Resume Scanning

Project Overview

This project is an intelligent Resume Screening System that automatically analyzes and ranks candidate resumes against a given job description using Natural Language Processing (NLP) and Machine Learning techniques.

The system extracts text from PDF resumes, cleans and processes the content using NLP, converts text into numerical vectors using TF-IDF, and calculates similarity scores using Cosine Similarity to determine how well each resume matches the job requirements.

It classifies resumes into categories such as Highly Matched, Moderately Matched, and Not Matched, helping recruiters quickly identify the most suitable candidates and drastically reduce manual screening time.

Screenshots

Key Features

Automatic PDF Resume Text Extraction
Advanced Text Cleaning using NLP (Stopword removal, punctuation removal)
TF-IDF Vectorization for text representation
Cosine Similarity based resume ranking
Smart classification into match categories
Batch processing of multiple resumes from a folder
Generates a CSV report with match percentage
Fast, accurate, and fully automated screening process

How It Works

The system reads a Job Description PDF.
It extracts text from multiple resume PDFs stored in a folder.
Text data is cleaned using NLP techniques:
- Lowercasing
- Punctuation removal
- Stopword removal
The cleaned text is converted into numerical form using TF-IDF Vectorizer.
Cosine Similarity is calculated between the Job Description and each resume.
Based on similarity score:
≥ 70% → Highly Matched
40–69% → Moderately Matched
< 40% → Not Matched
Results are sorted and saved into a CSV file for recruiters.