Hair-Oriented Data Model for Spatio-Temporal Data Mining

Abstract

Spatio-temporal data are complex in terms of number of attributes for spatial and temporal values, and the data are changing towards time. Traditional method to mining the spatio-temporal data is the fact that the data is stored in data warehouse in un-normalization form as union of spatial and temporal data know as tabular data warehouse. A Hair-Oriented Data Model (HODM) has been proved as a suitable data model for spatio-temporal data. It has reduced the file size and decreased query execution time. The spatio-temporal data stored using the HODM known as Hair-Oriented Data warehouse. However, this paper aims to presents a method to develop spatio-temporal data mining model using the Hair-Oriented data warehouse. The Hair-Oriented data model also provide with various functions for easy maintenance on spatio-temporal data warehouse. Experiment conducted using Climate-change spatio-temporal data set benchmark. Two Climate-change spatio-temporal models been developed using regression and k-nearest neighbor techniques. The performance of the Hair-Oriented Data Warehouse is evaluated by comparing its performance with traditional tabular data warehouse. The result shows that developing data mining spatio-temporal model using Hair-Oriented data warehouse is faster compare using the tabular data warehouse, therefore it can be concluded that the Hair-Oriented Data Model is suitable for Spatio-temporal data mining.