Research On IT Vocational Education Based On Data Mining

With the continuous development of Internet technology,information technology has been involved in all walks of life.With the rise of information technology,the shortage of IT talents has become more and more serious in many fields.The shortage of IT talent is an important factor affecting the rapid adoption of information technology.Due to the shortage of IT talents and the needs of the market,IT vocational education and training institutions have emerged in the market.This paper takes 1342 learners from 18 classes in the Java training business section of Beijing Baizhi Education Technology Co.,Ltd.as the research object.Based on the induction and analysis,the data mining method is used to compare K nearest neighbors(KNN,KNearest Neighbor algorithm,support vector machine(SVM)algorithm and the advantages and disadvantages of combining K nearest neighbor algorithm and support vector machine algorithm into SVM-KNN new algorithm,constructing classification prediction model,and realizing the training institution through the constructed prediction model Prediction of graduation status.According to the results of analysis and prediction,find out the key factors affecting learning in the learner attribute characteristics,which can provide IT vocational education learners and educators with good suggestions for improving learning outcomes: help students find relevant problems in learning as early as possible.Improve academic performance;and improve the overall quality of graduates.The dataset attributes collected in this paper include: professional,academic level,three stages in the school period-core Java stage,WEB stage,and framework stage.The characteristics of the five aspects of the academic performance,in-depth analysis and discussion,get The corresponding conclusions.The purpose of the research is to predict the graduation status of each student through IT vocational education data,establish a predictive model through data mining algorithms,and evaluate the model prediction by measuringseveral main indicators of machine learning prediction model--accuracy,accuracy and recall rate.effect.The specific work is as follows:(1)Based on the IT vocational education dataset,the statistical analysis of the learner's attribute characteristics,through the statistical analysis of the attribute characteristics and the use of SPSS tools to analyze the learner's attribute characteristics and graduation status Pearson correlation analysis,after analysis Outcome: There is a significant correlation between professional attributes,academic level attributes,and grade attributes in the data set and graduation status,while age attributes and gender attributes have weak or no correlation with graduation status.Through Pearson correlation analysis,it provides the basis for further preprocessing of data,and more importantly,provides suitable selection of attributes for building models.(2)In order to construct a more accurate classification algorithm prediction model,the data set is preprocessed by data preprocessing method—data integration,data cleansing and feature transformation,so as to prepare for the subsequent establishment of the algorithm prediction model.(3)Establish a KNN classifier to predict the graduate status of the learner.The prediction results of the classifier based on KNN algorithm on accuracy,precision and recall rate are90.31%,89.33% and 70.31%,respectively.The best K value is determined by the F1 score indicator.When K=7,the F1 score is the highest,which is 78.69%.(4)Establish SVM classifier,predict the graduation state,and build a prediction model based on this algorithm,and evaluate the performance of the model by training the model.The prediction results of the classifier based on the SVM algorithm for accuracy,accuracy and recall are 84.22%,74.97% and 89.95%,respectively.(5)Contrasting and analyzing the prediction results of the two classifiers mentioned above,summarizing the advantages and disadvantages of the KNN algorithm and the SVM algorithm,and then combining the two algorithms into a new algorithm-SVM-KNN algorithm.The prediction results of the classifier based on SVM-KNN algorithm on accuracy,precision and recall rate are 90.43%,88.15% and 88.01%,respectively.Compared to the KNN classifier,the SVM-KNN classifier reduces the prediction accuracy by 1.18%;compared to the SVM classifier,the SVM-KNN classifier reduces the recall rate by 1.84%.However,the overall performance improvement of the SVM-KNN algorithm,especially the prediction recall rate based on the KNN classifier and the prediction accuracy based on the SVM classifier are particularly significant.Compared with the classifier established by SVM algorithm,the classifier established by SVM-KNN algorithm improves the prediction accuracy by 13.18%;compared with the classifier established by KNN algorithm,the classifier established by SVM-KNN algorithm is in the recall rate.The forecast has increased by 17.7%.The experimental results of this paper show that the attribute characteristics of IT vocational education learners can be used to predict their graduation status,which can provide a basis for educators to optimize decision-making and optimize teaching methods to improve the overall quality of graduates;Suitable and efficient learning methods to improve their learning efficiency and learning effects.