Saturday, August 05, 2006

What is the difference between data-based, rule-based, and knowledge-based systems? (FAQ #15)

The chart below summarizes the key differences between data-based, rule-based, and knowledge-based systems:

Click to enlarge

The problem with legacy data-based systems is that they are hard-coded and limited to processing data and outputting information. It's still up to the human being to analyze all the information to make decisions and recommendations. The result is often information-overload and costly mistakes.

Rule-based systems process data and output information, but they also process rules and make decisions. They are good at processing lots of simple business rules with broad logic. They are commonly used for real-time decisioning systems, straight-thru processing (STP) systems, and compliance systems.

Knowledge-based systems also process data and rules to output information and make decisions. In addition, they also process expert knowledge to output answers, recommendations, and expert advice. They are good at processing deep logic and very complex business rules. They are commonly used for advising systems, expert systems, and knowledge automation.

Database vs. Rulebase vs. Knowledgebase

Deciding between a Rule-Based or Knowledge-Based solution

List of Applications where Rule-Based and Knowledge-Based solutions are Most Effective

Business Rules Knowledge Base - Articles



At Monday, August 07, 2006 5:33:00 PM , Gene Weng said...

This is a great comparison. Another dimension could be schema. In data base design, schema is a complicated issue. But in rule base, most likely it's not that complicated.

What do you think?


At Monday, August 07, 2006 10:03:00 PM , Andrei Palskoi said...

In a rule-based system the analog of a database schema is object model - and it can be very complex. In addition, rules can extend this model even more by establishing new concepts and facts derived from the original (raw) data. However, much of that complexity is usually hidden in hierarchical object structures, enumerated values, etc. In a database same things have to be explicitly defined as foreign keys and lookup tables, which makes the schema look complicated.


Post a Comment

Links to this post:

Create a Link

<< Home