A Data Mining Method For The Solution Of Fluid-Flow Problem

Discovering market baskets is an attractive topic of data mining theory. In most studies the researchers considered the purchases or the transactions of the customers as a set of items. In other words, the transactions are analysed on conceptual level. In some recent studies quantitative approach was proposed. The researchers are interested not only in the items or products chosen by the customers, but also in the quantity of each item or product in the transactions. This study is a continuation of a previous research. The transactions are characterized as a set of quantified items. This allows us to use algebraic methods as efficient tools to establish a formal description of market basket model. The advantage of this approach is that we can have better insight to the natural order between transactions and the role of network structure on them. In this general model the well-known results of lattice theory are applied. The explicit representation of frequent transactions or market baskets, as well as of the association rules are tested by a program. The results of testing show that the algorithms finding frequent transactions and confident association rules are efficient and can be used to explore frequent customers’ baskets, the association rules between the products, as well as the natural associations between different flows of the economic activities.