A New Approach for Mining Representative Patterns

Abstract

With the revolution of science and technology, we can accumulate huge amount of data which requires to be manipulated efficiently since the amount of data is expanding hence scarcity of knowledge is also increasing. Therefore analysis for more useful and interesting knowledge is on demand. Representative patterns can be a solution to represent data in a more concise way. Different efficient methods for mining frequent and erasable patterns exist in representative pattern mining field that are regarded as significant. We have proposed a new type of pattern called decaying pattern. These patterns are characterized as those patterns that were frequent for a time being and then decayed with time. These patterns of declining nature can give us the opportunity to analyze reasons behind items’ decrease such as extinct animals, finding unsolved accidental news, analysis of buying behavior of customers etc. that require further inspection.