Clustering and Feature Analysis of Smartphone Data for Depression Monitoring.

Modern advancements have allowed society to be at the most innovative stages of technology which involves the possibility of multimodal data collection. Dartmouth dataset is a rich dataset collected over 10 weeks from 60 participants. The dataset includes different types of data but this paper focuses on 10 different smartphone sensor data and a Patient Health Questionnaire (PHQ) 9 survey that monitors the severity of depression. This paper extracts key features from smartphone data to identify depression. A multi-view bi-clustering (MVBC) algorithm is applied to categorize homogeneous behaviour subgroups. MVBC takes multiple views of sensing data as input. The algorithm inputs three views: average, trend, and location views. MVBC categorizes the subjects to low, medium and high PHQ-9 scores. Real-world data collection may have fewer sensors, allowing for less features to be extracted. This creates a focus on prioritization of features. In this body of work, minimum redundancy maximum relevance (mRMR) is applied to the sensing features to prioritize the features that better distinguish the different groups. The resulting MVBC are compared to literature to support the categorized clusters. Decision Tree (DT) 10-fold cross validation shows that our method can classify individuals into the correct subgroups using a reduced number of features to achieve an overall accuracy of 94.7±1.62%. Achieving high accuracies with reduced features allows for focus on low power analysis and edge computing applications for long-term mental health monitoring using a smartphone.

Feature Analysis and Hierarchical Classification of Anxiety Severity during early COVID-19.

Distress, confusion, and anger are common responses to COVID-19. Statistics Canada created the Canadian Perspectives Survey Series (CPSS) to understand social issues and effects of COVID-19 on the Canadian labour force (LF). The evaluation of the health and health-related behaviours were done through surveys collected between April and July. Features are composed of 4600 participants and 62 questions, which include the General Anxiety Disorder (GAD)-7 questionnaire. This work proposes the use of CPSS2 survey data characteristics to identify the level of anxiety within the Canadian population during early stages of COVID-19 and is validated with the use of GAD-7 questionnaire. Minimum redundancy maximum relevance (mRMR) is applied to select the top 20 features to represent user anxiety. During classification, decision tree (DT) and support vector machine (SVM) are used to test the separation of anxiety severity. Hierarchical classification was used which separated the anxiety severity labels into different test sets and classified accordingly. We employ SVM for binary classification with 10-fold cross validation to separate the labels of Minimal and Severe anxiety to achieve an overall accuracy of 94.77±0.05%. After analysis, a subset of the reduced feature set can be represented as pseudo passive (PP) data, which are passive sensors that can augment qualitative data. The accurate classification provides proxy on what gives rise to anxiety, as well as the ability to provide early interventions. Future works can implement passive sensors to augment PP data and further understand why people cope this way.