Machine Learning Technique for Predicting Location

Abstract

In the current era of internet and mobile phone usage, the prediction of a person's location at a specific moment has become a subject of great interest among researchers. As a result, there has been a growing focus on developing more effective techniques to accurately identify the precise location of a user at a given instant in time. The quality of GPS data plays a crucial role in obtaining high-quality results. Numerous algorithms are available that leverage user movement patterns and historical data for this purpose. This research presents a location prediction model that incorporates data from multiple users. To achieve the most accurate predictions, regression techniques are utilized for user trajectory prediction, and ensemble algorithmic procedures, such as the random forest approach, the Adaboost method, and the XGBoost method, are employed. The primary goal is to improve prediction accuracy. The improvement accuracy of proposed ensemble method is around 21.2%decrease in errors, which is much greater than earlier systems that are equivalent. Compared to previous comparable systems, the proposed system demonstrates an approximately 15% increase in accuracy when utilizing the ensemble methodology.