PyHasse

Q

What is the meaning of "mis"?

A

mis is an acronym (m)ulti (i)ndicator (s)ystems and the first part of a namespace package. You may regard it as the umbrella for all tools used to analyse data sets simultaneously taking several indicators into account. In contrast to multivariate statistics the focus is on the support of decisions. For now only the PyHasse namespace is a container for similar systems.

Q

Why you are not using numpy, scipy or other mathematical modules?

A

These tools and libraries are very suitable for the partial order problems. But on one hand most of the functionality is not needed for PyHasse and therefore to heavy and on the other hand for some users it was often too difficult to install them. We like the unix philosophy of one tool for one task.

Q

What is a representative element?

A

Ah, okay: Clearly I have to explain this. Think of three cars. Two are the same type, same colour but belong to different persons. The third is another type, has another colour, etc. Then the first two can be considered as equivalent, and one of these two can be selected as a "representative element". This means that a ranking study could be performed by selecting for example the first car, and the third car, concentrating on relevant information.

Q

An often used term is the reduced list, what is a reduced list for?

A

The list of representative elements has less elements than the original list of objects; it is a "reduced list".

Q

With the car example, is it likely to explain the term "poset"?

A

Yes -- no problem: If there is an idea to order cars according to their colour and other criteria then the simultaneous consideration of all these criteria will typically lead to a Partially Ordered SET (poset). Consider the example with the presents: they cannot be identically ordered based on all criteria.

Q

Is a calculation always resulting in a nice looking Hasse diagram? And how about problems different to the car example?

A

A Hasse diagram is often helpful, therefore most modules automatically draw a Hasse diagram. But Hasse diagrams often raise many additional questions, which are statistical in their very nature. A simple example is with how many other objects a given object can be compared to and how many it cannot be compared to. But it should be noted that often Hasse diagrams become pretty complicated, you may call it messy, if datasets with many objects are studied. We then may often lose a clear picture because of the multitude of lines or because no object can be compared with another. Nevertheless many information pieces can be extracted from the posetic view, which means the way to look on a set of objects in terms of comparability and incomparability, or in other words: what can be compared and what cannot be compared.

Q

What is meant with incomparable?

A

Think of two houses at different locations. One is large, but near a noisy road, the other one is small but in a peaceful environment. There is a conflict: large versus small but noisy versus peaceful. Without additional information or preferences you cannot solve this conflict. You cannot compare both houses based on the given information. Technically spoken the two houses are "incomparable". Another example is apples and bananas. They are different in shape, color, taste and possible price. They are thus incomparable.

| About Rainer Bruggeman | About Helmut Hasse | About PyHasse | Glossar |