, #168)
    • ✨ enhancement Throw exception when libsvm command fails to run (#200, #202)
    • ✨ enhancement Normalize composer.json and sort packages (#214, #210)
    • ✨ enhancement Rewrite DBSCAN (#185)
    • πŸ›  fix phpunit include tests path (#230)
    • πŸ›  fix support of a rule in Apriori (#229)
    • πŸ›  fix apriori generates an empty array as a part of the frequent item sets (#224)
    • πŸ›  fix backpropagation random error (#157)
    • πŸ›  fix logistic regression implementation (#169)
    • πŸ›  fix activation functions support (#163)
    • πŸ›  fix string representation of integer labels issue in NaiveBayes (#206)
    • πŸ›  fix the implementation of conjugate gradient method (#184)
    • πŸ“š typo, tests and documentation fixes (#234, #221, #181, #183, #155, #159, #165, #187, #154, #191, #203, #209, #213, #212, #211)
  • v0.5.0 Changes

    November 14, 2017
    • ⬆️ general [php] Upgrade to PHP 7.1 (#150)
    • πŸ“„ general [coding standard] fix imports order and drop unused docs typehints
    • πŸ”‹ feature [NeuralNetwork] Add PReLU activation function (#128)
    • πŸ”‹ feature [NeuralNetwork] Add ThresholdedReLU activation function (#129)
    • πŸ”‹ feature [Dataset] Support CSV with long lines (#119)
    • πŸ”‹ feature [NeuralNetwork] Neural networks partial training and persistency (#91)
    • πŸ”‹ feature Add french stopwords (#92)
    • πŸ”‹ feature New methods: setBinPath, setVarPath in SupportVectorMachine (#73)
    • πŸ”‹ feature Linear Discrimant Analysis (LDA) (#82)
    • πŸ”‹ feature Linear algebra operations, Dimensionality reduction and some other minor changes (#81)
    • πŸ”‹ feature Partial training base (#78)
    • πŸ”‹ feature Add delimiter option for CsvDataset (#66)
    • πŸ”‹ feature LogisticRegression classifier & Optimization methods (#63)
    • πŸ”‹ feature Additional training for SVR (#59)
    • optimization Comparison - replace eval (#130)
    • πŸ’… optimization Use C-style casts (#124)
    • optimization Speed up DataTransformer (#122)
    • πŸ› bug DBSCAN fix for associative keys and array_merge performance optimization (#139)
    • πŸ› bug Ensure user-provided SupportVectorMachine paths are valid (#126)
    • πŸ› bug [DecisionTree] Fix string cast #120 (#121)
    • πŸ› bug fix invalid typehint for subs method (#110)
    • πŸ› bug Fix samples transformation in Pipeline training (#94)
    • πŸ› bug Fix division by 0 error during normalization (#83)
    • πŸ› bug Fix wrong docs references (#79)
  • v0.4.1 Changes

    March 05, 2017
    • AdaBoost improvements by Mustafa Karabulut
    • One-v-Rest Classification technique applied to linear classifiers by Mustafa Karabulut
  • v0.4 Changes

    February 23, 2017
    • πŸ”‹ feature [Classification] - Ensemble Classifiers : Bagging and RandomForest by Mustafa Karabulut
    • πŸ”‹ feature [Classification] - RandomForest::getFeatureImportances() method by Mustafa Karabulut
    • πŸ”‹ feature [Classification] - Linear classifiers: Perceptron, Adaline, DecisionStump by Mustafa Karabulut
    • πŸ”‹ feature [Classification] - AdaBoost algorithm by Mustafa Karabulut
    • πŸ› bug [Math] - Check if matrix is singular doing inverse by Povilas Susinskas
    • optimization - Euclidean optimization by Mustafa Karabulut
  • v0.3.0 Changes

    February 04, 2017
    * feature [Persistency] - ModelManager - save and restore trained models by David MonllaΓ³
    * feature [Classification] - DecisionTree implementation by Mustafa Karabulut
    * feature [Clustering] - Fuzzy C Means implementation by Mustafa Karabulut
    * other small fixes and code styles refactors