/predicting-disruption

Supervised Learning: Methods for predicting disruptive behavior in preschool children

Primary LanguageJupyter Notebook

Predicting Disruption

Supervised Learning: predicting behavior patterns of preschool children based on screen habits and demographics.

Start with the Overview Here

In this project, I use data collected by a Deakin University Study to see if screen time, outdoor time, and other behavioral measures can be used to predict children with a high probability of being disruptive.

PCA on binary variables

Also included is a sidebar on Primary Components Analysis of binary variables. A series of visualizations are created with simple, fake data to help data scientists build an intuition around what this does and its benefits and limitations.