CUSTOMIZED EXPERT SYSTEM FOR CULTIVATION OF POND TO REDUCE KNOWLEDGE ACQUISITION EFFORT
Ega Dioni Putri1*, Masayu Leylia Khodra1
1Laboratory of Graphics and Artificial Intelligence, Institut Teknologi Bandung, Indonesia
*Corresponding author: email@example.com
An expert system for helping farmer in cultivation of white shrimp pond, particularly in water quality management, has been developed. This system can be developed to be another pond expert system, even for the more general simple classification expert system. As the early step, we begin to focus on improving system as customized expert system specific for cultivation of pond. This work is intended to reduce knowledge acquisition effort which often meet bottleneck. Customization of the system is essentially done by replacement of knowledge base. In case ponds with species other than white shrimp interested in using expert system, there is no need to build new application. Experiment results that with customizing system components such quality standard, pond element, and rule, the expert system is able to help in compliance the need of expert in other ponds.
Keywords: expert system, pond, knowledge acquisition, simple classification, customization 1. INTRODUCTION
This research was initiated by development an expert system that is used to assist or even substitute expert in white shrimp pond namely Vannacues (Vannamei Cultivation Expert System) . White shrimp ponds or other ponds in general are artificial environment that always require human assistance to keep the balance of its forming elements. Pond water, as habitat for living things to interact with their environment and form food chain, is unable to achieve natural control mechanism itself as well as water in the natural environments . Imbalance problem in pond is caused by inharmonious ecological process of elements either chemically, physically, or biologically. Aquaculture expert is needed to provide the optimal solution of this problem by using his knowledge about linkages between elements. Nevertheless, not all ponds are able to provide the expert in cultivation. Vannacues was built as an alternative in compliance the need of expert.
Knowledge acquisition of Vannacues resulted conclusion that pond problem is a kind of construction problem. It is closely related to element linkages issue that leads the expert to formulate final solution of all possible solutions for each poor condition. For example, when the pond water becomes muddy, possible solution is “water replacement” or “siphon” , but both can be useless without looking at solution of problem occurred to linked element. However, the number of combination of possible poor conditions in pond is not too large and cases often recurrent, thus the solution given can still be enumerated. Therefore, we found that classification approach is feasible to be used. Through further steps of knowledge acquisition, we specified approach into simple classification. Surprisingly the expert involved in our research managed to map his knowledge and experience into several groups of cases. The result of knowledge mapping underlay us to select decision table as knowledge representation. Decision table used in Vannacues was modified from the standard form, which commonly known in expert system studies, with involving set-value attribute to represent condition and storing priority order of solution action. The knowledge base was built using database. It relied on look up operation into table and pattern matching the knowledge with user inserted facts.
Vannacues was developed limited for white shrimp pond, whereas the need of expert is not only in that kind of pond. To expand the utility, we developed Vannacues as a customized expert system that allows other ponds use the same system by replacing knowledge within it. We convinced it can be implemented because ponds in general have similar characteristics of problem....
References: 1] Arie Ben-David, Eibe Frank (2009) Accuracy of machine learning models versus “hand crafted” expert systems – A credit scoring case study. Expert Systems with Applications Journal 36: 5264-5271.
2] B. G. Buchanan (1989) Can Machine Learning Offer Anything to Expert Systems? Machine Learning 4: 251-254.
3] Dan D. Baliao, Siri Tokwinas (2002) Best Management Practices for a Mangrove-Friendly Shrimp Farming. Iloilo: SEAFDEC.
4] Ega Dioni Putri (2010) Expert System for Cultivation of White Shrimp Pond. Proc. International Conf. on Advanced Computer Science and Information System 2010 (ICACSIS 2010): 213-218.
5] J. Cullen, A. Bryman (1988) The Knowledge Acquisition Bottleneck: Time for Reassessment? Expert System 5:216-225.
6] Joseph Giarratano, Gary Riley (1998) Expert Systems: Principles and Programming. China: PWS Publishing Company 3th edition.
7] Marie J. Vlaanderen (1990) Automated Knowledge Acquisition for Expert System An Overview. Erasmus Universiteit Rotterdam, PhD: 95.
8] Marindro (2007) Pengelolaan Kualitas Air Tambak 01. http://bit.ly/hljM1a, Jan 4th 10.
9] Mark A. Musen (1989) Automated Generation of Model-based Knowledge Acquisition Tools. Computers 293: 23-26.
Please join StudyMode to read the full document