Back ... ... ... ... ... ... ... ... ... ... ... Send comments to review committee. ... ... ... ... ... ... ... ... ... ... ... Forward

The bead game is underdevelopment, however part of the function of the game comes from clicking the forward and back links above and from game players sending in Remarks.

Indexing

Note on Indexing Types

There is a spectrum of algorithms that produce an indexing of invariance in data.

The following basic classes exists:

1) Statistical

2) Rule based

3) Distributed

However, creating taxonomy for indexing algorithms is difficult.

In systems theory it would be claimed that the natural-type of problems to be solved have not found clear winners. On the other hand we can make the following distinction between types of problem.

Class 1: Many indexing problems are solved with some variation of statistical analysis leading to probabilities of relationships, followed by rule formation and hard coding of relationships.

Class 2: Other indexing problems are too non-stationary for statistical analysis. Distributed systems with recontextualization are generally applied. The distributed systems provide a transfer (learning) function and will work well if proper context is supplied by external source - sometimes but not always by a user.

Class 3: Some indexing problems are characterized by incomplete or uncertain information and data. Inference systems and knowledge acquisition methodology is used in these cases.