Name Matching using Random Forest

This project performs name matching using a trained Random Forest model.

---

Setup Instructions

1. Create a Virtual Environment

Make sure you have Python and virtualenv installed.

pip install virtualenv

# Create a virtual environment
virtualenv myenv

# Activate the environment
source myenv\Scripts\activate

# Install the requirements file given
pip install -r requirements.txt

Run the file

Training the model

python name_matching.py train --train_data_path data/STrain.csv --truth_labels_path data/G.csv --model_output_path rf_model.pkl

Arguments:

• --train_data_path: path to the test dataset (CSV, txt)
• --truth_labels_path: path to the groundtruth labels (CSV, txt)
• --model_output_path: path where the model pickle will be saved (.pkl)

Predict with the model

python name_matching.py predict --test_file_path data/STest.csv --truth_labels_path data/G.csv --model_path rf_model.pkl --output_path predictions.txt --num_entries 10000

Arguments:

• --test_file_path: path to the test dataset (CSV, txt)
• --truth_labels_path: path to the groundtruth labels (CSV, txt)
• --model_path: path to the saved trained model pickle file (.pkl)
• --output_path: File where the predictions will be saved
• --num_entries: Number of test entries to evaluate