| title | permalink | layout |
|---|---|---|
Research and Presentations |
/research/ |
page |
The goal of this project is about improving forecasting accuracy of electricity load through the cluster analysis of consumers (or prosumers) and time series representations (see extended abstract of this project: Improving Forecasting Accuracy Through the Influence of Time Series Representations and Clustering). Ami Solution focused on three interesting areas of data mining in the Advanced Metering Infrastructure domain:
- Time series analysis
- Clustering
- Forecasting and regression
The area of time series analysis consists of a research in (and also proposals of new) time series representations, specifically efficient dimensionality reduction of time series of electricity consumption that will input to a clustering algorithm. Ami Solution used an R package called TSrepr that involves various representations methods and is available on my GitHub repository: github.com/PetoLau/TSrepr.
The clustering task is about classification (clustering) consumers into more predictable (forecastable) groups of consumers. The challenge is to develop an algorithm that will be adaptable to a behavior of multiple data streams of electricity load. Results of clustering are then used in statistical time series analysis and regression methods to improve forecasting accuracy of aggregate (global) or individual (end-consumer) electricity load. Results of clustering can be also used for smart grid monitoring, anomaly (outlier) detection, and an extraction of typical patterns of electricity consumption.
The research scope of the forecasting and regression part focuses on methods that will benefit the most from clustering of consumers. Forecasting and regression methods have to incorporate to a model a seasonality and a trend, and they have to be adaptable to a concept drift appearance. Here is a promising approach – ensemble learning that combines multiple forecasts from various forecasting and regression methods.
One can read research papers related to this project on Google Scholar and ResearchGate (as is also mentioned in the About P.Laurinec).
Works (papers) that were presented by P.Laurinec at sevral conferences, workshops and meetup are listed below.
##### **Time Series Data Mining - from PhD to Startup** **Conference:** [SatRday-Belgrade'2018](https://belgrade2018.satrdays.org/#programme)
Date: 27.10.2018
Where: Belgrade, Serbia
Download the presentation (PDF)
##### **Time Series Representations for Better Data Mining** **Conference:** [eRum'2018](http://2018.erum.io)
Date: 15.5.2018
Where: Budapest, Hungary
Download the presentation (PDF)
##### **New Clustering-based Forecasting Method for Disaggregated End-consumer Electricity Load Using Smart Grid Data** **Conference:** [Informatics'2017](https://informatics.kpi.fei.tuke.sk/)
Date: 14.11.2017
Where: Poprad, Slovakia
Download the presentation (PDF)
##### **Is Unsupervised Ensemble Learning Useful for Aggregated or Clustered Load Forecasting?** **Workshop:** [ECML-PKDD NFMCP'2017](http://www.di.uniba.it/~loglisci/NFmcp17/program.html)
Date: 22.9.2017
Where: Skopje, Macedonia
Download the presentation (PDF)
##### **Using Clustering Of Electricity Consumers To Produce More Accurate Predictions** **Meetup:** [R<-Slovakia](https://petolau.github.io/First-R-Slovakia-meetup/)
Date: 22.3.2017
Where: Bratislava, Slovakia
Download the presentation (PDF)
##### **Adaptive Time Series Forecasting of Energy Consumption using Optimized Cluster Analysis** **Workshop:** [ICDM DaMEMO'2016](http://www.covic.otago.ac.nz/DaMEMO16/index.html)
Date: 12.12.2016
Where: Barcelona, Spain
Download the presentation (PDF)
##### **Comparison of Representations of Time Series for Clustering Smart Meter Data** **Conference:** [WCECS ICMLDA'2016](http://www.iaeng.org/publication/WCECS2016/)
Date: 21.10.2016
Where: San Francisco, California, U.S.A.
Download the presentation (PDF)
{% include share-page.html share-page_identifier=page.share-page_identifier %} {% include disqus.html disqus_identifier=page.disqus_identifier %}