Analysis of machine learning algorithms for specific datasets using composite index (CROSBI ID 724717)
Prilog sa skupa u zborniku | sažetak izlaganja sa skupa | međunarodna recenzija
Podaci o odgovornosti
Kadoić, Nikola ; Oreški, Dijana
engleski
Analysis of machine learning algorithms for specific datasets using composite index
Many machine learning algorithms (MLAs) can be applied to analyze the datasets. The MLAs applied to the datasets with specific characteristics (meta-feature values) should be evaluated concerning different measures that refer to MLA´s model quality. Those measures are related to the accuracy, confusion matrix, mean squared error, reliability and/or training time. In practice, for a specific dataset, the MLAs applications results with respect to quality measures present a multi- criteria decision-making (MCDM) problem. This paper presents the analysis of a given MCDM problem using the composite index approach. The composite index approach is a base for the simple additive weighting method (SAW) and analytic hierarchy process (AHP). By applying the SAW or AHP, we can decide on the optimal MLA for the observed dataset. Further, it is needed to investigate if the optimal MLA is also optimal for other datasets with the same meta-features.v
Composite index ; multi-criteria decision-making ; machine learning ; meta-features.
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o prilogu
15-15.
2022.
objavljeno
Podaci o matičnoj publikaciji
Book of abstracts, 19th International conference on operational research KOI 2022
Podaci o skupu
19th International Conference on Operational Research KOI 2022
predavanje
28.09.2022-30.09.2022
Šibenik, Hrvatska