A knowledge-based SDSS for Medfly area-wide control: development, validation and effectiveness evaluation
Cohen, Y., Cohen, A., and Broday, D. 2010.
In Decision Support Systems in Agriculture, Food and the Environment: Trends, Applications and Advances. Eds. Manos, B., Matsatsinis, N., Paparrizos, K., and Papathanasiou, J. Pages: 21- 46. IGI-Global, PN, USA.
The aim of this chapter is to describe the motivation to develop a knowledge-based spatial decision support system (KB-SDSS) for medfly control in citrus in Israel, its development approach and procedure, its validation, and the steps towards its assimilation among the zone-managers. Development of the KB-SDSS for medfly control in Israel, also known as MedCila involved four main phases:
1. Acquisition of expert and domain knowledge related to the control decision process;
2. Identification of the relevant criteria and modeling each criterion and the overall decision-making procedure;
3. Its integration into a GIS environment; and
4. Its performance evaluation by an expert-panel considering four aspects: verification, validation, acceptance and effectiveness. Comparison with data-mining approach exemplifies the effectiveness of the KB approach in this case. Our results show that the MedCila may reduce spraying actions by at least 8%, which is estimated to save ca. 13.3 ton/year of chemicals.
Peer reviewed articles
Spatial decision support system for Medfly control in citrus
Cohen, Y., Cohen, A., Hetzroni, A., Alchanatis, A., Broday, D., Gazit, Y., and Timar D. 2008.
Computers and Electronics in agriculture, 62(2):107-117
A spatial decision support system (SDSS), designated MedCila was developed for controlling Medfly on citrus in Israel. The development involved four main phases:
(1) Acquisition of relevant expert and domain knowledge;
(2) Identification of the relevant criteria and modeling each criterion and the overall decision-making procedure;
(3) Integration of the MedCila into a GIS environment; and
(4) Initial evaluation of MedCila performance. The criteria found to be most relevant for control decision-making were the number of flies and the presence of a ‘blue eye’ in the nearest trap, the host-species susceptibility, the relative development of the Medfly based on accumulative day-degree model, the history of trapping, and the Medfly population in the nearby traps. Binary, linear, logarithmic and biological-based models were developed for the criteria identified. The overall decision-making procedure of the MedCila was based on the Stanford Certainty Theory integrated with a rule-based decision tree. Initial evaluation of the MedCila performance was done by retrospective comparison between the MedCila recommendations and the coordinator decisions. It was shown that the Med- Cila provides recommendations that are generally accepted by the coordinators; it reduces the number of unnecessary spray actions in the absence of a Medfly threat in space and in time; and it reduces the number of plots for which the coordinator needs to make a decision for.
Peer reviewed articles
Performance and acceptance evaluation of a knowledge-based SDSS for Medfly area-
Cohen, A., Cohen, Y., Broday, D., and Timar D. 2008.
Journal of Applied Entomology 132, 734-745.
The performance evaluation of a spatial decision support system (SDSS) for medfly control in citrus in Israel, known as ‘MedCila’, is described. The performance was evaluated by considering four aspects: verification, validation, acceptance and effectiveness. For the performance evaluation of MedCila, a representative set of 420 real cases was collected. In addition, an expert panel was gathered to serve as ‘a reference level of expertise’ against which MedCila recommendations were compared. Various types of comparisons and cross-sections were made using the MedCila recommendations and the decisions of the zone managers to evaluate its performance. In general, MedCila was validated to provide an 80% proportion of success. For complex cases, it was validated to provide only 75% proportion of success. The results show that MedCila significantly reduces spraying actions by at least 8%, which is estimated
to save 13.3 tons of chemicals and 250 000 US$ each year. MedCila may improve the effectiveness of the decision-making process of the zone managers by highlighting the trivial cases and by providing them with clear-cut recommendations. Moreover, decision-making in complex cases by using MedCila provides the option of using more data and information that may increase the willingness to take a risk and reduce spraying actions.
Article in symposia proceedings
Developing a Learning Mechanism for SDSS for Medfly Control in Citrus
Cohen, A., Y. Cohen, D. Broday, A. Hezroni, V. Alchanatis, D. Timar, Y. Gazit. 2007. Stanford J.V. ed.
European Conference in Precision Agriculture, June 2007, Skiathos, Greece. Pages: 723-730.
An initial Spatial Decision Support System (SDSS) for Medfly infestation was developed to improve spraying actions. Beside the spray model that produces a spraying recommendations map, the SDSS has a learning mechanism that aims at realizing the ability of tuning the spray model or indicating data gaps. The learning mechanism compares between the SDSS recommendations and the expert decision, and selects appropriate cases for learning. The learning mechanism algorithm is described, and initial results are presented. The results indicate a high degree of agreement between the SDSS recommendations and the expert decisions. However, disagreements between the SDSS and the expert decision support the need for such a learning model and indicate for ways of tuning the SDSS.