V.V. Shapovalov, A.G. Korestalev, A.V. Tishkov

In order to make a managerial decision a manager of a medical institution uses various information given by an information system (IS). As soon as a standard set of reports about services rendered, profits and losses is obtained, a decision-maker inevitably regards numerical values received as qualitative, looks for a reason and acts according to his knowledge and experience. Implementing an engine for determining influence of one or another factor on the situation and ordering a list of factors will not only reduce time for analysis by decision-maker, but allow configuring the IS so that the most relevant information would be shown to user first of all. Today there are many methods of intelligent data analysis, which implement a search for hidden dependencies and solve a connected with that problem of reducing dimension: classical correlation analysis and main components method, cluster analysis and various implementations of searching for the nearest neighbor, Data Mining methods and so forth. All of them require sufficiently large amount of resources and time. Critics against these methods are also well-known: many of them are weakly applicable to one area while quite efficient in the other. In this paper we take cause-effect relations obtained by one or another method as input data and propose a quite easy but nonetheless efficient approach to assessing activities influence on taking a decision, based upon information-quantitative calculation of information amount for each activity. Top rated activities are priority-driven to be shown in IS and should be displayed primarily to an analyst and in reports.