DOCSLIB.ORG

Application of Support Vector Machine to Synthetic Earthquake Prediction∗

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Application of Support Vector Machine to Synthetic Earthquake Prediction∗ Earthq Sci (2009)22: 315−320 315 Doi: 10.1007/s11589-009-0315-8 Application of support vector machine to ∗ synthetic earthquake prediction Chun Jiang 1, Xueli Wei 2 Xiaofeng Cui 1 and Dexiang You 2 1 Earthquake Administration of Tianjin Municipality, Tianjin 300201, China 2 Tianjin University of Technology, Tianjin 300191, China Abstract This paper introduces the method of support vector machine (SVM) into the field of synthetic earthquake pre- diction, which is a non-linear and complex seismogenic system. As an example, we apply this method to predict the largest annual magnitude for the North China area (30°E−42°E, 108°N−125°N) and the capital region (38°E−41.5°E, 114°N−120°N) on the basis of seismicity parameters and observed precursory data. The corresponding prediction rates for the North China area and the capital region are 64.1% and 75%, respectively, which shows that the method is feasible. Key words: support vector machine; seismicity parameter; precursory data; synthetic earthquake prediction CLC number: P315.75 Document code: A 1 Introduction 2 Basic principle of SVM regression The support vector machine (SVM) has not only a algorithm (Vapnik, 1995) strict theory basis, but also a strong generalization (pre- It is firstly to consider the linear regression model. diction) capacity, which can better solve the practical Suppose that a sample with k cases is described as problems of small samples, nonlinearity, higher dimen- ∈n × sion and local minimum point. At present, SVM are (x1 ,y 1 ), ( x 2 , y 2 ),L ,(xk ,y k )RR , (1) widely used in text classification, handwriting recogni- the linear discriminant function is tion, image classification and bioinformatics, and now it =• + (2) has been extended to synthetic estimates and prediction f ()x w x b of time series. Wang et al (2005, 2006) introduced the and assume that all the training data can be fitted as a method to predict strong earthquakes in Chinese linear function without error under the accuracy of ε, mainland and studied the non-linear relations between that is, the time series of strong seismicity in China and the −• + ≤ ε ⎧yi ()w xi b global large earthquakes as well as sunspot activities. ⎨ i=1, 2, , k . (3) • + − ≤ ε L His applications have obtained some meaningful results. ⎩()w xi b yi In view of that earthquake preparation-occurrence Considering the case of permissible fitting error, is a complex nonlinear dynamic process, we introduce * the relaxation indexes ξi≥0 and ξ ≥0 are introduced. SVM into a kind of synthetic earthquake prediction. The i Then equation (3) turns into the following equation: preliminary research by this method on the basis of seismic parameters has obtained a better result (Jiang et ⎧y −()w• x +b ≤ε + ξ i i i = ⎨ ∗ i1,2,L , k , (4) al, 2006). In this paper, SVM is used for a synthetic • + − ≤ε + ξ ⎩()w xi b yi i earthquake prediction in the North China area and the ξ ≥ ξ * ≥ capital region on the basis of both seismicity parameters where i 0 and i 0. That the SVM used for regres- and observed precursory data. sion estimate is minimizing function (5) under con- straints (4). ∗ Received 21 October 2008; accepted in revised form 2 February 2009; published 10 June 2009. Corresponding author. e-mail: [email protected] 316 Earthq Sci (2009)22: 315−320 k φ •φ * 1 * If K(xi, xj)= (xi) (xj), then equation (9) is turned ξ ξ =• +ξ + ξ R(,,)w w w C∑ ()i i (5) 2 i=1 into k The first term of equation (5) is to make the regression ω~ α α * = −1 α − α* α − α * ⋅ (,) * ( )( ) w,,,b ξ ξ ∑ i i j j function more flat, so as to enhance the generalization 2 i, j= 1 ability. The second term is for reducing errors, the con- k k −α + α* ε + α − α * K(,)()()xi x j ∑ i i ∑ i iy i , (10) stant C>0 controls the punishment of samples whose i=1 i=1 errors are more than ε, where ε is a positive constant. If where − | f (xi) yi|<ε, then it will be ignored, otherwise, the error k =α − α * φ counted as | f (xi)−yi|−ε. w ∑ ()(i ix i ). (11) i=1 With fewer samples, we generally use the dual φ theory to find a solution for SVMs and turn it into a Denoting w• (x)=w0, f (x) can be expressed as k quadratic programming problem. Then the Lagrangian =α − α * + = + f ()()(,)x ∑ i iK x i x j b w0 b. (12) function is introduced: i=1 * ** 1 The theory of SVM uses the scalar product K(x , x ) = L(,,,,,,,)w b ξ ξ α α γ γ =w• w + i j φ • φ 2 (xi) (xj) in the high-dimensional eigenspace rather k k ξ+ ξ* − α ξ + ε − + − than uses function φ directly, thus the problem that w C∑ ()[i i ∑ i i yi f (xi )] i=1 i=1 cannot be expressed with the unknown φ is solved clev- k k α** ξ+ ε + − −ξ γ + ξ** γ erly, where K(xi, xj) is called kernel function. It has been ∑ i[ i yi f (xi )]∑ (i i i i ). (6) i=1 i=1 proved that a symmetric function can be a kernel as long α α * γ γ * ⋅⋅⋅ as it meets the Mercer condition. The common kernel where, i, i , i and i (i=1, 2, , k) are not less than functions are: 1 Polynomial kernel function K(xi, xj) = zero. Then the dual function of the Lagrange function (6) d (x • x +c) , c>0, d =1, 2, ⋅⋅⋅; 2 Radial basis kernel func- is i j = − − 2σ 2 tion (RBF) K(xi , x j ) exp( ||xi x j || 2 )； 3 Sig- k * 1 * * ω~ α α = −α − α α − α • − (,) * ( )( )(x x ) w,,,b ξ ξ ∑ i i j j i j moid kernel function K(x , x )= tanh[b (x• x )+ c ] . 2 i, j= 1 i j i j k k It should be noted that these kernel functions have α+ α* ε + α − α * ∑()()i i ∑ i iy i . (7) their own input ranges, which should be taken a i=1 i=1 data-scale transformation before specific applications. The dual solution to this problem is maximizing equa- Selecting kernel function needs certain prior knowledge tion (7) under constraints (8) (Xi, 1983) and there is no general conclusion at present. Scholkopf k α− α * = et al (1998) discussed the selection and construction of ∑ (i i ) 0, (8) i=1 kernel functions. * According to the Karush-Kuhn-Tucker (KKT) where 0≤αi≤C and ≤ 0≤α ≤ C , and i=1, 2, ⋅⋅⋅, k. i conditions (Xi, 1983), at the optimal solution we have The basic idea of non-linear approximation is: α ε+ ξ − + = firstly, mapping the input space into the ⎧ i[ iy i f (xi )] 0 ⎨ i = 1, 2, , k (13) high-dimensional eigen-space by the non-linear trans- α* ε+ ξ * + − = L ⎩ i[ iy i f (xi )] 0 formation φ (x); secondly, doing linear approximation in and the high-dimensional eigenspace, that is f (x)=w•φ (x)+b; ξ γ = then obtaining the non-linear regression result in the ⎧ i i 0 ⎨ i=1, 2, , k . (14) original space. Thus the issue of non-linear regression is ξ** γ = L ⎩ i i 0 turned into maximizing function α α * = k From equation (13) we have i i 0 , and ω~ α α * = −1 α − α* α − α * ⋅ (,) * ( )( ) w,,,b ξ ξ ∑ i i j j = = −ε − • φ 2 i, j 1 ⎧b yi w() xi k k ⎨ . (15) * * =ε + − • φ φ• φ −α + α ε + α − α ⎩b yi w() xi [ (xi ) ( x j )]∑ (i i )∑ (i i )y i (9) i=1 i=1 The above-mentioned algorithms can be achieved easily under constraints (8). by using the existing optimized software packages. Earthq Sci (2009)22: 315−320 317 3 Application of SVM to synthetic area based on the analysis of 210 earthquake cases from Earthquake Cases in China (Zhang, 1988, 1990a, b, earthquake prediction 1999, 2000; Chen, 2002a, b, c) and determine the fol- According to the incomplete figures, dozens of lowing rules for selecting earthquake cases and seismic seismicity parameters and precursors are used in moni- precursory abnormal indexes: toring and predicting earthquakes in China at present, 1) Earthquake cases with relatively concentrated such as hydrochemistry, water level, geoelectricity, seismic precursory anomalies in the observation items geomagnetism, electromagnetic wave, crustal deforma- should be the first choice. The seismic precursory ab- tion, gravity, stress and so on. However, earthquake normal indexes of M5.0−5.9, M6.0−6.9 and M≥7.0 preparation is a non-linear and unstable process, ac- earthquakes should make up 15%−20%, 30%−40% and companying a large number of different random phe- ≥50% of the total observation items, respectively. nomena. And seismicity parameters and earthquake 2) Seismic precursory anomalies relatively concen- precursors often show their single or group abnormal trated in a larger scope in the vicinity of epicentral re- features in different forms. No matter in the synthetic gion are selected. The seismic precursory anomalies of earthquake prediction based on different groups of seis- M5.0−5.9, M6.0−6.9 and M≥7.0 earthquakes should lo- mic precursory anomalies by the same method or based cate within a radius of 200 km, 300 km and 500 km to on the same group of seismic precursory anomalies by the epicenters, respectively. different methods, the conclusions drawn from different 3) Single seismic precursory anomaly with a phases or the same phase of seismic activities are often high-R score should be firstly selected.
Load more

Recommended publications
  • Some Experiences from Damages of Embankments During Strong Earthquakes China
    Missouri University of Science and Technology Scholars' Mine International Conference on Case Histories in (1988) - Second International Conference on Geotechnical Engineering Case Histories in Geotechnical Engineering 01 Jun 1988, 1:00 pm - 5:30 pm Some Experiences from Damages of Embankments During Strong Earthquakes China Shen Choggang Institute of Water Conservancy and Hydroelectric Research (IWHR), Beijing, China Follow this and additional works at: https://scholarsmine.mst.edu/icchge Part of the Geotechnical Engineering Commons Recommended Citation Choggang, Shen, "Some Experiences from Damages of Embankments During Strong Earthquakes China" (1988). International Conference on Case Histories in Geotechnical Engineering. 21. https://scholarsmine.mst.edu/icchge/2icchge/2icchge-session4/21 This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License. This Article - Conference proceedings is brought to you for free and open access by Scholars' Mine. It has been accepted for inclusion in International Conference on Case Histories in Geotechnical Engineering by an authorized administrator of Scholars' Mine. This work is protected by U. S. Copyright Law. Unauthorized use including reproduction for redistribution requires the permission of the copyright holder. For more information, please contact [email protected]. Proceedings: Second International Conference on Case Histories In Geotechnical Engineering, June 1-5, 1988, StLouis, Mo., Paper No. 4.03 Some Experiences from Damages of Embankments During Strong Earthquakes in China Shen Choggang Adviser, Senior Engineer, Institute of Water Conservancy and Hydroelectric Research (IWHR), Beijing, China SYNOPSIS: This paper gives a series of case histories of earth dams which have been damaged by strong earthquakes in the past 25 years in China.
    [Show full text]
  • Shahid Yusuf and Kaoru Nabeshima Contrasting Development Paths For
    Public Disclosure Authorized Shahid Yusuf and Kaoru Nabeshima TWO Public Disclosure Authorized DRAGON HEADS Contrasting Development Public Disclosure Authorized Paths for Beijing and Shanghai Public Disclosure Authorized Two Dragon Heads Two Dragon Heads Contrasting Development Paths for Beijing and Shanghai Shahid Yusuf Kaoru Nabeshima Washington, D.C. © 2010 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: [email protected] All rights reserved 1 2 3 4 12 11 10 09 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com.
    [Show full text]
  • 2014 Annual Report
    Stock Code: 1398 USD Preference Shares Stock Code: 4603 EUR Preference Shares Stock Code: 4604 RMB Preference Shares Stock Code: 84602 Annual Report 2014 2014 Annual ReportAnnual 中國北京市西城區復興門內大街55號 郵編：100140 No.55 Fuxingmennei Avenue, Xicheng District, Beijing, PRC Post Code: 100140 www.icbc.com.cn, www.icbc-ltd.com Company Profile Industrial and Commercial Bank of China Limited, and services to 5,090 thousand corporate customers formerly known as Industrial and Commercial Bank and 465 million personal customers by virtue of the of China, was established on 1 January 1984. On 28 distribution channels consisting of 17,122 domestic October 2005, the Bank was wholly restructured to a institutions, 338 overseas institutions and 2,007 joint-stock limited company. On 27 October 2006, the correspondent banks worldwide, as well as through its Bank was successfully listed on both SSE and SEHK. E-banking network comprising a range of Internet and telephone banking services and self-service banking Through its continuous endeavor and stable centers, forming a diversified and internationalized development, the Bank has developed into the top operating structure focusing on commercial banking large listed bank in the world, possessing an excellent business and maintaining a leading position in the customer base, a diversified business structure, strong domestic market in the commercial banking sector. innovation capabilities and market competitiveness. The By virtue of its outstanding performance, the Bank’s Bank has its presence in six continents, and its overseas brand image and international influence have been network has expanded to 41 countries and regions. consolidated. The Bank has become one of most The Bank provides comprehensive financial products valuable financial brands in the world.
    [Show full text]
  • Pathways to Earthquake Resilience in China
    Report Pathways to earthquake resilience in China October 2015 Overseas Development Institute 203 Blackfriars Road London SE1 8NJ Tel. +44 (0) 20 7922 0300 Fax. +44 (0) 20 7922 0399 E-mail: [email protected] www.odi.org www.odi.org/facebook www.odi.org/twitter Readers are encouraged to reproduce material from ODI Reports for their own publications, as long as they are not being sold commercially. As copyright holder, ODI requests due acknowledgement and a copy of the publication. For online use, we ask readers to link to the original resource on the ODI website. The views presented in this paper are those of the author(s) and do not necessarily represent the views of ODI. © Overseas Development Institute 2015. This work is licensed under a Creative Commons Attribution-NonCommercial Licence (CC BY-NC 3.0). ISSN: 2052-7209 Cover photo: Photo by GDS, Children receiving the GDS disaster risk reduction kit, Shaanxi Province, China Contents Acknowledgements 9 About the authors 9 Glossary of terms 11 Acronyms 11 1. Introduction 13 John Young 2. Earthquake disaster risk reduction policies and programmes in China 16 Cui Ke, Timothy Sim and Lena Dominelli 3. Current knowledge on seismic hazards in Shaanxi Province 23 By Feng Xijie, Richard Walker and Philip England 4. Community-based approaches to disaster risk reduction in China 30 Lena Dominelli, Timothy Sim and Cui Ke 5. Case study: World Vision’s community disaster response plan in Ranjia village 42 William Weizhong Chen, Ning Li and Ling Zhang 6. Case study: Gender Development Solution’s disaster risk reduction in primary education 46 Zhao Bin 7.
    [Show full text]
  • Chapter 1: Seismic Activity and Geological Background
    CHAPTER 1: SEISMIC ACTIVITY AND GEOLOGICAL BACKGROUND SEISMIC ACTIVITY IN TANGSHAN AND ITS SURROUNDING AREAS Zhu Chuanzhen* The Tangshan earthquake is not an isolated and unexpected event. It has a breeding and formation process. In this paper some inherent observations of the Tangshan earthquake are summarized based on the historical and recent seismic activities in Tangshan and its surrounding areas as well as the characteristics of the Tangshan earthquake sequences itself. Fundamental data necessary for the analysis and study of the damage in the Tangshan earthquake are provided. Meanwhile, some seismic precursors prior to the Tangshan earthquake are also mentioned briefly. I. Summary of Historical Earthquakes China is a country of active seismicity and has also the longest historical earthquake record in the world. The statistics and analysis of historical earthquakes for more than 3000 years show that the distribution of strong earthquakes in China are characterized by the belt shape in space and the reoccurrence in time. Moreover, the stress accumulation and release are non-uniform in time and space within individual seismic zones, and the seismicity is also characterized by having different active periods (Shi Zhenliang et al., 1974). Therefore, it is necessary to investigate the distribution features of strong earthquakes on a larger time and space scale in order to study the processes of the Tangshan earthquake. 1. Strong earthquakes in North China The seismic activity in North China can be considered for the whole area according to the epicentral distribution, focal mechanism, direction of long axis of isoseismals of historical strong earthquakes, as well as the mean crust thickness, geological structure and geomorphology.
    [Show full text]
  • Crustal Deformation Before Strong Earthquakes and Utility of Earthquake Prediction by Geodetic Measurements: Extended Abstract
    J. Phys. Earth, 34, Suppl., S111-S127, 1986 CRUSTAL DEFORMATION BEFORE STRONG EARTHQUAKES AND UTILITY OF EARTHQUAKE PREDICTION BY GEODETIC MEASUREMENTS: EXTENDED ABSTRACT Zusheng ZHANG,Shaofen YING, and Yongjian SHEN Geodetic Brigade, State Seismological Bureau, Tianjin, China (Received June 22, 1985) 1. Introduction The preseismic deformation data of 10 major earthquakes that occurred in mainland China are collected and analyzed in this paper. We hope to clarify the characteristics of crustal deformation in the preparatory process of strong earthquakes and to evaluate the monitoring and predicting of earthquakes by available geodetic means. Earthquake studies show that preparatory process and occurrence of a major shallow earthquake always seem to be accompanied by crustal surface deformation. The defor- mation data obtained in large areas by available geodetic means have now been widely used in research on crustal motion and earthquake prediction. Therefore, the characteristics of crustal deformation prior to a large earthquake and whether an earthquake can be correctly detected and predicted with geodetic data, have become a matter of great interest. In this paper, the data of crustal deformation preceding 10 major Chinese earthquakes (Table 1 and Fig. 1; MA et al., 1982) have been collected and checked or reprocessed. We hope to obtain some characteristics of crustal deformation before an earthquake by analyzing these data and making an appropriate evaluation of the role played by crustal deformation measurements in earthquake prediction. 2. Preseismic Deformation 2.1 The Dongchuan earthquakes of February 5 and 13, 1966 (MS=6.5, 6.2) The Dongchuan earthquakes occurred in the Xincun Quaternary graben. On the west margin of the basin, the Xiaojiang deep fault outcrops discontinuously.
    [Show full text]
  • Seismicity Acceleration Model and Its Application to Several Earthquake
    Vol.12 No.1 (35~45) ACTA SEISMOLOGICA SINICA Jan., 1999 Seismicity acceleration model and its applica- tion to several earthquake regions in China WEN-ZHENG YANG (~ ~i~) LI MA (~ ~) Center for Analysis and Prediction, China Seismological Bureau, Beijing i 00036, China Abstract With the theory of subcritical crack growth, we can deduce the fundamental equation of regional seismicity accel- eration model. Applying this model to intraptate earthquake regions, we select three earthquake subplates: North China Subplate, Chuan-Dian Block and Xinjiang Subplate, and divide the three subplates into seven researched regions by the difference of seismicity and tectonic conditions. With the modified equation given by Somette and Sammis (1995), we analysis the seismicity of each region. To those strong earthquakes already occurred in these region, the model can give close fitting of magnitude and occurrence time, and the result in this article indicates that the seismicity acceleration model can also be used for describing the seismicity of intraplate. In the article, we give the magnitude and occurrence time of possible strong earthquakes in Shanxi, Ordos, Bole-Tuokexun, Ayinke- Wuqia earthquake regions. In the same subplate or block, the earthquake periods for each earthquake region are similar in time interval. The constant a in model can be used to describe the intensity of regional seismicity, and for the Chinese Mainland, a is 0.4 generally. To the seismicity in Taiwan and other regions with complex tectonic conditions, the model does not fit well at present. Key words: seismicity acceleration model subcritical crack growth China earthquake region fit Introduction Before the occurrence of a strong regional earthquake, the seismicity in the region is com- monly changed from weak to strong, this phenomenon bespeaks the process of energy release in the region.
    [Show full text]
  • Mitrfaj4fi STRUCTURE and TECTONICS of NORTH CHINA
    STRUCTURE AND TECTONICS OF NORTH CHINA by Kaye M. Shedlock B.S. University of Maryland (1973) M.S. Johns Hopkins University (1978) Submitted to the Department of Earth, Atmospheric, and Planetary Sciences in Partial Fulfillment of the Requirements of the Degree of DOCTOR OF PHILOSOPHY at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY March 1986 O Massachusetts Institute of Technology 1986 Signature of Author Departmdt of-arth, Atmospheric, and Planetary Sciences March 27, 1986 Certified by Keiiti Aki Thesis Supervisor Certified by Peter Molnar Thesis Supervisor Certified by Steven Roecker Thesis Supervisor Accepted byk-/ Irmane entCommittee on Graduate Students MITrFAJ4fi STRUCTURE AND TECTONICS OF NORTH CHINA Kaye M. Shedlock Submitted to the Department of Earth, Atmospheric, and Planetary Sciences in partial fulfillment of the requirement for the degree of Doctor of Philosophy at the Massachusetts Institute of Technology. Abstract North China, a structurally diverse and tectonically active region, emcompasses the Yinchuan, Hetao, Weihe, and North China basins, the Bohai (Bo Sea), the Ordos Plateau, the Shanxi Grabens, and at least six mountain ranges. The North China block, the oldest craton in contempory China, has been the site of active tectonic deformation since Mesozoic times. Remnants of Mesozoic sediments provide hints of older basins, but the contemporary sedimentary basins in North China formed during Cenozoic times. The largest sedimentary basin in North China, the North China (or Bohai) basin, is composed of six major depressions and is surrounded by five mountain ranges. The North China basin is bounded in part by the 2400 km long Tanlu strike-slip fault system. The intraplate rifting that formed the North China basin appears to be the result of an approximately 30% extension and thinning of the lithosphere that began in the early Tertiary.
    [Show full text]
  • Report on the State of the Ecology and Environment in China 2017
    Report on the State of the Ecology and Environment in China 2017 The 2017 Report on the State of the Ecology and Environment in China is hereby announced in accordance with the Environmental Protection Law of the People’s Republic of China. Minister of Ministry of Ecology and Environment, the People’s Republic of China May 22, 2018 2017 Report on the State of the Ecology and Environment in China 目 录 Summary.................................................................................................1 Atmospheric Environment....................................................................8 Freshwater Environment....................................................................18 Marine Environment...........................................................................32 Land Environment...............................................................................36 Natural and Ecological Environment.................................................37 Acoustic Environment.........................................................................41 Radiation Environment.......................................................................43 Climate and Natural Disasters............................................................46 Infrastructure and Energy.................................................................52 Data Sources and Explanations for Assessment ...............................54 1 Report on the State of the Ecology and Environment in China 2017 Summary The year 2017 is a milestone in the development of the
    [Show full text]
  • The 1976 Tangshan, China Earthquake
    NSF /RA-800554 Earthquake Engineering EE Research Institute B:J: P 13 82-1231 75 THE 1976 TANGSHAN, CHINA EARTHQUAKE Papers Presented at the 2nd U.S. National Conference on Earthquake Engineering Held at Stanford University August 22-24, 1979 March 1980 INFORMATION RESOURCES NATIONAL SCIENCE FOUNDATION Published by The Earthquake Engineering Research Institute. a non-profit corporation for the development and dissemination of knowledge on the problems of destructive earthquakes. THE 1976 TANGSHAN, CHINA EARTHQUAKE Papers Presented at the 2nd U.S. National Conference on Earthquake Engineering Held at Stanford University August 22-24, 1979 Introduction by James M. Gere and Haresh C. Shah March 1980 Any opinions, findings, conclusions or recommendations expressed in this publication are those of the author(s) EE and do not necessarily reflect the views of the National Science Foundation. 1t:J: Earthquake Engineering Research Institute with support from \ -Ov The National Science Foundatic;, 50272 -101 3. Recipient's Accession No. REPORT DOCUMENTATION 11. REPORT NO. 12. PAGE NSF /RA-800554 PB82 1 23 1 7 5 1-4-.-T-itl-e -an-d-S-u-bt~it~le 5. Report Date 1976 Tangshan, China Earthquake (Papers Presented at the 2nd U.S. March 1980 National Conference on Earthquake Engineering Held at Stanford ~~_.ft8-------------~ University, August 22-~4, 1979) _ ______ .'L. IUc99065 7. Author(s) B. Performing Organization Rept. No. J .A. Bl ume 9. Performing Organization Name and Address 10. Project/Task/Work Unit No. Earthquake Engineering Research Institute ----- --~~--~~---~-l 2620 Telegraph Avenue 11. Contract(C) Or Grant(G) No. Berkeley, CA 94704 (C) (G) CEE8019240 12.
    [Show full text]
  • Application of Support Vector Machine to Synthetic Earthquake Prediction∗
    Earthq Sci (2009)22: 315−320 315 Doi: 10.1007/s11589-009-0315-8 Application of support vector machine to ∗ synthetic earthquake prediction Chun Jiang 1, Xueli Wei 2 Xiaofeng Cui 1 and Dexiang You 2 1 Earthquake Administration of Tianjin Municipality, Tianjin 300201, China 2 Tianjin University of Technology, Tianjin 300191, China Abstract This paper introduces the method of support vector machine (SVM) into the field of synthetic earthquake pre- diction, which is a non-linear and complex seismogenic system. As an example, we apply this method to predict the largest annual magnitude for the North China area (30°E−42°E, 108°N−125°N) and the capital region (38°E−41.5°E, 114°N−120°N) on the basis of seismicity parameters and observed precursory data. The corresponding prediction rates for the North China area and the capital region are 64.1% and 75%, respectively, which shows that the method is feasible. Key words: support vector machine; seismicity parameter; precursory data; synthetic earthquake prediction CLC number: P315.75 Document code: A 1 Introduction 2 Basic principle of SVM regression The support vector machine (SVM) has not only a algorithm (Vapnik, 1995) strict theory basis, but also a strong generalization (pre- It is firstly to consider the linear regression model. diction) capacity, which can better solve the practical Suppose that a sample with k cases is described as problems of small samples, nonlinearity, higher dimen- ∈ n × sion and local minimum point. At present, SVM are (x1, y1), (x2 , y2 ),L,(xk , yk ) R R, (1) widely used in text classification, handwriting recogni- the linear discriminant function is tion, image classification and bioinformatics, and now it = • + (2) has been extended to synthetic estimates and prediction f (x) w x b of time series.
    [Show full text]
  • Thirsty China Its Key Resource Constraint Is Water
    Thirsty China Its key resource constraint is water Investment strategy Summer 2006 Thirsty China Contents This report was written by Executive summary........................................................................... 3 Andy Rothman, CLSA’s China Macro Strategist [email protected] . nor any drop to drink .................................................................. 4 (8621) 23066000 Julia Zhu provided research assistance Paying a hefty price .........................................................................16 Solutions to avert a crisis ................................................................23 Impact on the investment environment ...........................................36 Water and other key commodities This January 2006 Blue Book examined Our spring 2005 investment strategy the global freshwater scarcity problem detailed China’s appetite for commodities 2 [email protected] Summer 2006 Executive summary Thirsty China Thirsty China We all know that a lack of natural resources constrains China’s ability to continue its blistering pace of growth. But, with our focus on globally traded commodities, we have neglected the biggest resource constraint: water. China has about one-quarter of the average per capita water supply in the world, and 136 of its cities face severe water shortages. The problem is especially bad in the north of the country, where the 43 million people of Beijing, Tianjin and Shanghai each have access to less water per capita than residents of Yemen, Jordan or Israel. Intense levels of pollution aggravate the shortage. More than half of China’s rivers are too polluted to serve as a source for drinking water, and one- quarter are so polluted they cannot even be used for industry or irrigation. The impact is clear. Water shortages have reduced industrial output by an average of US$25 billion per year, and agricultural output by US$19bn, according to the Chinese government.
    [Show full text]