Floating search methods in feature selection
Pattern Recognition Letters
Machine Learning
Feature Selection: Evaluation, Application, and Small Sample Performance
IEEE Transactions on Pattern Analysis and Machine Intelligence
Introduction to Machine Learning (Adaptive Computation and Machine Learning)
Introduction to Machine Learning (Adaptive Computation and Machine Learning)
Pattern Recognition and Machine Learning (Information Science and Statistics)
Pattern Recognition and Machine Learning (Information Science and Statistics)
A Direct Method of Nonparametric Measurement Selection
IEEE Transactions on Computers
ICT for green: how computers can help us to conserve energy
Proceedings of the 1st International Conference on Energy-Efficient Computing and Networking
Data Mining: Practical Machine Learning Tools and Techniques
Data Mining: Practical Machine Learning Tools and Techniques
Towards automatic classification of private households using electricity consumption data
BuildSys '12 Proceedings of the Fourth ACM Workshop on Embedded Sensing Systems for Energy-Efficiency in Buildings
Occupancy Detection from Electricity Consumption Data
Proceedings of the 5th ACM Workshop on Embedded Systems For Energy-Efficient Buildings
Hi-index | 0.00 |
Interest in analyzing electricity consumption data of private households has grown steadily in the last years. Several authors have for instance focused on identifying groups of households with similar consumption patterns or on providing feedback to consumers in order to motivate them to reduce their energy consumption. In this paper, we propose to use electricity consumption data to classify households according to pre-defined "properties" of interest. Examples of these properties include the floor area of a household or the number of its occupants. Energy providers can leverage knowledge of such household properties to shape premium services (e.g., energy consulting) for their customers. We present a classification system - called CLASS - that takes as input electricity consumption data of a private household and provides as output the estimated values of its properties. We describe the design and implementation of CLASS and evaluate its performance. To this end, we rely on electricity consumption traces from 3,488 private households, collected at a 30-minute granularity and for a period of more than 1.5 years. Our evaluation shows that CLASS - relying on electricity consumption data only - can estimate the majority of the considered household properties with more than 70% accuracy. For some of the properties, CLASS's accuracy exceeds 80%. Furthermore, we show that for selected properties the use of a priori information can increase classification accuracy by up to 11%.