sageml 0.2.0

SageML is library for automatic one-shot algorithm selection for tabular data using tda for extracting information about dataset.

SageML

SageML is an out-of-the-box AutoML solution designed to simplify the machine learning workflow. With minimal user input, SageML automates model selection, hyperparameter optimization, and provides a trained machine learning model ready for deployment.

Table of Contents

• SageML
  • Table of Contents
  • Features
  • Installation
  • Quick Start
  • Usage
    • Data Preprocessing
    • Model Selection
    • Hyperparameter Optimization
    • Model Evaluation
  • Documentation
  • Contributing
  • License

Features

• Automatic Model Selection: Chooses the best algorithm based on data characteristics with pre-trained neural network.
• Hyperparameter Optimization: Utilizes Optuna for efficient hyperparameter tuning.
• Data Preprocessing: Handles missing values, categorical encoding, and feature scaling automatically.
• Interactive Interface: User-friendly terminal interface with tutorials and step-by-step guidance.
• Extensibility: Modular architecture allows for easy customization and extension.
• Compatibility: Supports a wide range of algorithms from scikit-learn, CatBoost, XGBoost, and more.

Installation

SageML is available on PyPI. You can install it using pip:

pip install sageml

Note: For the latest features and updates, you might want to install from the GitHub repository.

Quick Start

Here's how you can get started with SageML in just a few lines of code:

from turbo_ml import SageML
import pandas as pd
# Initialize SageML with your dataset
sageML = SageML(pd.read_csv('classified/data.csv'), target='target')

# Make predictions
predictions = sageML.predict(pd.read_csv('not/classified/data.csv'))

Usage

Data Preprocessing

SageML automatically preprocesses your data to make it suitable for machine learning algorithms.

• Handles missing values with appropriate imputation methods.
• Encodes categorical variables using techniques like One-Hot Encoding.
• Scales numerical features for algorithms sensitive to feature scales.

Model Selection

• Analyzes data characteristics (e.g., number of features, class balance).
• Selects suitable algorithms from a pool that includes scikit-learn classifiers/regressors, CatBoost, XGBoost, etc.
• Supports both classification and regression tasks.

Hyperparameter Optimization

• Utilizes Optuna for efficient hyperparameter optimization.
• Employs advanced features like pruning to reduce computation time.

Model Evaluation

• Allows selection of evaluation metrics from scikit-learn or custom weighted sums.
• Supports cross-validation and hold-out validation strategies.

Documentation

Detailed documentation should be available soon.

Contributing

We welcome contributions from the community!

• Bug Reports & Feature Requests: Use the GitHub Issues to report bugs or suggest features.

License

SageML is licensed under the GNU General Public License v3.0.

---

Disclaimer: This project is under active development. Features and interfaces are subject to change.