Information flow estimation: A study of news on Twitter

Quantifying polarization in online political discourse

“…In the literature, it is possible to find several entropy measurements to estimate the information flow between the users of a network. In [52], the authors show how, using the entropy rate estimator (h), originally proposed in [53], some other entropy measurements could be derived. The entropy rate is defined as follows:…”

“…From that basic information flow estimator, the authors propose the time-synchronized cross-entropy metric [52], which leverages the usage of entropy to calculate the information flow between two users taking into account the dynamics of the conversation, which can be considered useful in the context of social networks. Such a measurement is mathematically defined below:…”

“…where T refers to the target, S refers to the source (for instance, target user and source user), N refers to the length of the text (where the sub-index determines whether it is the length of the source text or the length of the target text), and i refers to the longest subsequence in the target text that appears as contiguous in the source text, taking into consideration the time when the posts were published, if the metric is applied in the context of social networks' information flow measurement. Although the time-synchronized cross-entropy is already a valid metric for measuring information flow, we decided to measure the information flowing through a network using an entropy metric derived from it, called Neighbor Normalized Information Flow (NNIF), also defined in [52]. The reason behind this decision is simple: based on the benchmark carried out by the authors in [52], NNIF is able to measure a network's information flow in a much more reliable way than the time-synchronized cross-entropy, and other entropy measurements derived from it.…”

“…Although the time-synchronized cross-entropy is already a valid metric for measuring information flow, we decided to measure the information flowing through a network using an entropy metric derived from it, called Neighbor Normalized Information Flow (NNIF), also defined in [52]. The reason behind this decision is simple: based on the benchmark carried out by the authors in [52], NNIF is able to measure a network's information flow in a much more reliable way than the time-synchronized cross-entropy, and other entropy measurements derived from it. In fact, such an entropy estimator has already been used in the literature (as it was aforementioned) achieving good results [54].…”

Using Social Media in Response to COVID-19 Crisis in Health-Related Organizations in Iran: Short Communication

“…1 According to theories (uses and gratifications) and findings of studies, media satisfaction can be effective in forming or changing ideas and beliefs, as well as the type of response to a problem or crisis in such interactions. 2 Providing helpful resources including information and support from social media networks may protect people against stress. 3 Uses and gratification theory research was chosen as a framework 4 to gain a better perception of the factors influencing online communications (health-related behaviors) among Iranians working in health-related organizations during the COVID-19 pandemic.…”