/time-series-classification-on-sensor-data

Time series classification to identify the state of human activity.

Primary LanguageJupyter Notebook

time-series-classification-on-sensor-data

Data

The data is from a Kaggle competition Tabular Playground Series April 2022. The multivariate time series consists of 13 time series, each of which has 60 data points.

Goal

This is a binary classification problem, so we need to predict the probability of an input being in group 1.

Models

I tried various approaches, both in the feature extraction and in the modelling stages.

  • Feature extraction
    • ROCKET
    • Shapelets
    • None
  • Modelling
    • Gradient Boosting
    • Neural Network (the basic, a fully connected layer)
    • LSTM
    • Logistic regression

Three notebooks are included in this repository. The first two notebooks have their own repo, but I also put them here to complete the collection.

  1. tps_apr2022_rocket focus on ROCKET as a feature extractor. The video is here
  2. shapelet_tslearn focus on Shapelets to transform the dataset. The video is here. Note that I didn't set the seed for learning shapelets, so the shapelets visualizations in this notebook may look a little bit different than what you see in the video. But that doesn't affect the model performance.
  3. The video for models with low AUC is here. In that video I also share my opinion about what we can learn from this project.