An Attempt to Diagnose Learners' Understanding Levels Considering Confidence

Abstract

Intelligent Tutoring Systems require a high diagnosis ca- pability of learners' understanding levels. Confidence which is subjective and fuzzy data can be expected for im- provement of the diagnosis ability, but the utilization is not easy. This paper describes a diagnosis method with confidence, which uses Self Organization Map(SOM). The demonstration results showed usefulness of using confi- dence in estimating understanding level. Organization Map(SOM)(10) as diagnosing tool. SOM is a kind of artificial neural network based on competitive learning. It has input and competitive layers. In addition, it is able to reduce high dimensions of data vectors by its efficient behavior. We can discrimi- nate high dimensional data sets with the feature. This is remarkable function of SOM. Hence, SOM is very widely used in medical science, architectonics, economics, engi- neering, and so forth. In education field, a learning sup- port system by using self organized feature map has been presented(11). We have already reported SOM diagno- sis system based on objective data, and demonstrated its capability(12). This paper describes the method for diagnosing learner's understanding level by using subjective/objective data and SOM, and demonstrates the usefulness of the sys- tem in diagnosis experiments for 29 learners.