- Hujun Yin, University of Manchester (UK)
- Michal Wozniak, Wroclaw University of Technology (Poland)
The International Conference on Intelligent Data Engineering and Automated Learning (IDEAL) is an annual international conference dedicated to emerging and challenging topics in intelligent data analysis, data mining and their associated learning systems and paradigms. Its core themes include: the Big Data challenges, Machine Learning, Data Mining, Information Retrieval and Management, Bio- and Neuro-Informatics, Bio-Inspired Models (including Neural Networks, Evolutionary Computation and Swarm Intelligence), Agents and Hybrid Intelligent Systems, and Real-world Applications of Intelligent Techniques. Other related and emerging themes and topics are also welcome.
After recent successful events, IDEAL 2014 (Salamanca, Spain), IDEAL 2013 (Hefei, China), IDEAL 2012 (Natal, Brazil), IDEAL 2011 (Norwich, England), IDEAL 2010 (Paisley, Scotland), IDEAL 2009 (Burgos, Spain), IDEAL 2008 (Daejeon, South Korea) and IDEAL 2007 (Birmingham, England), the 16th edition, IDEAL 2015, will be held in Wroclaw, Poland on October 14-16, 2015, Wroclaw, Poland
The conference provides a unique opportunity and stimulating forum for presenting and discussing the latest theoretical advances and real-world applications in Computational Intelligence and Intelligent Data Analysis. Authors and researchers are warmly invited to submit their latest findings and research work to the conference. Special session organizers are warmly invited to submit their proposals to the organizers.
Proceedings of IDEAL 2015, to be published by Springer in its prestigious LNCS series, which is indexed in ISI Conference Proceedings Citation Index - Science (CPCI-S), included in ISI Web of Science, EI Engineering Index (Compendex and Inspec databases),ACM Digital Librar, dblp, Google Scholar, IO-Port, MathSciNet, Scopus, Zentralblatt MATH EI. In addition, selected papers will be invited for special issues in leading international journals in the field, including the International Journals of Neural Systems (IJNS) (2013 Impact Factor 6.065).
|Special session proposal||1st april 2015|
|Paper submission deadline||
|Decision||5th july 2015|
|Registration and final submission||10th july 2015|
|Conference Presentation||14-16 october 2015|
Nazwa odbiorcy: Politechnika Wrocławska, ul. Wybrzeże Wyspiańskiego 27, 50-370 Wrocław
Bank Zachodni WBK S.A. 16 Oddział Wrocław
50-373 Wrocław, ul. Norwida 1/3
37 1090 2402 0000 0006 1000 0434
Tytuł przelewu: 488291/K0402; IDEAL 2015; Paper ID
Receiver: Wrocław University of Technology, ul. Wyb. Wyspiańskiego 27, 50-370 Wrocław, Poland
Bank’s name: Bank Zachodni WBK S.A.
Branch: 16 Oddzial Wroclaw
Bank’s address: 50-373 Wroclaw, ul. Norwida 1/3
Account number: PL 91 1090 2402 0000 0006 1000 0626
SWIFT WBK PPL PP
Transfer title/description: 488291/K0402; IDEAL 2015; Paper ID
The conference will take place in Conference Center of Wroclaw University of Technology.
The University of Technology
D 20 building
Janiszewskiego 8 street
Prof. Jerzy Stefanowski
Prof. Leszek Rutkowski
Prof. Vaclav Snasel
|11:00||11:30||Coffee break||11:00||11:30||Coffee break||11:00||12:30||Session 13
Prof. Xin Yao
Prof. Manuel Grana
|16:30||17:00||Coffee break||16:30||17:00||Coffee break|
Imperial Room, Centennial Hall
Session 1 Special Session on Discovering Knowledge from Data
14.10.2015, 11:30—13:00, room 10C
chairman: Alberto Fernandez
|1||Addressing Overlapping in Classification with Imbalanced Datasets: A First Multi-Objective Approach for Feature and Instance Selection||Alberto Fernandez, Maria Jose Del Jesus and Francisco Herrera|
|2||Deterministic Extraction of Compact Sets of Rules for Subgroup Discovery||Juan L. Dominguez, Jacinto Mata and Victoria Pachón|
|3||Cost-Sensitive Neural Network with ROC-based Moving Threshold for Imbalanced Classification||Bartosz Krawczyk and Michal Wozniak|
|4||On Stability of Ensemble Gene Selection||Nicoletta Dessì, Barbara Pes and Marta Angioni|
14.10.2015, 11:30—13:00, room 10D
chairman: Robert Burduk
|1||A Particle Swarm Clustering Algorithm with Fuzzy Weighted Step Sizes||Alexandre Szabo, Myriam Regattieri Delgado and Leandro Nunes de Castro|
|2||A Bacterial Colony Algorithm for Association Rule Mining||Danilo Cunha, Rafael Xavier and Leandro de Castro|
|3||Improving the NSGA-II Performance with an External Population||Krzysztof Michalak|
|4||Local Search Based on a Local Utopia Point for the Multiobjective Travelling Salesman Problem||Krzysztof Michalak|
Session 3 Intelligent Applications in Real-world Problems
14.10.2015, 15:00—16:30, room 10C
chairman: Dariusz Frejlichowski
|1||Application of Cascades of Classifiers in the Vehicle Detection Scenario for the ’SM4Public’ System||Dariusz Frejlichowski, Katarzyna Gościewska, Adam Nowosielski, Pawel Forczmanski and Radosław Hofman|
|2||A learning Web platform based on a fuzzy linguistic recommender system to help students to learn recommendation techniques||Carlos Porcel, Mjesus Lizarte, Juan Bernabé-Moreno and Enrique Herrera-Viedma|
|3||NMF and PCA as applied to gearbox fault data||Anna Bartkowiak and Radoslaw Zimroz|
|4||Multi-manifold Approach to Multi-view Face Recognition||Shireen Mohd Zaki and Hujun Yin|
|5||The belief theory for emotion recognition||Halima Mhamdi, Med Salim Bouhlel and Hnia Jarray|
14.10.2015, 15:00—16:30, room 10D
chairman: Bartosz Krawczyk
|1||Hybrid Evolutionary Algorithm with Adaptive Crossover, Mutation and Simulated Annealing Processes to Project Scheduling||Virginia Yannibelli and Analía Amandi|
|2||Building an Efficient Evolutionary Algorithm for Forex Market Predictions||Rafal Moscinski and Danuta Zakrzewska|
|3||An Extension of Multi-label Binary Relevance Models Based on Randomized Reference Classifier and Local Fuzzy Confusion Matrix.||Pawel Trajdos and Marek Kurzynski|
|4||Effective realizations of biorthogonal wavelet transforms of lengths 2K + 1/2K − 1 with lattice structures on GPU and CPU||Dariusz Puchala, Bartłomiej Szczepaniak and Mykhaylo Yatsymirskyy|
Session 5 Special Session on Simulation-driven DES-like Modeling and Performance Evaluation
14.10.2015, 17:00—18:30, room 10C
chairman: Grzegorz Bocewicz
|1||Application of Fuzzy Logic Controller for Machine Load Balancing in Discrete Manufacturing System||Grzegorz Kłosowski, Arkadiusz Gola and Antoni Świć|
|2||Assessment of production system stability with the use of the FMEA analysis, simulation models and linguistic variables||Anna Burduk and Mieczyslaw Jagodzinski|
|3||Information Retrieval and Data Forecasting via Probabilistic Nodes Combination||Dariusz Jacek Jakóbczak|
|4||Knowledge Discovery in Enterprise Databases for Forecasting New Product Success||Marcin Relich and Krzysztof Bzdyra|
|5||A hybrid programming framework for resource-constrained scheduling problems||Paweł Sitek and Wikarek Jarosław|
14.10.2015, 17:00—18:30, room 10D
chairman: Vaclav Snasel
|1||Neural network-based user-independent physical activity recognition for mobile devices||Bojan Kolosnjaji and Claudia Eckert|
|2||Reduction of Signal Strength Data for Fingerprinting-Based Indoor Positioning||Maciej Grzenda|
|3||Pattern Password Authentication based on Touching Location||Orcan Alpar and Ondrej Krejcar|
|4||Natural Gesture Based Interaction with virtual heart in Augmented Reality||Rawia Frikha, Ridha Ejbali, Mourad Zaied and Chokri Ben Amar|
|5||Using a Portable Device for Online Single-Trial MRCP Detection and Classification||Ali Hassan, Usman Ghani, Farhan Riaz, Saad Rehman, Mads Jochumsen, Denise Taylor and Imran Khan Niazi|
15.10.2015, 11:30—13:00, room 10C
chairman: Krzysztof Walkowiak
|1||A simulated annealing heuristic for a branch and price-based routing and spectrum allocation algorithm in elastic optical networks||Mirosław Klinkowski and Krzysztof Walkowiak|
|2||Simulated Annealing Algorithm for Minimization of Bandwidth Fragmentation in Elastic Optical Networks with Multicast and Unicast Flows||Piotr Nagły and Krzysztof Walkowiak|
|3||Tabu search algorithm for routing and spectrum allocation of multicast demands in elastic optical networks||Róża Goścień|
|4||Multi Population Pattern Searching Algorithm for solving Routing Spectrum Allocation with Joint Unicast and Anycast problem in Elastic Optical Networks||Michal Przewozniczek|
|5||Tabu-search algorithm for optimization of elastic optical network based distributed computing systems||Marcin Markowski|
15.10.2015, 11:30—13:00, room 10D
chairman: Salvador Garcia
|1||Multi-Agent Reinforcement Learning for control systems: Challe nges and Proposals||Manuel Grana and Borja Fernandez-Gauna|
|2||Managing Monotonicity in Classification by a Pruned Random Forest||Sergio González, Francisco Herrera and Salvador García|
|3||Ensemble Selection Based on Discriminant Functions in Binary Classification Task||Robert Burduk and Paulina Baczyńska|
|4||Intelligent Automated Design of Machine Components using Antipatterns||Wojciech Kacalak, Maciej Majewski and Zbigniew Budniak|
|5||Fusion of Self-Organizing Maps with different sizes||Leandro Pasa, Jose Alfredo Ferreira Costa and Marcial Guerra Medeiros|
15.10.2015, 15:00—16:30, room 10C
chairman: Marcin Markowski
|1||ICA for detecting artifacts in a few channel BCI||Izabela Rejer and Paweł Górski|
|2||Early Alzheimer’s Disease Prediction in Machine Learning Setup: Performance Analysis with Missing Value Imputation||Sidra Minhas, Aasia Khanum, Farhan Riaz, Atif Alvi and Shoab A. Khan|
|3||Description of Visual Content in Dermoscopy Images Using Joint Histogram of Multiresolution Local Binary Patterns and Local Contrast||Sidra Naeem, Farhan Riaz, Ali Hassan and Rida Nisar|
|4||Modeling the Behavior of Unskilled Users in a Multi-UAV Simulation Environment||Víctor Rodríguez Fernández, Antonio Gonzalez-Pardo and David Camacho|
15.10.2015, 15:00—16:30, room 10D
chairman: Manuel Grana
|1||Study of collective robotic tasks based on the behavioral model of the agent||Fredy Martínez, Edwar Jacinto and Fernando Martínez|
|2||Minimalist artificial eye for autonomous robots and path planning||Omar Espinosa, Luisa Castañeda and Fredy Martínez|
|3||15 DOF robotic hand fuzzy-sliding control for grasping tasks||Edwar Jacinto, Holman Montiel and Fredy Martínez|
|4||An Empirical Evaluation of Robust Gaussian Process Models for System Identification||César Lincoln Mattos, José Santos and Guilherme Barreto|
|5||A Novel Recursive Solution to LS-SVR for Robust Identification of Dynamical Systems||José Santos and Guilherme Barreto|
16.10.2015, 9:00—10:30, room 10C
chairman: Bogusław Cyganek
|1||Data Streams Fusion by Frequent Correlations Mining||Radosław Ziembiński|
|2||Optimal Filtering for Time Series Classification||Frank Höppner|
|3||EVIDIST: A Similarity Measure for Uncertain Data Streams||Abdelwaheb Ferchichi, Mohamed Salah Gouider and Lamjed Ben Said|
|4||Multistep Forecast of FX Rates Using an Extended Self-Organizing Regressive Neural Network||Yicun Ouyang and Hujun Yin|
16.10.2015, 9:00—10:30, room 10D
chairman: Przemysław Ryba
|1||Web genre classification via hierarchical multi-label classification||Gjorgji Madjarov, Vedrana Vidulin, Ivica Dimitrovski and Dragi Kocev|
|2||OMAIDS: A Multi-agents Intrusion Detection System Based Ontology||Imen Brahmi and Hanen Brahmi|
|3||A Belief Function Reasoning Approach to Web User Profiling||Luepol Pipanmekaporn and Suwatchai Kamolsantiroj|
|4||Propagating disaster warnings on social and digital media||Stephen Kelly and Khurshid Ahmad|
|5||Data-driven simulation model generation for ERP and DES systems integration||Damian Krenczyk and Grzegorz Bocewicz|
16.10.2015, 11:00—12:30, room 10C
chairman: Jerzy Stefanowski
|1||Variable Transformation for Granularity Change in Hierarchical Databases in Actual Data Mining Solutions||Paulo J. L. Adeodato|
|2||A New Approach to Link Prediction in Gene Regulatory Networks||Turki Turki and Jason T. L. Wang|
|3||A distributed approach to flood prediction using a WSN and ML: a comparative study of ML techniques in a WSN deployed in Brazil||Gustavo Furquim, Gustavo Pessin, Pedro Henrique Gomes, Eduardo Mendiondo and Jó Ueyama|
|4||Throughput analysis of automatic production lines based on simulation methods||Slawomir Klos and Justyna Patalas-Maliszewska|
16.10.2015, 11:00—12:30, room 10D
chairman: Leszek Chmielewski
|1||The impact of news media and affect in financial markets||Stephen Kelly and Khurshid Ahmad|
|2||Qualitative and Quantitative Sentiment Proxies: Interaction between Markets||Zeyan Zhao and Khurshid Ahmad|
|3||Clusterization of indices and assets in the stock market||Leszek J Chmielewski, Maciej Janowicz, Luiza Ochnio and Arkadiusz Orłowski|
|4||Behaviour and Markets: the Interaction Between Sentiment Analysis and Ethical Values?||Jason Cook and Khurshid Ahmad|
Reinforcement learning minimizes the teacher feedback required for training. It has received a lot of attention in the control and robotics community, where the effort to generate the required teacher information for supervised training is unaffordable in many situations. The solutions and algorithms developed for single agent systems suffer a combinatorial explosion when translated into the realm of mult-agent systems, therefore new methods of decomposing the models and of speeding the learning process are very much desired and remain in some domains an interesting open problem. We will review fundamental ideas, as well as new developments and applications, which may be of use in the new paradigm of Industry 4.0. Specifically, multi-agent reinforcement learning for multi-robot system control problems will be described.
The lecture is concerned with two challenging problems in data mining, namely content based image retrieval (CBIR) and data streams. The CBIR process consists of retrieving the most visually similar images to a given query image from a database of images. First a review of available techniques for various CBIR problems is presented and new algorithms are described. Next we focus on classification problems for data streams. Unlike the static dataset, data stream is of infinite size. Data elements arrive to the system continuously, often with very high rates. Moreover, the concept of data can evolve in time, what is known as the concept drift. For these reasons, commonly known data mining algorithms cannot be directly applied to the data streams. We will present the latest algorithms developed for classification of data streams.
The Big Data paradigm is one of the main science and technology challenges of today. Big data includes various data sets that are too large or too complex for efficient processing and analysis using traditional as well as unconventional algorithms and tools. The challenge is to derive value from signals buried in an avalanche of noise arising from challenging data volume, flow and validity. The computer science challenges are as varied as they are important. Whether searching for influential nodes in huge networks, segmenting graphs into meaningful communities, modelling uncertainties in health trends for individual patients, controlling of complex systems, linking data bases with different levels of granularity in space and time, unbiased sampling, connecting with infrastructure involving sensors, and high performance computing, answers to these questions are the key to competitiveness and leadership in this field. The Big Data is usually modelled as point clouds in a high-dimensional space. One way to understand something about the data is to find a geometric object for which the data looks like a sampling of points. Then the geometric object is seen as an interpolation of the data. Main tool for studying of qualitative features of geometric objects is topology. Topology studies only properties of geometric objects which do not depend on the chosen coordinates, distance, but rather on intrinsic geometric properties of the objects.
Mining Big Data leads to a paradigm shift in machine learning methods for the supervised classification task. In particular, it concerns evolving data streams, where algorithms have to process considerable amounts of data using limited time and memory while analysing each incoming instance only once. Moreover, they face significant challenges caused by concept drift. The talk starts with taxonomy of concept drifts, focusing attention on differences between real and virtual drifts, their frequency, and concept transitions. Then, it overviews adaptive ensembles, which are quite often applied to evolving data streams. As many of these multiple classifiers are designed for adapting to a single type of drifts only, it is discussed how to construct new classifiers able to react to several types of drifts. Moreover, by studying properties of Accuracy Updated Ensemble, relations between block-based and on-line ensembles are discussed. Finally, the talk brings attention to open problems and identifies research challenges, also inspired by real world applications.
Many real-world classification problems have unbalanced classes, e.g., in fault detection and software defect prediction, where there are a large number of training examples for the normal class, but few for the abnormal classes. This talk gives an overview of some recent algorithms for dealing with class imbalance in machine learning, including ensemble approaches, sampling methods, evolutionary computation methods, and their combinations. First, we will discuss how diversity influences the classification performance, especially on the minority class, in ensemble classification algorithms. Then new ensemble algorithms are introduced and evaluated experimentally. Multi-class imbalance will be analysed and considered. The combination of ensemble learning and sampling techniques for dealing with class imbalance will be presented. Finally, we consider a new problem --- online class imbalance learning of data streams, where the majority and minority classes are not pre-defined and have to be learned and detected online. Some results are presented to demonstrate the effectiveness of the proposed algorithm.
Communication networks and the Internet are evolving from simple best effort packet forwarding-based infrastructures towards advanced platforms providing a rich set of various services like, e.g., cloud computing, content delivery networks, IP television, video streaming, Internet of Things. The complexity of such systems combined with ever-increasing demand for bandwidth, connection quality, and end-to-end interactivity make the optimization of communication networks very challenging. At the same time, the computational power of CPUs and GPUs is increasing every year opening new opportunities in developing more and more sophisticated algorithms. Apart from mathematical programming, which is a well-established tool for network modelling and optimization, such computationally intelligent methods as meta-heuristics are gaining much interest due to their effectiveness in providing either optimal or near-optima solutions to difficult optimization problems. Other recent trends that stimulate deployment of efficient optimization approaches for communication networks are software defined networking (SDN), network function virtualization (NFV) and Big Data concepts. The motivation of this Special Session is to provide a platform to present new ideas, achievements, and implementations in all aspects related to the application of intelligent computational techniques in network optimization. Research topics include network planning and operation, cross- and multi-layer design, network survivability, resource allocation, quality of service guarantees, energy efficiency aspects, service location, among others.
This special session aims at joining the contemporary innovations about knowledge discovery and data mining. Therefore, the presentation of works tackling theoretical issues and applications, from industry or academia, on machine learning is welcomed. The problem complexity is increasing and there is a wide variety of approaches to deal with it. Also a number of factors may be considered for the suitable choice of a concrete learner.
Discrete Event System (DES) simulation techniques provide tools enable verification and validation of models of wide range of systems from manufacturing and material handling systems, through city traffic systems, till computer systems, and computer networks. The simulation techniques supported by different formal frameworks ranging from classical discrete event simulation to declarative and mathematical programming ones can be employed in the course of modeling and performance of different DES such as flexible manufacturing systems, parallel processing systems, railway traffic networks and so on.
The methods employed are usually based on mathematical programming techniques, such as linear programming or quadratic programming; other problems, however, cannot be modeled using these techniques. The most general solvers that are applicable to a wide range of machine learning and data mining problems embedded in DES simulation techniques s are now gathered in the area of constraint programming. The most general solvers that are applicable to a wide range of machine learning and data mining problems embedded in DES simulation techniques s are now gathered in the area of constraint programming. In constraint programming, the user specifies the model, that is, the set of constraints to be satisfied and constraint solvers generate solutions. This raises the questions as to whether it is possible to (semi)-automatically learn such constraints or their formulations from data and experience, and then how standard constraint-programming techniques can be used in data mining and machine learning.
In that context the session provides an excellent forum for scientists, researchers, engineers and industrial practitioners to meet and share experiences, theoretical knowledge or application examples based on the latest trends in different kinds of DES as well as future directions and trends in dealing with the growing demand for novel large-scale robust simulation-driven modeling frameworks. Authors are invited to submit full papers describing original research work associated with artificial intelligence solutions for DES-like modeling and performance evaluation related problems (arising in transportation, telecommunication, manufacturing and other kinds of DES) in areas including, but not limited to,
This special session focuses on applications implemented via techiques under the umbrella of any perspective of artificial intelligence (reasoning, planning, machine learning). Therefore, the presentation of works tackling practical issues, especially from industrial or academic environments, is suitable. Open-source, commercial, research or demo implementations of one or a set of tools in a particular scope are especially useful for this session. We encourage to submit very recent applications and if possible unprecedented.
The first International Conference on Intelligent Data Engineering and Automated Learning (IDEAL) took place in 1998 in Hong Kong, China. Since 2002, it became an annual event. The proceedings are published by Springer in the popular and highly-recognized series Lecture Notes in Computer Science (LNCS).
Below you can find a list of the past IDEAL events. These events have also been indexed in Web of Science Proceedings, Scopus, Google Scholar, DBLP.
Manuscripts must be written in English and comply with the format of the LN CS/LNAI Series. The default page limit is 8 pages.
All submissions will be refereed by experts in the field based on originality, significance, quality and clarity. All contributions must be original, must not have been published elsewhere and must not be intended to be published elsewhere (conference or journal) during the review period. Accepted papers will be included in IDEAL 2015 Proceedings in the LNCS Series. All accepted papers will be considered for extension for possible publication in journal special issues dedicated to this conference.
Series: Lecture Notes in Computer Science Series Editors: Hutchison, D., Kanade, T., Kittler, J., Kleinberg, J.M., Kobsa, A., Mattern, F., Mitchell, J.C., Naor, M., Nierstrasz, O., Pandu Rangan, C., Steffen, B., Terzopoulos, D., Tygar, D., Weikum, G.
The series Lecture Notes in Computer Science (LNCS), including its subseries Lecture Notes in Artificial Intelligence (LNAI) and Lecture Notes in Bioinformatics (LNBI), has established itself as a medium for the publication of new developments in computer science and information technology research and teaching – quickly, informally, and at a high level. The cornerstone of LNCS’s editorial policy is its unwavering commitment to report the latest results from all areas of computer science and information technology research, development, and education. LNCS has always enjoyed close cooperation with the computer science R & D community, with numerous renowned academics, and with prestigious institutes and learned societies. Our mission is to serve this community by providing a most valuable publication service. LNCS commenced publication in 1973 and quite rapidly attracted attention, not least because of its thus far unprecedented publication turnaround times. The 1980s and 1990s witnessed a substantial growth in the series, particularly in terms of volumes published. In the late 1990s we developed a systematic approach to providing LNCS in a full-text electronic version, in parallel to the printed books. Another new feature introduced in the late 1990s was the conceptualization of a couple of color-cover sublines. Still, original research results reported in proceedings and postproceedings remain the core of LNCS.
The nominal length of each paper is 8 pages and any paper that exceeds this limit will incur extra page charges on the registration form. All papers must not exceed 12 pages. At least one author of each accepted paper must register before the 17th July with payment proof in order for the paper being included in the Proceedings.
Please follow strictly the author instructions of Springer-Verlag when preparing the final version.
Our publisher has recently introduced an extra control loop: once data processing is finished, they will contact all corresponding authors and ask them to check their papers. We expect this to happen shortly before the printing of the proceedings. At that time your quick interaction with Springer-Verlag will be greatly appreciated.
It is sufficient for one of the authors to sign the copyright form. You can scan the form into PDF or any other standard image format.
Please indicate your paper ID on the form.Go to submission system
IDEAL 2015 will take place in Wroclaw, the chief city in south-western Poland. Over the centuries the city has been either part of Poland, Bohemia, Austria, Prussia or Germany. Wroclaw is an excellent example of a multicultural metropolis situated at the interface of ethnically diverse areas. The architecture of Wroclaw reflects its history, which dates back almost one thousand years.
The conference will take place in Conference Center of Wroclaw University of Technology.
The University of Technology
D 20 building
Janiszewskiego 8 street