Research On Data Intelligence Analysis And Its Application

Nowadays,with the rapid progress of science and technology,especially computer science,the information explosion of the network age,we have already placed in the big data age.Faced with a massive amount of data at tens of thousands of terabytes,people seem intractable.Data mining technology is to solve this problem,it covers the statistical science,data science,machine learning and artificial intelligence and other fields of knowledge,to find a large number of potential and useful knowledge,to facilitate decision makers.The most significant two capabilities in data mining are the findings of hidden Association rules and the classification of unknown samples.Firstly,the thesis analyzes the theory basis and idea of association rule Mining and Decision tree classification,studies the strengths and shortcomings of association rule Mining Apriori algorithm and decision tree classification C4.5 algorithm,and analyzes the various improvement schemes proposed by previous scholars,and compares their improvement results.and tries to find the corresponding solution.Secondly,according to the shortcoming of frequent scanning database and establishing a large number of candidate itemsets in association rule Apriori algorithm,this paper proposes an improved algorithm of mining frequent itemsets combining the position index table of projection and order frequent itemsets,and verifies the data set by experimental analysis.It is found that it improves the time and space efficiency of searching frequent itemsets.Third,In order to solve the problem that the decision tree C4.5 algorithm needs a large number of logarithm operations and the disadvantage of high dimensional data,a method of combining binary particle swarm optimization(BPSO)and decision Tree C4.5 algorithm is proposed to predict the classification.Firstly,the binary particle Swarm optimization(BPSO)algorithm is used to select the attributes,and then the decision tree is established by the decision tree classification C4.5 algorithm.The experimental results prove that it enhances the efficiency and accuracy of the primitive C4.5 algorithm.has certain feasibility.At last,by obtaining the routing log information of the Internet behavior of teachers and students in Jiangnan University,the data is preprocessed,and the relevant models and algorithms of data mining are used to analyze the characteristics and rules of the online behavior of teachers and students,and through visual programming technology,the concrete information of users ' online behavior is displayed visually.In summary,this paper has made some progress in the research of association rule mining and Decision tree classification,which has some research value.And the design of a data analysis platform,intelligent and intuitive display of users' online behavior,has a certain practical significance.