Urgency.

   Successful promotion and development of website requires continual analysis. It is not easy to make a good site even if professionals make projects from the very beginning, but it is more complicated to have perfect site if it continuously renovates. The case is not only in qualification of executors but in short deadline, psychology of potential visitors and a lot of other reasons too. Some sites losing their visitors though they are working out all their major functions. What is with them? The following work with site must be done before trying to attract visitors: the site must be learned well, the same with it's auditory, and of course you should make researchment of the market niche that the site is trying to reserve, and at last but not least analysis of competing sites must be done. These will help you to realize what is already done and will give you the correct directions to the best result.

The problem to be solved.

    This work relates to the problems of analysis of existing methods of forecast, learning of target auditory and its classification.

Main target of the work:

   Analysis of target auditory of e-source and finding relations for predicting and learning visiting of the site and other features. Making of statistics scheme and prediction model of visiting of the source also.

Targets.

1. Development of system of data base administration through web-interface.
2. Gaining data about visiting statistics with help of administration system.
3. Features classification and building of data base on the basis of received information.
4. System and structural analysis of data and finding relations between factors.
5. Researchment of relations for prediction and learning quality of "popularity" of the site content.
6. Making statistics reports for fixed terms.

Research methods.

   During the process the following methods of system and structural analysis of data suggested: Statistics methods - correlation and regression analysis, statistics analysis of data distribution, stochastical methods of time rows analysis - autoregression method of integrated sliding average, imitation modeling of distributions, spectral analysis of Furye, development of object computer models for risk learning.
   During the process the following instruments are suggested:

• STATISTICA and MatLab systems - for data analysis;
• language of data base projecting MySQL;
• language РНР;
• language HTML;
• language ActionScript 2.0;
• object modeling systems for making computer models.