Twitter profile picture of Bojan Tunguz
Tweet by Bojan Tunguz @tunguz

Traditional Machine Learning is Still Relevant for Sleep Scoring

🧵 The TL;DR

Deep learning is not the only viable approach to machine learning and there are cases where traditional methods like linear and non-linear models can outperform them, such as when reevaluating sleep scoring on datasets. Preprocessing, feature extraction, and simple machine learning models are all necessary components of a conventional ML pipeline.

    🔑 Key Points

  • Deep learning is not the only viable approach to machine learning.
  • Better and more robust results can be obtained with simpler, easier to train and deploy classical ML algorithms.
  • Preprocessing, feature extraction, and simple machine learning models are all necessary components of a conventional ML pipeline.

View Tweet Thread