In an attempt to overcome the double problem of storage and
retrieval of fuzzy information a special sub-system (FUZZ)-was
designed and incorporated into the SLEMS. The FUZZ module
can recognize and process a number of fuzzy expressions such
as "slightly more than y" or "more or less y", where "y" isa
number. Others recognized by the system include "about y",
"roughly equal to y", "greater than y", "less than y", "much
greater than y", "much less than y", and "from y to x^". The
FUZZ sub-system therefore makes it possible for the LEARN
sub-system to query a semantic network of fuzzy object triplets.
The fuzzy knowledge processing operator is based on a new
method for processing fuzzy expressions based on the concept
of fuzzy geometric partition of the search space(Mtalo, 1990).
During a query session the FUZZ module parses and tests both
the query and examined database objects against a limited set of
fuzzy expressions. If no fuzzy expression is found, the module
passes the query to the normal query processor, otherwise, it
performs a fuzzy object comparison in order to locate the
matching database object.
Using this mechanism it is possible to store and query non-
precise information provided by soil erosion domain experts
without the loss of information associated with attempts to
translate fuzzy expressions into exact or precise facts.
4. CONCLUDING REMARKS.
This paper has explored the utility of the knowledge based
systems approach in the solution of soil erosion problems. The
paper discussed briefly data requirements and information
processing issues relevant to the introduction of the technology
in soil loss estimation and modelling. A feasible method for the
representation and manipulation of fuzzy information in the soil
erosion domain was also developed.
An experimental knowledge based system prototype was also
developed from basic principles and its application demonstrated
in the soil loss estimation and modelling applications. The
system, consisting of a unique combination of four easily
accessible information management tools, demonstrates the
viability of the knowledge based approach in general.
Although there has not been much progress in the development
and use of expert systems in the soil erosion domain, the
complexity of the problem beggars the adoption of knowledge
based systems in this area. Also, because of the broad nature of
the soil erosion problem, a multi-disciplinary approach to the
development of soil erosion expert systems is strongly
recommended.
In conclusion SLEMS four sub-systems offer easily manageable
functions for solving simple soil erosion related problems. Its
ability to manipulate vague information provides a partial
solution to the problem of handling fuzzy data and fuzzy
queries. In addition its is strongly argued that knowledge based
systems are a useful vehicle for inter-disciplinary technology
transfer. Finally, on the basis of the response from a multi-
disciplinary group of experts from Tanzania, the use of the
expert systems technology in developing countries is not only
viable but desirable.
REFERENCES.
Barr A.; E.A. Feigenbaum (1981), The Handbook of Artificial
Intelligence, Vol. I, Addison-Wesley Publishing. Co., Inc.,
Reading, Mass.
Barr A.; E.A. Feigenbaum (1982), The Handbook of Artificial
Intelligence, Vol. II, Addison-Wesley Publishing. Co., Inc.,
Reading, Mass.
Bondelid, T.R.; T.J. Jackson; R.H. McCuen (1980).
"Comparison of conventional and remote sensing estimates of
runoff curve numbers in Southern Pennsylvania," Technical
Papers of the 46th Annual Meeting of the American Society of
Photogrammetry, ASP-ACSM Convention, Washington D.C.,
February 22-27, Falls Church, pp. 440-449.
204
Ebrahimi, M. (1987) "A structured approach to expert systems
design," Proceedings of the Western Conference on Expert
Systems, WESTEX-87, June 2-4, Anaheim, California, pp. 18-
23.
Firebaugh, M.W. (1988). Artificial Intelligence: A Knowledge
Based Approach, Boyd & Fraser Publishing Company, Boston.
Goldman, S.J.; K. Jackson; T.A. Bursztynsky (1986). Erosion
and Sediment Control Handbook, McGraw Hill Book Co. N.Y.
Green, C.J.R.; M.M. Keyes (1987). "Verification and
validation of expert systems," Proceedings of the Western
Conference on Expert Systems, WESTEX-87, June 2-4,
Anaheim, California, pp. 38-43.
Kandel, A. (1986). Fuzzy Mathematical Techniques with
Applications, Addison-Wesley Publishing Company, Reading,
Massachussets.
Luger, G.F.; W.A. Stubblefield (1989). Artificial Intelligence
and the Design of Expert Systems, The Benjamin/Cummings
Publishing Company, Inc. Redwood City, NY.
Meyer, L.D. (1984). "Evolution of the universal soil loss
equation," Journal of Soil and Water Conservation, March-
April, pp. 99-104.
Morgan, R.P.C. (1986). Soil Erosion and its Control, Van
Nostrand Reinhold Company, N.Y.
Morrison, J.E.; S.H. Parker; A.C. Jones; R.R. Allen; D.E.
Wilkins; G.M. Powell; R. Grisco; D.C. Erbach; L.P. Herndon;
D.L. Murray (1989). "Expert system for selecting conservation
planting machines: PLANTING," Transactions of the ASAE,
Vol. 32, No.2, March-April, pp. 397-401.
Mtalo, E.G. (1990). SLEMS: A Knowledge Based Approach to
Soil Loss Estimation and Modelling, Msc Thesis, UNB,
Fredericton.
Schildt, H. (1987). Artificial Intelligence Using C: The C
Programmers Guide to AI Techniques, Osborne MacGraw-Hill,
Berkeley, California, pp. 239-265.
Stevens, A. (1987). C Data Base Development, Management
Information Source Inc., MIS Press.
Wischmeier, W.H. (1984). "The USLE: Some reflections,"
Journal of Soil and Water Conservation, March-April, pp. 105-
107.
Zadeh, L.A. (1989). "Knowledge representation in fuzzy
logic," IEEE Transactions on Knowledge and Data Engineering,
Vol.1, No.1,March, pp. 89-98.