The current study examines the factors influencing investor opinions on issues related to listed firms during the first day of Initial Public Offerings (IPOs), focusing on a sample of 350 fixed-priced IPOs listed on the Malaysian stock exchange (Bursa Malaysia) from 2004 to 2021. This research contributes to existing literature by employing various machine learning methods, which address the limitations of traditional linear regression models commonly used in previous studies. Specifically, five methods—extra tree regressor (ETR), single feature selection (SFS), reverse single feature (RSF), recursive feature elimination (RFE), and sequential modelling feature adding (SMFA)—are utilized to assess the importance of features in predicting the investor opinion gap within the dataset.The study's experiments indicate that these methods effectively mitigate noisy data, enhancing their reliability for this type of analysis. The findings provide valuable insights for regulators regarding safeguarding investors' rights to information disclosed in prospectuses.