A Study On The College Admission Scores Predictions In New College Entrance Examination Based On An Artificial Neural Network

Since the 21st century,several rounds of reforms of the college entrance examination in China has been completed,which is closely concerned with each family.As the admission score line can only be announced after the wish-registering,candidates can only finish the wish-registering based on the known college entrance examination scores,provincial rankings and enrollment plans and so on.To avoid wasting scores or slipping when perform the wish-registering,it is necessary to predict the admission score line of universities.After the reform of the new college entrance examination,the application mode has been changed from school-based to major-based form,and the score prediction needs to be adjusted accordingly.At present,some widely used methods to predict the admission scores,such as the line difference method,the average ranking method,and the linear regression method,etc.,are based on predicting the college entrance examination scores without making adjustments to the new college entrance examination.This paper used the artificial neural network model on the basis of collecting data from various universities and majors over the years,and proposed a method to predict the admission score concerning the new college entrance examination.First,a neural network model was constructed,which includes node design,network parameter settings,and then the model was optimized to prevent overfitting.Secondly,the data of majors and schools of each university in the past three years were collected,and then the original data was filtered,and feature encoding and selection were performed according to the needs of the model.At the same time,a control group was created in which the average ratio of majors in previous years was set.Finally,the model solution and comparison with traditional forecasting methods were performed.The experimental results showed that the predicted value obtained by using the neural network model is relatively close to the original value,indicating that the model can accurately predict the score.Secondly,by comparing the sum of squared errors of the neural network model with the average ranking method and the online percentile algorithm,it was found that the error of the neural network model was the lowest,and the average ranking method was the highest,which also showed that the neural network model can be used to provide reference for prediction with its higher accuracy.Finally,when other conditions remain unchanged,the absolute error of the control group was lower than that of the experimental group,which proved that adding the average ratio could help to improve the accuracy of the model.Through experiments,this paper confirmed that the neural network model can effectively predict the admission score.With the inclusion of reference values,the accuracy of the model can be improved.Therefore,it provided a reference for wish-registering of the new college entrance examination.However,there are still some limitations in the consideration of some parameters,such as the development of colleges and universities and changes in their enrollment plans.In the future,it can be solved by adding more parameters and forecasting combinations.