Application Of Time Series Clustering In Learning Behavior Analysis

With the development of MOOC technology,universities have started to establish their own cloud classroom learning platform,where there is a large amount of learning data in the cloud classroom.In the era of big data,data mining and analysis is particularly important.Cluster analysis is unsupervised learning.Fuzzy clustering has been widely used since it was proposed,and fuzzy c-means clustering is an important fuzzy clustering method.Time series analysis is an ancient subject which has been widely used in product field.However,it is seldom used in the field of education.The time series model uses curve to fit the time series data and can predict the future trend of the series.In this paper,the ARMA(p,q) model is adopted to fit the prediction.This paper analyzes the learning behavior data of cloud classroom in a university and finds its characteristics.When clustering the user data,the Euclidean distance between the sample point and the cluster center is replaced by the time series similar distance between the sample point and the cluster center.According to the data characteristics,DTW is used to calculate the time series similarity distance between sample points and clustering centers,and a fuzzy c-means clustering algorithm(DTW-FCM) based on DTW is designed.The fuzzy c-means clustering algorithm based on DTW is applied to the preprocessed user data to obtain the membership degree of each user.Then,the ARMA model based on dtw-fcm is applied to the cloud classroom learning behavior analysis,and the processed learning behavior data is fitted and predicted.That is,according to the membership degree and other characteristic classification obtained by dtw-fcm clustering,then the data were analyzed by ARMA model to obtain the predicted results and analyze the predicted results.The contrast experimental results show that the ARMA model based on DTW-FCM is better than the ARMA model based on traditional FCM.In this paper,the effective analysis of the cloud classroom learning data,the information obtained can provide a certain reference for teaching,has a certain guiding significance.