Study On K-modes Algorithm And Neural Network Use In Teaching Evaluation And Learning Prediction

Nowadays,big data,cloud computing,artificial intelligence and other new technologies are widely applied to the construction of digital campus and intelligent campus.Under education circumstances,using big data of education properly not only affects the operation orderly of the whole education,but also will certainly become the inexhaustible motive force to help colleges and universities to promote the reform and innovation of the education and teaching system.Based on data of teaching evaluation system and students' online learning system in Y university,this paper focuses on the teaching operation and students' self-learning ability.Improved K-modes algorithm is used to facilitate clustering analysis for classroom teaching operation,and the neural network algorithm based on machine learning is used to predict and compare the students' online course learning,in order to provide a meaningful reference for the construction,reform and innovation of teaching management system in colleges and universities.The main works are as follows:1.Firstly,through the preliminary analysis and transformation of the evaluation data of students in Y university,the improved cosine algorithm is used to eliminate the abnormal evaluation data and the normalization method is used to standardize the data.In addition,the traditional K-modes algorithm is used to analyze the teaching evaluation data,and three problems are pointed out.Finally,based on these three problems,the traditional K-Modes algorithm is improved from three aspects: the number of clustering families,determination of the initial clustering centers and the clustering distancemeasurement.The experimental results show that the improved algorithm is more reasonable and effective.2.Based on the normalization of online learning data of students in Y university,a neural network model is established.Then three situations are predicted which are different classes of the same teacher,different teachers in the same course and different classes of different teachers in different courses.The results show that the model has good adaptability;moreover,this model is compared with the the traditional regression analysis model.The results show prediction of neural network method is superior to the regression analysis method in indicators we choose.Although the real data of Y university is taken as the research object,this paper makes some meaningful explorations on teaching running states and students' online learning.Also some preliminary conclusions with certain reference value are drawn.However,due to the short time and limited ability of the author,no more systematic software has been formed.In the future,further research will be done to enable teaching managers,teachers and students to grasp more real information of learning in time and constantly improve learning efficiency.