IT depends on how you classify the data in the map. It can be natural breaks, Jenks, standard deviations, or an eccentric personal adjustment of the levels at which you distinguish between the mapped categories. There is a huge literature out there on these different approaches as I presume you know. Then in the maps legend you can label the categories high to low, if you wish.

Dear Batu,

Is it a technical or a theoretical question? If the latter, I agree with Tom Koch , there are several alternatives.The selection of the categorization method depends on your data, your aims, and your preference. If it is a technical question, then let as know which software you use for displaying the map. If R, I recommend using package sf. Its plotting function (plot.sf()) has an argument called "breaks" and it accept the following method names: "fixed", "sd", "equal", "pretty", "quantile", "kmeans", "hclust", "bclust", "fisher", "jenks" and"dpih". Please refer to the link below for description of these categorization methods:

https://rdrr.io/cran/classInt/man/classIntervals.html

HTH,

Ákos

Dear Batu Nacun

There are some classification methods in GIS including natural break, equal intervals, etc. You may select the best one based on your need. The best reference, however, for selecting the appropriate method is previous articles in the same field.

@ Ákos Bede-Fazekas

Dear Ákos

Thank you for your answer, it is technique question, I would classify the probability map in ArcGIS, but I have not find the suitable answer, all of your answers helpful for me, but I can not operate at my ability scope. @@

@ Mohammad Mehrabi

Dear Mohammd

Thank you for your answer, I try to classify it using ArcGIS, but I wonder how to describe the classify method in my paper.

Thank you for your suggestion. I will check it again.

Kind regards

Batu

Ákos Bede-Fazekas thank you for your answer.

Kind regards

Batu

When you go in ArcGIS to Symnbolize you’ll find there are several options. It will depend on if you’re dealing with point data or with data for census tracts or some other polygon structure. In Symbolize you’ll have choices—just try them—which let you setthe number of categories and the method by which the data is parsed, natural breaks, Jenks, etc. Try a few but the one you choose should be dictated by the data.