Research And Implementation Of Data Analysis Tool Based On Concept System

The purpose of data analysis is to help users extract valuable information from data,discover original knowledge,and it plays an important role in decision making,among which the representation and expression of knowledge is the key step.As an important knowledge representation pattern,association rules can effectively and intuitively enable people to conduct data analysis and make decisions,so it is necessary to visualize association rules in a clear and intuitive way.However,the existing visual analysis of association rules lacks the definition of users' personal interest points,ignores the knowledge background in the visualization process,and the analysis process is not suitable for users' analysis behavior pattern.As the main data structure in the theory of Formal Concept Analysis(FCA),concept lattice can directly show the generalization and instantiation of all concepts in form,and increase the understanding of users to knowledge.This paper studies and implements an analysis tool based on concept system,combines formal concept analysis theory with association rules,realizes visual analysis,enhances user interaction,and improves the accuracy of data analysis.The main research contents of this paper are as follows:(1)The visual representation of association rules based on concept lattice is studied deeply,overlay foreground knowledge on full conceptual background and a visualization method of association rules based on concept lattice is proposed,at the same time the user-centered heuristic interaction is realized to enhance the user's comprehensive knowledge and understanding on this basis of this paper.(2)To break through the limitations existing in the process of data analysis system used in the practice,this article explores and summarizes three characteristics of user data analysis : order,repeatability and multiplex,and on this basis,it further designs the basic data analysis functions such as diverse data sources,user interest point attribute definition,association rule visualization and so on.Not only in the functional structure of the level of research and design,but also the overall execution order and module processing logic is defined.Thus found more meaningful data analysis tools can be designed through the exploration and research of behavioral characteristics.(3)By combining the system with the practical application,the function of defining the attributes based on the user's interest points is designed.By defining a subset of interest on the data set,we can focus on the user's target requirements and narrow the target range.(4)Under the premise of following the software engineering method,the data analysis prototype system which can assist the user in data analysis is developed and realized independently.It not only designs and realizes the data analysis function that fits the user's analysis behavior characteristics,but also provides an open data interface to facilitate the algorithm embedding of the follow-up research results,so as to better meet the user's analysis needs.