Electricity usage type selection and model validation based on electricity usage measurements

Teemu Mutanen, Tomi Sorasalmi, Robert Weiss

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

    Abstract

    In a smart gird environment, the hourly electricity consumption measurements are used for detection of user segments and of a possible underlying distribution. Electricity usages are segmented according to the daily behaviour. The resulted segments indicate that electricity consumption is distributed differently in different segments. The results show that the electricity consumption is distributed based on a log-normal distribution in some segments and in some subsamples inside the segment. The information can be used as the basis when training other models
    Original languageEnglish
    Title of host publicationAdvances in Data Mining, Poster and Industry Proceedings 2012
    Pages52-60
    Publication statusPublished - 2012
    MoE publication typeA4 Article in a conference publication
    Event12th Industrial Conference on Advances in Data Mining, ICDM 2012 - Berlin, Germany
    Duration: 13 Jul 201220 Jul 2012

    Conference

    Conference12th Industrial Conference on Advances in Data Mining, ICDM 2012
    Abbreviated titleICDM 2012
    CountryGermany
    CityBerlin
    Period13/07/1220/07/12

    Keywords

    • high frequency data
    • electricity measurements
    • classication
    • log-normal distribution
    • model validation

    Fingerprint Dive into the research topics of 'Electricity usage type selection and model validation based on electricity usage measurements'. Together they form a unique fingerprint.

    Cite this