We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Tree-Based Learning in RNNs for Power Consumption Forecasting

Abstract: A Recurrent Neural Network that operates on several time lags, called an RNN(p), is the natural generalization of an Autoregressive ARX(p) model. It is a powerful forecasting tool when different time scales can influence a given phenomenon, as it happens in the energy sector where hourly, daily, weekly and yearly interactions coexist. The cost-effective BPTT is the industry standard as learning algorithm for RNNs. We prove that, when training RNN(p) models, other learning algorithms turn out to be much more efficient in terms of both time and space complexity. We also introduce a new learning algorithm, the Tree Recombined Recurrent Learning, that leverages on a tree representation of the unrolled network and appears to be even more effective. We present an application of RNN(p) models for power consumption forecasting on the hourly scale: experimental results demonstrate the efficiency of the proposed algorithm and the excellent predictive accuracy achieved by the selected model both in point and in probabilistic forecasting of the energy consumption.
Subjects: Machine Learning (cs.LG); Signal Processing (eess.SP); Statistical Finance (q-fin.ST)
Cite as: arXiv:2209.01378 [cs.LG]
  (or arXiv:2209.01378v1 [cs.LG] for this version)

Submission history

From: Pietro Manzoni [view email]
[v1] Sat, 3 Sep 2022 09:21:39 GMT (1087kb,D)

Link back to: arXiv, form interface, contact.