This project implements a complete computational framework for CAT, bridging the robust mathematical foundations of Item Response Theory (IRT) with the dynamic, data-driven logic of adaptive algorithms. By continuously selecting the most informative items for each individual, CAT achieves precise ability estimation with substantially fewer questions, reducing testing time by 50–70% while maintaining or enhancing measurement quality. The system integrates core components—from Fisher Information and Bayesian estimation to practical constraints like content balancing and exposure control—into a cohesive, operational engine for next-generation educational, psychological, and professional assessment. Through both theoretical exposition and functional R implementation, this work demonstrates how adaptive testing makes assessment not only shorter and fairer, but also more accurate and engaging for every examinee.

2026-02-04

This interactive web presentation explores key relationships between automotive performance metrics using R Markdown and Plotly. Through dynamic visualizations, discover how horsepower impacts fuel efficiency, how cylinder count affects MPG, and the distribution of vehicle weights. With fully interactive plots that allow zooming, panning, and hovering for detailed insights, this project demonstrates modern data visualization techniques for communicating complex automotive data in an engaging and accessible way.