关键词: 直觉模糊集, 隶属度, 非隶属度, 深度学习, 深层神经网络

Abstract: The determination of the membership, non-membership, and hesitation degrees is an important issue in intuitionistic fuzzy sets (IFS) research since it can directly influence the extendibility and application of relevant researches. However, there are many limitations in existing methods for determining the membership degree in IFS, such as subjectivity and inconsistency of standards. Most experiments are based on simulated data and are unsuitable for practical data. Aimed at the above problems and large-scale unstructured data, a method for determining membership, non-membership and hesitation in IFS based on deep learning is proposed. In this work, a Convolutional Neural Network and Long Short-Term Memory (CNN-LSTM) model is constructed based on text, and it can automatically calculate the membership, non-membership and hesitation degrees of the text after the training process. Besides, the new method can be applied to other fields by modeling and training. The proposed method overcomes the technical and thinking limitations of traditional methods, expands the research ideas of IFS, and provides possibilities for its practical application.

Key words: intuitionistic fuzzy sets, membership degree, non-membership degree, deep learning, deep neural network