Maintaining Adaptive Learning


Overview

Tasks

Included with

Tivoli Problem Management

Overview

How Adaptive Learning Works

Adaptive Learning is a description-based Diagnostic Aid. It uses the description of a caller's problem to find a possible solution. The solutions may be:

This Diagnostic Aid is called Adaptive Learning because as you use it to solve problems, it learns which solutions are most likely to solve particular types of problems. The more you work with Adaptive Learning, the better it becomes at finding solutions to callers' problems.

Note: To learn about Adaptive Learning from a user's perspective, see the Tivoli Problem Management 6.0 User's Guide.


Tasks

Finding Solutions with Adaptive Learning

To find solutions, Adaptive Learning uses a process that involves several steps. Understanding these steps is key to understanding how to configure Adaptive Learning to function most efficiently. The steps are as follows:
  1. A help desk analyst types a description of a caller's problem in the Call Registration dialog box.
    Example: "Microsoft Word locks up when I try to print."
  2. Adaptive Learning searches the problem description for keywords. Keywords are any words or phrases that might be helpful in finding a solution.
    Example: Keywords include "Microsoft," "Word," "lock," "print," and "system failure."
  3. Adaptive Learning uses the Thesaurus to match keywords with concepts. Concepts are synonyms for the whole list of keywords with which they are associated.
    Example: The keyword "lock" might be associated with the concept "GPF." Other keywords associated with the concept "GPF" might include "freeze," "crash," "system failure," and so on.
  4. Adaptive Learning retrieves any solutions related to the concepts it matched with the keywords in the description.
    Example: Adaptive Learning retrieves any solutions related to the concept "GPF." In addition, Adaptive Learning can use the problem's SCIM information to identify possible solutions.
  5. If one of the solutions solves the caller's problem, the help desk analyst activates that solution. Using the Learn Utility, Adaptive Learning learns that this solution is likely to solve other problems involving the same concepts. The next time Adaptive Learning searches for solutions to a similar problem, the solution that worked previously has a higher score than before. That is, the probability is good that it is the correct solution.

Configuring Adaptive Learning

The system administrator configures and maintains Adaptive Learning. This involves several tasks: