Skip to main content

Statistical Methods vs DL Methods

Comparing outputs from multivariate multi-step forecasting using traditional methods and ML based methods

Statistical and Machine Learning forecasting methods: Concerns and ways forward claim that statistical methods outperform machine learning based methods on various tasks.

Multiple papers have different claims. Few claim that statistical methods outperform machine learning based methods. While other claim that machine learning based methods outperform statistical methods.

References:

  1. Statistical and Machine Learning forecasting methods: Concerns and ways forward (storage.googleapis.com) (S > ML)
  2. 1909.13316.pdf (arxiv.org) Machine Learning vs Statistical Methods for Time Series Forecasting: Size Matters (DL > S)
  3. Comparison of Time Series Methods and Machine Learning Algorithms for Forecasting Taiwan Blood Services Foundation's Blood Supply (hindawi.com) (S > ML)
  4. Comparing Deep Learning and Statistical Methods in Forecasting Crowd Distribution from Aggregated Mobile Phone Data (ML > S)
  5. A Comparative Study of Time Series Forecasting Methods for Short Term Electric Energy Consumption Prediction in Smart Buildings (ML > S)
  6. COVID-19: A Comparison of Time Series Methods to Forecast Percentage of Active Cases per Population (ML > S)
  7. An Empirical Comparison of Machine Learning Models for Time Series Forecasting: Econometric Reviews: Vol 29, No 5-6 (tandfonline.com) (comparison of only ML based models)