Skip to Main content Skip to Navigation

Opinions Mining from Posters’ Users in Social Networks

Abstract : Nowadays, the analysis of social networks is a rich field of research and the analysis of opinions and the detection of communities is becoming a problem because of the considerable development of social media While existing sentiment analysis methods focus only on the extraction of positive and negative opinions, in our work we aim to extract fuzzy opinions. That is why we propose to follow Fuzzy Support Vector Machine during our sentiment analysis process. We also present a new hierarchical classification approach for the detection of communities sharing the same opinions. Our mixed hierarchical clustering technique, based on the assumption that there exists an initial solution composed of k partitions and the combination of ascendants and descendants methods, does not the change of the number of partitions, modifes the repartition of the initial structure. At the end of the introduced clustering process, a fixed point, representing a local optimum of the cost function which measures the degree of importance between two partitions, is obtained. Consequently, the introduced combined model leads to the emergence of local community structure. To avoid this local optimum and detect community structure converged to the global optimum of the cost function, the detection of community structures, in this study, is not considered only as a clustering problem, but as an optimization issue. In addition, it is essential not only to determine communities but also to identify the modifications in the network structure over the time. Additionally, we track the evolution of events over time. We develop, in this study, an approach for natural hazard events news detection and danger citizen’ groups clustering . Analyzing the ambiguity and the vagueness of similarity of social publications plays a key role in event detection. This matter was ignored in traditional event detection techniques. To this end, we apply fuzzy sets techniques on the extracted events to enhance the clustering quality and remove the vagueness of the extracted information. Then, the defined degree of citizens’ danger is injected as input to the introduced citizens clustering method in order to detect citizens’ communities with close disaster degrees. The experimental analysis shows promising results made a set of comparisons between our proposed contributions and the existing state of the art.
Document type :
Complete list of metadata
Contributor : Radhia Toujani Connect in order to contact the contributor
Submitted on : Sunday, July 4, 2021 - 12:01:16 PM
Last modification on : Tuesday, July 6, 2021 - 2:11:21 PM
Long-term archiving on: : Tuesday, October 5, 2021 - 6:08:12 PM


Files produced by the author(s)


  • HAL Id : tel-03277617, version 1


Radhia Toujani. Opinions Mining from Posters’ Users in Social Networks. Informatique [cs]. isg Tunis, 2021. Français. ⟨tel-03277617⟩



Record views


Files downloads