In this paper we present clustering method is very sensitive to the initial center values ,requirements on the data set too high, and cannot handle noisy data the proposal method ...
Feature selection is a critical procedure in many pattern recognition applications. There are two distinct mechanisms for feature selection namely the wrapper methods and the filte...
Michal Haindl, Petr Somol, Dimitrios Ververidis, C...
This paper demonstrates so-called sentinels in the TARGIT BI Suite. Sentinels are a novel type of rules that can warn a user if one or more measure changes in a multi-dimensional ...
Metaheuristics represent an important class of techniques to solve, approximately, hard combinatorial optimization problems for which the use of exact methods is impractical. In t...
Alexandre Plastino, Erick R. Fonseca, Richard Fuch...
Previous study has shown that mining frequent patterns with length-decreasing support constraint is very helpful in removing some uninteresting patterns based on the observation t...