Research Of College Entrance Examination Scores Prediction Method Based Artificial Neural Network

Since the 21 st century,The higher education in China has been transformed from "elite education" to "popular education",It affects many families and their future.Completing the volunteer as a key link in the college entrance examination,Needless to say how important it is.Due to the lack of knowledge of the process and related information in the college entrance examination volunteered,Most candidates will filled in unreasonable volunteered.The result is that candidates lose their ranks or wasted their scores.In recent years,the reform of the Chinese college entrance examination system has been implemented,and now several provinces have already completed the merger of the two or three admission batches.Therefore,candidates will face many new problems in volunteering.They not only need to understand how to fill in the college entrance examination voluntary after the batch merger,but also make adjustments in school selection and score line prediction.In the current market,most of the prediction of the college entrance examination score use the conventional forecasting method,few adjustments have been made to the current admission merger.Even though a few fractional line projections take into account the merger situation,But the prediction is not clear enough.Arounding by college application fill out,used the current popular artificial neural network technology to predict college admission scores on the basis of collecting college entrance exam admission score data over the years.This article mainly includes the following aspects.Analysis of feasibility of the project: Analyze the application process and the status of merger admission.Combined with artificial neural network technology,this paper puts forward the prediction method of the college entrance examination score line based on artificial neural network in the background of admission merger.Design and generate data sets that need to be used when constructing artificial neural network model.The influence factors of the fractional line are extracted by analyzing the existing methods of the college entrance examination score line prediction and the parallel voluntary filling process.Design the characteristics of the year of voluntary merger in order to support the merger policy.Collect the influence factors of the college entrance examination score line on the Internet by crawling techniques,manual collection,etc.,Converted collected data into which artificial neural network can received and processed by means of eigenvalue conversion,missing value exclusion and normalization.Design and build a college entrance scores prediction model: According to the collected college entrance examination related data,design an appropriate artificial neural network architecture such as input layer,hidden layer,and output layer.Configure the development environment and implement the neural network through the TensorFlow framework after completing the design of the artificial neural network architecture.Optimized artificial neural network model parameters through experiments: Using artificial neural network model of merger admission province colleges and universities admit fractional line to study,continuously optimize model parameters by experiment and get united in admissions under the background of the university entrance exam score prediction model.Perform performance test on the artificial neural network model: By setting up test standard for artificial neural network model to tested the accuracy,recall and in Second and third admission batches merge.Through the above series of analysis and experiments,the college entrance examination score line prediction model obtained in this study can achieve an prediction accuracy of about 80% in the cross-validation test with an error value of ±2 points.The college entrance score data obtained in this paper was not used in previous studies.By adopting the new data,the new way for predicting the college entrance examination breaks the traditional way which using point difference,provides new ideas for the prediction of college entrance examination scores.It enriched the application of artificial neural networks in China on practical issues,and provided reference for the community to understand more deeply about volunteered and admission merger.