Investigating depression signs among older adults using Swedish National Registry Data
| Title | Investigating depression signs among older adults using Swedish National Registry Data |
|---|---|
| Summary | Investigating depression signs among older adults using Swedish National Registry Data |
| Keywords | |
| TimeFrame | June 2025 |
| References | |
| Prerequisites | |
| Author | |
| Supervisor | Mahmoud Rahat |
| Level | Master |
| Status | Open |
In this project, you get the chance to work with Statistikkonsulenterna Väst AB and researchers/clinicians at Sahlgrenska University Hospital to apply machine learning methods on a unique dataset from the Swedish National Registry. A few interesting aspects about the dataset: 1) The data size is enormous. 2) The dataset is highly imbalanced. 3) It includes temporal information for multiple years.
The thesis will be a continuation of a recently finished master’s thesis titled “Suicide prediction among older adults using Swedish National Registry Data”. In this thesis, we explored predictive modeling of suicidal behavior among older adults in Sweden using machine learning and survival analysis. We explored models like Cox Proportional Hazards, Random Survival Forest, and LSTM networks in a two-layer architecture to predict suicide risk.
A hypothesis suggests that a large group of elderly are underdiagnosed and, therefore, do not receive proper treatment for their well-being. This is believed to be connected to the mental illness or depression of seniors, who have received less attention compared to the younger. The goal is to employ AI methods on the data to recognize such cases and react.
