Changelog¶

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

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]¶

Plotting contour function, plotting heatmap function and plot 2d scatter diagram function for decomposition
Prediction for the training set
Dropping the rows with missing values by specific columns
Summary folder to include all produced artifacts in run’s output
New Models:
- Regression Models
  - Ridge Regression
- Clustering Models
  - Affinity Propagation Clustering
New Mode:
- Anomaly Detection
  - Isolation Forest
Docs:
- Mind map of all options in README
- Citation info
- Anomaly detection algorithm example

Showing formula function for linear models in both regression and classifiction in terms of the number of the target values’ type
Built-in inferenc data only for regression and classification
Docs:
- Installation manual
- Clustering algorithm example

Missing value process with three options
Fixed random state for all models
New Models:
- Regression Models
  - Bayesian Ridge Regression
- Clustering Models
  - Agglomerative Clustering

MLOps core of model inference in command line interface using transformer pipeline
Multi-class label and binary label training for all classification models
CSV data file import
Reduced data storage in decomposition
Data selection function with null, space and Chinese parentheses dection functionality
label customization in classification
Feature selection function
Design diagrams of the whole project
Feature scaling for unsupervised learning
Built-in inference dataset loading
Silhouette score frequency diagram for all clustering model
Two clustering model score for all clustering model
New Models:
- Regression Models
  - Elastic Net
  - Stochastic Gradient Regression
- Classification Models
  - Gradient Boosting
  - K-Nearest Neighbors
  - Stochastic Gradient Descent

Colourful command line interface to highligh importance stuffs.
Standardization of run-driven operation for an experiment.
Specialized storage mechanism to achieve the MLOps core of machine learning lifecycle management using MLflow
Online documentation, including project section, user section, developer section.
New Models:
- Regression Models
  - Lasso Regression
  - Gradient Boosting
  - K-Nearest Neighbors
- Decomposition Models
  - T-SNE
  - MDS
Docker deployment configuration.
Continuous intergration (CI) before git commit using pre-commit.

Manual hyper parameters selection and automated hyper parameter selection using FLAML and Ray for every existed models
New Models:
- Classification Models
  - Multi-layer Perceptron
  - Extra Trees

End-to-end cutomized automated machine learning training pipeline with specialized design pattern to achieve the MLOps core of continuous training in command line interface.
New Models
- Regression Models
  - Linear Regression
  - Polynomial Regression
  - Decision Tree
  - Extra Trees
  - Random Forest
  - XGBoost
  - Support Vector Machine
  - Multi-layer Perceptron
- Classification Models
  - Decision Tree
  - Random Forest
  - XGBoost
  - Support Vector Machine
  - Logistic Regression
- Clustering Models
  - KMeans
  - DBSCAN
- Decomposition Models
  - Principle Component Analysis
Build up continuous integration (CI) after git commit using Git Action