Depression is a pressing global issue that impacts millions of individuals worldwide. This prevailing psychologicaldisorder profoundly influences the thoughts and behavior of those who suffer from it. We have developed DepressMind, a versatile screening tool designed to facilitate the analysis of social network data. This automated tool explores multiple psychological dimensions associated with clinical depression and estimates the extent to which these symptoms manifest in language use. Our project comprises two distinct components: one for data extraction and another one for analysis.The data extraction phase is dedicated to harvesting texts and the associated meta-information from social networks and transforming them into a user-friendly format that serves various analytical purposes.For the analysis, the main objective is to conduct an in-depth inspection of the user publications and establish connections between the posted contents and dimensions or traits defined by well-established clinical instruments.Specifically, we aim to associate extracts authored by individuals with symptoms or dimensions of the Beck Depression Inventory (BDI).

Automatic classification of behaviour change language can enhance conversational agents’ capabilities to adjust their behaviour based on users’ current situations and to encourage individuals to make positive changes. However, the lack of annotated language data of change-seekers hampers the performance of existing classifiers. In this study, we investigate the use of semi-supervised learning (SSL) to classify highly imbalanced texts around behaviour change. We assess the impact of including pseudo-labelled data from various sources and examine the balance between the amount of added pseudo-labelled data and the strictness of the inclusion criteria. Our findings indicate that while adding pseudo-labelled samples to the training data has limited classification impact, it does not significantly reduce performance regardless of the source of these new samples. This reinforces previous findings on the feasibility of applying classifiers trained on behaviour change language to diverse contexts.