# 148

Dear JAKUB WI ˛ECKOWSKI , JAROSŁAW W ˛ATRÓBSKI , and WOJCIECH SAŁABUN

I read your paper:

Inaccuracies in expert judgment: Comparative analysis of RANCOM and AHP methods in housing location selection problem

My comments

1- I am puzzled with the title of this paper, because to speak about inaccuracies, you need to know what is accurate. If you do not have that yardstick, how can you compare?

In the abstract you say “The effective extraction of expert knowledge is paramount for ensuring reliable results in decision support systems, especially considering…”

First time I hear about ‘effective extraction’ . What do you mean with this?

2- “Identifying criteria weights based on expert knowledge is essential for reducing the impact of emerging uncertainties”

This needs translation, because I do not understand how ‘emerging uncertainties’ depend on weights.

3- “comprehensive verification of the performance of the RANCOM and AHP methods, combined with selected Multi-Criteria Decision Analysis (MCDA) methods”

I confess that I am lost, for I always understood the AHP isa MCDM method. How can you combine it with itself?

4- “Key factors include evaluation criteria, their relevance to the problem”

Not in my opinion. Criteria do not play any role in alternatives evaluation.

Allow me to put an example: Suppose that you have two alternatives and three criteria, you can easily draw this problem in a Cartesian 2D coordinate system with the two alternatives as axis and the three criteria as straight lines.

The solution of the problem is in one of the vertices of the polygon formed by the intersection of these three lines. The solution is a function of the slopes of the lines and these are consequence of the dispersion of the values in each criterion, as entropy shows. This discrepancy is what evaluate alternatives, not the weights.

What do weights do?

When you use weights, you multiply each performance value of a criterion for the same number. This means that the criteria lines will displace parallel to themselves in different distances (because the different weights), and in so doing, may modify the solution, but this modification is due to a physical translation of the criteria, not to an evaluation; it is simply geometry, and is not related with evaluation.

5- Page 4 “Petrovic et ´ al. used the Entropy method combined with the TOPSIS method to evaluate the annual operational efficiency of passenger and freight road transport in Serbia “

This is the correct way in my opinion. Look at the difference. Entropy weights are obtained as a function of data, while subjective weights do not have any relationship to data. How can they evaluate them, when they have been computed without considering the alternatives?

6- “However, they may be limited by the quality and completeness of the available data”

And where is the quality in subjective weights? If a data is incomplete for instance in a row, you compute the entropy of that row in the usual way, but using for average a different Napierian logarithm of the number of variables, that is ln(n-1), instead of ln(n)

7- “Moreover, they often fail to fully capture the nuances of expert knowledge or contextual factors specific to the decision problem”

This is true, but you can always consider the DM opinion at the end of the computation process, where, with data obtained without human intervention, the DM can make as many changes as he/she wants, and analyze results.

8- page 5 “he decision-maker or expert determines the position of each criterion relative to other factors, assigning lower values to more significant criteria”

And which are those other factors?

These are some of my comments, I hope they can be useful

Nolberto Munier