Research And Application Of Analysis And Forecast System Of Students' Experimental Behavior

With the development of digital information technology,massive amounts of data are also constantly produced,and today's society has entered the era of big data.In the field of education and teaching,big data is also playing an increasingly important role.In many universities today,students always need to take many courses and conduct relevant course experiments during their college careers.They will generate a large amount of experimental behavior data during the experiments.These experimental behavior data are mainly reflected in students' experimental time,experimental location,experimental operation commands,etc.,and these experimental behavior data are often not effectively collected and utilized.If the relevant experimental behavior data can be collected and then be analyzed by using data mining technology,it is possible to analyze and predict the correlation between some data of students' experiments and students' achievement and teachers' teaching,which can help teachers improve their teaching quality and provide a basis for the school's teaching reform.On the basis of analyzing the relationship between experimental teaching and big data-related technology,this thesis designs and implements a relatively complete system for analyzing and predicting students' experimental behaviors.The system is mainly composed of students' online experiment platform and experimental behavior analysis system.Among them,the students' online experiment platform is based on a Linux server,which provides college students with a real-time online course experiment environment,and can also record and collect students' experimental behavior data.The main body of the experimental behavior analysis system includes: a front-end client developed by using the Layui framework,a back-end server developed by using the SpringBoot framework,and a big data distributed computing framework based on the Hadoop-Spark framework.Based on the consideration of the system's analysis and forecasting needs,the system designed and implemented in this thesis adopts the method of combining clustering algorithm and regression algorithm to analyze and predict the students' performance.In this thesis,the K-means ++ algorithm is used to quickly cluster the students' experimental behavior data,and then the KNN regression prediction algorithm is used to perform regression prediction modeling on collected and processed multidimensional data such as students' information,experimental location,experimental time,and experimental command.And finally achieve the effect of relatively accurately predicting students' experimental behavior performance and assisting teachers to evaluate students' achievements.Finally,combined with the functions of the main body of the system studied in this thesis,the results of experimental performance prediction and other experimental behavior analysis results can be visualized on the front page in the form of charts.It has been verified that the system facilitates teachers and school leaders to conduct research and analysis of students' experimental behaviors,plays a role in assisting teachers in teaching,and also provides a reference basis for college teaching reform.