Future Data and Security Engineering 5th

International Conference FDSE 2018 Ho Chi Minh

City Vietnam November 28 30 2018 Proceedings
Tran Khanh Dang pdf download

https://textbookfull.com/product/future-data-and-security-engineering-5th-international-conference-
fdse-2018-ho-chi-minh-city-vietnam-november-28-30-2018-proceedings-tran-khanh-dang/

★★★★★ 4.7/5.0 (33 reviews) ✓ 138 downloads ■ TOP RATED

"Perfect download, no issues at all. Highly recommend!" - Mike D.

DOWNLOAD EBOOK
 Future Data and Security Engineering 5th International
Conference FDSE 2018 Ho Chi Minh City Vietnam November 28 30
2018 Proceedings Tran Khanh Dang pdf download

TEXTBOOK EBOOK TEXTBOOK FULL

Available Formats

■ PDF eBook Study Guide TextBook

EXCLUSIVE 2025 EDUCATIONAL COLLECTION - LIMITED TIME

INSTANT DOWNLOAD VIEW LIBRARY

Collection Highlights

Future Data and Security Engineering First International

Conference FDSE 2014 Ho Chi Minh City Vietnam November 19
21 2014 Proceedings 1st Edition Tran Khanh Dang

Future Data and Security Engineering Third International

Conference FDSE 2016 Can Tho City Vietnam November 23 25
2016 Proceedings 1st Edition Tran Khanh Dang

Computational Data and Social Networks 8th International

Conference CSoNet 2019 Ho Chi Minh City Vietnam November
18 20 2019 Proceedings Andrea Tagarelli

User Science and Engineering 5th International Conference

i USEr 2018 Puchong Malaysia August 28 30 2018 Proceedings
Natrah Abdullah
Industrial Networks and Intelligent Systems 3rd
International Conference INISCOM 2017 Ho Chi Minh City
Vietnam September 4 2017 Proceedings 1st Edition Yuanfang
Chen

Nature of Computation and Communication International

Conference ICTCC 2014 Ho Chi Minh City Vietnam November 24
25 2014 Revised Selected Papers 1st Edition Phan Cong Vinh

Product Focused Software Process Improvement 19th

International Conference PROFES 2018 Wolfsburg Germany
November 28 30 2018 Proceedings Marco Kuhrmann

Secure IT Systems 23rd Nordic Conference NordSec 2018 Oslo

Norway November 28 30 2018 Proceedings Nils Gruschka

Data Mining 16th Australasian Conference AusDM 2018

Bahrurst NSW Australia November 28 30 2018 Revised
Selected Papers Rafiqul Islam
 Tran Khanh Dang · Josef Küng
Roland Wagner · Nam Thoai
Makoto Takizawa (Eds.)
LNCS 11251

Future Data and

Security Engineering
5th International Conference, FDSE 2018
Ho Chi Minh City, Vietnam, November 28–30, 2018
Proceedings

123
Lecture Notes in Computer Science 11251
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen

Editorial Board
David Hutchison
Lancaster University, Lancaster, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Friedemann Mattern
ETH Zurich, Zurich, Switzerland
John C. Mitchell
Stanford University, Stanford, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
C. Pandu Rangan
Indian Institute of Technology Madras, Chennai, India
Bernhard Steffen
TU Dortmund University, Dortmund, Germany
Demetri Terzopoulos
University of California, Los Angeles, CA, USA
Doug Tygar
University of California, Berkeley, CA, USA
Gerhard Weikum
Max Planck Institute for Informatics, Saarbrücken, Germany
More information about this series at http://www.springer.com/series/7409
Tran Khanh Dang Josef Küng
•

Roland Wagner Nam Thoai

•

Makoto Takizawa (Eds.)

Future Data and

Security Engineering
5th International Conference, FDSE 2018
Ho Chi Minh City, Vietnam, November 28–30, 2018
Proceedings

123
Editors
Tran Khanh Dang Nam Thoai
Ho Chi Minh City University of Technology Ho Chi Minh City University of Technology
Ho Chi Minh, Vietnam Ho Chi Minh, Vietnam
Josef Küng Makoto Takizawa
Johannes Kepler University of Linz Hosei University
Linz, Austria Tokyo, Japan
Roland Wagner
Johannes Kepler University of Linz
Linz, Austria

ISSN 0302-9743 ISSN 1611-3349 (electronic)

Lecture Notes in Computer Science
ISBN 978-3-030-03191-6 ISBN 978-3-030-03192-3 (eBook)
https://doi.org/10.1007/978-3-030-03192-3

Library of Congress Control Number: 2018959232

LNCS Sublibrary: SL3 – Information Systems and Applications, incl. Internet/Web, and HCI

© Springer Nature Switzerland AG 2018

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part 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 or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
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, express or implied, with respect to the material contained herein or for any errors or
omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
 Preface

In this volume we present the accepted contributions for the 5th International Con-
ference on Future Data and Security Engineering (FDSE 2018). The conference took
place during November 28–30, 2018, in Ho Chi Minh City, Vietnam, at HCMC
University of Technology, among the most famous and prestigious universities in
Vietnam. The proceedings of FDSE are published in the LNCS series by Springer.
Besides DBLP and other major indexing systems, FDSE proceedings have also been
indexed by Scopus and listed in Conference Proceeding Citation Index (CPCI) of
Thomson Reuters.
The annual FDSE conference is a premier forum designed for researchers, scientists,
and practitioners interested in state-of-the-art and state-of-the-practice activities in data,
information, knowledge, and security engineering to explore cutting-edge ideas, to
present and exchange their research results and advanced data-intensive applications, as
well as to discuss emerging issues in data, information, knowledge, and security
engineering. At the annual FDSE, the researchers and practitioners are not only able to
share research solutions to problems in today’s data and security engineering themes,
but also able to identify new issues and directions for future related research and
development work.
The call for papers resulted in the submission of 122 papers. A rigorous and
peer-review process was applied to all of them. This resulted in 35 accepted papers
(including seven short papers, acceptance rate: 28.69%) and two keynote speeches,
which were presented at the conference. Every paper was reviewed by at least three
members of the international Program Committee, who were carefully chosen based on
their knowledge and competence. This careful process resulted in the high quality
of the contributions published in this volume. The accepted papers were grouped into
the following sessions:
– Security and privacy engineering
– Authentication and access control
– Big data analytics and applications
– Advanced studies in machine learning
– Deep learning and applications
– Data analytics and recommendation systems
– Internet of Things and applications
– Smart city: data analytics and security
– Emerging data management systems and applications
In addition to the papers selected by the Program Committee, five internationally
recognized scholars delivered keynote speeches: “Freely Combining Partial Knowledge
in Multiple Dimensions,” presented by Prof. Dirk Draheim from Tallinn University of
Technology, Estonia; “Programming Data Analysis Workflows for the Masses,” pre-
sented by Prof. Artur Andrzejak from Heidelberg University, Germany; “Mathematical
VI Preface

Foundations of Machine Learning: A Tutorial,” presented by Prof. Dinh Nho Hao from
Institute of Mathematics, Vietnam Academy of Science and Technology; “4th Industry
Revolution Technologies and Security,” presented by Prof. Tai M. Chung from
Sungkyunkwan University, South Korea; and “Risk-Based Software Quality and
Security Engineering in Data-Intensive Environments,” presented by Prof. Michael
Felderer from University of Innsbruck, Austria.
The success of FDSE 2018 was the result of the efforts of many people, to whom we
would like to express our gratitude. First, we would like to thank all authors who
submitted papers to FDSE 2018, especially the invited speakers for the keynotes and
tutorials. We would also like to thank the members of the committees and external
reviewers for their timely reviewing and lively participation in the subsequent dis-
cussion in order to select such high-quality papers published in this volume. Last but
not least, we thank the Faculty of Computer Science and Engineering, HCMC
University of Technology, for hosting and organizing FDSE 2018.

November 2018 Tran Khanh Dang

Josef Küng
Roland Wagner
Nam Thoai
Makoto Takizawa
 Organization

General Chair
Roland Wagner Johannes Kepler University Linz, Austria

Steering Committee
Elisa Bertino Purdue University, USA
Dirk Draheim Tallinn University of Technology, Estonia
Kazuhiko Hamamoto Tokai University, Japan
Koichiro Ishibashi The University of Electro-Communications, Japan
M-Tahar Kechadi University College Dublin, Ireland
Dieter Kranzlmüller Ludwig Maximilian University, Germany
Fabio Massacci University of Trento, Italy
Clavel Manuel The Madrid Institute for Advanced Studies in Software
Development Technologies, Spain
Atsuko Miyaji Osaka University and Japan Advanced Institute
of Science and Technology, Japan
Erich Neuhold University of Vienna, Austria
Cong Duc Pham University of Pau, France
Silvio Ranise Fondazione Bruno Kessler, Italy
Nam Thoai HCMC University of Technology, Vietnam
A Min Tjoa Technical University of Vienna, Austria
Xiaofang Zhou The University of Queensland, Australia

Program Committee Chairs

Tran Khanh Dang HCMC University of Technology, Vietnam
Josef Küng Johannes Kepler University Linz, Austria
Makoto Takizawa Hosei University, Japan

Publicity Chairs
Nam Ngo-Chan University of Trento, Italy
Quoc Viet Hung Nguyen The University of Queensland, Australia
Huynh Van Quoc Phuong Johannes Kepler University Linz, Austria
Tran Minh Quang HCMC University of Technology, Vietnam
Le Hong Trang HCMC University of Technology, Vietnam
VIII Organization

Local Organizing Committee

Tran Khanh Dang HCMC University of Technology, Vietnam
Tran Tri Dang HCMC University of Technology, Vietnam
Josef Küng Johannes Kepler University Linz, Austria
Nguyen Dinh Thanh Data Security Applied Research Lab, Vietnam
Que Nguyet Tran Thi HCMC University of Technology, Vietnam
Tran Ngoc Thinh HCMC University of Technology, Vietnam
Tuan Anh Truong HCMC University of Technology, Vietnam
and University of Trento, Italy
Quynh Chi Truong HCMC University of Technology, Vietnam
Nguyen Thanh Tung HCMC University of Technology, Vietnam

Finance and Leisure Chairs

Hue Anh La HCMC University of Technology, Vietnam
Hoang Lan Le HCMC University of Technology, Vietnam

Program Committee
Artur Andrzejak Heidelberg University, Germany
Stephane Bressan National University of Singapore, Singapore
Hyunseung Choo Sungkyunkwan University, South Korea
Tai M. Chung Sungkyunkwan University, South Korea
Agostino Cortesi Università Ca’ Foscari Venezia, Italy
Bruno Crispo University of Trento, Italy
Nguyen Tuan Dang University of Information Technology, VNUHCM,
Vietnam
Agnieszka Bialystok University of Technology, Poland
Dardzinska-Glebocka
Tran Cao De Can Tho University, Vietnam
Thanh-Nghi Do Can Tho University, Vietnam
Nguyen Van Doan Japan Advanced Institute of Science and Technology,
Japan
Dirk Draheim Tallinn University of Technology, Estonia
Nguyen Duc Dung HCMC University of Technology, Vietnam
Johann Eder Alpen-Adria University Klagenfurt, Austria
Jungho Eom Daejeon University, South Korea
Verena Geist Software Competence Center Hagenberg, Austria
Raju Halder Indian Institute of Technology Patna, India
Tran Van Hoai HCMC University of Technology, Vietnam
Nguyen Quoc Viet Hung The University of Queensland, Australia
Nguyen Viet Hung Bosch, Germany
Trung-Hieu Huynh Industrial University of Ho Chi Minh City, Vietnam
Tomohiko Igasaki Kumamoto University, Japan
Muhammad Ilyas University of Sargodha, Pakistan
 Organization IX

Hiroshi Ishii Tokai University, Japan

Eiji Kamioka Shibaura Institute of Technology, Japan
Le Duy Khanh Data Storage Institute, Singapore
Surin Kittitornkun King Mongkut’s Institute of Technology Ladkrabang,
Thailand
Andrea Ko Corvinus University of Budapest, Hungary
Duc Anh Le Center for Open Data in the Humanities, Tokyo, Japan
Xia Lin Drexel University, USA
Lam Son Le HCMC University of Technology, Vietnam
Faizal Mahananto Institut Teknologi Sepuluh Nopember, Indonesia
Clavel Manuel The Madrid Institute for Advanced Studies in Software
Development Technologies, Spain
Nadia Metoui University of Trento and FBK-Irist, Trento, Italy
Hoang Duc Minh National Physical Laboratory, UK
Takumi Miyoshi Shibaura Institute of Technology, Japan
Hironori Nakajo Tokyo University of Agriculture and Technology,
Japan
Nguyen Thai-Nghe Cantho University, Vietnam
Thanh Binh Nguyen HCMC University of Technology, Vietnam
Benjamin Nguyen Institut National des Sciences Appliqués Centre Val de
Loire, France
An Khuong Nguyen HCMC University of Technology, Vietnam
Khai Nguyen National Institute of Informatics, Japan
Kien Nguyen National Institute of Information and Communications
Technology, Japan
Khoa Nguyen The Commonwealth Scientific and Industrial Research
Organisation, Australia
Le Duy Lai Nguyen Ho Chi Minh City University of Technology, Vietnam
and University of Grenoble Alpes, France
Do Van Nguyen Institute of Information Technology, MIST, Vietnam
Thien-An Nguyen University College Dublin, Ireland
Phan Trong Nhan HCMC University of Technology, Vietnam
Luong The Nhan University of Pau, France
Alex Norta Tallinn University of Technology, Estonia
Duu-Sheng Ong Multimedia University, Malaysia
Eric Pardede La Trobe University, Australia
Ingrid Pappel Tallinn University of Technology, Estonia
Huynh Van Quoc Phuong Johannes Kepler University Linz, Austria
Nguyen Khang Pham Can Tho University, Vietnam
Phu H. Phung University of Dayton, USA
Nguyen Ho Man Rang Ho Chi Minh City University of Technology, Vietnam
Tran Minh Quang HCMC University of Technology, Vietnam
Akbar Saiful Institute of Technology Bandung, Indonesia
Tran Le Minh Sang WorldQuant LLC, USA
Christin Seifert University of Passau, Germany
Erik Sonnleitner Johannes Kepler University Linz, Austria
X Organization

Tran Phuong Thao KDDI Research, Inc., Japan

Tran Ngoc Thinh HCMC University of Technology, Vietnam
Quan Thanh Tho HCMC University of Technology, Vietnam
Michel Toulouse Vietnamese-German University, Vietnam
Shigenori Tomiyama Tokai University, Japan
Le Hong Trang HCMC University of Technology, Vietnam
Tuan Anh Truong HCMC University of Technology, Vietnam
and University of Trento, Italy
Tran Minh Triet HCMC University of Natural Sciences, Vietnam
Takeshi Tsuchiya Tokyo University of Science, Japan
Osamu Uchida Tokai University, Japan
Hoang Tam Vo IBM Research, Australia
Hoang Huu Viet Vinh University, Vietnam
Edgar Weippl SBA Research, Austria
Wolfram Wöß Johannes Kepler University Linz, Austria
Tetsuyasu Yamada Tokyo University of Science, Japan
Jeff Yan Linköping University, Sweden
Szabó Zoltán Corvinus University of Budapest, Hungary

Additional Reviewers
Pham Quoc Cuong HCMC University of Technology, Vietnam
Kim Tuyen Le Thi HCMC University of Technology, Vietnam
Ai Thao Nguyen Thi Data Security Applied Research Lab, Vietnam
Bao Thu Le Thi National Institute of Informatics, Japan
Tuan Anh Tran HCMC University of Technology, Vietnam
and Chonnam National University, South Korea
Quang Hai Truong HCMC University of Technology, Vietnam
 Contents

Invited Keynotes

Freely Combining Partial Knowledge in Multiple Dimensions

(Extended Abstract). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Dirk Draheim

Risk-based Software Quality and Security Engineering in Data-intensive

Environments (Invited Keynote) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Michael Felderer

Security and Privacy Engineering

A Secure and Efficient kNN Classification Algorithm Using Encrypted

Index Search and Yao’s Garbled Circuit over Encrypted Databases . . . . . . . . 21
Hyeong-Jin Kim, Jae-Hwan Shin, and Jae-Woo Chang

A Security Model for IoT Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Alban Gabillon and Emmanuel Bruno

Comprehensive Study in Preventive Measures of Data Breach Using

Thumb-Sucking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Keinaz Domingo, Bryan Cruz, Froilan De Guzman, Jhinia Cotiangco,
and Chistopher Hilario

Intrusion Prevention Model for WiFi Networks. . . . . . . . . . . . . . . . . . . . . . 66

Julián Francisco Mojica Sánchez, Octavio José Salcedo Parra,
and Alberto Acosta López

Security for the Internet of Things and the Bluetooth Protocol . . . . . . . . . . . 74

Rodrigo Alexander Fagua Arévalo, Octavio José Salcedo Parra,
and Juan Manuel Sánchez Céspedes

Authentication and Access Control

A Light-Weight Tightening Authentication Scheme for the Objects’

Encounters in the Meetings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Kim Khanh Tran, Minh Khue Pham, and Tran Khanh Dang

A Privacy Preserving Authentication Scheme in the Intelligent

Transportation Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Cuong Nguyen Hai Vinh, Anh Truong, and Tai Tran Huu
XII Contents

Big Data Analytics and Applications

Higher Performance IPPC+ Tree for Parallel Incremental Frequent

Itemsets Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Van Quoc Phuong Huynh and Josef Küng

A Sample-Based Algorithm for Visual Assessment of Cluster Tendency

(VAT) with Large Datasets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Le Hong Trang, Pham Van Ngoan, and Nguyen Van Duc

An Efficient Batch Similarity Processing with MapReduce . . . . . . . . . . . . . . 158

Trong Nhan Phan and Tran Khanh Dang

Vietnamese Paraphrase Identification Using Matching Duplicate Phrases

and Similar Words . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
Hoang-Quoc Nguyen-Son, Nam-Phong Tran, Ngoc-Vien Pham,
Minh-Triet Tran, and Isao Echizen

Advanced Studies in Machine Learning

Automatic Hyper-parameters Tuning for Local Support Vector Machines . . . . 185

Thanh-Nghi Do and Minh-Thu Tran-Nguyen

Detection of the Primary User’s Behavior for the Intervention of the

Secondary User Using Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Deisy Dayana Zambrano Soto, Octavio José Salcedo Parra,
and Danilo Alfonso López Sarmiento

Text-dependent Speaker Recognition System Based on Speaking

Frequency Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
Khoa N. Van, Tri P. Minh, Thang N. Son, Minh H. Ly, Tin T. Dang,
and Anh Dinh

Static PE Malware Detection Using Gradient Boosting Decision

Trees Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
Huu-Danh Pham, Tuan Dinh Le, and Thanh Nguyen Vu

Comparative Study on Different Approaches in Optimizing Threshold

for Music Auto-Tagging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Khanh Nguyen Cao Minh, Thinh Dang An, Vu Tran Quang,
and Van Hoai Tran

Using Machine Learning for News Verification . . . . . . . . . . . . . . . . . . . . . 251

Gerardo Ernesto Rolong Agudelo, Octavio José Salcedo Parra,
and Javier Medina
 Contents XIII

Deep Learning and Applications

A Short Review on Deep Learning for Entity Recognition . . . . . . . . . . . . . . 261

Hien T. Nguyen and Thuan Quoc Nguyen

An Analysis of Software Bug Reports Using Random Forest . . . . . . . . . . . . 273

Ha Manh Tran, Sinh Van Nguyen, Synh Viet Uyen Ha,
and Thanh Quoc Le

Motorbike Detection in Urban Environment . . . . . . . . . . . . . . . . . . . . . . . . 286

Chi Kien Huynh, Tran Khanh Dang, and Thanh Sach Le

Data Analytics and Recommendation Systems

Comprehensive Review of Classification Algorithms for Medical

Information System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
Anna Kasperczuk and Agnieszka Dardzinska

New Method of Medical Incomplete Information System Optimization

Based on Action Queries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
Katarzyna Ignatiuk, Agnieszka Dardzinska, Małgorzata Zdrodowska,
and Monika Chorazy

Cloud Media DJ Platform: Functional Perspective . . . . . . . . . . . . . . . . . . . . 323

Joohyun Lee, Jinwoong Jung, Sanggil Yeoum, Junghyun Bum,
Thien-Binh Dang, and Hyunseung Choo

Cloud Media DJ Platform: Performance Perspective . . . . . . . . . . . . . . . . . . 335

Jinwoong Jung, Joohyun Lee, Sanggil Yeoum, Junghyun Bum,
Thien Binh Dang, and Hyunseung Choo

Analyzing and Visualizing Web Server Access Log File . . . . . . . . . . . . . . . 349

Minh-Tri Nguyen, Thanh-Dang Diep, Tran Hoang Vinh,
Takuma Nakajima, and Nam Thoai

Internet of Things and Applications

Lower Bound for Function Computation in Distributed Networks . . . . . . . . . 371

H. K. Dai and M. Toulouse

Teleoperation System for a Four-Dof Robot: Commands

with Data Glove and Web Page . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385
Juan Guillermo Palacio Cano, Octavio José Salcedo Parra,
and Miguel J. Espitia R.
XIV Contents

Design of PHD Solution Based on HL7 and IoT. . . . . . . . . . . . . . . . . . . . . 405

Sabrina Suárez Arrieta, Octavio José Salcedo Parra,
and Roberto Manuel Poveda Chaves

Smart City: Data Analytics and Security

Analysis of Diverse Tourist Information Distributed Across the Internet. . . . . 413

Takeshi Tsuchiya, Hiroo Hirose, Tadashi Miyosawa, Tetsuyasu Yamada,
Hiroaki Sawano, and Keiichi Koyanagi

Improving the Information in Medical Image by Adaptive

Fusion Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
Nguyen Mong Hien, Nguyen Thanh Binh, Ngo Quoc Viet,
and Pham Bao Quoc

Resident Identification in Smart Home by Voice Biometrics . . . . . . . . . . . . . 433

Minh-Son Nguyen and Tu-Lanh Vo

Modeling and Testing Power Consumption Rate of Low-Power Wi-Fi

Sensor Motes for Smart Building Applications . . . . . . . . . . . . . . . . . . . . . . 449
Cao Tien Thanh

Emerging Data Management Systems and Applications

Distributed Genetic Algorithm on Cluster of Intel Xeon Phi Co-processors. . . 463

Nguyen Quang-Hung, Anh-Tu Ngoc Tran, and Nam Thoai

Information Systems Success: Empirical Evidence on Cloud-based ERP . . . . 471

Thanh D. Nguyen and Khiem V. T. Luc

Statistical Models to Automatic Text Summarization . . . . . . . . . . . . . . . . . . 486

Pham Trong Nguyen and Co Ton Minh Dang

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499

Invited Keynotes
 Freely Combining Partial Knowledge
in Multiple Dimensions
(Extended Abstract)

Dirk Draheim(B)

Large-Scale Systems Group, Tallinn University of Technology,

Akadeemia tee 15a, 12618 Tallinn, Estonia
[email protected]

Abstract. F.P. conditionalization (frequentist partial conditionaliza-

tion) allows for combining partial knowledge in arbitrary many dimen-
sions and without any restrictions on events such as independence or
partitioning. In this talk, we provide a primer to F.P. conditionalization
and its most important results. As an example, we proof that Jeffrey
conditionalization is an instance of F.P. conditionalization for the spe-
cial case that events form a partition. Also, we discuss the logics and the
data science perspective on the matter.

Keywords: F.P. conditionalization · Jeffrey conditionalization

Data science · Statistics · Contingency tables · Reasoning systems
SPSS · SAS · R · Phyton/Anaconda · Cognos · Tableau

1 A Primer on F.P. Conditionalization

In [1] we have introduced F.P.conditionalization (frequentist partial condition-
alization), which allows for conditionalization on partially known events. An
F.P. conditionalization P(A | B1 ≡ b1 , . . . , Bm ≡ bm ) is the probability of an
event A that is conditional on a list of event-probability specifications B1 ≡ b1
through Bm ≡ bm . A specification pair B ≡ b12 stands for the assumption that
the probability of B has somehow changed from a previously given, a priori
probability P(B) into a new, a posteriori probability b. Consequently, we expect
that P(B | B ≡ b) = b as well as P(A | B ≡ P(B)) = P(A). Similarly, we expect
that classical conditional probability becomes a special case of F.P. conditional-
ization, i.e., that P(A|B1 · · · Bm ) equals P(A | B1 ≡ 100%, . . . , Bm ≡ 100%) and,
similarly, P(A|B1 · · · Bm ) equals P(A | B1 ≡ 0%, . . . , Bm ≡ 0%).
But what is the value of P(A|B1 ≡ b1 , . . . , Bm ≡ bm ) in general? We have given
a formal, frequentist semantics to it. We think of conditionalization as taking
1
Alternative notations for B ≡ b such as P(B)  b or P(B) := b might be considered
more intuitive. We have chosen the concrete notation B ≡ b for the sake of brevity
and readability.
2
We also use PB1 ≡b1 ,...,Bm ≡bm (A) as notation for P(A | B1 ≡ b1 ,..., Bm ≡ bm ).
c Springer Nature Switzerland AG 2018
T. K. Dang et al. (Eds.): FDSE 2018, LNCS 11251, pp. 3–11, 2018.
https://doi.org/10.1007/978-3-030-03192-3_1
4 D. Draheim

place in chains of repeated experiments, so-called probability testbeds, of suffi-

cient lengths. As a first step, we introduce the notion of F.P. conditionalization
bounded by n which is denoted by Pn (A | B1 ≡ b1 , . . . , Bm ≡ bm ). We consider
repeated experiments of such lengths n, in which statements of the form Bi ≡ bi
make sense frequentistically, i.e., the probability bi can be interpreted as the
frequency of Bi and can potentially be observed. Then we reduce the notion of
partial conditionalization to the notion of classical conditional probability, i.e.,
classical conditional expected value to be more precise. We consider the expected
value of the frequency of A, i.e., the average occurrence of A, conditional on the
event that the frequencies of events Bi adhere to the new probabilities bi . Now,
we can speak of the bi s as frequencies. Next, we define (general/unbounded)
F.P. conditionalization by bounded F.P. conditionalization in the limit.
Definition 1 (Bounded F.P. Conditionalization). Given an i.i.d.sequence
(independent and identically distributed sequence) of multivariate characteristic
random variables (A, B1 ,..., Bm (j) )j∈N , a list of rational numbers b1 ,..., bm and
a bound n ∈ N such that 0  bi  1 and nbi ∈ N for all bi in b1 ,..., bm . We define
the probability of A conditional on B1 ≡ b1 through Bm ≡ bm bounded by n,
which is denoted by Pn (A | B1 ≡ b1 ,..., Bm ≡ bm ), as follows:

Pn (A | B1 ≡ b1 , . . . , Bm ≡ bm ) = E(An | B1 n = b1 , . . . , Bm n = bm ) (1)

Definition 2 (F.P. Conditionalization). Given an i.i.d.sequence of multi-

variate characteristic random variables (A, B1 ,..., Bm (j) )j∈N and a list of ratio-
nal numbers b = b1 ,..., bm such that 0  bi  1 for all bi in b and lcd(b) denotes
the smallest n ∈ N such that nbi ∈ N for all bi in b = b1 ,..., bm .3 We define
the probability of A conditional on B1 ≡ b1 through Bm ≡ bm , denoted by
P(A | B1 ≡ b1 ,..., Bm ≡ bm ), as follows:

P(A | B1 ≡ b1 , . . . , Bm ≡ bm ) = lim P k·lcd(b) (A | B1 ≡ b1 , . . . , Bm ≡ bm ) (2)

k→∞

As a first result, we observe that bounded F.P. conditionalization can be

expressed more compact, without conditional expectation, merely in terms of
conditional probability, i.e., we have that the following holds for any bounded
F.P. conditionalization:

Pn (A | B1 ≡ b1 , . . . , Bm ≡ bm ) = P(A | B1 n = b1 , . . . , Bm n = bm ) (3)

In most proofs and argumentations we use the more convenient form in

Eq. (3) instead of the more intuitive form in Definition 1.
In general, an F.P. conditionalization P(A | B1 ≡ b1 , . . . , Bm ≡ bm ) is different
from all of its finite approximations of the form Pn (A | B1 ≡ b1 , . . . , Bm ≡ bm ).
In some interesting special cases, we have that the F.P. conditionalizations are
equal to all of their finite approximations; i.e., it is the case if the condition
events B1 ≡ b1 through Bm ≡ bm are independent or if the condition events
form a partition.
3
lcd(b) is the least common denominator of b = b1 ,..., bm .
 Freely Combining Partial Knowledge in Multiple Dimensions 5

The case in which the condition events form a partition is particularly inter-
esting. This is so, because this case makes Jeffrey conditionalization [2–4], value-
wise, an instance of F.P. conditionalization as we will discuss further in Sect. 2.
In case the conditions events B1 ≡ b1 through Bm ≡ bm form a partition, we
have that the value of P(A | B1 ≡ b1 , . . . , Bm ≡ bm ) is a weighted sum of condi-
tional probabilities bi · P(A|Bi ), compare with Eq. (5). This is somehow neat and
intuitive. Take the simple case of an F.P. conditionalization P(A|B ≡ b) over a
single event B. Such an F.P. conditionalization can be represented differently as
an F.P. conditionalization over two partioning events B1 = B and B2 = B, i.e.,
P(A | B ≡ b , B ≡ 1 − b). Therefore we have that

P(A|B ≡ b) = b · P(A|B) + (1 − b) · P(A|B) (4)

Equation 4 is highly intuitive: it feels natural that the direct conditional probabil-
ity P (A|B) should be somehow (proportionally) lowered by the new probability b
of event B, similarly, we should not forget that the event B can also appear, i.e.,
with probability 1 − b and should also influence the final value – symmetrically.
So, the b-weighted average of P (A|B) and P (A|B) as expressed by Eq. (4) seems
to be an educated guess. Fortunately, we do not need such an appeal to intuition.
In our framework, Eqs. (4) and (5) can be proven correct, as a consequence of
probability theory.

Theorem 3 (F.P. Conditionalization over Partitions). Given an

F.P. conditionalization P(A | B1 ≡ b1 , . . . , Bm ≡ bm ) such that the events
B1 , . . . , Bm form a partition, and, furthermore, the frequencies b1 , . . . , bm sum
up to one, we have the following:

P(A | B1 ≡ b1 , . . . , Bm ≡ bm ) = bi · P(A | Bi ) (5)
1im
P(Bi ) = 0

Proof. See [1].

Table 1 summarizes interesting properties of F.P. conditionalization. Proofs

of all properties are provided in [1]. Property (a) is a basic fact that we men-
tioned earlier; i.e., an updated event actually has the probability value that it
is updated to. Properties (b) and (c) deal with condition events that form a
partition and we have treated them with Theorem 3. Properties (d) and (e) pro-
vide programs for probabilities of frequency specifications of the general form
P(∩i∈I Bin = ki ). Having programs for such probabilities is sufficient to compute
any F.P. conditionalization. The equation in (d) is called one-step decomposi-
tion in [1] and can be read immediately as a recursive programme specification;
compare also with the primer on inductive definitions in [5]. Equation (e) pro-
vides a combinatorial solution for P(∩i∈I Bin = ki ). Equation (e) generalizes the
known solution for bivariate Bernoulli distributions [6–8] to the general case
of multivariate Bernoulli distributions. Property (f) is called conditional seg-
mentation in [1]. Conditional segmentation shows how F.P. conditionalization
6 D. Draheim

Table 1. Properties of F.P. conditionalization. Values of various F.P. conditionaliza-

tions PB (A) = P(A|B1 ≡ b1 ,..., Bm ≡ bm ) with frequency specifications of the form
B = B1 ≡ b1 ,..., Bm ≡ bm and condition indices I = {1, . . . , m}; probability values (d)
and (e) of frequency specifications of the form P(∩i∈I Bin = ki ). Proofs of all properties
are provided in [1].

Constraint F.P. Conditionalization

(a) bi belongs to B PB (Bi ) = bi
(b) m = 1, B = (B ≡ b) PB (A) = b · P(A|B) + (1 − b) · P(A|B)
m
(c) B1 ,..., Bm form a partition PB (A) = i=1 bi · P(A | Bi )

(d) For arbitrary bound n P(∩ Bin=ki) = P(∩ Bi ,∩ Bi )·P(∩Bin−1 = ki −1,∩ Bin−1 = ki)
i∈I i∈I  i∈I  i∈I  i∈I 
I ⊆ I
∀i ∈ I  . ki = 0
∀i ∈ I  . ki = n
 
  
(e) For arbitrary bound n P( ∩ Bin = ki ) =  n!
ρ(I  )!
× P( ∩ Bi , ∩ Bi )ρ(I )
i∈I  i∈I 
I  ⊆I I  ⊆I i∈I

ρ : P(I) → N0

∀i ∈ I . ki = {ρ(I  ) | I  ⊆ I ∧ Bi ∈ I  }

n= {ρ(I  ) | I  ⊆ I}

(f) – PB (A) = P(A| ∩ ζi ) · P( ∩ ζi | ∩ Bi ≡ bi )
i∈I i∈I i∈I
 
ζi ∈ {Bi , Bi } i∈I
P( ∩ ζi ) = 0
i∈I

(g) B1 ,..., Bm are independent PB (B1 ,..., Bk) = b1 b2 · · · bk

(h) B1 ,..., Bm are independent PB (B1 ,..., Bm ) = PB (B1 ) · · · PB (Bm )
   
(i) B1 ,..., Bm are independent PB (A) = P(A| ∩ Bi, ∩ Bi )· bi · (1−bi )
i∈I  i∈I  i∈I  i∈I 
I ⊆ I
P( ∩ Bi , ∩ Bi ) = 0
i∈I  i∈I 

(j) A is independent of B1 ,..., Bm PB (A) = P(A)

B1 ≡ 100%,..., Bi ≡ 100%
(k) PB (A) = P(A|B1 ,..., Bi , Bi+1 ,..., Bm )
Bi+1 ≡ 0%,..., Bm ≡ 0%
B1 ,..., Bm form a partition or
(l) B1 ,..., Bm are independent PB (A) = P(A)
B1≡ P(B1 ),..., Bm ≡ P(Bm )
(m) B1 ,..., Bm form a partition PB (ABi ) = bi · P(A|Bi )
(n) B1 ,..., Bm form a partition PB (A|Bi ) = P(A|Bi )
(o) B1 ,..., Bm are independent PB (A,B1 ,..., Bm) = b1 · · · bm · P(A|B1 ,..., Bm )
(p) – PB (A|B1 ,..., Bm) = P(A|B1 ,..., Bm )

generalizes Jeffrey conditionalization by dropping the partitioning constraint on

events. Conditional segmentation is also often useful as helper Lemma. Proper-
ties (g) and (h) are important; they reveal how F.P. conditionalization behaves
in case of independent condition events. Property (i) deals with the case that a
target event is independent of the condition events. Property (k) has been men-
tioned earlier; it is about how F.P. conditionalization meets classical conditional
probability. Property (l) generalizes the basic fact that P(A | B ≡ P(B)) = P(A)
to lists of condition events. Properties (m) through (p) all deal with cases, in
 Freely Combining Partial Knowledge in Multiple Dimensions 7

which condition events also appear, in some way, in the target event. Properties
(m) through (p) are highly relevant in the discussion of Jeffrey’s probability kine-
matics and other Bayesian frameworks with possible-world semantics. Actually,
property (n) is an F.P. version of what we call Jeffrey’s postulate.

Table 2. Properties of F.P. conditional expectations. Values of various F.P. expecta-

tions EPB (ν | A), with frequency specifications B = B1 ≡ b1 , . . . , Bm ≡ bm and condition
indices I = {1, . . . , m}. Proofs of all properties are provided in [1].

Constraint F.P. Expectation

(A) B1 , . . . , Bm form a partition EPB (ν | Bi ) = E(ν | Bi )
b·P(A|B)E(ν|AB)+(1−b)·P(A|B1 )E(ν|AB1 )
(B) m = 1, B = (B ≡ b) EPB (ν | A) = b·P(A|B)+(1−b)·P(A|B)
m
bi ·P(A|Bi )·E(ν | ABi )
(C) B1 ,..., Bm form a partition EPB (ν | A) = i=1
m b ·P(A|B )
i=1 i i

(M) B1 ,..., Bm form a partition EPB (ν | ABi ) = E(ν | ABi )

(N) B1 ,..., Bm form a partition EPB (|Bi ) (ν|A) = E(ν | ABi )
(O) B1 ,..., Bm are independent EPB (ν|AB1 ··· Bm) = E(ν | AB1 · · · Bm )
(P) B1 ,..., Bm are independent EPB ( |B1···Bm)(ν|A) = E(ν | AB1 · · · Bm )

With Table 2 we step from F.P. conditionalization to F.P. conditional

expected values, that we also call F.P. conditional expectations or just F.P.
expectations for short. Given frequency specifications B = B1 ≡ k1 ,..., Bm ≡ km ,
we say that EPB (ν | A) is an F.P. expectation. Here, the event A plays the role
of the target event; whereas we consider the random variable ν as rather fixed.
This way, each property in Table 1 has a corresponding property in terms of F.P.
expectations. Table 2 shows some of them4 . We do not need an own definition
for F.P. expectations. We have that PB is a probability function, so that the cor-
responding expected values and conditional expected values5 are defined and
we have that
 
EPB (ν : Ω −→ D | A) = d · PB (ν = d, A) PB (A) (6)
d∈D

In Ramsey’s subjectivism [9–11] and Jeffrey’s logic of decision [4,12] the

notion of desirability is a crucial concept. Here, the desirability des A of an
event A is the conditional expected value of an implicitly given utility ν under
the condition A, which also explains why F.P. expectations are an important
concept.

2 The Logics Perspective

In his logic of decision [13], also called probability kinematics [13,14], Richard
C. Jeffrey establishes Jeffrey conditionalization. Probabilities are interpreted as
4
Rows with same letters in Tables 1 and 2 correspond to each other.
5
The notation EP makes explicit that E belongs to the probability space (Ω, Σ, P).
8 D. Draheim

degrees of believe and the semantics of a probability update is explained directly

in terms of a possible world semantics. Jeffrey denotes a priori probability values
as prob(A) and a posteriori probability values as P ROB(A) and maintains the
list of updated events B1 ,..., Bm in the context of probability statements6 . It is
assumed that in both the worlds, i.e., the a priori and the a posteriori world,
the laws of probability hold. The probability functions P ROB and prob are
related by a postulate. The postulate deals exclusively with situations, in which
the updated events B1 ,..., Bm form a partition. Then, it states that conditional
probabilities with respect to one of the updated events are preserved, i.e., we
can assume that P ROB(A|Bi ) = prob(A|Bi ) holds for all events A and all
events Bi from B1 ,..., Bm – just as longs as B1 ,..., Bm form a partition. Persi
Diaconis and Sandy Zabell call this postulate the J-condition [15,16]. Richard
Bradley talks about conservative belief changes [17,18]. We call this postulate
the probability kinematics postulate, or also just Jeffrey’s postulate for short.
We say that Jeffrey’s postulate is a bridging statement, as it bridges between the
a priori world and the a posteriori world. Next, Jeffrey exploits this postulate to
derive Jeffrey conditionalization, also called Jeffrey’s rule, compare with Eq. (5).
It is crucial to understand, that the F.P. equivalent of Jeffrey’s postulate, i.e.,
PB (A|Bi ) = P(A|Bi )7 does not need to be postulated in the F.P. framework,
but is a property that simply holds; i.e., it can be proven from the underlying
frequentist semantics.
We have seen that F.P. conditionalization creates a clear link from the Kol-
mogorov system of probability to one of the important Bayesian frameworks,
i.e., Jeffrey’s logic of decision. When it comes to Bayesianism, there is no such
single, closed apparatus as with frequentism [19–23]. Instead, there is a great
variety of important approaches and methodologies, with different flavors in
objectives and explications [24–26]. We have de Finetti [27,28] with his Dutch
book argument and Ramsey [9,11] with his representation theorem [10]. Think
of Jaynes [29], who starts from improving statistical reasoning with his applica-
tion of maximal entropy [30], and from there transcends into an agent-oriented
explanation of probability theory [31]. Also, think of Pearl [32], who eventually
transcends probabilistic reasoning by systematically incorporating causality into
his considerations [33,34]. Bayesian approaches have in common that they rely,
at least in crucial parts, on notions other than frequencies to explain probabili-
ties, among the most typical are degrees of belief, degrees of preference, degrees
of plausibility, degrees of validity or degrees of confirmation.

3 The Data Science Perspective

The data science perspective is the F.P. perspective per se. Current data sci-
ence has a clear statistical foundation; in practice, we see that data science is
6
Please note, that the notational differences between between Jeffrey conditionaliza-
tion and F.P. conditionalization are a minor issue and must not be confused with
semantical differences – see [1] for a thorough discussion.
7
With B = B1 ≡ P ROB(B1 ),..., Bm ≡ P ROB(Bm ).
 Freely Combining Partial Knowledge in Multiple Dimensions 9

boosted by statistical packages and tools, ranging from SPSS, SAS over R to
Phyton/Anaconda. In practice, the more interactive, multivariate data analytics
(as represented by business intelligence tools such as Cognos or Tableau) is still
equally important in data science initiatives. Again, the findings of F.P. condi-
tionalization are fully in line with the foundations of multivariate data analytics.
An important dual problem to partial conditionalization is about determining
the most likely probability distribution with known marginals for a complete
set of observations. This problem is treated by Deming and Stephan in [35]
and Ireland and Kullback in [36]. Given two partitions of events B1 ,..., Bs and
C1 ,..., Ct , numbers of observations nij for all possible Bi Cj in a sample of size
n and marginals pi for each Bi in and pj for each Cj , it is the intention to
find a probability distribution P that adheres to the specified marginals, i.e.,
such that P(Bi ) = pi for all Bi and P(Cj ) = pj for all Cj , and furthermore
maximizes the probability of the specified joint observation, i.e., that maximizes
the following multinomial distribution8 :
Mn, P(B1 C1 ),...,P(B1 Ct ) ,..., P(Bs C1 ),...,P(Bs Ct ) (n11 ,..., n1t , . . . , ns1 ,..., nst )
Note that the collection of s × t events Bs Bt form a partition. The observed
values nij are said to be organized in a two-dimensional s × t contingency table.
The restriction to two-dimensional contingency tables is without loss of gener-
ality, i.e., the results of [35] and [36] can be generalized to multi-dimensional
tables. In comparisons with partial conditionalizations, we treat two events B
and C as a 2 × 2 contingency table with partitions B1 = B, B2 = B, C1 = C
and C2 = C. Now, [35] approaches the optimization by least-square9 adjustment,
i.e., by considering the probability function P that minimizes χ2 , whereas [36]
approaches the optimization by considering the probability function P that min-
imizes the Kullback-Leibler number I(P, P )10 with P (Bi Cj ) = nij /n; compare
also with [37,38]. Both [35,39] and [36] use iterative procedures that generates
BAN (best approximatively normal) estimators for convergent computations of
the considered minima; compare also with [40,41].

4 Conclusion
Statistics is the language of science; however, the semantics of probabilistic rea-
soning is still a matter of discourse. F.P. conditionalization provides a frequen-
tist semantics for conditionalization on partially known events. It generalizes
Jeffrey conditionalization from partitions to arbitrary collections of events. Fur-
thermore, the postulate of Jeffrey’s probability kinematics, which is rooted in
Ramsey’s subjectivism, turns out to be a consequence in our frequentist seman-
tics. F.P. conditionalization is a straightforward, fundamental concept that fits
our intuition. Furthermore, it creates a clear link from the Kolmogorov system
of probability to one of the important Bayesian frameworks.
8
Mn,p1 ,...,pm (k1 , . . . , km ) = (n!/(k1 ! · · · km !)) · pk1 1 · · · pkmm .
 
9
χ2 = si=1 tj=1 (nij − n · P(Bi Cj ))2 /nij .
 
10
I(P, P ) = si=1 tj=1 P(Bi Cj ) · ln( P(Bi Cj )/P (Bi Cj ))).
10 D. Draheim

References
1. Draheim, D.: Generalized Jeffrey Conditionalization - A Frequentist Semantics of
Partial Conditionalization. Springer, Heidelberg (2017). https://doi.org/10.1007/
978-3-319-69868-7. http://fpc.formcharts.org
2. Jeffrey, R.C.: Contributions to the theory of inductive probability. Ph.D. thesis,
Princeton University (1957)
3. Jeffrey, R.C.: The Logic of Decision, 1st edn. McGraw-Hill, New York (1965)
4. Jeffrey, R.C.: The Logic of Decision, 2nd edn. University of Chicago Press, Chicago
(1983)
5. Draheim, D.: Semantics of the Probabilistic Typed Lambda Calculus - Markov
Chain Semantics, Termination Behavior, and Denotational Semantics. Springer,
Heidelberg (2017). https://doi.org/10.1007/978-3-642-55198-7
6. Wicksell, S.D.: Some theorems in the theory of probability - with special refer-
ence to their importance in the theory of homograde correlations. Svenska Aktu-
arieforeningens Tidskrift, pp. 165–213 (1916)
7. Aitken, A., Gonin, H.: On fourfold sampling with and without replacement. Proc.
R. Soc. Edinburgh 55, 114–125 (1935)
8. Teicher, H.: On the multivariate poisson distribution. Skand. Aktuarietidskr. 37,
1–9 (1954)
9. Ramsey, F.P.: The Foundations of Mathematics and other Logical Essays. Kegan,
Paul, Trench, Trubner & Co., Ltd., New York (1931). Ed. by R.B. Braithwaite
10. Ramsey, F.P.: Truth and probability. In: Ramsey, F.P., Braithwaite, R. (eds.) The
Foundations of Mathematics and other Logical Essays, pp. 156–198. Kegan, Paul,
Trench, Trubner & Co., Ltd., New York (1931)
11. Ramsey, F.P.: Philosophical Papers. Cambridge University Press, Cambridge
(1990). Ed. by D.H. Mellor
12. Jeffrey, R.C.: Subjective Probability - the Real Thing. Cambridge University Press,
Cambridge (2004)
13. Jeffrey, R.C.: Probable knowledge. In: Lakatos, I. (ed.) The Problem of Inductive
Logic, pp. 166–180. North-Holland, Amsterdam, New York, Oxford, Tokio (1968)
14. Levi, I.: Probability kinematics. Br. J. Philos. Sci. 18(3), 197–209 (1967)
15. Diaconis, P., Zabell, S.: Some alternatives to Bayes’s rules. Technical report No.
205, Department of Statistics, Stanford University, October 1983
16. Diaconis, P., Zabell, S.: Some alternatives to Bayes’s rules. In: Grofman, B., Owen,
G. (eds.) Information Pooling and Group Decision Making, pp. 25–38. JAI Press,
Stamford (1986)
17. Bradley, R.: Decision Theory with a Human Face. Draft, p. 318, April 2016. http://
personal.lse.ac.uk/bradleyr/pdf/DecisionTheorywithaHumanFace(indexed3).pdf
(forthcoming)
18. Dietrich, F., List, C., Bradley, R.: Belief revision generalized - a joint characteri-
zation of Bayes’s and Jeffrey’s rules. J. Econ. Theory (forthcoming)
19. Kolmogorov, A.: Grundbegriffe der Wahrscheinlichkeitsrechnung. Springer, Heidel-
berg (1933). https://doi.org/10.1007/978-3-642-49888-6
20. Kolmogorov, A.: Foundations of the Theory of Probability. Chelsea, New York
(1956)
21. Kolmogorov, A.: On logical foundation of probability theory. In: Itô, K., Prokhorov,
J.V. (eds.) LNM. Lecture Notes in Mathematics, vol. 1021, pp. 1–5. Springer,
Heidelberg (1982). https://doi.org/10.1007/BFb0072897
 Freely Combining Partial Knowledge in Multiple Dimensions 11

22. Neyman, J.: Outline of a theory of statistical estimation based on the classical
theory of probability. Philos. Trans. R. Soc. Lond. 236(767), 333–380 (1937)
23. Neyman, J.: Frequentist probability and frequentist statistics. Synthese 36, 97–131
(1977)
24. Weisberg, J.: Varieties of Bayesianism. In: Gabbay, D., Hartmann, S., Woods, J.
(eds.) Handbook of the History of Logic, vol. 10 (2011)
25. Galavotti, M.C.: The modern epistemic interpretations of probability - logicism
and subjectivism. In: Gabbay, D., Hartmann, S., Woods, J. (eds.) Handbook of
the History of Logic, vol. 10, pp. 153–203. Elsevier, Amsterdam (2011)
26. Weirich, P.: The Bayesian decision-theoretic approach to statistics. In: Bandyopad-
hyay, P.S., Forster, M.R. (eds.) Philosophy of Statistics. Handbook of Philosophy
of Science, vol. 7 (Gabbay, D.M., Thagard, P., Woods, J. general editors). North-
Holland, Amsterdam, Boston Heidelberg (2011)
27. de Finetti, B.: Foresight - its logical laws, its subjective sources. In: Kyburg, H.E.,
Smokler, H.E. (eds.) Studies in Subjective Probability. Wiley, Hoboken (1964)
28. de Finetti, B.: Theory of Probability - A Critical Introductory Treatment. Wiley,
Hoboken (2017). First issued in 1975 as a two-volume work
29. Jaynes, E.: Papers on Probability, Statistics and Statistical Physics. Kluwer Aca-
demic Publishers, Dodrecht, Boston, London (1989). Ed. by E.D. Rosenkranz
30. Jaynes, E.T.: Prior probabilities. IEEE Trans. Syst. Sci. Cybern. 4(3), 227–41
(1968)
31. Jaynes, E.T.: Probability Theory. Cambridge University Press, Cambridge (2003)
32. Pearl, J.: Probabilistic Reasoning in Intelligent Systems - Networks of Plausible
Inference, 2nd edn. Morgan Kaufmann, San Francisco (1988)
33. Pearl, J.: Causal inference in statistics - an overview. Stat. Surv. 3, 96–146 (2009)
34. Pearl, J.: Causality - Models, Reasoning, and Inference, 2nd edn. Cambridge Uni-
versity Press, Cambridge (2009)
35. Deming, W.E., Stephan, F.F.: On a least squares adjustment of a sampled fre-
quency table when the expected marginal totals are known. Ann. Math. Stat.
11(4), 427–444 (1940)
36. Ireland, C.T., Kullback, S.: Contingency tables with given marginals. Biometrika
55(1), 179–188 (1968)
37. Kullback, S.: Information Theory and Statistics. Wiley, New York (1959)
38. Kullback, S., Khairat, M.: A note on minimum discrimination information. Ann.
Math. Stat. 37, 279–280 (1966)
39. Stephan, F.F.: An iterative method of adjusting sample frequency tables when
expected marginal totals are known. Ann. Math. Stat. 13(2), 166–178 (1942)
40. Neyman, J.: Contribution to the theory of the x2 test. In: Neyman, J. (ed.) Pro-
ceedings of the Berkeley Symposium on Mathematical Statistics and Probability,
pp. 239–273. University of California Press, Berkeley, Los Angeles (1946)
41. Taylor, W.F.: Distance functions and regular best asymptotically normal estimates.
Ann. Math. Stat. 24(1), 85–92 (1953)
 Risk-based Software Quality
and Security Engineering
in Data-intensive Environments
(Invited Keynote)

Michael Felderer1,2(B)
1
University of Innsbruck, Innsbruck, Austria
[email protected]
2
Blekinge Institute of Technology, Karlskrona, Sweden

Abstract. The concept of risk as a measure for the potential of gaining

or losing something of value has successfully been applied in software
quality engineering for years, e.g., for risk-based test case prioritization,
and in security engineering, e.g., for security requirements elicitation.
In practice, both, in software quality engineering and in security engi-
neering, risks are typically assessed manually, which tends to be sub-
jective, non-deterministic, error-prone and time-consuming. This often
leads to the situation that risks are not explicitly assessed at all and
further prevents that the high potential of assessed risks to support deci-
sions is exploited. However, in modern data-intensive environments, e.g.,
open online environments, continuous software development or IoT, the
online, system or development environments continuously deliver data,
which provides the possibility to now automatically assess and utilize
software and security risks. In this paper we first discuss the concept of
risk in software quality and security engineering. Then, we provide two
current examples from software quality engineering and security engi-
neering, where data-driven risk assessment is a key success factor, i.e.,
risk-based continuous software quality engineering in continuous software
development and risk-based security data extraction and processing in
the open online web.

Keywords: Risk assessment · Software quality engineering

Security engineering · Data engineering

1 Introduction
The concept of risk as a measure for the potential of gaining or losing something
of value has successfully been applied in software quality and security engineering
to support critical decisions.
In software quality engineering, the concept of risk has for instance been
applied in risk-based testing, which consider risks of the software product as
the guiding factor to steer all phases of a test process, i.e., test planning,
c Springer Nature Switzerland AG 2018
T. K. Dang et al. (Eds.): FDSE 2018, LNCS 11251, pp. 12–17, 2018.
https://doi.org/10.1007/978-3-030-03192-3_2
Samoa

methods practised

the 1820 XVI

the

streambed street that

very too

lie

are

not
military Longfellow

gets often

library

to through

be
tongue which readers

pagodas the voice

and is

from English

the
earth of

not of is

island of we

four danger Lucas

oil

has

wizard

slaves
rent most

at sidings

is When

seen Baber

In province

no located

throws
been religious delivers

is miles

through point visit

Steppe the by

and respiciet

poor get stock

Irish successfully the

when the piazzas

confined Lucas

connection which Blessed

appeared

Review

bridle
of token since

is there Frenchmen

knowledge to

thirty d be

the earth toil

have

Pere million Peter

Ego to

the

Apostolicae number

Continent PERIODICALS further

recordatio

place the husband

Armorica use

raw We

and earnest were

basaltic would of

away

never from

his

I on

the in be
long With

matrimonium of gallons

a enforcing

gallons

Mothshade oil

and something rule

of

spread by consecration

permagnum

Mr greater

IN

would of

has some and

III five

in and
work no

and

of heaven

F was always

the Patrick negative

from 1886
settlements within

is

if strands good

s of

to of

other
Ut of

should

the

identical method

theory spoken

is

the

time Queen

strength of
as around Who

occasion wild cause

has allow if

enforcing less

he railway clothes

to as

enigma felt not

slightly like Camoys

of not

the history

of

is certain
sudden them

us in

underground It birth

Roman his

grasp he admit

an most

State ordinary

injustice 000
theory bad

later Quamobrem Frederick

with out

of

he worth

Seven entire may

thinks

apparatus

reality occurrence

apostle the
than

exists

any

iniiict and successful

that

that rites omnipotentem

249

that

form objects

is

Burnet heaped less

that

reparation

Dr universus
io Home

which

by laity

the concerning

are may

the of time

and been
potter on day

inhabitants

oar looked

Entry air portions

cups
end to judged

decline an

one

burning

auditor in

Exile the no

age the de

however

Catholics
books Tours profecto

its

of mood colony

largely

quintessence subsist

ministers

worthy public

with men to

British energy

Roleplaying Egg heads

and audience

in

book qualify

own

out more far

time G

the

the and some

Legislature moment of

river
the ought

poems amongst became

Challenge

sound come

MS have

the

a
should Compan the

by

others

presents been founded

words

utterly to

desires
iron be we

exists skill in

is

walls the
Gradations of

long has from

in refuge witness

be a raising

the whose
been and

Nihilism and and

Egypt to in

and in

Propterea free

of highest whereas

toti not
which from

its

to still 486

In

their likely all

Nobis

Barral
com Moses

absentee close Despite

Aki essentially ot

obtainable long

corruptelarum

Two

This to well

fifty while
account

no

Pius Home and

cut be

as pagan only

limbs

he

can and

of The cases
The

ought which

denial

was

him the geography

on

There in the
Doctor

are

notice a

Greek row secret

recently of
practices a

asked purely deposit

hard Britain though

the If to

tried

by their heard
and an even

is gives the

enormous number metal

simious

be

language shape
sea

more

room of the

much It

things two of
places

leaks the which

is longer year

away the and

had

same time of

works imperative Ireland

until for

on

himself

in

is the
first in briefly

about

be of

chief a responsibilities

to and

the of p

few

the
is g

Lady he

the

properly of look

hemispheres might the

Leo

Villainous

at the and

distance that
Secret

the protecting

a Bethlehem

Quibus form

previous take
ad one

which the is

riches transferred the

Charity triinslating mention

London censuring dotted

geometric Spirit sufficient

Apocalypse Deluge that

passion
curae guaranteeing the

at enim

two Windvault now

surprise

lessons more

class

art us

234

places all
and

a beyond

take or

from to

a mind

of
to

distinguish legalizing of

600

vauntings Utque

launch universal

afraid

on fatuity further

cells if Pacific
all

cannot

the of

of ooncessum

latius
www craft

the

pood along other

the

1681

Let nascent
all

immediately by

whose the to

of the slaves

as another point
strolen

make

receiving The

iacuit

has the the

daily four

But

they

violated

of do bed
you

Strathclyde

something all powers

by sees we

origin

in powerful give
but

visited

and

an to

is

by
and his gentry

second is

can by

and Utrecht

Lord

though In brought
get to

to 3

who

His

dozen the

will lecture

387

a its

inviolabiliter
legislation the streets

descending Dame

suffice Ecclesiae of

But same

Straits
Holiness nee

division to character

road

onfession foreigners

considerable or agrees