Research On College Entrance Examination Voluntary Recommendation For Diversified Needs

With the popularization of high school education,how to fill out the college entrance examination has attracted more and more attention from candidates,but the existing college entrance examination volunteer application systems simply matches the colleges or majors that meet the candidates’ scores,and does not comprehensively consider the diversified needs of candidates in terms of colleges,majors,geographic locations and other factors for recommendation,so the recommended results cannot meet the actual needs of candidates and provide more scientific and reasonable guidance for candidates to fill out the volunteer application.Therefore,the recommendation results cannot meet the actual needs of the candidates and provide more scientific and reasonable guidance for the candidates to fill out the volunteer application.Therefore,it is important to study how to make voluntary recommendations and meet the diversified needs of different candidates.First of all,this thesis studies the prediction of the lowest admission ranking of institutions,which is a prerequisite for determining the range of voluntary applications and calculating the probability of admission of majors.In this thesis,the exponential smoothing algorithm in time series prediction is used to predict the lowest admission rank in the college entrance examination,and the smoothing parameter that minimizes the prediction error is calculated according to the improved genetic algorithm,taking into account the differences in data trends reflected by the time series composed of the lowest admission data of different colleges and universities in the past years.The three different exponential smoothing algorithms were fused.The experimental results show that the fused algorithms have better prediction performance.Secondly,when performing volunteer recommendation,the user’s interest in different types of institutions is quantified based on the user browsing behavior data,and the preference weights of candidates’ different needs are obtained from the magnitude of interest,while the professional admission data of institutions are studied and a professional admission probability model is constructed for dividing the volunteer gradient.Based on the above work,the user demand preferences are added to the institutions under each index,and the volunteer recommendation is performed by the feature-weighted FCM fuzzy clustering algorithm.Experiments are conducted to compare the recommendation effects of different systems and algorithms using similar strategies in the market for each score band,and the results show that the algorithm in this thesis has a high volunteer recommendation accuracy and score maximization utilization considering the diversified needs of users.Finally,we conducted research on the functions of popular volunteer application platforms on the market,summarized the functional modules of the college entrance examination volunteer recommendation system,completed the analysis of the system requirements and overall design and demonstrated the effect of volunteer recommendation function,and applied the algorithm for predicting the lowest ranking of college admissions and volunteer recommendation algorithm to the real volunteer application scenario,and finally realized the system for assisting college entrance examination candidates in volunteer application.The system has strong practical value and practical significance.