Computational Risk Management David L. Olson Georg Lauhoff Descriptive Data Mining Second Edition Computational Risk Management Editors-in-Chief Desheng Dash Wu, RiskLab, University of Toronto, Toronto, ON, Canada David L. Olson, Department of Supply Chain Management and Analytics, University of Nebraska-Lincoln, Lincoln, NE, USA John Birge, University of Chicago Booth School of Business, Chicago, IL, USA Risksexistineveryaspectofourlivesandriskmanagementhasalwaysbeenavital topic.Mostcomputationaltechniquesandtoolshavebeenusedforoptimizingrisk management and the risk management tools benefit from computational approaches. Computational intelligence models such as neural networks and support vector machines have been widely used for early warning of company bankruptcy and credit risk rating. Operational research approaches such as VaR (valueatrisk)optimizationhavebeenstandardizedinmanagingmarketsandcredit risk, agent-based theories are employed in supply chain risk management and varioussimulationtechniquesareemployedbyresearchersworkingonproblemsof environmental risk management and disaster risk management. Investigation of computational tools in risk management is beneficial to both practitioners and researchers.The Computational Risk Managementseries isahigh-qualityresearch book series with an emphasis on computational aspects of risk management and analysis.Inthisseries,researchmonographsaswellasconferenceproceedingsare published. More information about this series at http://www.springer.com/series/8827 David L. Olson Georg Lauhoff (cid:129) Descriptive Data Mining Second Edition 123 DavidL. Olson Georg Lauhoff Collegeof Business SanJose, CA,USA University of Nebraska–Lincoln Lincoln, NE,USA ISSN 2191-1436 ISSN 2191-1444 (electronic) Computational Risk Management ISBN978-981-13-7180-6 ISBN978-981-13-7181-3 (eBook) https://doi.org/10.1007/978-981-13-7181-3 LibraryofCongressControlNumber:2019934798 ©SpringerNatureSingaporePteLtd.2017,2019 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission orinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar methodologynowknownorhereafterdeveloped. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfrom therelevantprotectivelawsandregulationsandthereforefreeforgeneraluse. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained hereinorforanyerrorsoromissionsthatmayhavebeenmade.Thepublisherremainsneutralwithregard tojurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations. ThisSpringerimprintispublishedbytheregisteredcompanySpringerNatureSingaporePteLtd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore Preface Knowledge management involves the application of human knowledge (episte- mology)withthetechnologicaladvancesofourcurrentsociety(computersystems) andbigdata,bothintermsofcollectingdataandinanalyzingit.Weseethreetypes of analytic tools.Descriptive analytics focus on the reports ofwhat has happened. Predictive analytics extend statistical and/or artificial intelligence to provide forecastingcapability.Italsoincludesclassificationmodeling.Diagnosticanalytics can apply analysis to sensor input to direct control systems automatically. Prescriptiveanalyticsappliesquantitativemodelstooptimizesystems,oratleastto identify improved systems. Data mining includes descriptive and predictive mod- eling. Operations research includes all the three. This book focuses on descriptive analytics. Lincoln, USA David L. Olson San Jose, USA Georg Lauhoff v Book Concept The book seeks to provide simple explanations and demonstration of some descriptive tools. This second edition provides more examples of big data impact, updatesthecontentonvisualization,clarifiessomepoints,andexpandscoverageof associationrulesandclusteranalysis.Chapter1givesanoverviewofthecontextof knowledge management. Chapter 2 discusses some basic software support to data visualization.Chapter3coversfundamentalsofmarketbasketanalysis,andChap.4 provides a demonstration of RFM modeling, a basic marketing data mining tool. Chapter 5 demonstrates association rule mining. Chapter 6 has more in-depth cov- erage ofclusteranalysis. Chapter 7 discusses link analysis. Models are demonstrated using business-related data. The style of the book is intended to be descriptive, seeking to explain how methods work, with some citations, but without deep scholarly references. The data sets and software are all selected for widespread availability and access by any reader with computer links. vii Contents 1 Knowledge Management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Computer Support Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Examples of Knowledge Management . . . . . . . . . . . . . . . . . . . . . . . . . 4 Data Mining Descriptive Applications . . . . . . . . . . . . . . . . . . . . . . . . . 7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2 Data Visualization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Data Visualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 R Software. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Loan Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Energy Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Basic Visualization of Time Series . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 3 Market Basket Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Co-occurrence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Demonstration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Fit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Profit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Lift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Market Basket Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4 Recency Frequency and Monetary Analysis . . . . . . . . . . . . . . . . . . . 45 Dataset 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Balancing Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Lift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Value Function. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 ix x Contents Data Mining Classification Models . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Logistic Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Decision Tree. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Neural Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Dataset 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 5 Association Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 The Apriori Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Association Rules from Software. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 Non-negative Matric Factorization. . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 6 Cluster Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 K-Means Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 A Clustering Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 Loan Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Clustering Methods Used in Software . . . . . . . . . . . . . . . . . . . . . . . . . 81 Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 R (Rattle) K-Means Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Other R Clustering Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 KNIME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 WEKA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 7 Link Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Link Analysis Terms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Basic Network Graphics with NodeXL . . . . . . . . . . . . . . . . . . . . . . . . 114 Network Analysis of Facebook Network or Other Networks . . . . . . . . . 118 Link Analysis of Your Emails. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Link Analysis Application with PolyAnalyst (Olson and Shi 2007) . . . . 125 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 8 Descriptive Data Mining. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 About the Authors DavidL.Olson istheJames&H.K.StuartChancellor’sDistinguishedChairand FullProfessorattheUniversityofNebraska.Hehaspublishedresearchinover150 refereed journal articles, primarily on the topic of multiple objective decision- making, information technology, supply chain risk management, and data mining. He teaches in the management information systems, management science, and operationsmanagementareas.Hehasauthoredover20books.HeisMemberofthe DecisionSciencesInstitute,theInstituteforOperationsResearchandManagement Sciences, and the Multiple Criteria Decision Making Society. He was a Lowry Mays endowed Professor at Texas A&M University from 1999 to 2001. He was named the Raymond E. Miles Distinguished Scholar award for 2002, and was a James C. and Rhonda Seacrest Fellow from 2005 to 2006. He was named Best Enterprise Information Systems Educator by IFIP in 2006. He is a Fellow of the Decision Sciences Institute. Georg Lauhoff is Technologist at Western Digital Corporation and carries out R&D in materials science and its application in data storage devices and uses the techniquesdescribed inthisbookforhiswork.Heco-authored38refereed journal articles and over 30 conference presentations, primarily on the topic of materials science, data storage materials, and magnetic thin films. He was awarded schol- arshipsandresearchgrantsintheUKandJapan.HewastheClerkMaxwellScholar from 1995 to 1998 and is a Fellow of the Cambridge Philosophical Society. He studiedphysicsatAachen(Diplom)andCambridgeUniversity(MasterandPh.D.) specializing in the field of materials science and magnetic thin films and sensors. After graduating, he moved to Japan and held a faculty position in Materials ScienceandEngineeringattheToyotaTechnologicalInstituteandthencarriedout research in the sequencing of DNA using magnetic sensors at Cambridge University before moving in 2005 to the recording industry in the Bay area. xi