Considering User Interests to Provid an Event Base News Stream Framework
Mohammad Reza Hassanpour Charmchi, Hamid Hassanpour, Bagher Rahimpour Cami · 2022
With the wide spread use of social networks, news, as the most important product of online media, has a significant effect in creating news streams and guiding public opinion. By creating a news stream, news agencies sensitize people’s minds to a specific issue. One of the most important components of creating a news stream is generating and recommending the appropriate text. In this article, we use the fine-tuned GPT-2 model based on a deep neural network to generate text related to the target topic. We also recommend the generated text to users using recommender systems. In this research, the performance of the system was evaluated using the dataset collected from a Persian question-and-answer (Q& A) social network. The test results indicate that the relevance of the generated text to the topic by the fine-tuned GPT-2 model, has increased by 4% compared to the basic model.