近年來，由於傳感器技術的進步，使用者可以很容易地收集家電的使用數據。然而，在產生巨量資料的情況下，如何從大數據探勘家電的使用行為徵是具有挑戰性的。現今探勘使擁著行為模式的研究主要集中在靜態探勘，忽略了挖掘結果的動態維護。在本文中，我們提出了一個漸增式探勘演算法：DCMiner，在智慧家居環境中有效率地探勘和維護電器間的使用關係特徵樣式。此外，一些最佳化的方法，能提出有效地減少搜索空間。實驗結果顯示出，DCMiner不僅能有效率，具有極大的可擴展性。我們並且使用了的一個真實的測資來驗證漸增式探勘電器中使用關聯性的實用性。

Abstract: Recently, due to the great advent of sensor technology, residents can collect household appliance usage data easily. However, in general, usage data are generated progressively; visualizing how appliances are used from huge amount of data is challenging. Thus, an algorithm is needed to incrementally discover appliance usage patterns. Prior studies on usage pattern discovery are mainly focused on mining patterns while ignoring the incremental maintenance of mined results. In this paper, a novel method, Dynamic Correlation Miner (DCMiner), is developed to incrementally capture and maintain the usage correlations among appliances in a smart home environment. Furthermore, several optimization techniques are proposed to effectively reduce the search space. Experimental results indicate that the proposed method is efficient in execution time and possesses great scalability. Subsequent application of DCMiner on a real dataset also demonstrates its practicability.

Table of Contents

Chinese Abstract-----------------------------------------------------------------------------I

Abstract---------------------------------------------------------------------------------------II

Table of Contents---------------------------------------------------------------------------III

List of Figures--------------------------------------------------------------------------------V

List of Tables--------------------------------------------------------------------------------VI

Chapter 1 INTRODUCTION--------------------------------------------------------------1

Chapter 2 related works---------------------------------------------------------------------4

Chapter 3 Problem Definition--------------------------------------------------------------6

Definition 1------------------------------------------------------------------------------6

Definition 2------------------------------------------------------------------------------6

Definition 3------------------------------------------------------------------------------7

Definition 4------------------------------------------------------------------------------8

Problem Definition ---------------------------------------------------------------------10

Chapter 4 DCMiner Algorithm------------------------------------------------------------11

4.1  DCMiner----------------------------------------------------------------------------12

Definition 5-------------------------------------------------------------------------------12

Definition 6-------------------------------------------------------------------------------14
Chapter 5 Experiments------------------------------------------------------¬-----------------19
5.1 Runtime Performance and Memory Usage on Synthetic Dataset ----------------20
	
5.2 Scalability Study of DCMiner ---------------------------------------------------------21

5.3 The Influence of Incremental Scenarios----------------------------------------------22

5.4 The Improvement of Search Reduction and Pruning Strategies-------------------23

5.5 Real Dataset Analysis-------------------------------------------------------------------24

Chapter 6 Conclusion ---------------------------------------------------------------------------25

Reference ------------------------------------------------------------------------------------------26       

List of Figures
Fig 1 Concept of usage database growing and a daily usage record -----------------------2

Fig 2 An example of usage database and its usage representation -------------------------7

Fig 3 Fig 2 An example of usage database and its usage representation -----------------10

Fig 4 The concept of DCMiner algorithm ----------------------------------------------------12

Fig. 5: The search space reduction on FPTDB of database DB in Fig. 2--------------------15

Fig 6 The runtime and memory usage of two algorithms on D100k-C20-N10k --------21

Fig 7 The scalability test and the influence of increment ratio ----------------------------22

Fig 8 The study of improvement of the search reduction and two pruning strategies
test, and the  performance test on real dataset [15] ----------------------------------23

   



List of Tables

Table 1: Parameters of synthetic data generator ------------------------------------------20

Table 2: Part of discovered correlation patterns from REDD dataset [15]-------------24

Reference

[1]	J. Allen, Maintaining Knowledge about Temporal Intervals. Communications of ACM, vol.26, issue 11, pp.832–843, 1983.
[2]	O. Aritoni, V. Negru. A Methodology for Household Appliances Behavior Recognition in AmI Systems Integration. 7th International Conference on Automatic and Autonomous Systems  (ICAS’11), pp. 175–178, 2011.
[3]	F. Chen, J. Dai, B. Wang, S. Sahu. Naphade, M., Lu, C.T. Activity Analysis Based on Low Sample Rate Smart Meters. 7th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD’11), pp. 240–248, 2011.
[4]	Y. Chen, J. Jiang, W. Peng, S. Lee. An Efficient Algorithm for Mining Time Interval-based Patterns in Large Databases. Proceedings of 19th ACM International Conference on Information and Knowledge Management (CIKM’10), pp. 49–58, 2010.
[5]	Y. Chen, W. Peng, W. Lee. A Novel System for Extracting Useful Correlation in Smart Home Environmen.  The Sixth International Workshop on  Domain Driven Data Mining (DDDM’13), pp. 357–364, 2013.
[6]	Y. Chen, C. Chen, W. Peng, W. Lee. Mining Correlation Patterns among Appliances in Smart Home Environment. 18th Pacific-Asia Conference in Knowledge Discovery and Data Mining, Advances in Knowledge Discovery and Data Mining (PAKDD’14), pp. 222–233, 2014.
[7]	Y. Chen, Y. Ko, W. Peng, W. Lee. Mining Appliance Usage Patterns in a Smart Home Environment. 17th Pacific-Asia Conference in Knowledge Discovery and Data Mining, Advances in Knowledge Discovery and Data Mining (PAKDD’13), pp. 99–110, 2013.
[8]	L. Farinaccio, R. Zmeureanu. Using a Pattern Recognition Approach to Disaggregate the Total Electricity Consumption in a House into the Major End-uses. Energy and Buildings, vol. 30, no. 3, pp. 245–259, 1999.
[9]	H. Goncalves, A. Ocneanu, M. Bergés. Unsupervised Disaggregation of Appliances Using Aggregated Consumption Data. The 1st KDD workshop on Data Mining Applications in Sustainability (SustKDD’11), 2011.
[10]	M. Ito, R. Uda, S. Ichimura, K. Tago,  T. Hoshi,  Y. Matsushita. A Method of Appliance Detection Based on Features of Power Waveform. The 4th IEEE Symposium on Applications and the Internet (SAINT’04), pp. 291–294, 2004.
[11]	V. Jakkula, D. Cook. Using Temporal Relations in Smart Environment Data for Activity Prediction. Proceedings of the 24th International Conference on Machine Learning (ICML'07), pp. 1–4, 2007.
[12]	V. Jakkula, D. Cook, A. Crandall. Temporal Pattern Discovery for Anomaly Detection in a Smart Home. Proceedings of the 3rd IET Conference on Intelligent Environments (IE’07), pp. 339–345, 2007.
[13]	T. Kato, H. Cho, D. Lee, T. Toyomura, T. Yamazaki. Appliance recognition from electric current signals for information-energy integrated network in home environments. Ambient Assistive Health and Wellness Management in the Heart of the City, vol. 5597, pp. 150–157, 2009.
[14]	H. Kim, M. Marwah, M. Arlitt, G. Lyon, J. Han, J. Unsupervised disaggregation of low frequency power measurements. 11th SIAM International Conference on Data Mining (SDM’11), pp. 747–758, 2011.      
[15]	J. Kolter, M. Johnson. REDD: A Public Dataset for Energy Disaggregation Research. The 1st KDD workshop on Data Mining Applications in Sustainability (SustKDD’11), 2011.                                                                 
[16]	G. Lin, S. Lee, J. Hsu, W. Jih. Applying power meters for appliance recognition on the electric panel. The 5th IEEE Conference on Industrial Electronics and Applications (ISIEA’10), pp. 2254–2259, 2010.
[17]	H. Matthews, L. Soibelman, M. Berges, E. Goldman. Automatically disaggregating the total electrical load in residential buildings: a profile of the required solution. Intelligent Computing in Engineering, pp. 381–389, 2008.
[18]	J. Pei, J. Han, B. Mortazavi-Asl, H. Pito, Q. Chen, U. Dayal, M. Hsu. PrefixSpan: Mining Sequential Patterns Efficiently by Prefix-Projected Pattern Growth. Proceedings of 17th International Conference on Data Engineering (ICDE’01), pp. 215–224, 2001.
[19]	A. Prudenzi. A Neuron Nets Based Procedure for Identifying Domestic Appliances Pattern-of-use from Energy Recordings at Meter Panel. IEEE Power Engineering Society Winter Meeting, vol. 2, pp.491-496, 2002.

[20]  K. Suzuki, S. Inagaki, T. Suzuki, H. Nakamura, K. Ito. Nonintrusive appliance
     load monitoring based on integer programming. International Conference on
     Instrumentation, Control and Information Technology (SICE’08), pp. 2742–2747, 
     2008.