Data Science – Analytics and Applications: Proceedings of the 1st International Data Science Conference – iDSC2017 (German and English Edition)

The iDSC Proceedings reports on state-of-the-art results in Data Science research, development and business. Topics and content of the IDSC2017 proceedings are • Reasoning and Predictive Analytics • Data Analytics in Community Networks • Data Analytics through Sentiment Analysis • User/Customer-centric Data Analytics • Data Analytics in Industrial Application Scenarios Advances in technology and changes in the business and social environment have led to an increasing flood of data, fueling both the need and the desire to generate value from these assets. The emerging field of Data Science is poised to deliver theoretical and practical solutions to the pressing issues of data-driven applications. The 1st International Data Science Conference (iDSC2017 / http://www.idsc.at) organized by Salzburg University of Applied Sciences in cooperation with Information Professionals GmbH, established a new key Data Science event, by pro viding a forum for the international exchange of Data Science technologies and applications. Preface 5 Future TDM Symposium Recap 6 Organisation 9 Sponsors of the conference 11 Table of Content 12 German Abstracts 14 I Full Papers – Double Blind Reviewed 21 Reasoning and Predictive Analytics 22 1 Circadian Cycles and Work Under Pressure: A Stochastic Process Model for E-learningPopulation Dynamics 23 I. INTRODUCTION 23 II. MODEL DEFINITION 24 III. METHODOLOGY 26 IV. RESULTS 27 V. CONCLUSIONS 28 REFERENCES 28 2 Investigating and Forecasting User Activities in Newsblogs: A Study of Seasonality, Volatility andAttention Burst 29 I. INTRODUCTION 29 II. RELATED WORK 29 III. DATA SET DESCRIPTION 30 IV. EMPIRICAL OBSERVATIONS 30 V. TIME SERIES ANALYSIS 31 VI. MAIN RESULTS 32 VII.CONCLUSIONS AND FUTURE WORK 33 REFERENCES 34 VIII. APPENDIX 34 3 Knowledge-Based Short-Term Load-Forecasting forMaritime Container Terminals 35 I. INTRODUCTION 35 II. DAILY OPERATIONS AND POWER CONSUMPTION AT AMARITIME CONTAINER TERMINAL 36 III. CASE-BASED REASONING APPROACH 37 IV. ARTIFICIAL NEURAL NETWORK APPROACH 38 V. EVALUATION 39 VI. CONCLUSION & FURTHER WORK 40 REFERENCES 40 II Data Analytics in Community Networks 41 4 Beyond Spectral Clustering: A Comparative Studyof Community Detection for Document Clustering 42 I. INTRODUCTION 42 II. METHODOLOGY 42 III. EQUIVALENCE TO NCUT 44 IV. RESULTS 45 V. CONCLUSIONS 46 REFERENCES 47 5 Third Party Effect: Community Based Spreading inComplex Networks 48 I. INTRODUCTION 48 II. A COMMUNITY BASED SIR MODEL 49 III. ANALYZING A SYNTHETIC NETWORK 49 IV. ANALYZING REAL NETWORKS 51 V. CONCLUSIONS AND FUTURE WORK 52 REFERENCES 52 6 Cosine Approximate Nearest Neighbors 53 I. INTRODUCTION 53 II. PROBLEM STATEMENT 54 III. CONSTRUCTING THE SIMILARITY GRAPH 54 IV. EXPERIMENT SETUP 55 V. RESULTS & DISCUSSION 56 VI. CONCLUSION 58 REFERENCES 58 III Data Analytics through Sentiment Analysis 59 7 Information Extraction Engine for Sentiment-TopicMatching in Product Intelligence Applications 60 I. INTRODUCTION 60 II. RELATED WORK 60 III. ARTICLE AND INFORMATION EXTRACTION ENGINE 61 IV. PRODUCT INTELLIGENCE WITH ARIE 62 V. EVALUATION 63 VI. CONCLUSION 64 ACKNOWLEDGMENT 64 REFERENCES 64 8 Towards German Word Embeddings: A Use Casewith Predictive Sentiment Analysis 65 I. INTRODUCTION 65 II. CONTRIBUTION 65 III. RELATED WORK 65 IV. DATA 66 V. EXPERIMENTS 66 VI. CONCLUSION AND FUTURE WORK 68 REFERENCES 68 IV User/Customer-centric Data Analytics 69 9 Feature Extraction and Large Activity-SetRecognition Using Mobile Phone Sensors 70 I. INTRODUCTION 70 II. MOBILE PHONE BASED ACTIVITY RECOGNITION: OVERVIEW 70 III. ACTIVITY RECOGNITION PROCESS 71 IV. EXPERIMENTAL REULTS 74 V. CONCLUSIONS AND FUTURE WORK 75 REFERENCES 75 10 The Choice of Metric for Clustering of ElectricalPower Distribution Consumers 76 I. INTRODUCTION 76 II. RELATED WORK 77 III. LOAD TYPE CREATION ALGORITHM 77 IV. ANALYZED SIMILARITY MEASURES 78 V. EXPERIMENTAL RESULTS 79 VI. CONCLUSION 80 REFERENCES 81 11 Evolution of the Bitcoin Address Graph 82 I. INTRODUCTION 82 II. RELATED WORK 83 III. BITCOIN 83 IV. ANALYSIS 84 V. CONCLUSION 87 VI. ACKNOWLEDGEMENT 87 REFERENCES 87 V Data Analytics in Industrial Application Scenarios 88 12 A Reference Architecture for Quality Improvement inSteel Production 89 I. INTRODUCTION 89 II. RELATED WORK 90 III. DOMAIN OF STEEL MANUFACTURING 90 IV. ARCHITECTURE 91 V. CONCLUSION AND FUTUREWORK 94 REFERENCES 94 13 Anomaly Detection and Structural Analysis inIndustrial Production Environments 95 I. INTRODUCTION 95 II. RELATED WORK 95 III. METHOD 96 IV.PROCESSMODEL& IMPLEMENTATION 98 V. CONCLUSION AND FUTURE WORK 98 REFERENCES 99 14 Semantically Annotated Manufacturing Data to support Decision Making in Industry 4.0:A Use-Case Driven Approach 100 I. INTRODUCTION 100 II. MODELLING, STORING AND PUBLISHINGSEMANTICALLY ANNOTATED MANUFACTURING DATA 101 III. ANALYSIS OF MANUFACTURING DATA BASED ON SEMANTICALLY ANNOTATED DATA: USE-CASES 103 IV. CONCLUSION AND DISCUSSION 105 V. REFERENCES 105 VI Short Papers and Student Contributions 106 15 Improving Maintenance Processes with Data Science 107 I. A FULLY INTEGRATED MAINTENANCE CYCLE 107 II. THE STATE OF DATA IN INDUSTRY 108 III. DETECTION OF MAINTENANCE DEMAND 108 IV. OUTLOOK 109 REFERENCES 109 16 ouRfrrame 110 I. INTRODUCTION 110 II. DESCRIPTION OF THE TOOL 110 III. APPLICATION EXAMPLE 111 IV. OUTLOOK 111 ACKNOWLEDGMENT 111 REFERENCES 111 17 Sentiment Analysis 112 I. INTRODUCTION 112 II. DEFINITION OF SENTIMENT ANALYSIS 112 III. BASIC THEORY 112 IV. PROBABILISTIC CLASSIFIERS 113 V. TOOLKIT AND LIBRARIES 113 VI. ALTERNATIVES FOR NON-CODERS 113 VII. CONCLUSION 113 REFERENCES 113 The iDSC Proceedings reports on state-of-the-art results in Data Science research, development and business. Topics and content of the IDSC2017 proceedings are " Reasoning and Predictive Analytics " Data Analytics in Community Networks " Data Analytics through Sentiment Analysis " User/Customer-centric Data Analytics " Data Analytics in Industrial Application Scenarios Advances in technology and changes in the business and social environment have led to an increasing flood of data, fueling both the need and the desire to generate value from these assets. The emerging field of Data Science is poised to deliver theoretical and practical solutions to the pressing issues of data-driven applications. The 1st International Data Science Conference (iDSC2017 / http://www.idsc.at) organized by Salzburg University of Applied Sciences in cooperation with Information Professionals GmbH, established a new key Data Science event, by providi ng a forum for the international exchange of Data Science technologies and applications. Editors FH-Ass. Prof. DI (FH) DI Peter Haber is professor for analog and digital signal processing and responsible coordinator of system theory and electrical engineering at the University of Applied Sciences Salzburg. He has been working as a researcher and project manager in national and international research projects; managing and coordinating both national and international projects, such as the two Leonardo da Vinci EU projects POOL and P2B. He is a member of the international advisory boards for the ICERI, INTEND and EDULEARN conferences since 2009. DI (FH) Dr. Thomas Lampoltshammer, M.A. MSc is currently working as a Senior Scientist at the Department for E-Governance and Administration at Danube University Krems/Austria in the fields of Security Studies and associated societal challenges. His research experience covers nation al and EU-fund ed proj ects in ICT-related topics, such as Geoinformatics, Semantics, Social Media, Legal Informatics, and Open Data. He has a strong background in the design and implementation of expert and decision-making systems, data analytics, as well as semantic-based reasoning. FH-Prof. Ing. MMag. Dr. Manfred Mayr is professor for IT management and department head at Salzburg University of Applied Sciences. He is a lecturer at international conferences and author of various publications in the field of IT-management and business informatics. He was the responsible coordinator and project manager of the two international "EU-Leonardo Da Vinci research projects" "POOL Project Organization OnLine" and Pool2Business (P2B) and various other national projects. The editors are the conference chairs of the International Data Science Conference Front Matter ....Pages I-XXIII Front Matter ....Pages 9-10 Circadian Cycles and Work Under Pressure: A Stochastic Process Model for E-learning Population Dynamics (Christian Backhage, César Ojeda, Rafet Sifa)....Pages 13-18 Investigating and Forecasting User Activities in Newsblogs: A Study of Seasonality, Volatility and Attention Burst (Christian Bauckhage, César Ojeda, Rafet Sifa)....Pages 19-24 Knowledge-Based Short-Term Load-Forecasting for Maritime Container Terminals (Norman Ihle, Axel Hahn)....Pages 25-30 Front Matter ....Pages 31-31 Beyond Spectral Clustering: A Comparative Study of Community Detection for Document Clustering (Christian Backhage, Kostadin Cvejoski, César Ojeda, Rafet Sifa)....Pages 33-38 Third Party Effect: Community Based Spreading in Complex Networks (Christian Bauckhage, César Ojeda, Rafet Sifa, Shubham Agarwal)....Pages 39-43 Cosine Approximate Nearest Neighbors (David C. Anastasiu)....Pages 45-50 Front Matter ....Pages 51-51 Information Extraction Engine for Sentiment-Topic Matching in Product Intelligence Applications (Cornelia Ferner, Werner Pomwenger, Stefan Wegenkittl, Martin Schnöll, Veronika Haaf, Arnold Keller)....Pages 53-57 Towards German Word Embeddings: A Use Case with Predictive Sentiment Analysis (Eduardo Brito, Rafet Sifa, Kostadin Cvejoski, César Ojeda, Christian Bauckhage)....Pages 59-62 Front Matter ....Pages 63-63 Feature Extraction and Large Activity-Set Recognition Using Mobile Phone Sensors (Wassim El Hajj, Ghassen Ben Brahim, Cynthia El-Hayek, Hazem Hajj)....Pages 65-70 The Choice of Metric for Clustering of Electrical Power Distribution Consumers (Nikola Obrenović, Goran Vidaković, Ivan Luković)....Pages 71-76 Evolution of the Bitcoin Address Graph (Erwin Filtz, Axel Polleres, Roman Karl, Bernhard Haslhofer)....Pages 77-82 Front Matter ....Pages 83-83 A Reference Architecture for Quality Improvement in Steel Production (David Arnu, Edwin Yaqub, Claudio Mocci, Valentina Colla, Marcus Neuer, Gabriel Fricout et al.)....Pages 85-90 Anomaly Detection and Structural Analysis in Industrial Production Environments (Martin Atzmueller, David Arnu, Andreas Schmidt)....Pages 91-95 Semantically Annotated Manufacturing Data to support Decision Making in Industry 4.0: A Use-Case Driven Approach (Stefan Schabus, Johannes Scholz)....Pages 97-102 Front Matter ....Pages 103-103 Improving Maintenance Processes with Data Science (Dorian Prill, Simon Kranzer, Robert Merz)....Pages 105-107 ouRframe (Marco Gruber, Elisabeth Birnbacher, Tobias Fellner)....Pages 109-110 Sentiment Analysis (Hofer Dominik)....Pages 111-112