![Pattern recognition algorithm for determining days of the week with similar energy consumption profiles [An article from: Energy & Buildings]](https://www.ebooknetworking.net/books/B00/0RR/bigB000RR2W3Y.jpg)
Pattern recognition algorithm for determining days of the week with similar energy consumption profiles [An article from: Energy & Buildings]
Book Details
Ratings & Reviews
No reviews yet — be the first!
No reviews yet.
Description
This digital document is a journal article from Energy & Buildings, published by Elsevier in 2005. The article is delivered in HTML format and is available in your Amazon.com Media Library immediately after purchase. You can view it with any web browser.
Description:
This paper describes a pattern recognition algorithm for determining days of the week with similar energy consumption profiles. The algorithm determines energy use features, such as average daily consumption or peak daily consumption, from time series of energy use. Features are transformed to remove the effects of seasonal variation that may be present in time series data. Then, the transformed features are grouped by day of the week into seven clusters. Univariate and multivariate outlier analysis methods are used to remove unusual data from the seven clusters. Finally, a modified agglomerative hierarchical clustering algorithm determines days of the week with similar energy consumption profiles. Knowledge of days of the week with similar energy consumption profiles can be used in the following ways: (1) supervisory control strategies that use forecasting algorithms, and (2) methods for detecting abnormal energy consumption in buildings. This paper contains field tests results from three buildings.
Description:
This paper describes a pattern recognition algorithm for determining days of the week with similar energy consumption profiles. The algorithm determines energy use features, such as average daily consumption or peak daily consumption, from time series of energy use. Features are transformed to remove the effects of seasonal variation that may be present in time series data. Then, the transformed features are grouped by day of the week into seven clusters. Univariate and multivariate outlier analysis methods are used to remove unusual data from the seven clusters. Finally, a modified agglomerative hierarchical clustering algorithm determines days of the week with similar energy consumption profiles. Knowledge of days of the week with similar energy consumption profiles can be used in the following ways: (1) supervisory control strategies that use forecasting algorithms, and (2) methods for detecting abnormal energy consumption in buildings. This paper contains field tests results from three buildings.
