Time Series Forecasting with Extreme Learning Machines

[jnewhouse003]

Title

Time Series Forecasting with Extreme Learning Machines

URL

https://www.analyticsvidhya.com/blog/2021/12/time-series-forecasting-with-extreme-learning-machines/

Summary

A walkthrough on how to use Extreme Learning Machines with datasets to conduct time series forecasting

Key Points

The ELM algorithm can be described as follows :

1. The weights and the biases of the hidden layer are randomly assigned.
2. The hidden layer output matrix (H) is calculated by multiplying the inputs with the randomly assigned weights, adding biases, and finally applying an activation function on the output.
3. The output weight matrix is calculated by multiplying the Moore Penrose inverse of H (hidden layer output matrix) with the training data matrix (T).
4. The output weight matrix is finally used to make predictions on new data.

ELMs can be a really powerful tool. Statistical approaches to times series modeling like ARIMA models can only be applied to stationary time series (whose properties do not depend on the time at which the series is observed), they fail to capture seasonality.

Citation

“Time Series Forecasting with Extreme Learning Machines -.” Analytics Vidhya, 13 Dec. 2021, https://www.analyticsvidhya.com/blog/2021/12/time-series-forecasting-with-extreme-learning-machines/.

Repo link

[yanbingsong]

Hello，We do not have access to the website in your article.