Herramienta software para el análisis de canasta de mercado sin selección de candidatos

Tools leading to useful knowledge being obtained for supporting marketing decisions being taken are currently needed in the ecommerce environment. A process is needed for this which uses a series of techniques for data-processing; data-mining is one such technique enabling automatic information discovery. This work presents the association rules as a suitable technique for discovering how customers buy from a company offering business to consumer (B2C) e-business, aimed at supporting decision-making in supplying its customers or capturing new ones. Many algorithms such as A priori, DHP, Partition, FP-Growth and Eclat are available for implementing association rules; the following criteria were defined for selecting the appropriate algorithm: database insert, computational cost, performance and execution time. The development of a software tool is also presented which involved the CRISP-DM approach; this software tool was formed by the following four sub-modules: data pre-processing, data-mining, results analysis and results application. The application design used three-layer architecture: presentation logic, business logic and service logic. Data warehouse design and algorithm design were included in developing this data-mining software tool. It was tested by using a FoodMart company database; the tests included performance, functionality and results' validity, thereby allowing association rules to be found. The results led to concluding that using association rules as a data mining technique facilitates analysing volumes of information for B2C e-business services which represents a competitive advantage for those companies using Internet as their sales' media.