Age Group Detection in Stochastic Gas Smart Meter Data Using Decision-Tree Learning

William Hurst*, Casimiro A. Curbelo Montanez, Dhiya Al-Jumeily

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference paperAcademicpeer-review

1 Citation (Scopus)


Smart meters are the next generation gas and electricity meters where the meter readings are presented digitally and accurately to the consumer via an In-Home Display unit. Access to the data sets generated by smart meters is becoming increasingly prevalent. As such, this paper presents an approach for detecting age groups from aggregated smart meter data. The benefits of achieving this range from healthcare cluster mapping for smart resource allocation and intelligent forecasting, to anomaly detection within age-range groups. The technique proposed and presented in this paper uses a cloud analytics platform for the data processing. Using this approach, the classification is able to achieve a 75.1% AUC prediction accuracy using a two-class decision forest and a 74.6% AUC with a boosted two-class decision tree. A two-class linear regression model, which is able to achieve a 53.7% accuracy, is applied as a benchmark for comparison with the decision tree approach.

Original languageEnglish
Title of host publicationIntelligent Computing Methodologies - 15th International Conference, ICIC 2019, Proceedings
EditorsDe-Shuang Huang, Zhi-Kai Huang, Abir Hussain
PublisherSpringer Verlag
Number of pages13
ISBN (Print)9783030267650
Publication statusPublished - 2019
Externally publishedYes
Event15th International Conference on Intelligent Computing, ICIC 2019 - Nanchang, China
Duration: 3 Aug 20196 Aug 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11645 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference15th International Conference on Intelligent Computing, ICIC 2019


  • Machine learning
  • Smart meter
  • Variability

Fingerprint Dive into the research topics of 'Age Group Detection in Stochastic Gas Smart Meter Data Using Decision-Tree Learning'. Together they form a unique fingerprint.

Cite this