Welcome to pranaam’s documentation!

pranaam is a Python package for predicting religion from names using machine learning models trained on Bihar Land Records data. The package supports both Hindi and English names and provides high accuracy predictions.

Contents:

Note

Interactive Jupyter notebook examples are available in the docs/examples/ directory:

  • Basic Usage (basic_usage.ipynb): Learn fundamentals with single and batch predictions

  • Pandas Integration (pandas_integration.ipynb): DataFrame processing workflows

  • CSV Processing (csv_processing.ipynb): File processing for real-world datasets

  • Performance Benchmarks (performance_benchmarks.ipynb): Optimization strategies

These notebooks are fully executable and demonstrate best practices for production use.

Overview

Pranaam uses machine learning models trained on 4M unique records from Bihar Land Records data to predict religion (currently Muslim/not-Muslim) from names. The package supports:

  • High Accuracy: 98% accuracy on unseen names for both Hindi and English

  • Multiple Languages: Support for Hindi and English names

  • Easy to Use: Simple API with pandas DataFrame output

  • Pre-trained Models: Models are automatically downloaded and cached

Quick Example

from pranaam import pred_rel

# English names
names = ["Shah Rukh Khan", "Amitabh Bachchan"]
result = pred_rel(names)
print(result)

# Hindi names
hindi_names = ["शाहरुख खान", "अमिताभ बच्चन"]
result = pred_rel(hindi_names, lang="hin")
print(result)

Installation

Install pranaam using pip:

pip install pranaam

For development:

pip install -e .[dev]

Indices and tables