| Autor: |
Gruzman, Vyacheslav, Karelova, Riya, Kazunin, Roman |
| Předmět: |
|
| Zdroj: |
AIP Conference Proceedings; 2022, Vol. 2456 Issue 1, p1-4, 4p |
| Abstrakt: |
For process analysis, there is a need to identify the input factors that most significantly affect the output factors. To do this, there is an expert ranking method. This method involves collecting a group of people competent in the area under analysis. These experts rank each of the input factors based on their experience. This method has a number of drawbacks: it is necessary to spend resources on the search and collection of experts, and each of them can be subjective in setting a rank to a particular factor. Given the above, there was a need to automate the ranking process. The article describes mathematical tools for ranking input factors of any subject areas' data set. The new method is based on correlation analysis, which reflects the degree of dependence of the output data on any selected factor. Data is conjugated into groups by the same 'frozen' input factors. In each such group, it is necessary to find a correlation of the output parameter from the unfrozen (mobile) factor. For each factor, it is necessary to search for a weighted average correlation value, which is the basis for ranking. This rank reflects the degree of influence of each input factor on the output. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
