Research And Implementation Of Intelligent Recommendation Technology For College Entrance Examination Based On Machine Learning

The new college entrance examination was officially launched in 2014,bringing new modes and policies for filling in the application forms,making the application modes diverse and the number of applications far exceeding the past,which requires candidates to master more information.Therefore,using big data technology to process and predict college entrance examination information,and provide candidates with an effective and reliable application recommendation scheme is not aimless.The main work is as follows:(1)This thesis takes the probability of candidates being admitted by colleges and majors as the evaluation criterion,comprehensively considers multiple data features such as the nature of running schools,school types,whether they are double first-class universities,etc.,predicts the admission scores of colleges and universities,and provides reliable application recommendations for candidates based on this.This thesis crawls the admission situation of various colleges and majors in the country in the previous 4 years and the “one score one segment table”from the Internet,and uses two machine learning algorithms with the best results for prediction,and then uses linear regression to obtain the final prediction score.(2)This thesis recalls the major library,uses jaccard similarity coefficient and fastText to further recall the basic major pool,and obtains the major candidate pool.This thesis sorts the major candidate pool,uses the predicted score and LSTM model to predict the admission probability of each major,and obtains two ranking results according to the admission probability and multiple feature weights.This thesis comprehensively analyzes the two ranking results,and recommends according to the personalized needs of candidates.(3)This thesis develops a college application recommendation system,which aims to provide personalized support and efficient help for candidates when filling in their applications,and solve the problems of information overload and decision difficulty under big data environment.