Machine Learning

In a comprehensive analysis of water quality data across multiple locations and years, this project applies the Informer architecture to forecast key environmental indicators. Its scalable design enables efficient handling of large, multivariate time-series data with strong predictive accuracy. Benchmarking against traditional methods highlights the advantages of deep learning in capturing long-term patterns. The findings show how such predictive tools can support effective environmental monitoring and sustainable resource management.  

Through detailed exploration of student academic and demographic data, this project investigates the key factors that influence performance outcomes. Multiple machine learning models were trained and evaluated to determine which best predicts student success, offering a comparative view of their effectiveness. By combining predictive modeling with data-driven insights, the project aims to support strategies for academic improvement and early intervention.

With in-depth analysis of sensor-based motion data, this project focuses on recognizing and classifying human activities with machine learning. Multiple models were trained and compared to evaluate their effectiveness in capturing subtle patterns in movement. The study highlights how features derived from accelerometer and gyroscope signals influence classification accuracy. By combining predictive modeling with behavioral insights, the project demonstrates applications in health monitoring, fitness tracking, and human-computer interaction.