Tuesday, 11 September 2012

Group B, Session 9-10, 11th sept'12


Today we studied about MDS (Multi-Dimensional Scaling) and Permap (Perceptual Mapping).

PERMAP takes object-to-object proximity values (also called similarities, dissimilarities, correlations, distances, interactions, psychological distances, dependencies, confusabilities, preferences, joint or conditional probabilities, etc.), or up to 30 object attribute values which can be used to calculate one of the forgoing types of proximities, and uses MDS to make a map that shows the relationships between the objects. Succinctly, PERMAP makes classical metric and nonmetric MDS analyses in one, two, three … or eight dimensions, for one-mode two-way or two-mode two-way data, with up to 1000 objects and with missing values allowed.
Its fundamental purpose is to uncover any "hidden structure" that might be residing in a complex data set.

Perceptual mapping is of two types: Overall Similarity and Attribute Based

In Overall Similarity, we try to look at a broader vision. Its main advantage is that the attributes which are hidden also can come out while mapping. But its biggest disadvantage is that in order to interpret data, we need to have overall information of the data.

In Attribute Based mapping, we try to look at the different attributes. Its advantage is that more precision can be put in the analysis using individual attributes. But its disadvantage is that we may forget to consider any attribute that may be critically important for the analysis to happen. If that happens, the analysis may be wrong.

It also allows complete on-line control of the badness function, distance metric, attributes-to-proximities formula, attribute set composition, mapping weights, and metric or nonmetric MDS, again, while the solution is running and the developing map is being displayed. The solution can be mirrored, rotated, translated, or zoomed.

On the left hand side of the map, there is control box called On-Line Controls which has four Controls: MDS type, Badness, Dimensions, and Attribute Fns.

First option is the MDS type option. PREMAP can make ratio, interval, or ordinal analyses, as well as ratio and interval analyses using MDS type.

The second option is a badness function. A Badness Function is simply a definition of what it is that makes the positions of a pair of objects bad.  Fundamentally, the object pair should be separated by a distance that is consistent with the pair's dissimilarity.  Thus, the badness should involve a measure of the mismatch between the two objects. There are several valid ways of defining this mismatch. PREMAP offers the five most common types of badness measure, Stress, Stress1, SStress, Multiscale, and Fractional. The choice of the badness measure is not always important. But, each of the five badness types must be adjusted for the type of MDS being used.

In short, Permap is easy to use, PC-based, visually oriented, and allows real-time interaction with the analysis.  It has been designed to have an intuitive interface and it avoids many of the arcane alternatives that are seen in the research literature but are never used in practice.

Author: Vivek Agarwal (11020841119)
Team B, Operations

No comments:

Post a Comment