Theses

An End-to-End AI Surveillance System Using Multi-Source Data Fusion and Visual Intelligence for Public Safety
Advisor: Dr. Guobin Xu
May, 2026.

Teamwork Success Prediction.
Advisor: Dr. Aya Salama
Jul, 2023.

Publications

Kim VanExel, Kevin Quintero, Aliyu Bello Aliyu, Jin Guo, Guobin Xu, Xing Liu, and Lin Deng. Distributed High-Performance Computing for Accelerating Image Classification Workloads. Accepted at ICCCN 2026.

Aliyu Bello Aliyu, Jin Guo, Guobin Xu, and Lin Deng. An AI Surveillance Framework with Multi-Source Fusion and Temporal Validation for Public Safety. Accepted at SERA 2026.

Academic Papers

Predictive Analysis for Hazardous Material Transportation Incidents: Applied advanced data analytics techniques, including geospatial mapping and machine learning, to analyze transportation incidents. Leveraged a Decision Tree model for predictive insights, identifying factors influencing incident likelihood, and provided actionable recommendations for improved safety and logistics.

Comparative Analysis of Machine Learning Models on Fashion MNIST Dataset: Conducted a comparative analysis of SVM, Random Forest, XGBoost, and CNN models on the Fashion MNIST dataset using TensorFlow and Scikit-learn. CNN achieved the highest accuracy (92%), while XGBoost offered the best balance between speed and performance.

Database Management System (DBMS): Compared relational and graph databases using SQL Server and Neo4j to assess scalability and query performance. Found relational DBMS excelled on small datasets, while Neo4j outperformed in large, complex network queries.

Time Series Analysis for Smart Water Management: Performed time series forecasting on the Acea Smart Water dataset using Auto-ARIMA to model groundwater depth influenced by rainfall and temperature. Auto-ARIMA delivered the most accurate results, effectively capturing temporal groundwater trends.

Big Data, Machine Learning, and Hadoop: The study examines the interplay between machine learning and Hadoop in managing massive datasets, emphasizing Hadoop’s scalability and machine learning’s analytical power. It concludes that combining both technologies enables efficient data processing, pattern discovery, and smarter decision-making in Big Data environments.

Smart Contracts and Cryptocurrency In Blockchain: Analyzed blockchain applications in finance and data systems, showing smart contracts automate secure transactions while cryptocurrencies support transparent, decentralized payments.

Face Recognition Techniques: Compared 3D and thermal facial recognition methods, finding 3D more precise geometrically and thermal imaging superior in low-light, together enhancing biometric security.

Smart Supply Chain: Conducted data mining and visualization on the DataCo Smart Supply Chain dataset using R with clustering, PCA, and discriminant analysis. Identified six key clusters and strong variable associations for supply chain optimization.