Research On Improved Optimization Algorithm Of Weighted Concept Lattice

Concept lattice named Galois lattice proposed by the German mathematician R.Wille is a knowledge representation model. It is also known as Formal Concept analysis. As one of the effective tools of data statistics and analysis concept lattice has two advantages. Firstly, each node in the concept lattice is composed with two different parts objects set and attributes set. Concept lattice is the binary relation established between the two hierarchies. It shows the concepts of generalization and specialization relationship between two nodes. Secondly, concept lattice has realized the visualization of data, a vivid expression of the various concepts between relations by their corresponding Hasse diagram. Since 1982, the concept lattice proposed by R.Wille, experts all over the world made a more in-depth research on the construction algorithm of concept lattice, extract the association rules and attribute reduction. Some of them discussed the combine of concept lattice and other disciplines theory. Due to its strong visibility and combination of other theory concept lattice has been widely used in pattern recognition, expert system, data analysis, decision analysis, data mining, information retrieval, and many other fields.At present, most of the studies are based on the assumption that the attributes of the concept lattice are equally important. However, in real life for the users the attributes tend to have different meaning and value. In the concept lattice witch constructed with equal important attributes contains too many redundant nodes. The redundant nodes cause the low efficiency of the construction of concept lattice. Related knowledge extracted based on such concept lattice is not only considerable and low-quality but also with higher complexity of time and space. The related rules extracted according to the minimum support and minimum confidence of association rules set by users can't satisfy the needs of the practical problems. The interactivity is so poor that sometimes the results may be deviate from the actual meaning.In order to make the weights value of attribute weights in weighted concept lattice more accurate and the association rules extracted to meet user demand this paper has done works as follows. In this article, on the attribute weight assignment methods of concept lattice and the construction algorithm of concept lattice are studied at the first of all. Then the intension weight value acquisition of weighted concept lattice based on information entropy and deviation is optimized to improve. The algorithm of association rules mining is proposed based on the improved weighted concept lattice. Finally the optimal weighted association rules mining algorithm of improved weighted concept lattice is applied to explore the risk assessment for Vibrio parahaemolyticus more accurately and effectively. This main research of this article and innovative ideas are as follows:1) This paper introduces the basic theory of classic concept lattice first of all. The introduction of related concepts and definition of weighted concept lattice are made to lay the foundation of further studies. In the view of different properties of weight assignment method and construction algorithm of concept lattice are studied. Intension weight value acquisition of weighted concept lattice based on information entropy and deviance is discussed emphatically.2) The improved algorithm is consisted by three different parts including single attribute intent weight value, the multiple attribute intent weight value and the adaptive threshold value interval. The weight optimization algorithm of concept lattice is proposed to improve the efficiency of constructing concept lattice. It makes that the users can achieve more real and reliable results of knowledge extraction.3) Through to the study on main construction algorithms of concept lattice mutual information is introduced to quantify the amount of information to guide policy on index tree structure optimization. In the construction process, the existing node does not satisfy importance threshold range to be deleted, so as to achieve the purpose of saving concept lattice structure of space and time.4) Weighted Concept Lattice structure optimization will optimize risk assessment algorithm for mining weighted association rules apply to Vibrio parahaemolyticus. Vibrio parahaemolyticus as a primary pathogen of foodborne illness caused by microorganisms brings great harm to people's health. According to support threshold and minimum confidence set by the users associated knowledge mining is done from pollution data including different sales, different quarters and different storage methods. The application is a great significance of aquatic food safety management and Vibrio parahaemolyticus risk assessment.