A Study On Predicting The Admission Ranking Of College Applicants And Candidate Volunteer Recommendation Under The New College Entrance Examination

In recent years,the new college entrance examination reform has been implemented in batches in various provinces.The changes in the new and old college entrance examinations have led to significant errors in predicting the admission ranking of college applicants when filling out their applications.Moreover,various factors,such as the annual college enrollment plan,the number of candidates for the college entrance examination,and the school’s reputation,will all affect the number of university applicants.When the position of candidates meets the requirements of colleges and universities,the multi-level preference of candidates for colleges and universities will determine the final voluntary choice.In the era of the information explosion,how to select colleges and universities from the vast amount of college information within a limited time combined with their preferences is also a problem faced by candidates.In order to solve the above problems,this article takes the Jiangsu New College Entrance Examination,which has undergone significant changes in the new and old college entrance exams and is difficult to fill out voluntary applications,for example.Firstly,a new prediction model for college admission positions is proposed to capture the comprehensive impact of multiple-factor changes on college admission positions in recent years.When training the model,the feature of selected subject grades is added to narrow the difference between the scoring rules of the Jiangsu New and Old College Entrance Examinations and further improve the accuracy of admission position prediction.Then,a volunteer recommendation model based on the multi-level preferences of candidates is proposed,and based on the actual scores of candidates and their multi-level preferences,college volunteer recommendation is carried out.The above-proposed model for predicting the ranking of applicants and the voluntary recommendation model is applicable to the Jiangsu New College Entrance Examination and can be extended to other provinces for use in the New College Entrance Examination.The specific research content of this article is as follows:(1)Collect and organize data from universities in recent years and preprocess it.Firstly,collect and organize data on the ranking of universities in recent years,and carry out targeted missing value processing.Then,use label encoding to quantify the characteristics of the select grades of old Jiangsu universities in the college entrance examination.Finally,standardize the deviation of different strength attribute scores of universities in the comprehensive strength ranking data obtained using crawler technology and OCR(Optical Character Recognition)technology.The collected and preprocessed data will be used for subsequent experiments on admission ranking of college applicants prediction and voluntary university recommendation.(2)Propose a clustering method based on the trend of university investment positions in recent years,called T(Trend)clustering.Based on T clustering,propose a combination model of BP network and GM(1,1)based on position trend clustering(Trend Clustering BP&GM(1,1),TCx BPGM).The TCx BPGM model can capture the comprehensive impact of multiple factors on the ranking of universities in recent years.Under this influence,the ranking of universities over the years shows different trends.The model uses different methods to combine ranking predictions for universities with different ranking categories,improving prediction accuracy.And when training the model,the quantified characteristics of subject selection level are added to narrow the difference between the new and old college entrance exams,further improving the accuracy of prediction.(3)Propose a volunteer recommendation model based on candidates’ multi-level preferences.Firstly,the model combines the actual scores of candidates and the ranking data of universities over the years to divide them into "rush","stable",and "guarantee" volunteer gradients.Then,through questionnaire surveys,the model collects candidate performance and multi-level preference information and quantifies preference data.Finally,the volunteer recommendation model calculates a recommendation index based on the candidate’s multi-level preferences and recommends a certain number of universities to candidates with different volunteer gradients.