Related Demand Expression And Recommendation Model Based On Concept Lattice

With the rapid development of technology, and the progress of economy, the pace of life inthe information age is accelerated. In such business environment, improving customer satisfactionand maintain competitive advantage becomes the primary issues that businesses concern about. Inaddition, in the business environment with high consumption, customers’ diversification demandspromote the businesses to increase emphasis on integrated services. The related demand is comeup to meet this market background. Integrating the related demands, and mining associationdemands with the customer’s individual requirements will provide customer integrated solutions.On the one hand, the time that customers spent to buy other products/services can be saved. Onthe other hand, the integrated related demand can enhance customer value, thus improvingcustomer satisfaction. This helps companies gain competitive advantages in the highlycompetitive market. The concept lattice which was made by Wille, a German professor, reflectsthe relationship between objects and attributes, and relationships with specialized generalization.It is a concept hierarchy, and can be used in related demand expression and mining. Therefore, inthis paper, based on the concept lattice theory, the related demand expression and mining, and therecommendation of related demands were studied. The main work is as follows:(1) The analysis of characteristics of related demand. With systems analysis, studyingtheories of consumer behavior and micro economics, the paper accesses to relevant literature toanalyze and summarize the characteristics of related demand.(2) The establishment of two types of expression and mining models of related demandbased on concept lattice. The paper establishes a general concept lattice expression and conceptlattice expression with interest measure. The former only considers the frequency of needs in thedatabase, but the latter also takes the value of demand into consideration. The paper then minesthe association demands on each model based on concept lattice theories.(3) The establishment of two types of related demand recommendation models. Use AnalyticHierarchy Process to establish two types of related demand recommendation models. One is withdynamic programming method and the other is with greedy algorithm. Both are aiming tomaximize customer value, and are constrained to budget. The former through repeated iterationsand backtracking must be able to get the optimal solution, but when the demand quantity is large,it is more suitable to choose simple, intuitive, and efficient greedy algorithm. (4) The analysis of case study of customer related demand. Using an example to verify thefeasibility of the model established in this paper and the accuracy of the model results andanalysis. It provides practical guidance for the future research of related demand.