Comparison of Chinese character correct and error classifier for overseas students based on handwriting motion characteristics

Abstract

The quality of Chinese characters written by foreign students is reflected not only in the result of writing, but also in the writing movement. 25 characteristics of handwriting movement were selected from 39 parameters, such as time, space, movement and pressure. Based on the feature of handwriting motion, 6 classifiers, such as Decision Tree and Artificial Neural Network, are selected to compare the performance of them. Experiments show that the accuracy of DT algorithm is the best regardless of whether the target characters are known or not. Handwriting movement features are effective representations of the quality of Chinese characters written by foreign students.