Study On The Application Of GEP Technology In Universities Student Achievement Analysis

It is the main content of the data mining technology that using different algorithms to dig out useful information from the large number of homogeneous or heterogeneous data to provide guidance for decision analysis. There are a large number of scholars in studying the data mining technology in many areas, and also got a lot of research results. Association rule mining is a new data mining technology which can discover the the useful relationships between data attributes from huge amounts of data, and using these relationships to decision analysis for guidance, it has been widely used in many fields, and achieved a lot in recent years.Evolutionary algorithm is a branch of data mining technology which is inspired by Darwin’s theory of evolution and Mendel’s genetics. The laws of competition between different species, survival of the fittest, are introduced to solve problems by generating an initial solution, and then the initial solution for breeding, competition, genetic process until approaching the optimal solution. Gene expression programming, as a new member of evolutionary algorithm, combines the original members of the evolutionary genetic algorithms with genetic programming advantages have outstanding contributions in solving complex problems in many areas.The concepts of data mining techniques, algorithms, applications, and classification were introduced firstly, and the relevant principles of the association rule, the algorithm processes, algorithms, category of algorithms were described in details. Then the traditional gene expression programming was introduced which has five parts:coding, fitness function, genetic manipulation, numerical constants, and carried out two improvements to the traditional methods. The initial population of traditional GEP algorithm is random, which may lead to uneven distribution of the first generation population easily, and then, may result in the process of limited genetic diversity of individuals, so that the algorithm converges to failure or problems of local optima, so carried out the improvement which homogenized population initialization for the algorithm; against the improve individual fitness function value calculated when the traditional GEP individualized evaluation needs to create gene expression tree, traversal, release constantly in solving complex problems, consuming a lot of time and space, so carried out the improvement which individual fitness function value calculating without expression tree. Analysing and contrasting the algorithm before and after the improvements by experiments. Combining gene expression programming with association rules to implement the algorithm of association rules based on gene expression programming in this article. Finally, making full use of association rules based on GEP algorithm to student achievement for data mining, analysing the information of teaching, especially the analysing of student achievement data in different courses which can provide guidance for decision-making to improve student achievement, teaching quality, and analysing the relationship between courses from the scores can verificate the rationality of the curriculum plan.