wulfnovak Goto Github PK

meta-transaction-risk-classification

Classifies Meta transaction risk. Notebooks include, EDA, Feature Engineering, comparison of Halving Grid Search and Bayesian Optimization hyperparameter selection on validation set, an ML pipeline with scikit-learn, and voting + stacking ensemble models composed of input pipeline models.

rising-or-declining-country-emissions

This project categorizes countries as having rising or declining emissions from 2000-2022. Two methods are used, one which relies on there being a significant linear trend over the time period, and another that relies on longitudinal clustering. Both give similar results.

time_series_model_comparison

With minimal feature engineering, test dataset across several time series models.

wulfnovak.github.io