Application of data mining for young children education using emotion information

Abstract

The current preschool education is still facing many difficulties. First, unlike primary and secondary schools which have test scores, there are few good ways to assess the learning situation of young children. Second, it is hard for teachers and parents to give thoughtful care all the time to each child. In particular, recent incidents of child abuse have been exposed frequently, causing social panic. However, as research of young children's mental health and emotion is still in its infancy due to the lack of relevant data, methods that focus on measuring and analyzing young children's mental health and emotional states are lacking. Furthermore, on the one hand, the principal goal of preschool education is to stimulate interest and enhance cognitive ability. On the other hand, assimilation and accommodation which are two specific stages of cognitive development require an active learner, not a passive one. Thus, emotion analysis can solve these problems to some extent.In this paper, we design an intelligent system, obtaining video clips in a kindergarten's classroom and managing to leverage the emotion data to portray cognitive learning rules and mental states of young children. At the same time, with the augmentation by data analysis, it brings broader applications developing educational policy and teaching practice. For example, children with abnormal behaviors such as violent mood swings and unusual time nodes during courses can be detected and notified to parents and teachers. And it is meaningful to measure the children's acceptance and preference of course activities and content which has been explained in our experiments.