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


5.1.2. GIS and Expert Systems

An example of integrating GIS and expert systems is given by Loh et al. (1994). In this example involving the USDA Forest Service, both expert systems, GIS and relational database management systems (RDBMS), have been introduced for maintaining resource inventories, GIS for organizing and analyzing spatially referenced data, and expert systems for capturing the essence of standards and guidelines of management plans and the accumulated knowledge of experts in the Service. To achieve this objective, some problems have been identified, namely: data sharing between the GIS and expert system components, and fitting the two components into a common problem-solving framework. The example application has succeeded in solving these problems using ARC/INFO for GIS, CLIPS for the expert system, and ORACLE for RDBMS. However, some other issues have been raised that require more research: the enhancement of the expert systems components to reason on spatial rules. This necessitates the inclusion of a spatial operator such as "intersect with" or "adjacent to." To accommodate such operators, a macro would have to be developed on the GIS side to communicate with the expert system shell. The integration of GIS and the expert system is fairly new, and more problems will be envisaged when more applications using both technologies are performed.

5.1.3. Multimedia and Expert Systems

Integrating multimedia with expert systems is a hot topic that is booming currently. The integration with images was frequently done to more efficiently acquire the user inputs, whereas other types of media such as sound and video are also addressed. An example of integrating multimedia with expert systems is given by Rafea et al. (1995).

It was found that describing symptoms in words is very difficult and sometimes very confusing. Therefore, images are identified to be used for two main purposes: describing a disorder symptom, and confirming the diagnosis of the cause of a certain disorder. Detailed images for all symptoms, and unique images that confirm the occurrence of disorders at different stages, should be collected.

Although images are very useful in acquiring the user inputs, the uncertainty problem is still there. Therefore, giving the user the option to select an image with a degree of certainty should be provided. Providing more than one picture for the same symptom can reduce the user uncertainty, but this will lead to exerting more effort in collecting and classifying the images.

As the output of an expert system for crop production management is a set of agricultural operations, describing how to perform an agricultural operation in words is very hard, and one can never guarantee that the user can understand what has been written. Displaying a video for a professional doing the recommended operation would be very educational.

The sound is essential because sometimes it is not easy to write terminology used by growers in daily life. In addition, combining the video with sound is also recommended in order to comment on how the operation is done.

Although the given example proved the possibility of integrating multimedia with expert systems, there are still some problems that need further research: the intelligent selection of the appropriate media for presentation, taking into consideration the user level; getting input data from images, for example providing the expert system with an infected leaf of a plant for diagnosis; and/or enhancements of all input devices interfaces, which is a general research issue, in order to provide expert system with data in different forms.

5.2. KNOWLEDGE SHARING AND REUSE

Knowledge sharing and reuse is one of the topics that has attracted the attention of the AI researchers in the last few years. The research in knowledge sharing in agriculture can be directed toward identifying common knowledge that can be shared among different expert systems such as identification of agriculture ontology, knowledge related to common resources, namely: soil, water and climate, knowledge related to the same taxonomic category of a set of crops, etc. The research in knowledge reuse can be directed toward building a library of domain-specific tasks in agriculture such as: irrigation, fertilization, integrated pest management, etc.

5.3. INTELLIGENT RETRIEVAL OF AGRICULTURAL DATA

Meteorological data are very important for agronomists as forecasting weather data helps in giving recommendations to growers. A system that serves as an intelligent assistant for meteorologists to locate and analyze historical situations of interest has been developed using case-based reasoning (Jones and Roydhouse, 1995). This work was oriented to be used by meteorologists, but not agronomists. The basic idea could be investigated to retrieve historical situations that are important for managing different crops.

5.4. AUTOMATIC KNOWLEDGE ACQUISITION

Automatic knowledge acquisition is a general research issue in expert systems development. In agriculture, research can be directed toward using the knowledge models developed for specific tasks in building tools to acquire the domain knowledge interactively with the domain experts. Another approach is to use machine learning to automatically generate and refine knowledge bases. Although this approach was one of the first approaches in building expert systems (Michalski et al., 1983), no elaboration on this approach was pursued. A recent application that uses machine learning to generate agriculture data is given by McQueen et al. (1995). This approach can be very useful when data sets are available in certain areas.


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.