Paper Details

Functional Association rule form is novel form of association rules (ARs) that do not requirediscretization of continuous variables or the use of interval values in either sides of the rule. This rule form capturesnonlinear relationships among continuous variables, and provides an alternative pattern representation for miningessential relations hidden in a given data. A new neural network based, co-operative, coevolutionary algorithm ispresented for FAR mining. Conventionally, ARM is majorly concerned with categorical data sets. When it is used toprocess continuous variables, it converts the values of the variables into intervals. This discretization process determinesthe granularity of the ARs being generated and generates granularity levels for ARs. In contrast, the FAR proposed canhandle nonlinearity in the relationship and can deal with continuous variables directly and without converting them intointervals. Deep Learning approach is used to increase the accuracy of FAR mining. A new measure for accuracy isintroduced. K-means clustering used for normalization of the continuous data.

Ms. Pooja Kulkarni, Mrs. Mrs. V. L. Kolhe, “Functional Association Rule Mining Using Cooperative Coevolutionary Deep Neural Networks”, International Journal of Advance Engineering and Research Development (IJAERD), Vol. 4, Issue 7, pp. 341-346, July 2017.