Research Papers - Dept of Information Technology

Permanent URI for this collectionhttps://rda.sliit.lk/handle/123456789/593

Browse

Filters

2020 - 20222

Advanced Search

Filter by

Settings

Subject: search.filters.subject.Measuring

Search Results

Now showing 1 - 2 of 2
  • Thumbnail Image
    PublicationEmbargo
    Measuring Psychological Stress Rate Using Social Media Posts Engagement
    (IEEE, 2022-08-15) Perera, W.T. H; Lanerolle, T. Y; Andrado, Y. D. S; Wickramasinghe, W.A.P.C; Bandara, P.S; Kishara, J
    In psychology, stress is a feeling of feelings and pressure. Stress is a type of psychological pain. Literature has showcased that mental health stages like anxiety and depression might be identified by the social media post captions, emojis, and the way users communicate with others. Among the main underlying causes and correlates of illnesses and mental health problems is stress. In this study, we explore the conclusions and posts of psychological stress using the data of social media users, who use and share their Facebook accounts. In the first step, a user who are stressed often post about exhaustion, losing control, increasing self-focus, and physical pain using their post captions, emojis, and post images they usually post on Facebook. Collect and read all the posts that are fetched via the social networks and then measure the stress level against different factors. Then the system demonstrates how the user interacts with the intelligent custom virtual AI counselor application thus innovated can be trained and be scaled to measure against the factors. Data can be collected by using Graph API, followed by machine learning techniques and natural language processing (NLP) techniques, and an intelligent custom AI virtual application to measure stress levels by different factors. Also, use AI techniques to build health guidance plans for everyone with the help of the above collections. And reacting to the simple games is another factor to measure a highly accurate result in stress level. Natural Language Processing (NLP) is commonly used to implement smart communication virtual counselor agents. Scaled social media-based stress measurements outperform survey-based stress measurements, held up against involving a combination of social and demographic factors such as gender, age, race, income, and education. A discussion of the implications of using social media as a new tool for monitoring stress levels and developing health-related advice for individuals is presented in the conclusion.
  • Thumbnail Image
    PublicationEmbargo
    Shilpa: A Novel Neural Based Approach for Measuring Human Stress Level
    (IEEE, 2020-11-04) Perera, B. T. N; Jayarathne, B. G. D. N; Dharmakeerthi, T. G. G. M; Thanthilage, K. T. D. D. K
    21st century is far more advanced than the 20th century because of its new innovations along with the relevant technological mappings. Technology makes our day to day work easy. However, this has been led our simple life to be very complex. We have become really busy, money minded and most importantly we don't have time to spend with our families or thinking about ourselves. As Millennials form our childhood what we have experienced is the stress to be the best. The competition which has been generated around and among us cannot be handled; so, people have been depressed and this would let them even committing suicide. Therefore, a Learning Assistant for advanced level students, which has been named as “Shilpa”, would be a practical remedy and a companion to overcome such difficulties. Shilpa has been stepped forward to monitor students' stress levels and to help them understand their weak areas considering the curriculum of a particular subject. Once identifying a particularly weak area, Shilpa navigates the user to the summarized version of a weakly identified content of a lesson in which the user doesn't have to go through the entire course curriculum to improve his/her weak areas. The implemented system has been tested considering all the novel components, and an overall value of 0.81 has been experimented as per the precision. It can be concluded that this novel approach has achieved an overall 81% accuracy over the existing state-of-the-art baselines.