Table 15 Dıfference and Importance of the Other State of Art Studies^a focusing on the application domain

19

Does not require a mathematical model of system dynamics, but only knowledge of (un) desired values for process variables.

Improved on the current state of the art in resilience evaluation by providing experimental data showing that it is possible to summarize the resilience of a system through numerical indexes that ensure model freedom and generality.

20

–

–

23

–

–

24

–

–

25

The existing systems were not evaluated in Big Data

The proposed system outperformed other similar systems

26

In traditional machine learning approaches, the features and goals are independent. The traditional RNN model has defects such as vanishing gradient problem (gradient approximates zero) and exploding gradient problem (gradient is very high). The vanishing gra- dient will make the learning process difficult to converge, and exploding gradient will lead to the instability of the learning process.

Due to the shortcomings of the traditional RNN model, Long Short-Term Memory (LSTM) and Gated Recurrent Units (GRU) were proposed.

It is important for public opinion supervision to identify sensitive information topics through topic models and con- duct sentiment analysis based on sensitive information topics.

27

Quality commitments are used as the first dimension of brand authenticity. In addition to the four dimensions of brand authenticity−quality commitment, heritage, uniqueness, and symbolism−we added an alter- native option for our coders to choose from. When they found that a tweet is related to brand authenticity but the four dimensions are not able to capture it, they chose the none of them category.

Fill the current gaps in the literature, and contribute both in terms of better precision, and of providing firms with valuable insights about the way people interact with their brands.