| Popis: |
Classification allows us to handle the large amount of data that is available nowadays. In our work, we use the classification features to divide employees into the several classes and examine the differences between the classical and flexible classification. We also emphasize the advantages of classical classification as well as the disadvantages, and how we can solve them by fuzzy logic. Fuzzy rule-based systems are explainable and therefore interpretable because the rules are defined by linguistic variables. Design of a more complex system is a tedious task. To resolve this, we examine interpretability criteria for fuzzy rule-based systems. We examine this topic on the examples with two classification attributes because it is easily illustrated graphically. To use more attributes is mathematically possible, but it is harder to visualize for users in a three and more dimensional spaces. In our work, we propose how to create an explainable design for classification and propose possibilities how to expand it. |
