Video Game Recommendation via NLP ML Model

Disciplines

Other Computer Engineering

Abstract (300 words maximum)

This project explores the use of Natural Language Processing and Machine Learning to recommend video games by analyzing player reviews from the Steam platform.

We plan to collect review data coming from the Steam API database, mainly from games released between early 2024 and September 2025. The text from the reviews will be processed in a NLP model, and it will be tested to give games a positive, negative, or possibly even mixed reactions.

For the evaluation of our model, we will be using metrics like precision and RMSE to help analyze and combine the data using keywords to mark a review as the appropriate rating for the game. This will give people who use the Steam platform more of an edge to choose what games to buy and what is worth playing in the current landscape.

In the end, the goal of the project is to help better the recommendation system on the Steam platform and give people who use it a better way of knowing if something is good or not. Of course, there will be difficulties with applying this with things akin to review bombing or random reviews made by users of the platform. However, with those in mind, the model will analyze the text in each game and help assist gamers with more of a choice in their games.

Use of AI Disclaimer

no

Academic department under which the project should be listed

CCSE – Computer Science

Primary Investigator (PI) Name

Dr. Hafiz Khan

This document is currently not available here.

Share

COinS
 

Video Game Recommendation via NLP ML Model

This project explores the use of Natural Language Processing and Machine Learning to recommend video games by analyzing player reviews from the Steam platform.

We plan to collect review data coming from the Steam API database, mainly from games released between early 2024 and September 2025. The text from the reviews will be processed in a NLP model, and it will be tested to give games a positive, negative, or possibly even mixed reactions.

For the evaluation of our model, we will be using metrics like precision and RMSE to help analyze and combine the data using keywords to mark a review as the appropriate rating for the game. This will give people who use the Steam platform more of an edge to choose what games to buy and what is worth playing in the current landscape.

In the end, the goal of the project is to help better the recommendation system on the Steam platform and give people who use it a better way of knowing if something is good or not. Of course, there will be difficulties with applying this with things akin to review bombing or random reviews made by users of the platform. However, with those in mind, the model will analyze the text in each game and help assist gamers with more of a choice in their games.