Predicting Ratings of Indian IPOs from Red Herring Prospectus

EasyChair Preprint 15779

Abstract

In recent years, Initial Public Offerings (IPOs) of Indian companies have emerged as a popular investment opportunity, with many investors seeking quick returns through listing gains. However, analysing lengthy prospectuses to make informed, data-driven investment decisions can be a cumbersome task. To address this
challenge, we propose a task to mine red herring prospectuses of companies planning to go public and classify them into four categories: Apply, Neutral, May Apply, or Avoid. This method aims to streamline the decision-making process for investors by providing clear and concise recommendations based on the prospectus data.In addition to introducing two new datasets, we propose a novel method for predicting ratings of Indian IPOs that surpasses the performance of existing state-of-the-art Large Language Models.

Keyphrases: Financial texts, Indian IPO, Natural Language Processing, large language models

@booklet{EasyChair:15779,
  author    = {Sohom Ghosh and Sudip Kumar Naskar},
  title     = {Predicting Ratings of Indian IPOs from Red Herring Prospectus},
  howpublished = {EasyChair Preprint 15779},
  year      = {EasyChair, 2025}}