Project information
- Scope:
Data cleaning, AI algorithm research and development, Model deployment, Backend & Frontend development, Product launch and maintenance, Model validation - Technologies:
Python, Pandas, Scikit-Learn, XGBoost, Flask, Node.js, D3.js, Docker, Heroku - WebApp link: ELSA
- Publication link: SSRN archive
- Awards:
Poster award, 36th European College of Neuropsychopharmacology (ECNP) Congress, 10/2023
Finalist, Max Planck Foundation Startup Product Demo day, 11/2022
Using Machine Learning to identify personal Depression & Anxiety risk factors to help general-public strengthen their mental well-being
The challenge
The burden of Depression & Anxiety is a global socioeconomic issue. The development of these disorders are strongly influenced by social, environmental and physical factors. In particular, individuals from Low-income and resource-limited communities are more vulnerable, yet lack the mental healthcare access and resources to help them strengthen their mental well-being.
The solution
Social, environmental, and physical conditions can be easily assessed beyond clinical environment and are largely associated with the presence of Depression & Anxiety. We developed and validated a machine learning model using these conditions to detect the presence of Depression & Anxiety with transparent model result explanation, which was further deployed into an easy-to-use web application with quantified personalised risk factor identification applicable for the general-public outside of clinical settings.
The end result
The tool offers valuable insights into personal mental health risks,
empowering individuals to tackle issues relevant to their mental health.
This is particularly helpful for individuals from low-/middle-income countries,
who are often vulnerable but lack the access to mental healthcare.
The tool has been validated in Low-and-Middle Income country as well as High-Income country.