Brought to you by EarthWeb
IT Library Logo

Click Here!
Click Here!

Search the site:
 
EXPERT SEARCH -----
Programming Languages
Databases
Security
Web Services
Network Services
Middleware
Components
Operating Systems
User Interfaces
Groupware & Collaboration
Content Management
Productivity Applications
Hardware
Fun & Games

EarthWeb Direct EarthWeb Direct Fatbrain Auctions Support Source Answers

EarthWeb sites
Crossnodes
Datamation
Developer.com
DICE
EarthWeb.com
EarthWeb Direct
ERP Hub
Gamelan
GoCertify.com
HTMLGoodies
Intranet Journal
IT Knowledge
IT Library
JavaGoodies
JARS
JavaScripts.com
open source IT
RoadCoders
Y2K Info

Previous Table of Contents Next


4. METHODOLOGIES AND APPLICATIONS

In this section, it is not the intention to describe in detail different expert system methodologies, as this should be covered elsewhere. The attempt here is to analyze the examples given in the previous section from two aspects: the methodological aspect and the domain application aspect. The two main methodology categories are first-generation and second-generation expert systems. The first-generation expert system methodology is based on using commercial expert system shells after acquiring the knowledge through traditional knowledge acquisition techniques and using rapid prototyping methods. The second-generation expert system methodology is primarily based on the principle of knowledge level, which means developing a knowledge model at the human level, problem-solving approach, not at the computational level approach. The domain application aspect will be analyzed taking the agriculture area and the task type to classify the given application. A domain-specific methodology will be presented in the last subsection .

4.1. METHODOLOGICAL ASPECTS

Most of the expert systems developed in agriculture apply the first-generation expert system methodology. Examples of this methodology can be found in Warkentin et al. (1990), Rafea et al. (1991), and Mohan and Arumugam (1995). The first two applications use the EXSYS shell, and a documentation of this methodology based on a modified Waterfall method is described in (Rafea et al. 1993). The third example application uses the CLIPS shell while the fourth example uses the VP-Expert shell.

Another approach that can be considered as an intermediate between the first- and second-generation expert systems is the usage of expert system shells in the implementation of a higher level of reasoning. An example of this approach can be found in Gerevini et al. (1992). In this application, the shell KEE is used but the system is explicitly divided into two parts. The first part is concerned with helping the user in classifying the findings, whereas the second part is an interactive abductive module that suggests explanation of those findings.

Few applications have been developed following the second-generation expert system approach. The two methodologies used are the KADS methodology and the generic task methodology. The KADS methodology was used for developing expert systems for cucumber and citrus management (Rafea et al. 1994; 1995), whereas the generic task methodology was used for developing an expert system for wheat management (Schroeder et al., 1994; Schulthess et al., 1996). In the second-generation expert systems methodology, there are generic models for different types of tasks such as diagnosis, planning, design, etc. In the agriculture domain, we find that two main generic models are used, namely: the diagnosis (or more broadly the classification), and the scheduling tasks. In KADS methodology, there is a library of expertise model for each one of these two tasks (Rafea et al., 1994), whereas in the generic task methodology, the hierarchical classification model is found suitable for the diagnosis applications and the routine design model is found suitable for the scheduling application (Kamel et al., 1994).

4.2. DOMAIN APPLICATION ASPECTS

The agriculture domain can be classified into subdomains, namely: plant production, animal production, and management of natural resources related to the agricultural operations such as soil and water. Expert systems have been applied in the three subdomains; however, the concentration here is on the plant production part, as most of the expert systems in agriculture have been developed in this subdomain.

Expert systems for field crops are implemented for: diagnosis of soybean diseases (Michalski et al., 1983), crop management for cotton (Lemon, 1986; Plant, 1989), and weed identification for wheat (Schulthess et.al., 1996). Expert systems were also implemented for horticulture crops: apple orchard management (Roach et al., 1985; Gerevini et al., 1992), and cucumber production management (Rafea et al., 1995). Agroforestry is another area in plant production where expert systems have been developed (Warkentin et al., 1990). Some other applications cannot be categorized commodity wise, for example, the expert system developed for selecting evapotranspiration estimation methods (Mohan and Arumugam, 1995).

Another way of classifying agricultural expert systems is the domain-specific task that this system performs, such as: irrigation, fertilization, pest management, diagnosis of plant diseases, etc. There is mapping between the domain-specific tasks and the generic tasks. For example, irrigation application is mainly a special case of scheduling, whereas diagnosis is a special case of classification.


Previous Table of Contents Next

footer nav
Use of this site is subject certain Terms & Conditions.
Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details.