Research On Data Mining Method In College Teaching Evaluation

Education big data era is coming.Using educational data mining and learning analysis technology to thoroughly study the massive education data will be the development trend of education.Educational data mining and learning analysis can be used to const ruct educational-related models.They also can be used to explore the relationship between educational variables.Teaching evaluation plays an important role in school teaching activities.Most of the existing teaching evaluation results use a q uantitative calculation method.The calculation results can only reflect part of the teaching situation.And it is difficult for educators to get more valuable information from the evaluation data.Therefore,this paper focus on exploiting knowledge discovery method to mining the teaching evaluation data.And using the method mine the hidden knowledge in the teaching evaluation data.Using the knowledge provides advice to teaching manager and teacher.This paper mainly completed the following aspects of the work:(1)Analysis and preprocessing of teaching evaluation data.The data analysis work mainly includes analyzing the data structure of the teaching evaluation data and constructing the related table conceptual model.Data preprocessing work concludes data cleaning,data integration,data transformation and data normalization.(2)Research on outlier detection algorithm for teaching evaluation data.This paper analyzes the shortcomings of the existing outlier detection algorithm based on categorical data,and proposes an improved outlier detection algorithm for teaching evaluation data.Applying the algorithm to the real teaching evaluation data,so as to detect the noise data in the data.Experiments show that the algorithm has high detection accuracy and efficiency.(3)Research on association rule mining algorithm for teaching evaluation data.This paper analyzes the advantages and disadvantages of frequent itemsets mining algorithm based on compression matrix.Although the algorithm is more efficient than the Apriori algorithm.But it is not entirely applicable to quantitative attribute data.The algorithm generates invalid connection calculations during the mining process.Therefore,this paper proposes an Frequent Itemsets Mining Algorithm for quantitative attribute data Based on Compression Matrix.Using the algorithm to find out what links betwee n teaching factors and what factors will affect the teaching evaluation results.Finally,providing advice to teachers to improve their teaching.Experiments show that the algorithm can effectively avoid the invalid connection operation,improve the efficiency of mining.