• Version
• Download 17
• File Size 799.87 KB
• File Count 1
• Create Date 6 August 2025
• Last Updated 6 August 2025

Predicting Online Gaming Engagement Levels Using Machine Learning Models

 

 

1M.N.KEERTHI ,2GORIBIDDI SHRIKANTH

1Assistant Professor, Department Of MCA, 2MCA Final Semester,

1Master of Computer Applications,

1Sanketika Vidya Parishad Engineering College, Vishakhapatnam, Andhra Pradesh, India

 

 

Abstract:

This Streamlit application, titled "Online Gaming Engagement Level Predictor," is designed to predict player engagement levels using machine learning. It allows users to upload datasets containing player demographics, gameplay statistics, and attributes like game genre and difficulty. The app performs thorough Exploratory Data Analysis (EDA) with visual tools such as histograms, boxplots, and count plots to understand data distribution and detect anomalies. It also provides insights through correlation heatmaps and pair plots. Preprocessing involves handling outliers, encoding categorical variables using one-hot encoding, label mapping, and target encoding with K-Fold cross-validation. Stratified sampling is used to split the dataset into balanced training and testing sets. Multiple models including Random Forest, Gradient Boosting, LightGBM, and CatBoost are trained and evaluated. Model performance is assessed using accuracy, ROC-AUC scores, classification reports, and confusion matrices. Feature importance is also visualized for tree-based models. The best-performing model is selected based on ROC-AUC, helping game developers make data-driven decisions to enhance player engagement.
Index Terms: Streamlit Application, Online Gaming, User Engagement Prediction, Machine Learning, Exploratory Data Analysis (EDA), Categorical Encoding, Random Forest, Gradient Boosting, ROC-AUC Score, Feature Importance, Game Analytics, Predictive Modeling, Classification Models, Stratified Sampling, Data Visualization..

Download