An Analysis And Prediction Model Of The Influencing Factors Of Student Achievement Based On Data Mining Technology

For the current students,data mining is an enduring discipline,and for the workers engaged in data mining,it is a profound understanding of the strong development prospects of data mining.This paper is mainly divided into two parts.The first part mainly analyzes the factors that affect students' performance,and finds out the most obvious factors that affect students' scores.Because of the large number of data dimensions used,the operation of establishing multidimensional data sets is adopted for the data set,in which the contents of multidimensional data information database include students' name,student number,ID card number,admission card number and various subjects Objective examination results,source of students,junior high school graduation results,teachers' teaching situation,etc.After the establishment of multi-dimensional data information base,K-means clustering method is applied to deal with outliers of some data,and principal component analysis(PCA)is also used to reduce the dimension of some data.Finally,this paper analyzes the impact on students' performance from three aspects: the source of students,entrance scores,and teachers' teaching.In the second part of this paper,we mainly establish a model to predict the performance of a subject.After getting the data set,we first complete the data preprocessing operation.Because the methods used in this paper are two kinds of decision tree algorithms,namely ID3 and C4.5 algorithm,we must first discretize the data used,and here we do the same width box processing for the students' scores,experimental results and the number of students' questions.Then the two algorithms are evaluated and C4.5 algorithm is used as the modeling algorithm.In order to establish the characteristic factors that affect students' performance,this paper compares the influence of family relationship,the number of students doing programming questions,learning time after class,age and parents' education level on students' performance,and finds out that the characteristic factors are: the number of programming questions,the scores of programming questions and the achievement of each chapter.Due to the different degree of influence of each characteristic factor on the score,the concept of weight is introduced when the C4.5 algorithm is used to complete the modeling,that is,the weight is added to the C4.5 algorithm to calculate the information gain rate,so as to complete the improvement of the C4.5 algorithm.Finally,the model evaluation shows that the improved algorithm is obviously feasible because of the improved algorithm.This paper mainly explores the factors that have the most obvious impact on students' performance,and selects the appropriate algorithm to establish the prediction model based on these characteristic factors,so as to complete the prediction and analysis of students' performance.This paper mainly explores the internal relationship between students' performance and various factors.It is of great significance for teachers' education and improvement of students' scores.