Prototype of Pilot Knowledge Evaluation by an Intelligent CAI (Computer -Aided Instruction) System Using a Bayesian Diagnostic Model

摘要

This thesis describes the Testing system, a Computer-Aided Instruction system which monitors the progress of a student and automatically adjusts the lesson to suit his/her understanding of the material being presented. Each lesson is divided into three segments. The first segment consists of questions critical to the basic understanding of the lesson, the second questions complementary to the lesson, and the third question supplementary. After the first segment, the student's performance is evaluated and a decision is made whether or not he/she should continue to the complementary segment of the lesson or go on to the next lesson. Similarly, in the second segment, a decision whether to continue to the supplementary segment or to the next lesson is made. Thus, a proficient student can progress through the material quickly, a slow learner slowly. The questions are grouped into types, and each of them is ranked as an easy or difficult concept to provide more precise guidance for the student's improvement. An example lesson is provided as an illustration.