Research On The Key Technology And System Implementation Of The Application For College Entrance

Higher education enrollment aspiration filling is related to the fate of candidates,yet many of candidates face problems of ‘high score but not to be enrolled proportionately' or‘high score but fail to be enrolled'.Higher education enrollment aspiration filling is a large and complex system engineering.Parents and students cannot fill out scientific and reasonable aspirations due to tight time and amounts of information.Therefore,developing a set of aspiration filling guidance system is of great practical significance.The key problem to aspiration filing is accurate prediction of minimum admission line of colleges and universities.There are three commonly used prediction methods at present:two-line difference method?admission difficulty coefficient method and score ranking method.These three methods are relatively simple to calculate,but they are all approximated by linear functions which cause big error existing in prediction results and real admission line.This paper presents a prediction method based on neural network,also designs and realizes a set of aspiration filling guidance system.The paper mainly includes two aspects in data prediction:(1)Prediction of admission line bases on neural network.Aiming at the highly nonlinear characteristic of admission data in previous years and the results of the year,this paper presents a method realizing prediction of admission line based on neural network.In this method particle swarm algorithm which is used to dynamically map the particle's own historical ability and global environmental cognition in the iterative process were firstly adopted to optimize the neural network model,solving the problem that the neural network algorithm is easy to fall into the local minimum.Through experiments on the data from 2012-2015 in Jiangsu province,we get the results showing that compared with the line method,the accuracy rate of prediction error that less than 1 point ?2 points can improve 12%?14% respectively,the prediction error that less than 3 points of the thirdbatch universities and higher vocational colleges improve 43% by using neural network method directly,which illustrates that adopting neural network to predict has a higher accuracy rate.Compared with directly using neural network,the optimized neural network model further improved the accuracy rate 4% and 8% of prediction error of first-batch and second-batch university's admission line that less than 1 point?2 point and 4% of the prediction error of the third-batch universities and higher vocational college that less than 3 points,the results show that the optimized model has higher forecast accuracy.(2)Six-stage professional recommendation method based on C4.5.Higher education enrollment aspiration filling rules stipulate every candidate fills 8 colleges and universities,each of which choose 6 majors in the same batch and the same class.Aiming at the shortcomings of the six-stage professional recommendation based on the line difference method,the sixth-stage professional recommendation method based on the C4.5 algorithm is proposed.The method has a significant improvement in the accuracy rate than based on the line difference method.On the basis of this,we use the equivalent infinitesimal theory to optimize the attribute of selection split of C4.5,which makes the calculation time greatly shorter than that of the original C4.5 algorithm.The result shows that compared with methods based on the line difference,the method based on C4.5 improves 11%?17%?20% and 26% respectively in the range of 0 stage for the first batch,second batch,third batch and higher vocational universities and colleges.Comparing the recommendation method based on optimized C4.5 with C4.5,the time complexity is changed from O(9)27)2)2n)to O(9)2),and the efficiency is obviously improved in the case of keeping the accuracy unchanged.In this paper,the implementation of the system includes two aspects:(1)In order to meet the business needs of the system and the expansion of the latter part of the business,this paper puts forward the method of data table partition in the database design for the problem of too many data items;It proposes a method of subcontracting according to the data fluctuation frequency aiming at the problem of data fluctuation;puts forward the method of how to make the child parent association with different enrollment codes in order to solve the problem that a number of enrollment codes exist in the same school in the same class;puts forward the scheme of the college entrance examination policy and voluntary reporting rules in view of the problem of the regular reform of the college entrance examination policy.(2)To improve the efficiency of Web access for Web requests,this paper proposes a Web efficient access control scheme based on Apache Shiro.The scheme encapsulates the access control module in the Web application and forms the form of the filter chain,allowing the communication of any connected modules in the chain,thus improving the access efficiency of the system.The experimental results show that when the number of concurrent threads reaches 5000,the throughput is increased from 24 rps to 41 rps,which improves 70% and the average response time is reduced from 1700 ms to 1100 ms,which is 35% lower.