A Framework to Improve Reuse in Weather-Based DSS Based on Coupling Weather Conditions

• Published in: Big Data Analytics Vol. 10721; pp. 40 - 58
• Main Authors: Mamatha, A., Krishna Reddy, P., Mondal, Anirban, Ninomiya, Seishi, Sreenivas, G.
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2017; Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Agro-informatics; Agromet service; Data mining; Decision support systems; Reuse; Weather patterns
• ISBN: 9783319724126; 3319724126
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-319-72413-3_3

In weather-based decision support system (DSS), the domain experts provide suggestions to carry out appropriate measures to improve the efficiency of the respective domain by analyzing both the forecasted and observed weather values. In this paper, to provide suggestions for a given combination of forecasted and observed values, we have proposed a framework to exploit reuse of the suggestions which have been prepared for the past combinations of observed and forecasted values over the years. We define the notion of coupled weather condition (CWC) which represents the weather conditions of two consecutive durations for a given combination of weather variables. By employing the domain-specific categories, the proposed framework exploits the reuse of CWCs for the given domain. We have applied the proposed framework by considering the case study of agromet advisory service of India Meteorological Department (IMD). The extent of reuse has been computed by considering 30 years of weather data from Rajendranagar, Hyderabad, Telangana State, based on the weather categories data provided by IMD. The reuse over 30 years is computed by considering the period of year and crop seasons of a year. Period is defined as portion of time of the year(s) that is considered to analyze the similarity. The results are very positive. The results show that the percentage of reuse of CWCs with three weather variables for the period of year is about 77% after five years. The results provide the scope to develop automatic weather-based DSS in various domains with minimal human intervention and improve the utilization of the generated content.