| Summary: | This paper reviews and compares theories of fuzzy sets and soft sets from the perspective of transformation, and a machine learning model-SF-ANN (the soft sets and fuzzy sets based artificial neural network) is proposed. Liu et al. proved that every fuzzy set on a universe U can be considered as a soft set, and show that any soft set can be regarded as even a fuzzy set. Inspired by this idea, we construct a neuron-like structure based on soft sets and fuzzy sets, and we get a more practical fuzzy learning model-SF-ANN. In practical applications, it can be used as a general methodology for establishing the membership function of fuzzy sets, and it also can be applied to pattern recognition, decision-making, etc. In general, it provides a new perspective to observe the relationship between soft sets and fuzzy sets, and it is easy to relate soft set theory and fuzzy set theory to machine learning methods. To a certain extent, it reveals that the research of fuzzy sets and artificial neural networks do lead to the same destination.
|
|---|