WP6 will develop and apply machine learning and other data analytic methods for answering the research questions of the project. It will focus on modelling the relations between exposome and mental health in children. In these models we will look at intermediate factors as restoration (sleep), stress and self-regulation
18 month update
The aim of WP6 is to learn and apply how machine learning and other statistical methods can be used to model the relation between the exposome and mental health and cognitive development in children.
Within the first 18 months, WP6 looked into how the data available within the Equal-Life project can be combined and compared (“validated”) for use throughout the project. Machine learning will be used to determine the relevant information in selected studies (“untargeted analyses”). ‘Random Forest’ was proposed as a suitable method and a Random Forest based computer code was thus developed and shared. This code can be applied to individual studies available in Equal-Life to identify the important predictors of the mental health and cognitive development in children, next to evidence from the literature and theoretical considerations (from WP1).
Simulation studies were performed to investigate whether the proposed approach for the validation of the (harmonized) variables would lead to unbiased estimates and outcomes. WP6 will continue with the development of a methodology for targeted analyses. This methodology will be based on testable hypotheses translated from the conceptual models, path models and their accompanied research questions developed by WP1.