Morphological analysis of data mining and prediction algorithms

This post is to proclaim the upload of two Morphological analysis tables I made a month ago. They were a (small) part of my presentation last month at the Wolfram Technology Conference 2013. The tables were made in order to provide summarization, guidance, and insight of the data mining and prediction algorithms of the Mathematica for Prediction project at GitHub.

More about Morphological analysis can be found at these links: Wikipedia, goodreads.com, “General Morphological Analysis”.

The first Morphological analysis table provides breakdown of algorithms and data along two axes:
1. data type,
2. algorithm name or family.
The entries of the first table tell how the algorithms utilize the data. An empty entry means that the corresponding algorithm cannot be applied to the corresponding data.

The second Morphological analysis table provides a breakdown along the axes:
1. problem type,
2. algorithm name or family.
The entries of the second table tell how the problems are answered by the algorithm. An empty entry means that corresponding algorithm is not applicable to the corresponding problem.

Clearly, this classification can be expanded in (at least) couple of ways. One is to provide more problems and algorithms to the axes; another is to provide details and examples for the classification entries. (And I hope I will do these extensions soon.)

Morphological analysis is something I studied 21 years ago while pursuing a degree in the field of Engineering Ergonomy and Industrial Design. My professor, Nicola Orloev, taught Morphological analysis mostly as a method to derive new, attractive, “good”, and fitting designs for industrial devices and objects (electrical, mechanical, architectural). He also demonstrated how he successfully utilized it in other contexts.

I used Morphological analysis
1. to explain and communicate the functionalities of mathematical algorithm frameworks,
2. for design of domain specific languages,
3. for defining tasks within unfamiliar and unknown problem areas,
4. for summarizing the complex interaction of user behavior and mathematical algorithms, interface layouts and functionalities.