A brief introduction to DSS's - Part 3

Introduction

    As was mention in previous articles (Part 1 & Part 2), the DSS implementation will try to represent what we know about the human decision process. So, just to start we will apprise a rule base system and then we will cover the rule base system implementation. A rule base system is a system that use rules to make deductions and help experts to make a decision on a particular topic.

Components

A rule based system has mainly four components:

• 1. A list of rules (which is a specific type of knowledge).

  2. An inference engine, which infers information or takes action based on the interaction of the input and the list of rules.

  3. Temporrary working memory.

  4. User interface.

How the inference engine works

    There are basically two types of inference, forward chaining and backward chaining. In the first one data it is put in the working memory and this triggers rules that match with the new data. These rules can then perform actions, and these actions could add new data to memory. This is called also as data directed inference because inference is triggered by the arrival of new data. In the backward chaining inference the system needs to know the value of a piece of data. It searches for rules whose conclusions mention this data. This is also colled goal directed inference, because inferences are not performed until the system is made to prove a particular goal.

    This two types of inferences are basic, there are other systems that have probability factors attached to rules and others that store their rules in frames, and trigger them when an object is recognized as matching that frame, but are out of the scope of this article.

Rule types

It seems to be four types or rules

• 1. Derivation rules, each rule express knowledge that if one set of statements happens to be true, then some other set must also be true.

  2. Transformation rules, each rule relates truth in one domain to truth in other.

  3. Integrity constraints, these are rules of the form “it must be true that....”

  4. Reaction rules, these involve a notion of action.

Where the rules come from?

In a cognitive model the rules comes or could be derived from previous knowledge, but in an expert system the rules will be derived as descriptions of behaviors of one or more human expert: by taking protocols, conducting interviews, setting classification tasks. This is known as “knowledge acquisition”

References

• 1. Decision support systems: concepts and resources for managers. By Daniel J. Power

  2. Handbook on Decision Support Systems 2: Variations. By Frada Burstein, Clyde W. Holsapple

  3. Intelligent Decision Making: An AI-Based Approach. By Gloria Phillips-Wren, Nikhil Ichalkaranje, Lakhmi Jain

  4. Comparing Rule-Based Systems (http://www.w3.org/2000/10/swap/doc/rule-systems)

  5. Rule Base Systems (http://en.wikipedia.org/wiki/Rule-based_system)

  6. Expert Systems (http://www.j-paine.org/students/lectures/lect3/node10.html)

This article is the first of a serie that will cover the DSS subject, from history to best practices to implement DSS in a real context. 

The author ________ is part of the Hercules Research Project.