Player behavior prediction for in game purchase using ML

In today's gaming industry, understanding player behaviour and predicting in-game purchases are crucial for enhancing user experience and maximizing revenue. This project, titled "Player Behaviour Prediction for In-Game Purchases Using Machine Learning," aims to develop a predictive model that can accurately forecast both player behaviour and their purchasing decisions. The absence of real-world data necessitates the creation of synthetic data that mirrors typical player interactions and purchase patterns in gaming environments.

The project will employ a range of machine learning algorithms, including Decision Tree, Random Forest, Logistic Regression, and XGBoost Classifier, to build robust predictive models. By analysing player actions, engagement levels, and historical purchase data, the system will be capable of predicting whether a player is likely to make an in-game purchase and identifying the behavioural patterns leading to such decisions.

The backend of this system is developed using Python, leveraging its powerful libraries for machine learning and data processing. The frontend is designed using HTML, CSS, and JavaScript, providing a user-friendly interface for inputting player data and visualizing predictions.

By implementing this project, we aim to enhance the gaming experience and provide valuable tools for optimizing in-game purchases, ultimately contributing to the success of modern gaming enterprises.

Keywords: Decision Tree 2. Random forest 3. Logistic regression 4. XGBoost classifier, supervised learning, K-Best feature selection.