Header menu link for other important links
Hybrid Model for Stress Detection in Social Media by Using Dynamic Factor Graph Model and Convolutional Neural Networks
Published in Springer Science and Business Media Deutschland GmbH
Volume: 692
Pages: 101 - 107
In present days, mental pressure is getting to be real medical problems. It distinguishes to recognize pressure opportune in proactive way. By extending the predominance digital life, person using to sharing their step-by-step activities and speaking to companions by means of electronic systems administration locales in various ways, so it makes viable to utilize online electronic life data for stress acknowledgment. Here, we discover clients' stress state which is immovably related to that of companions in their Web/public activity, and also, we use different sizes of dataset in various states which is identified with public activity to productively look at the relationship of clients' stress states and connecting within public activity. First, we identify the clients' stress states like visual pictures, writings and group properties with alternate points of view and further move on novel model to show. Dynamic factor graph model (DFGM) is combined with convolutional neural network (CNN) to utilize tweets and group cooperation information for stress disclosure and furthermore which can identify some dependencies locally and scale of invariance in discourse of speech recognitions and also in image recognition (i.e., smileys). The most part of FGM focuses on mood cast technique which is dependent on unique persistent in demonstrating and predicts the user’s emotions in social life. By further examining the social life of the user’s data, we can likewise find many scenarios and the amount of social models in form of scattered connection of stress related to users and also with non-stressed users. © 2021, The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About the journal
JournalData powered by TypesetLecture Notes in Electrical Engineering
PublisherData powered by TypesetSpringer Science and Business Media Deutschland GmbH