سوران قادري

In this article, we will examine the early concepts of previous works with the early detection of trends in social networks. The main focus of this research is on how to deal with emerging trends and the use of cognitive models to analyze the data of unexpected hot trends. We examine the characteristics necessary for the transformation of a phenomenon into a rendering of explosive nodes that lead to its decline. Instead of using motor sensor systems, which is common in cognitive models, our proposed model uses indexers' feedback to receive feedback data. These indexers provide real-time outputs of the top trends, we manually filter the top trends and improve and enhance interactive indexing with the help of users. This record helps us to achieve the desired goal in different stages. Our pager is also an upgrade of previous models such as beep and semi-beep. In this regard, by maintaining the main structure of the model, we have made changes in the memory architectures of the cognitive models. These changes allow us to exploit the indexer's tracking data. Finally, our implemented model provides new capabilities in social trend prediction analysis and can help to improve the accuracy and speed of trend detection in early moments. In the next step, the BECOG model is introduced as a new record and compared with the main models of existing common architectures. The purpose of this comparison is to investigate the effect of using indexing models on the multi-stage improvement of trends on the accuracy of the models' performance. The specific results of this eva‎luation show that the proposed model has on average increased the F1 criterion by +2, accuracy by +4 and recall by +3 compared to the basic models. Also, compared to the previous works in this field, on average, the F1 criterion has increased by +4 and recall by +8. These results show the effectiveness of the BECOG model in diagnosing and predicting hot trends in social networks and confirm the capabilities of this model in improving the accuracy and speed of trend analysis.

پيش بيني , شناختي , شبكه اجتماعي

prediction , social network , cognitive

soran ghaderi

mohamdreza kangavari