Scientific publications

The paper considers problems of developing the expert system (ES) "Lakes of Karelia" using artificial intelligence methods and the mathematical apparatus of fuzzy logic. Characteristics of the geographic information system, data bases and banks for the system have been described. Results of classifying lakes by multivariate statistics, factor, cluster and logical information analysis based on multidimensional scaling are presented. Statistical distributions of some parameters of Karelia's lakes are far from normal, making classical methods of statistical analysis inapplicable in their classification. It is suggested that methods of nonparametric statistics and the mathematical apparatus of fuzzy sets and fuzzy logic are used to this end. Relationships between order characteristics and ratings of the lakes were studied using logical information analysis, which involves comparison of the entropy of distribution of a natural parameter in the absence of any information about the study object with the entropy of the same parameter when additional information is available. Parameters of a lake are represented as fuzzy sets, and fuzzy logic is applied when certain parameters are used to determine others. Properties of so-called acquired knowledge were assigned to the expert system. Relying thereon, calculations were carried out in the expert system. Trophic status estimates for many lakes were fuzzy. Classification of Karelia's lakes using the ES revealed high uniformity and absence of sharp change from lake to lake, which is a characteristic feature of the region's lakes.
The proposed expert system for lakes of Karelia, based on the GIS "Lakes and Rivers of Republic of Karelia", is currently applicable for more comprehensive assessment of resources of a lake, and can also be used as a management tool for conservation and sustainable management of lakes and their resources.