Research And Implementation Of College Entrance Examination Scores Prediction System Based On Machine Learning

College entrance examination(CEE) as the most important way of admission to universities in China, has an important position. But most of the candidates every year to fill the voluntary, basically in accordance with the previous year's college entrance examination scores, and the accuracy of this approach has repeatedly proved to be not enough.Therefore, how to predict the target college and target professional admission score more accurately, has become the current college entrance examination to fill the voluntary in the urgent need to solve the problem, which is the focus of this paper.This paper presents a method of using Adaboot to predict college entrance examination scores, and puts forward the method of using deep learning to predict college admission scores.Firstly, we study the method of using Aaboost to predict college entrance examination scores. It can be found that Adaboost algorithm has a good performance compared with random forest algorithm by a lot of experiments. Through a series of feature engineering, such as adjustment, processing and combination of eigenvalues, Parameter adjustment, etc., making the model for the prediction of college entrance examination scores have a good effect, the error value in about five points. Then, the paper puts forward the method of predicting college admission scores using deep learning method. In this paper, the neural network model of multi-hidden layer structure is used. Through the improvement of the model's input layer and the adjustment of the hidden layer number, and the data, the best prediction model has been obtained.The result shows that the prediction result of the professional admission score is well.In the end, this paper designs and implements a prediction system of college entrance examination score based on machine learning. The system in the design and implementation to take into account the high concurrency, large amount of data, but also taking into account the data security situation. This system is easy to use and has accurate forecasting results. It can provide good forecasting service for candidates and their parents.