Информационная система "Конференции"

Abstracts

The modern economic is stable and predictive, so business owners and top-managers state the problem of effective planning for organization financial and operative functioning. Any planning is based on the analysis of the past experience and the future dynamics prediction. The organization data bases collect the past results and gives us the good (or the only) origin for building predictive model. It's clear, that data is useless information if one doesn't perform the productive data proceeding. To proceed the "raw" data, one needs the following:

1. The data administration core for query and sample generation in short time period without data distortion.

2. The "raw" data analysis tools for useful knowledge discovery.

Nowadays there are many special technologies for data collecting, storage, static proceeding, administration. But there is a missing of the effective methods and tools for data analysis that are able to discover potentially useful, but not obvious information.

The OLAP and Data Mining are used in the data analysis problem more often. The OLAP methods were designed for data administration using the multidimension hyper-cube representation model. Also, OLAP provides a kind of analytical data proceeding in a form of test of hypothesis about the information stored in the "raw" data. The Data mining is the set of methods, which provide discovering of not obvious, previously unknown and potentially useful information in large data storage. The Data mining methods extract knowledge without human-user participation, because of using intelligent algorithms and technologies that do not require any assumption about knowledge structure.

The author has realized the complex analysis of the common known computer-based systems that are solving OLAP and Data mining problems and the ERP-systems that are aimed to budgeting problem solving. The 12 most popular in Russia systems were reviewed. The list of advantages and disadvantages of given systems was formed.

The author develops the following specification to the automated decision support system for budget control problems, which analyze the information (stored in data bases) using intelligent information technologies of Data mining:

- The data source is text table, as almost all EPR-systems have data bases import in text.

- DSS would be developed in a form of separate modules for separate budget problems (so one needs not to have large computer-based system, but only necessary tools).

- The system algorithms are generated and adjusted without human-user participation so there is no need of data mining specialist and the system becomes comprehensible for anyone.

- The core of DSS is genetic algorithms, which are able to automatically generate and adjust the effective intelligent information technologies of Data mining such as neural networks, fuzzy logic, genetic programming.

For this DSS project the author states the specific problems of budget control with its description of solution effect (benefit), mathematical statements and methods of solution.

The main advantages of given approach is the opportunity of effective solving the budgeting problems using Data mining methods. The DSS uses the unique schema of combined working of evolutionary algorithms and Data mining procedures, which automatically generates and adjust new intelligent technology for the certain budget control problem.

Webmaster: www@www-sbras.nsc.ru