sisense-automl

sisense-automl is a Python package for automating machine learning processes using Sisense and AutoML. It provides easy-to-use interfaces for data preprocessing, model training, and evaluation.

⚠️ Compatibility Warning

This package requires Python 3.9 to function correctly. Please ensure you are using the correct Python version.

Why only Python 3.9?

• Auto-sklearn does not support Python 3.10 or newer.
• Scikit-learn 0.24.0 (required by Auto-sklearn) works best with Python 3.9.
• NumPy 1.23+ removed numpy.distutils, which some dependencies rely on.

Features

• Preprocesses data by handling duplicates, splitting numerical and categorical features, and encoding them appropriately.
• Splits data into training and testing sets.
• Uses auto-sklearn for training classification and regression models.
• Saves trained models and datasets for later use.

Installation

You can install the package using pip:

pip install sisense-automl

System-Level Dependencies

Before installing the package, make sure you have the following system-level dependencies:

sudo apt-get install swig -y

Python-Level Dependencies

The package will automatically install the required Python dependencies as specified in setup.py.

Usage

Here is a basic example of how to use the 'sisense-automl' package:

import pandas as pd
from sisense_automl import AutoMl

# Load your dataset
data = pd.read_csv("your_dataset.csv")

# Define your target column and objective
target_column = "target"
objective = "classification"  # or "regression"

# Define the folder path to save models and data
folder_path = "your_folder_path"

# Create an AutoMl instance and run the process
automl = AutoMl(data, target_column, objective, folder_path)

License

This project is licensed under the MIT License. See the LICENSE file for more details.

Author

Himanshu Negi - himanshu.negi.08@gmail.com

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Please make sure to update tests as appropriate.