Kalimdor.js

Kalimdor.js is a Machine Learning library written in Typescript. It solves Machine Learning problems and teaches users how Machine Learning algorithms work.

User Installation

Using yarn

$ yarn add kalimdor

Using NPM

$ npm install --save kalimdor

Highlights

• Machine Learning on the browser and Node.js
• Learning APIs for users
• Low entry barrier

Development

We welcome new contributors of all level of experience. The development guide will be added to assist new contributors to easily join the project.

• You want to participate in a Machine Learning project, which will boost your Machine Learning skills and knowledge
• Looking to be part of a growing community
• You want to learn Machine Learning
• You like Typescript :heart: Machine Learning

Simplicity

Kalimdor.js provides a simple and consistent set of APIs to interact with the models and algorithms. For example, all models have follow APIs:

• fit for training
• predict for inferencing
• toJSON for saving the model's state
• fromJSON for loading the model from the checkpoint

Testing

Testing ensures you that you are currently using the most stable version of Kalimdor.js

$ npm run test

Supporting

Simply give us a :star2: by clicking on

Contributing

We simply follow "fork-and-pull" workflow of Github. Please read CONTRIBUTING.md for more detail.

Further notice

Great references that helped building this project!

• https://machinelearningmastery.com/
• https://github.com/mljs/ml
• http://scikit-learn.org/stable/documentation.html

Contributors

Thanks goes to these wonderful people (emoji key):

Jason Shin 📝 🐛 💻 📖 ⚠️	Jaivarsan 💬 🤔 📢	Oleg Stotsky 🐛 💻 📖 ⚠️	Ben 💬 🎨 📢 🐛 💻	Christoph Reinbothe 💻 🤔 🚇 👀	Adam King 💻 ⚠️ 📖

Change Log

All notable changes to this project will be documented in this file.

[2.1.4]

:bug: Browser bundle fix :bug:

• fix: Browser bundle is now correctly working but it now excludes SVM and dataset APIs during the bundling process

[2.1.3]

:feature: new API! :feature:

• feature: BaggingClassifier added by @OlegStotsky congratulations!

[2.1.2]

:bug: Installation issues :bug:

• fix: SVM APIs are now based on libsvm-ts, a new module managed by machinelearn.js! It solves various issues and introduces friendlier/easy-to-use APIs
• fix: fs-extra is no longer a production dependency

[2.1.1]

:rocket: new API! :rocket:

• feature: mean_squared_log_error by @VariableVasasMT thanks!
• enhance: Enhanced CI using AzurePipelines

[2.1.0]

:rocket: New API! :rocket:

• feature: Logistic Regression by @OlegStotsky
• feature: Heart Disease dataset by @OlegStotsky
• enhance: Various internal code enhancement

[2.0.2]

:bug: API consistency update :bug:

• enhance: MinMaxScaler now implements transform and fit_transform correctly accroding to ScikitLearn MinMaxScaler

[2.0.1]

:rocket: feature implementations :rocket:

• feature: Multivariate Linear Regression

[2.0.0]

:rocket: Version 2 release! :rocket:

• enhance: PCA refactored to tfjs-core
• enhance: GaussianNB is now refactored to tfjs-core @benjaminmcdonald thanks!!
• feature: MultinomialNB written in tfjs-core @benjaminmcdonald thanks again!

[1.2.6]

:rocket: enhancing typing system for SVMs :rocket:

• bug: Adding correct typings for SVMs. Thanks @adamjking3!

[1.2.5]

:bug: removing a module for a security reason :bug:

[1.2.4]

:rocket: bug fixes :rocket:

• docs: Removing an extra text from the site footer
• bug: Boston Dataset is not correctly exported from index.ts

[1.2.3]

:rocket: bug fixes :rocket:

• docs: all-contributors module for contribution recognisation
• devops: Husky precommit hooks to ensure all commits satisfy the linters. Thanks @OlegStotsky !!
• docs: Removed the codacy badge and its integration
• devops: Development environment now supports Windows environment. Thanks @LSBOSS !
• bug: PolynomialFeatures and normalize exports in preprocessing/index.ts
• bug: SGDClassifier and SGDRegressor exports in linear_model/index.ts

[1.2.2]

:rocket: Adding Examples page, more linear models and more datasets :rocket:

• docs: Examples page on the doc site - Titanic Datasets using RandomForest
• docs: Hiding utils functions from showing up in the doc
• feature: Adding PolynomialFeature extraction
• docs: KMeans clustering demo
• bug: Incorrect import statement on KMeans clustering example

[1.2.1]

:bug: Fixing the documentation bug :bug:

• docs: Updating the SGD documentations

[1.2.0]

:rocket: Introducing new features! :rocket:

Updates:

• feature: preprocessing/add_dummy_feature
• feature: linear_model/SGDClassifier
• feature: linear_model/SGDRegressor

[1.1.1]

Updates:

• enhance: ensemble/RandomForest
• docs: Better rendering for object and promise type return values
• enhance: Datasets APIs are now based on promise

[1.1.0]

:raised_hands: Minor release to deliver a couple improvements :raised_hands:

In this minor release, it focuses on the first enhancement of existing APIs and delivers new features. As the release contains a feature, GaussianNB, and an enhancement of DecisionTree, we thought a minor release is suitable for it.

Updates:

• feature: GaussianNB (Gaussian Naive Bayes Classifier) -- What is Naive Bayes? Please check out https://machinelearningmastery.com/naive-bayes-for-machine-learning/
• enhancement: DecisionTree -- Fixed the way that the DecisionTree returns the prediction results -- Instead of returning all the possible Leafs, it will return the most occurred target Leaf according to a voting process

[1.0.0]

:baby_chick: hooray! initial release :baby_chick:

In the first release of machinelearn.js, it aims to deliver the essential models and algorithms required to conduct the fundamental Machine Learning experiment and development. Some APIs still might be unstable but they will be improved over time as we receive feedback from people. You will be seeing new patches being delivered constantly to make gradual improvements until the library reaches a certain level of maturity.

• feature: cluster/KMeans
• feature: datasets/Iris
• feature: decomposition/PCA
• feature: ensemble/RandomForest
• feature: feature_extraction/CountVectorizer
• feature: linear_model/LinearRegression
• feature: metrics/accuracyScore
• feature: metrics/confusion_matrix
• feature: metrics/zeroOneLoss
• feature: model_selection/KFold
• feature: model_selection/train_test_split
• feature: neighbors/KDTree
• feature: neighbors/KNeighborClassifier
• feature: preprocessing/Binarizer
• feature: preprocessing/Imputer
• feature: preprocessing/LabelEncoder
• feature: preprocessing/MinMaxScaler
• feature: preprocessing/OneHotEncoder
• feature: svm/NuSVC
• feature: svm/NuSVR
• feature: svm/OneClassSVM
• feature: svm/SVC
• feature: svm/SVR
• feature: tree/DecisionTreeClassifier