Research Projects Using Growing Up Data

Longitudinal trajectory modelling using novel machine learning methods

Publication Date:
2022
Lead Organisation:
University of Auckland
Lead Researcher:
Susan Morton
Access Type:
Internal
Primary Classification:
Family and Whanau
Secondary Classification:
Culture and Identity
Education
Health and Wellbeing
Psych and Cog
SCONE

Objective: Investigate the utility of applying novel machine learning methodologies to explore the patterns in the longitudinal information collected from the Growing Up in New Zealand cohort.

Methods: We will apply and extend current machine learning methods including pre-processing, modelling, and postprocessing to the longitudinal data collected in Growing Up to understand how these methodologies might enhance or complement traditional longitudinal analyses. We will leverage and tailor current machine learning techniques while being cognisant of the biases in the dataset and being explicit about how machine learning can mitigate and specify these.

Anticipated Output: This methodological project is to explore the utility of machine learning within a longitudinal cohort study to prepare for analyses of existing data together with novel qualitative data being collected using the app being piloted within the Ministry of Business, Innovation & Employment "Our Voices" project. This work is designed to complement ongoing Growing Up in New Zealand trajectory analyses and allow comparisons with the application of machine learning with more traditional longitudinal statistical methodologies.