Analysis And Research Of Teaching Behavior Based On Campus Data

Teaching behavior analysis is a scientific research activity,which has regularity,principle and concurrency.It seeks teaching rules in the teaching process and discovers teaching needs and potential problems.Traditional teaching behavior analysis usually adopts the mode of questionnaire survey and path-dependent analysis.Its research direction is often focused on the quality of teaching of teachers or the academic performance of students.The object of analysis is single and lack of large-scale objective data demonstration.Data mining technology is gradually being widely used in the field of education.Teaching behavior analysis can help us take the initiative to master the characteristics and laws of teaching behavior.It is conducive to the formulation of educational planning and the improvement of teaching mode.This paper used the campus data platform to collect teaching behavior data such as teaching and surfing the Internet.In this work,we used the PauTa criteria and the method of feature screening to complete the preprocessing of teaching behavior data.It excavated the practical value of educational data in teaching management,teaching evaluation and subject learning.This paper has studied and summarized the relevant research on teaching behavior analysis.Based on traditional teaching behavior analysis techniques,we combined the residual calculation(RC)and the classification algorithm of support vector machine(SVM).And we developed different behavioral analysis models for two different groups of teachers and students.The main work of this paper is as follows:(1)This paper elaborated on the research status,background and significance of teaching behavior analysis,and we discussed the existing research methods and technologies at home and abroad.For the processing of abnormal data in campus data,we introduced the Laida criterion(Pauta).According to the needs of educational work,we set a standard threshold for teaching behavior data,if the sample value exceeds the threshold interval,it defined as a coarse error.In order to reduce the noise ratio of the data samples,we rejected the data with coarse errors.It improves accuracy and credibility for subsequent behavioral analysis efforts.(2)For the behavior of the teacher,we excavated the starting and ending time data of teachers' classroom teaching based on the multimedia teaching courses in colleges and universities.This paper proposed a teaching behavior evaluation model based on the point of time(PT)and the length of time(LT).We used the RC method to achieve a vector fit comparison of behavioral data with standard data,and we build an evaluation index system based on multimedia teaching time data.(3)According to the students learning behavior,this article took the students online data record as the starting point.We used the K-means algorithm to count and analyze the online behavior of different types of students.Then the binary classifier of “studying” and “nonstudying” is established according to the law of education.Each student sample can clearly understand the learning proportion of online behavior after SVM classification.And this paper used Pearson correlation coefficient to analyze the relationship between students' online learning behavior and online time,frequency and client type.It can help us understand the interest orientation of students' online behavior clearly.This paper finds that there are irregular differences in teaching time between different teachers and colleges.The behavior analysis results of RC are the normative representation of teachers' teaching time.And the proportion of students' online learning behavior is related to their online time,frequency and others,different student samples have different patterns of change.The results of this paper provided a real and objective data reference for the improvement of teaching management level in colleges and the improvement of campus network usage methods.