원문정보
초록
영어
More and more enterprises are switching over to Machine learning applications to improve their analyzing and predicting capabilities of their business. In this paper we propose a new outlook towards utility computing where public services can be view as a business. A public service can be better delivered by viewing it as a business model rather than a service model. The demand supply can be better analyzed and predicted by our model. This paper is about using efficient mining techniques on real time smart meter data for any utility like water, power or gas etc. The parameters that smart meters provide from time to time over a network can give us real time readings of the consumption which in itself adds enough intelligence to the service. Now by applying temporal mining techniques on this smart meter data we attempt to show how the Business intelligence can be improved by data analysis and analytics. Though there is an opposition from some point of views that smart meters are hazardous to health due to its RF technology we can only improve utility computing by smarter data so that the service in efficient and effective.
목차
1. Introduction
2. Temporal Data Types on Smart Meter Data
2.1. Temporal
2.2. Time Series
2.3. Second-order Headings
2.4. Sequences
3. Temporal Data mining Tasks that Yield Useful Inferences
3.1. Clustering
3.2. Classification
3.3. Association Rules
3.4. Prediction
3.5. Search and Retrieval
4. Temporal Data Mining Algorithms
4.1. Generalized Sequential Pattern (GSP) Algorithm
4.2. Sequential Pattern Discovery using Equivalence Classes (SPADE)
5. Comparison between GSP and SPADE
6. Conclusion and Future Scope
References