Research And Application Of College Entrance Examination Volunteer Personalized Recommendation Method Based On Machine Learning

It is particularly important for the majority of candidates to fill in and be admitted to the college entrance examination,which greatly affects the future development of candidates.In the past,the college entrance examination volunteer filling guidance was based on experience,which was very arbitrary,and the test survived at the risk of not being admitted,resulting in regret.There were some cases of candidates not being admitted,which caused regret.Many existing commercial college entrance examination voluntary recommendation systems attach too much importance to individuation,the actual admission rate is not high.Voluntary recommendation for college entrance examination needs to be scientific,intelligent and personalized,and needs to be transformed from human experience to machine learning.To solve these problems,a personalized recommendation method based on LSTM regression prediction and FCM cluster analysis is applied in this study.The main work of this paper includes the following parts:(1)A six-layer LSTM neural network was constructed,and neural network model training was carried out using four batches of liberal arts and science data from the second batch of undergraduate online.The university admission rank was predicted and model test was carried out,and Pearson correlation coefficient of this data set was calculated.Among them,the correlation degree of Class A data in the first batch was the highest,and its R2 value reached0.95.Since the second batch of undergraduate,due to the gradual increase of data dispersion,the prediction effect decreases.And found each batch of subjects with high probability of admission rank.The FCM cluster analysis is made on the range of admitted colleges and universities in a certain ranking segment.The cluster analysis is carried out from three perspectives of college category,college popularity and college region respectively,which provides an important basis for personalized recommendation.(2)The classification analysis is carried out on the examinees.The examinees with one science and engineering major and two literature and history majors and the Gunda line are selected as examples.The personalized recommendation method based on the combination of LSTM regression prediction and FCM cluster analysis is used to make personalized recommendation for these two examinees in combination with their personal voluntary inclination,and a higher admission rate is obtained.The results show that the personalized recommendation method proposed in this paper is guaranteed in terms of personalized recommendation and admission rate.