Limit search to available items
Book Cover
Book
Author Shasha, Dennis Elliott.

Title High performance discovery in time series : techniques and case studies / Dennis Shasha, Yunyue Zhu
Published New York : Springer, [2004]
©2004

Copies

Location Call no. Vol. Availability
 W'PONDS  519.55 Sha/Hpd  AVAILABLE
Description ix, 190 pages : illustrations ; 24 cm
Series Monographs in computer science
Monographs in computer science.
Contents Machine derived contents note: Part I Review of Techniques -- 1 Time Series Preliminaries 3 -- 1.1 High Performance Time Series Analysis 5 -- 2 Data Reduction and Transformation Techniques 9 -- 2.1 Fourier Transform 10 -- 2.1.1 Orthogonal Function Families 11 -- 2.1.2 Fourier Series 13 -- 2.1.3 Fourier Transform 15 -- 2.1.4 Discrete Fourier Transform 16 -- 2.1.5 Fast Fourier Transform 25 -- 2.1.6 Discrete Fourier Transform - The Bottom Line 26 -- 2.2 Wavelet Transform 29 -- 2.2.1 From Fourier Analysis to Wavelet Analysis 29 -- 2.2.2 Haar Wavelet 32 -- 2.2.3 Multiresolution Analysis 34 -- 2.2.4 Wavelet Transform 37 -- 2.2.5 Discrete Wavelet Transform 41 -- 2.2.6 Wavelets - The Bottom Line 46 -- 2.3 Singular Value Decomposition 52 -- 2.3.1 Existence and Uniqueness of Singular Value Decomposition . 52 -- 2.3.2 Optimality of Singular Value Decomposition 56 -- 2.3.3 Data Reduction in Singular Value Decomposition 57 -- 2.3.4 Singular Value Decomposition - The Bottom Line 58 -- 2.4 Sketches 61 -- 2.4.1 Euclidean Distance 62 -- 2.4.2 Lp Distance 65 -- 2.4.3 Sketches - The Bottom Line 67 -- 2.5 Comparison of Data Reduction Techniques 67 -- 2.6 Q uestions 69 -- 3 Indexing Methods 73 -- 3.1 B -tree 73 -- 3.2 K D -B -tree 76 -- 3.3 R -tree 78 -- 3.4 Grid Structure 81 -- 3.5 Indexes The Bottom Line 85 -- 3.6 Questions 85 -- 4 Flexible Similarity Search 87 -- 4.1 GEM INI Framework 88 -- 4.2 Shifting and Scaling 90 -- 4.3 Time Scaling 94 -- 4.4 Local Dynamic Time Warping 96 -- 4.5 Q uestions 100 -- Part II Case Studies -- 5 StatStream 103 -- 5.1 Introduction 103 -- 5.2 Data And Queries 105 -- 5.2.1 Time Series Data Streams 105 -- 5.2.2 Temporal Spans 105 -- 5.2.3 Statistics to M onitor 106 -- 5.3 Statistics Over Sliding Windows 106 -- 5.3.1 Single Stream Statistics 107 -- 5.3.2 Correlation Statistics 108 -- 5.3.3 Inner Product with Aligned Windows 108 -- 5.3.4 Inner Product with Unaligned Windows 110 -- 5.3.5 IO Performance 112 -- 5.3.6 Monitoring Correlations between Data Streams 112 -- 5.3.7 Parallel Implementation 116 -- 5.4 Software Architecture of the StatStream System 117 -- 5.5 Em pirical Study 118 -- 5.5.1 Speed Measurement 119 -- 5.5.2 Measuring Precision 120 -- 5.6 Related Work 123 -- 5.7 Conclusion 125 -- 5.8 Q uestions 126 -- 6 Query by Humming 127 -- 6.1 Introduction 127 -- 6.2 Related W ork 128 -- 6.2.1 Insights from HumFinder 130 -- 6.3 Quer y by Humming System 130 -- 6.3.1 User Humming: The Input Hum-query 130 -- 6.3.2 A Database of Music 131 -- 6.3.3 Indexing Databases for Efficient Hum Query Retrieval 133 -- 6.4 Indexing Scheme for Dynamic Time Warping 134 -- 6.5 Software Architecture of the HumFinder System 139 -- 6.6 Experim ents 140 -- 6.6.1 Quality of the Query by Humming System 141 -- 6.6.2 Experiments on the DTW Index 142 -- 6.6.3 Scalability Testing 144 -- 6.7 Conclusions 148 -- 6.8 Questions 150 -- 7 Elastic Burst Detection 151 -- 7.1 Introduction 151 -- 7.1.1 Problem Statement 152 -- 7.1.2 Insights of Omniburst 153 -- 7.2 Data Structure and Algorithm 154 -- 7.2.1 Wavelet Data Structure 154 -- 7.2.2 Shifted Binary Tree 155 -- 7.2.3 Streaming Algorithm 159 -- 7.2.4 Other Aggregates 161 -- 7.2.5 Extension to Two Dimensions 161 -- 7.3 Software Architecture of the OmniBurst System 162 -- 7.4 Empirical Results of the OmniBurst System 163 -- 7.4.1 Effectiveness Study 163 -- 7.4.2 Performance Study 164 -- 7.5 Related work 168 -- 7.6 Conclusions and Future Work 172 -- 7.7 Questions 172 -- 8 A Call to Exploration 175 -- A Answers to the Questions 177 -- A .2 Chapter 2 177 -- A .3 Chapter 3 178 -- A .4 Chapter 4 178 -- A .5 Chapter 5 179 -- A .6 Chapter 6 180 -- A .7 C hapter 7 180 -- R eferences 181 -- Index 189
Bibliography Includes bibliographical references (pages [181]-187) and index
Subject Time-series analysis.
Author Zhu, Yunyue.
LC no. 2004556616
ISBN 0387008578