Each year, more than 700,000 lives are lost to suicide worldwide. It is therefore incredibly important to implement effective prevention interventions. However, for most of these interventions it is crucial to know which subgroups in the population to target.
The first part of Berkelmans' thesis focuses entirely on this problem, and approaches it through the lens of big data. Using data from Statistics Netherlands he looked at demographic data, and whether he could identify groups of high risk by their demographic features. This led to finding many groups, such as men, those of middle age, those on benefits, and those living alone.
Berkelmans wanted to know whether there are intersections of these populations that are at higher risk than you would expect if these risk factors act independently. Again he found multiple unexpected groups such as male widowers, and people with a low level of education between ages 25 and 40. Subsequently he looked at medication usage, and found that a great deal of classes of medication were associated with a heightened risk of suicide.
The second part of this thesis focuses on the theoretical questions that arose in relation to the first part: how do you decide which features to include, is it possible to quantify dependence between observer variables? Berkelmans started out designing a measure of dependence which answers the second of these questions.
He showed it had a number of basic properties you would expect such a measure to have, and showed none of the reasonably commonly used measures have these properties. After that, he extended this to a measure of feature importance by considering how much a feature contributes to the dependency of the outcome on "coalitions" of features. His notion of feature importance satisfied all of them, whereas most other feature importance methods had less than half, with none having more than 13 of the 20 properties.
This research is featured on the website of Computable.
Information
Title: "Turing and Van Gogh walk into a bar: A computational approach to suicide research"
Supervisors:
- prof.dr. R.D. van der Mei
- prof.dr. S. Bhulai
Co-supervisors:
- dr. R. Gilissen
- dr. L. Schweren
