Resume Classification and Ranking System

Overview

This project is a Resume Classification and Ranking System that processes multiple resumes in PDF format, classifies them into predefined job categories, and ranks them based on relevance to a specified job role.

Features

• Extracts text from resumes in PDF format.
• Cleans and preprocesses text using regex.
• Tokenizes and sequences text for model input.
• Uses a deep learning model (CNN + LSTM) to classify resumes.
• Ranks resumes based on their softmax probability score for a given job role.
• Normalizes scores and sorts resumes in descending order of relevance.

Dependencies

Ensure you have the following Python libraries installed:

pip install numpy pandas tensorflow scikit-learn PyPDF2

Dataset

The model is trained using the UpdatedResumeDataSet.csv, which contains resumes and their corresponding job categories.

Model Architecture

• Embedding Layer: Converts words into dense vectors.
• Conv1D Layer: Captures local dependencies in text.
• MaxPooling1D Layer: Reduces dimensionality.
• LSTM Layer: Extracts long-term dependencies.
• Dropout Layer: Prevents overfitting.
• Dense Layer with Softmax Activation: Outputs probability distribution across job categories.

Usage

1. Place resumes in the Resumes folder.
2. Load the pre-trained model weights (deeprank_model.h5).
3. Run the script to classify and rank resumes for a given job role.
4. Example:

pdf_folder = "Resumes"
job_role = "Data Science"
ranked_resumes = process_resumes(pdf_folder, job_role)
print(ranked_resumes)

Output

The script returns a sorted DataFrame containing:

• Resume file name
• Job role probability score
• Normalized score (0-1 scale)

License

This project is licensed under the MIT License.

Author

Nilesh Ranjan Pal