| Summary: | 碩士 === 國立臺中教育大學 === 教育測驗統計研究所 === 102 === This study developed a computerized testing system of visual motor integration based on the basic Chinese characters. The test-takers aged from four to six used the tablet to take the test. Moreover, the reliability analysis and item response theory (IRT) were used to analysis the items about visual perception. On the other hand, for visual integration, some features such as distance, speed, slop, and angles, were designed and extracted for classifying the degree of disability automatically. From the experimental results, the item difficulties are relevant according to the Rasch model (one-parameter logistic model) for the items about visual perception. For visual integration, twelve to thirty four features are extracted according to the strokes of the basic Chinese characters. Comparing the classification results with these features by applying the support vector machine (SVM) and K nearest neighbor classifier (KNN) with expert’s judgments are consistent and the corresponding classification accuracies are between 80.31% and 97.59%.
|
|---|
