Page 38 - Invited Paper Session (IPS) - Volume 1

IPS35 Dinov I.D. et al.

Figure 5: Examples of recent Big health data analytic studies.

A study examining over 10,000 participants in the UKBB cohort identified
deep phenotypic traits in the population related to mental health using
unsupervised machine learning methods (Zhao, Zhao et al. 2019). The left
panel above shows the automated end-to-end computational pipeline
workflow deriving thousands of brain morphometric features. The panel on
the right shows a decision tree illustrating a simple clinical decision support
system providing machine guidance for identifying depression feelings based
on categorical variables and neuroimaging biomarkers. Each terminal node,
includes the percentage of subjects being labelled as “no” and “yes”, in this
case, answering the question “Ever depressed for a whole week.” The p-values
listed at branching nodes indicate the significance of the corresponding
splitting criterion.

4. Discussion and Conclusion
There are many remaining data science “open problems” including
establishing the fundamentals of data representation, modelling, and
analytics, quality control and data value metrics, and effectively strategies for
data wrangling, harmonization, aggregation, and joint understanding. There
also are terrific opportunities for scientific discoveries, basic science
developments, ubiquitous range of applications, development of effective
educational resources, and designing learning modules to engage a wider

27 | I S I W S C 2 0 1 9