Online Behaviour Indicator Extraction for Enhanced Cancer Patient Management using Real-World Data

• DOI: 10.1145/3617233.3617264
• Publication Date: 2023-12-30T11:05:32Z
• Authors: Georgia Pantalona; Anna Barachanou; Nikolaos Loukas; Lazaros Apostolidis; Filareti Tsalakan...; Symeon Papadopoulos

Mental health issues of breast cancer patients and survivors are detrimental to their recovery, reintegration normal living, quality life. With the increasing popularity social media platforms online browsing, combined with rise Natural Language Processing (NLP), there exists great potential in analyzing textual information various user activities infer important indicators about a person's behaviour emotional well-being. This work proposes NLP-based framework for emotion extraction from posts interest identification browser history data. We experimented contextual representations by fine-tuning pre-trained models like BERT fully connected layer, stacked LSTM Bi-LSTM layers, external resources form lexicons self-attention. evaluated our on curated datasets derived combination open achieved state-of-the-art performance. Overall, provides evidence non-obtrusive that reflect well-being patients.