Prediction And Application Research Of Teacher Training Results In Mixed Training

With the rapid development of China's education informatization,the application of information technology has improved,and various educational data have grown rapidly.In the face of educational big data,how to analyze the educational data and discover potential educational information and educational rules have spawned The birth of learning analysis technology and the research and application of data mining technology in education.From the current research status quo,the research on education data mining analysis is gradually moving from concept to concept level to analysis and application level.It has done a lot of research work in academic performance prediction and evaluation,but still faces many challenges.This study provides a data support for further optimization of the learning process by constructing a learner's academic performance prediction model and using data mining techniques to analyze learning behavior data and academic performance.This paper applies the idea of data mining,taking the relevant data of the participants in the national training program(2016)-model network training and school-based training integrated training program in a certain area as the research object,using decision tree,rule induction,K-Neighbors,artificial neural networks,naive Bayesian five classification algorithms and linear regression algorithms to construct prediction models,which are expected to provide decision support for teacher training participants and related managers.The main research contents mainly include:(1)The selection of academic factors,the analysis of factors affecting students'academic performance can be considered from a variety of perspectives,such as the entire curriculum implementation perspective,learner perspective,teacher perspective.In this study,from the perspective of learners,the behavioral data and final results of the teacher training staff on the teacher training platform were selected to construct the academic performance prediction model.(2)Collection and processing of learning data.A large amount of learning behavior data is the basis for educational data mining and learning analysis.During the learning process,students have recorded their learning behavior data during the use of the online platform,including online learning duration.,forum discussion,job submissions,etc.A series of data preprocessing operations such as conversion and cleanup of the collected data sets are introduced.Finally,1534 standard data samples are formed,and the academic behavior attributes used for performance prediction are determined and applied to the construction of the prediction model.(3)Education data modeling and analysis.In the modeling research,statistical methods are first used to analyze the dataset factors.Secondly,the artificial neural network algorithm or decision tree algorithm is used to predict the academic performance.Forecast research.(4)The application of academic achievement prediction model,the construction of six performance prediction models is constructed.Based on the overview of each algorithm,the generation process and prediction results of the prediction model are introduced.Among the five prediction models constructed,the most accurate one is the prediction model based on decision tree construction,and the lowest accuracy is the prediction model based on naive Bayesian construction.(5)Comparing the accuracy,advantages and disadvantages of the five classification prediction models,comprehensively considering the artificial neural network model as the prediction tool model,and using the Python language to create a performance prediction tool.The conclusion of the study is:Through the construction and operation of the prediction model,it is found that in the learning behaviors collected by the network training platform,the behaviors that have a greater impact on the total scores have excellent number of jobs,the number of workshops submitted and the number of workshop activities,and qualified operations.The number and the number of courses to be studied,the participating trainees can adjust their learning strategies to obtain better training results according to their own training situation,and the important learning behaviors that affect the total scores.The relevant managers and implementers can refer to the results to adjust.Develop implementation plans and optimize evaluation methods.The innovation of the thesis is:using a variety of data mining forecasting methods and predicting academic performance based on academic behavior in teacher training,and finding learning behavior factors that have an important impact on the final grade of the training,can provide certain participants for the teacher training program.Decision making reference.