🎓 Aspect-Based Sentiment Analysis (ABSA) for Open-Ended Survey Responses

This project implements an Aspect-Based Sentiment Analysis (ABSA) system specifically designed to analyze open-ended survey responses. The goal is to extract meaningful insights by identifying specific aspects mentioned in free-text answers and classifying the sentiment (positive, negative, or neutral) associated with each aspect.

🧠 Key Features

🔍 Aspect Extraction: Automatically detects key aspects/topics from open-ended responses.

💬 Sentiment Classification: Determines the sentiment (positive, negative) for each identified aspect.

🛠️ Text Preprocessing: Includes cleaning, case folding, tokenization, normalization + synonym, and stopword removal.

📊 Output: Sentiment summary per aspect for easier interpretation.

🧪 Use Case

Ideal for researchers or analysts who handle qualitative survey data and need to extract structured sentiment insights from open-text feedback. Common applications include:

Event feedback evaluation

⚙️ Tech Stack

Python (main language)

Scikit-learn,  PyTorch, BERT, etc 

NLTK, Sastrawi, etc (for preprocessing)

Flask (Deployed as a web app)