Fraction-score : a new support measure for co-location pattern mining

Abstract

Co-location data mining are popular on spatial objects with categorical labels, this brings about the interesting pattern that objects with various labels are frequently located in close geographic proximity. Similar to numerous item sets, co location patterns are defined based on a support measure that appraise the popularity or prevalence of a pattern candidate or a label set. Some support measures exist to define the co location patterns and they share an idea of counting the number of instances of a given set C as i ts support, where an instance of C is an object whose objects carry all the labels in C and are located close to one another. However, these support measures suffer from various weaknesses, some will fail to notice all possible instances while the other support measures will ignore some cases when multiple instances overlap . In this FYP , a new support measure will be proposed which is called the Fraction Score. Its main idea is to count instances fractionally if they overlap. Compared to the existing support measures, Fraction Score not only capture all possible instances, but also handles the cases where instances overlap appropriately so that the supports defined are more meaningful and consistent with the desirable anti-monotonicity property . Efficient algorithms have been developed which are significantly faster than a baseline that adapts the state of art to solve the co location pattern mining problem based on Fraction Score. Extensive experiments have been conducted as well using both real and synthetic datasets, which verified the outstanding performance of the Fraction Score and the efficiency of the developed algorithms.