Dbm-A Data Mining System

Dbm is a patented innovative data mining system which mines databases for knowledge undiscovered by users. The key to a successful data mining system is to design an algorithm to reduce the running time as much as possible. Dbm applies many new techniques to reduce the running time, thus enabling it to solve practical problems.

The figure shows a segment of rules which are generated by Dbm from a regular hospital database of heart disease. The database is nothing strange, but the generated rules are important, interesting, and unknown before. Different databases, such as business, science , engineering, medical, or even stock market databases can be mined by Dbm, and valuable results will be generated. For special requirements, we can customize the system for you.

Solving Problems

• Huge database - Dbm is fast enough to mine it.
• Statistical result required - Dbm uses certainty factor to solve it.
• Error data, missing data, and noisy data - Dbm can treat such data by statistical method.
• Special structured database - Dbm can read any database even that is created by yourself, using comma, space, tab, vertical bar or others as delimiter between fields.
• Fuzzy set - Dbm can take fuzzy set.
• Continuous data Dbm can create fuzzy set for such kind of data (making different levels).
• Threshold - Dbm can take threshold for certainty factor and positive count assigned by the user.
• Output to an expert system - Actually this is one of the design objects of Dbm. The output is translated to both rule format and table format by Dbm automatically. An expert system can read the output directly.

Dbm can be used independently or integrated with other parts (spreadsheet, database, etc.). The price of Dbm for VisualAge in Windows is $799.00.

MicroExpert 1928 Interface Lane, #104 Charlotte, NC, 28262 Phone: 1-704-597-5627 FAX: 1-704-597-5627 E-mail: hmchang@mindspring.com