LLM-Assisted Purpose Mining

Target Students

  • Computer Science (Bachelor/Master)
  • Psychology (Bachelor/Master)

Requirements

  • Computer Science:
    • Python programming
    • Text analysis/NLP fundamentals
  • Psychology:
    • Understanding of evaluation metrics
    • Statistical analysis of model

Description

This project aims to extract the underlying purpose of Ecommerce product reviews—such as personal use, creativity, professional use, or gaming—using a combination of manual annotation and large language models (LLMs). A user study will also be conducted to assess the practical utility of purpose-based review categorization.

The project combines manual annotation and large language models (LLMs) to extract and classify these purposes. In addition to building a purpose-extraction system, a user study will be conducted to evaluate how helpful it is to sort reviews based on purpose rather than just recency or rating.

The end goal is to enhance user experience in review browsing by offering more contextually relevant insights.

Kick-off meeting will be announced by email

Send an email to Kunal Kunal with the following information: registration number, first and last name, course of study and semester. Registration is open from October 1st, 2025 (first come, first served).