Cloud Energy Micro-Moment Data Classification: A Platform Study

TL;DR

This study benchmarks major cloud AI platforms for classifying energy micro-moments to enhance consumer energy behavior, demonstrating comparable performance but highlighting algorithmic limitations.

Contribution

It provides a comparative analysis of cloud platforms for energy micro-moment classification using various classifiers, aiding researchers in platform selection.

Findings

Cloud platforms show similar classification performance.

Algorithmic limitations affect training efficiency.

Benchmarking guides future energy behavior interventions.

Abstract

Energy efficiency is a crucial factor in the well-being of our planet. In parallel, Machine Learning (ML) plays an instrumental role in automating our lives and creating convenient workflows for enhancing behavior. So, analyzing energy behavior can help understand weak points and lay the path towards better interventions. Moving towards higher performance, cloud platforms can assist researchers in conducting classification trials that need high computational power. Under the larger umbrella of the Consumer Engagement Towards Energy Saving Behavior by means of Exploiting Micro Moments and Mobile Recommendation Systems (EM)3 framework, we aim to influence consumers behavioral change via improving their power consumption consciousness. In this paper, common cloud artificial intelligence platforms are benchmarked and compared for micro-moment classification. The Amazon Web Services, Google Cloud Platform, Google Colab, and Microsoft Azure Machine Learning are employed on simulated and real energy consumption datasets. The KNN, DNN, and SVM classifiers have been employed. Superb performance has been observed in the selected cloud platforms, showing relatively close performance. Yet, the nature of some algorithms limits the training performance.