An advanced, dual-interface Auto-ML platform designed for the comprehensive analysis of tabular data. This project now offers both a robust Command-Line Interface (CLI) and a professional Streamlit Web Dashboard.

It bridges the gap between traditional and modern AI by integrating:

• Supervised Learning: Regression & Classification (with XGBoost support).
• Time-Series Forecasting: Powered by Facebook Prophet.
• Reinforcement Learning: Trend prediction using PPO Agents.
• Unsupervised Learning: Clustering & PCA.

Engineered for performance, it features PyArrow-accelerated data loading, compressed model artifacts, and strict data leakage prevention using scikit-learn Pipelines.

A professional web interface built with Streamlit offering:

• Modern Dark UI: A sleek, glassmorphism-inspired dark theme designed for professional use.
• Robust Architecture: Features a "Safe Boot" system to handle heavy ML dependency loading without crashing.
• Data Upload: Drag-and-drop CSV/Excel/JSON files with instant PyArrow processing.
• EDA: Tabbed interactions for distributions, correlations, and feature overviews using Plotly.
• No-Code Training: Train XGBoost/Sklearn models with one click.
• Interactive Forecasting: Visual time-series prediction with dynamic zoom/pan.

• Gradient Boosting Powerhouse: Integrated XGBoost (Regressor/Classifier) alongside Random Forest and Gradient Boosting for state-of-the-art performance.
• Prophet Forecasting: specialized additive models for accurate time-series prediction, handling seasonality and trends automatically.
• RL Trend Predictor: Custom Gymnasium Environment trained with Stable Baselines3 (PPO) to predict future market/data directions.

• High-Performance I/O: Uses engine='pyarrow' for ultra-fast CSV reading and memory efficiency.
• Storage Optimization: Models are saved with level-3 compression to reduce artifact size without losing accuracy.
• Robust Pipelines: Automates scaling, imputation, and encoding within searchable pipelines to prevent data leakage.

1. Clone the Repository:

   git clone https://github.com/Shafiyullah/Predicto-AI.git

2. Create a Virtual Environment (Recommended):

   python -m venv venv

   • Windows: venv\Scripts\activate
   • Linux/Mac: source venv/bin/activate

3. Install Dependencies: Includes xgboost, prophet, streamlit, and plotly.

   pip install -r requirements.txt

Launch the full GUI experience:

streamlit run dashboard.py

Access the dashboard in your browser at http://localhost:8501

Run the classic terminal-based tool:

python main.py

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