Current content (in order of publication):
- Time Series: Getting Started (08/02/2021): This notebook contains some sample code to turn your tabular dataset into a time series dataset and run an exploratory analysis.
- Tabular Data: scikit-learn pipelines & LightGBM (26/02/2021): This notebook contains the famous titanic problem which we solve using scikit-learn pipelines in combination with LightGBM: a state of the art gradient boosting machine.
- Time Series: MINIROCKET classifier (04/06/2021): This notebook provides looks into how MINIROCKET — a novel feature extraction for time series classification — compares to feature extraction using tsfresh and Fourier transformation in terms of speed, performance, and explainabillity.
- Time Series: Labeling tools (07/06/2021): This document contains an overview of the currently available labeling tools for time series. What tools are the ones to be used, which ones you probably better avoid.
- Tabular Data: Tensorflow Data Validation (30/06/2021): This notebook showcases the functionality of TensorFlow Data Validation (TFDV), a tool developed by tensorflow to track anomalous changes in your data.
- Tabular Data: Serving decision forests with Tensorflow Serving (02/07/2021): This Readme documents the use of our custom TF Serving docker image built to serve Tensorflow Decision Forests.
- Tabular Data: 5 Tips to start working with imbalanced datasets (01/10/2021): Some pointers to keep in mind when you encounter (heavily) imbalanced ML problems.
- Tabular Data: Integrating physical knowledge into your neural networks (08/09/2022): How physical knowledge can be integrated into a neural network in order to solve your physical problem with machine learning.