Research On Educational Video Classification Based On Cognitive Style

Cognitive style,also known as cognitive method,which is the behavior pattern exhibited by the individual in the process of cognition and learning.According to the classical theory of pedagogy,for the same learner,using teaching methods and learning materials that match their cognitive style will effectively improve their learning efficiency.In recent years,the teaching reform based on video as the carrier of teaching content was deepening continuously.The activities of "one teacher and one class" of the Ministry of Education,and the construction of the resources for the courses in the colleges,has greatly enriched the number of teaching videos.It also encouraged more and more teachers to participate in educational video construction activities.Educational videos about the same knowledge vary widely in terms of content organization,language description,picture representation,etc.The core issue to be addressed in this paper is how to help learners to choose video that matches their own cognitive style in order to improve their learning efficiency.Due to the relationship between the teaching style of video and the students' cognitive style has not been studied in the classical pedagogy theory,so this paper will work from the following four aspects which based on the classical cognitive style theory proposed by Richard J.Riding.(1)Combining the cognitive style of field independence-field dependence and verbal-imagery,this paper has studies how to extract the digital features associated with cognitive style from video.From the three dimensions of text,image and audio,we have extracted the features of text,the teacher's head,the teacher's action,the color,the audio,the logical,etc.,which are proposed as the basis for the classification of the video teaching style.(2)According to the extracted video features,this paper has studies how to classify educational videos that based on cognitive style.Combining the cognitive style theory proposed by Richard J.Riding,videos are divided into four categories: verbal-field dependence,verbal-field independence,imagery-field dependence and imagery-field independence.Then,the multi class support vector machine classification algorithm is used to train the classifier for the small sample training set,so the automatic classification of educational video is completed and significantly reduce manual workload.(3)The paper has studies the matching problem between learners' cognitive style and video teaching style.According to the results of SVM classification,the four categories videos match the four cognitive styles in classical cognitive style theory.This paper has analyzed the relationship between students' cognitive style and video teaching style,and also has designed the experimental scheme of empirical test.(4)Empirical experiments were carried out in three secondary schools,a total of 387 students were selected from 9 classes.First,learners are classified by cognitive style tests.Then,for the same knowledge point,we let the learners study four different types of videos.Finally,test results are used to measure the learning effect of learners'.The empirical results show that when the students learn the same cognitive style type videos,the learning effect is the best.The learning effect is the worst when they learn the teaching videos which are completely different from the cognitive style types.The study shows that according to the learners' cognitive style,recommending the corresponding performance style of educational video will effectively improve their learning efficiency.The video classification method proposed in this paper is expected to be widely used in video search,educational content recommendation and other systems.