Rumale (Ruby machine learning)

The Rumale::MetricLearning module provides metric learning algorithms that follow the standard Rumale interface. Supported algorithms include:

• Fisher Discriminant Analysis
• Neighbourhood Component Analysis

Rumale::Preprocessing is a component of the Rumale machine learning library for Ruby. It provides various preprocessing techniques that follow the standard Rumale interface, allowing them to be integrated into machine learning workflows.

Supported techniques include:

• L2 normalization
• Standard scaling
• One-hot encoding

The Rumale::Ensemble module provides ensemble learning algorithms that follow the standard Rumale interface. This includes algorithms such as:

• AdaBoost
• Gradient Tree Boosting
• Random Forest

The Rumale::LinearModel package provides various linear model algorithms that follow the standard Rumale interface. Supported algorithms include:

• Logistic Regression
• Support Vector Machine (SVM)
• Lasso Regression
• Ridge Regression

Rumale::Manifold is a component of the Rumale machine learning library for Ruby. It provides data embedding algorithms that follow the standard Rumale interface.

Key algorithms included in this package:

• Multi-dimensional Scaling (MDS)
• t-distributed Stochastic Neighbor Embedding (t-SNE)

The Rumale::KernelMachine package provides kernel method-based machine learning algorithms that follow the standard Rumale interface. It includes implementations for:

• Kernel Support Vector Machine
• Kernel Principal Component Analysis (KPCA)
• Kernel Ridge Regression

The Rumale::Decomposition module provides matrix decomposition algorithms that follow the standard Rumale interface. It is used for dimensionality reduction and extracting latent structures from data. Supported algorithms include:

• Principal Component Analysis (PCA)
• Non-negative Matrix Factorization (NMF)
• Factor Analysis (FA)
• Independent Component Analysis (ICA)

The Rumale::FeatureExtraction module provides various methods for transforming raw data into feature vectors, compatible with the standard Rumale interface. Supported methods include:

• TF-IDF: Term Frequency-Inverse Document Frequency for text processing.
• Feature Hashing: A method to transform features into a fixed-size vector using a hash function.

The Rumale::KernelApproximation module provides algorithms to approximate kernel functions, allowing you to transform data into a high-dimensional feature space that can be used with linear models. This is useful for applying non-linear kernel methods (like RBF) efficiently.

Supported methods include:

• RBF feature mapping
• Nystroem method

All algorithms follow the standard Rumale interface.

The Rumale::NeuralNetwork module provides machine learning algorithms compatible with the standard Rumale interface. It includes classifiers and regression algorithms based on:

• Multi-layer perceptron
• Radial basis function network
• Random vector functional link network

The Rumale::Clustering package provides various cluster analysis algorithms that follow the standard Rumale interface. Supported algorithms include:

• K-Means
• Gaussian Mixture Model
• DBSCAN
• Spectral Clustering

For detailed API references, consult the Rumale API Documentation - Clustering.