Open AccessSCI-Tree: An Incremental Algorithm for Computing Support Counts of all Closed Intervals from an Interval Dataset
Author(s) -
Dwipen Laskar,
Naba Jyoti Sarmah,
Anjana Kakoti Mahanta
Publication year - 2019
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.i8009.0881019
Subject(s) - interval (graph theory) , algorithm , data mining , tree (set theory) , computer science , interval data , set (abstract data type) , interval arithmetic , interval tree , data set , decision tree , mathematics , tree structure , artificial intelligence , binary tree , measure (data warehouse) , mathematical analysis , combinatorics , bounded function , programming language
Interval data mining is used to extract unknown patterns, hidden rules, associations etc. associated in interval based data. The extraction of closed interval is important because by mining the set of closed intervals and their support counts, the support counts of any interval can be computed easily. In this work an incremental algorithm for computing closed intervals together with their support counts from interval dataset is proposed. Many methods for mining closed intervals are available. Most of these methods assume a static data set as input and hence the algorithms are non-incremental. Real life data sets are however dynamic by nature. An efficient incremental algorithm called CI-Tree has been already proposed for computing closed intervals present in dynamic interval data. However this method could not compute the support values of the closed intervals. The proposed algorithm called SCI-Tree extracts all closed intervals together with their support values incrementally from the given interval data. Also, all the frequent closed intervals can be computed for any user defined minimum support with a single scan of SCI-Tree without revisiting the dataset. The proposed method has been tested with real life and synthetic datasets and results have been reported.