Clariant's award-winning AmoMax-Casale catalyst selected for cutting-edge green ammonia project in Australia

Abstract

According to psychology, loneliness exists in two contexts: positive and negative. It is about “I am alone” or “I want to be alone.” The positive sense of loneliness is commonly known as solitude where the person wants to spend time with himself to feel good or want to introspect himself. In the negative sense, persons experience loneliness due to circumstances and, if it goes undetected, can lead to depression. Nowadays, people are proactive to share their feelings on social media. Using the web text shared by individuals on the social media, one can dig out the feelings and expressions wrapped in the text. The chapter presents techniques to detect whether a person is experiencing a feeling of solitude or loneliness. Following the basic preprocessing steps, four models are applied: GloVe embedding with BiLSTM, GloVe embedding with GRU, Word2Vec with Random Forest, and Word2Vec with XGBoost. It is observed that BiLSTM and GRU with GloVe embedding demonstrate the best accuracy and high recall. The proposed application can help medical professionals to detect depression at an early stage and provide timely support. Moreover, the application can be further extended to detect all levels of depression.