Release v0.3.1

Version bump to 0.3.1 with latest updates.

mlcli-toolkit v0.3.0

🚀 Release v0.3.0 - New ML Algorithms!

What's New

Gradient Boosting

• LightGBM: Fast gradient boosting with leaf-wise tree growth, early stopping, and feature importance
• CatBoost: Gradient boosting with excellent categorical feature handling

Clustering Algorithms

• K-Means: Partition-based clustering with silhouette metrics and automatic optimal K detection
• DBSCAN: Density-based clustering with automatic noise detection

Anomaly Detection

• Isolation Forest: Tree-based anomaly detection using isolation principle
• One-Class SVM: Novelty detection using support vector methods

Installation

pip install mlcli-toolkit

Quick Start

# Train LightGBM
mlcli train -d data.csv -m lightgbm --target label

# Clustering with K-Means
mlcli train -d data.csv -m kmeans

# Anomaly detection
mlcli train -d data.csv -m isolation_forest

mlcli-toolkit v0.2.0

🚀 Release v0.2.0 - Production Ready!

What's New

• Documentation: Comprehensive docs/, examples/, and tests/ folders
• Contributing: CONTRIBUTING.md, CODE_OF_CONDUCT.md, SECURITY.md
• Testing: pytest framework with trainer and tuner tests
• Examples: Sample configs for all models and tuning strategies
• Improved: GridSearchTuner parameter handling
• Cleaned: Production-ready structure

Features

• CLI Training: Train ML models from the command line
• Multiple Algorithms: Random Forest, XGBoost, SVM, Logistic Regression
• Deep Learning: TensorFlow DNN, CNN, RNN trainers
• Experiment Tracking: Track and compare experiment runs
• Hyperparameter Tuning: Grid, Random, and Bayesian optimization
• Model Explainability: SHAP and LIME explanations
• Data Preprocessing: StandardScaler, Normalization, Encoding, Feature Selection

Installation

pip install mlcli-toolkit