DATA ANALYSIS. The Ins and Outs of Solving Real Problems

Transcription

1 DATA ANALYSIS The Ins and Outs of Solving Real Problems

2 Competitive Methods in Operations Research and Data Analysis Series Editor: J. Janssen, Free University of Brussels, Brussels, Belgium Editorial Board: S. Osaki, Hiroshima University, Hiroshima, Japan R. L. Disney, Virginia Polytechnic Institute and State University, Blacksburg, Virginia DATA ANALYSIS: The Ins and Outs of Solving Real Problems Edited by J. Janssen, F. Marcotorchino, and J. M. Proth

3 DATA ANALYSIS The Ins and Outs of Solving Real Problems Edited by J.Janssen Free University of Brussels Brussels, Belgium F. Ma,coto,chino IBM Scientific Center Paris, France and J. M. P,oth I. N.R.I.A. Le Chesnay, France SPRINGER SCIENCE+BUSINESS MEDIA, LLC

4 Library of Congress Cataloging in Publication Data International Symposium on Data Analysis, the Ins and Outs of Solving Real Problems (4th: 1985: Brussels, Belgium) Data analysis. (Competitive methods of operations research and data analysis) Includes bibliographical references and index. 1. Multivariate analysis-data processing -Congresses. 2. Factor analysis-data processing-congresses. 1. Janssen, Jacques, II. Proth, Jean-Marie, III. Title. IV. Series. 0A278.J ' ISBN ISBN (ebook) DOI / Proceedings of the Fourth International Symposium on Data Analysis: The Ins and Outs of Solving Real Problems, held June 10-12, 1985, in Brussels, Belgium 1987 Springer Science+Business Media New York Originally published by Plenum Press, New York in 1987 Softcover reprint of the hardcover 1 st edition 1987 AII rights reserved No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or' otherwise, without written permission from the Publisher

5 PREFACE This book is the result of the fourth International Symposium on Data Analysis held on June 1985 at the Universite Libre de Bruxelles with the help'of the European Institute for Advanced Management. As the preceding ones, the organization of the Symposium started with a call for real life problems from which an International Committee selected six topics and asked for several solutions. These topics are : I) Multivariate and longitudinal data on growing children 2) Prehistoric assemblages and lithic artifacts from a small Westeuropean area 3) A comparison of results of European elections 4) Classification of heterogeneous data related to microcomputers 5) Group technology in production management 6) Juvenile gelinquency They are covered by the S1X chapters of this book in the following systematic way : a) firstly, a presentation of the problem is given in the original context of the relevant discipline (Medicine, archaelogy, politics, marketing, production and education); b) Secondly, we present the solution found by people who presents the problem; c) thirdly, we find the other retained solutions among the most significative ones; v

6 vi PREFACE d) finally, a short conclusion compares the different approaches. The diversity of the six selected problems clearly shows that Data Analysis can be used for solving a wide variety of problems. Moreover, the fact that each problem is approached by several different way - at least two - also shows that, in general, a "universal" statistical method does not exist. This is renforced by the strong complementary of proposed solutions, aside the fact that any "good" or judicious approach must detect the basis of the internal structure of the data. Finally let us mention that the strong interest from people both from University and Industry for the Symposium stimulates the Editors to reorganize, in an enlarged version in the topics not only in Data Analysis but also in Stochastic Modelling, such a meeting in the future. Finally, the Editors should like to mention the strong help given by IBM-France and the CADEPS (Centre d'analyse des Donnees et Processus Stochastiques, Ecole de Commerce SOLVAY, Brussels) for taking in hands the local organization and in particular the one of Mr P.CULLUS (CADEPS) for the help with final preparation of this volume. Jacques Janssen F. Marcotorchino J. M. Proth

7 CONTENTS CHAPTER 1. }illltivariate AND LONGITUDINAL DATA ON GROWING CHILDREN Section 1.1. Presentation of the French auxological survey.. 3 M.Sempe and Groupe d'auxologie Medico-Sociale Section 1.2. Introduction to the proposed solutions P.M.Kroonenberg Section 1.3. Solution using STATIS Ch.Lavit and M.-O.Pernin Section 1.4. Solution using SPECTRAMAP P.J.Lewi, G.Calomme and J.Van Hoof Section 1.5. Solution using LONG I J.Pontier and M.-O.Pernin Section 1.6. Solution using clustering method A.Mineo Section 1.7. Solutions using a three-mode principal component analysis and some comparison results with the other approaches P.M.Kroonenberg CHAPTER 2. PREHISTORIC ASSEMBLAGES AND LITHIC ARTIFACTS FROM A SMALL WEST-EUROPEAN AREA Section 2. I. Problems of chronological seriation in the epipaleoli thic ("meso li thi c") J.G.Rozoy vii

8 viii CONTENTS Section 2.2. A seriation through correspondence analysis with construction of a continuous gradient J.L.S1achmuy1der Section 2.3. Analysis of contingency data using SPECTRAMAP 153 P.J.Lewi, J.Van Hoof and G.Ca1omme Section 2.4. Conclusion J.G.Rozoy 169 CHAPTER 3. A COMPARISON OF RESULTS OF EUROPEAN ELECTIONS Section 3.1. Presentation of the problem J.Janssen, P.Cu11us and P.Van Brusse1 173 Section 3.2. The comparative study of results of elections 177 P.Cu11us Section 3.3. Similarity structure analysis of european elections... L.Guttman and S.Levy Section 3.4. Conclusion P.Van Brussel CHAPTER 4. CLASSIFICATION OF HETEROGENEOUS DATA RELATED TO MICROCOMPUTERS Section 4.1. Presentation of the problem S.Chah 209 Section 4.2. Tripled comparison ~n automatic classification. 213 S.Chah Section 4.3. Us~ng classical tools of data analysis to solve the problem P.Van Brusse1 Section 4.4. Classification of microcomputers reproduced by SPECTRAMAP P.J.Lewi, A.Quets and K.Van Reuse1 Section 4.5. Partial-order classification of microcomputers 255 S.Levy and L.Guttman

9 CONTENTS Section 4.6. Conclusion P.Cullus 269 ix CHAPTER 5. GROUP TECHNOLOGY IN PRODUCTION MANAGEMENT Section 5. I. The short horizon planning level-generalized issue.. H.Garcia and J.M.Proth Section 5.2. Spatial organization of machine tools in a workshop G.Caraux Section 5.3. Conclusions J.M.Proth CHAPTER 6. JUVENILE DELINQUENCY Section 6. I. Backgrounds of delinquent behavior M.Junger and J.Junger-Tas Section 6.2. Social integration or labelling M.Junger Section 6.3. Establishing causal order among variables in delinquency research failure of panel analysis 331 J.A.Hagenaars Section 6.4. Using LISREL to choose between contradictory criminological theories P.G.M.van der Heijden, F.Meijerink and A.Mooijaart Section 6.5. Some conclusions M.Junger, P.G.M.van der Heijden and J.A.Hagenaars KEYWORDS. 375 INDEX.. 377