Peacekeeper Interns Alex Rinaldi, Omar Oseguera and JoAnn Tuazon were recently accepted to present at the 19th International Conference on Human-Computer Interaction for their work in Machine Learning and Sentiment Analysis. The group started their project as part of their senior seminar at Cal State Bakersfield.
In their research, the team used a corpus of therapy transcripts, performed sentiment analysis (detection of emotion in text), and used machine learning to predict whether a transcript showed signs of suicidal ideation or not.
Additionally, the team has conducted research in using Recurrent Neural Networks for response generation. The team hopes to train the RNNs with their other work in order to create a suicide-prevention chat application.
Two poster abstracts will be presented in Vancouver, Canada this summer.