Data Mining, Statistics and Biometrics
Generalized Models from Beta(p, 2) Densities with Strong Allee Effect: Dynamical Approach
Sandra M. Aleixo, J. Leonel Rocha
Abstract
A dynamical approach to study the behaviour of generalized populational growth models from Beta(p, 2) densities, with strong Allee effect, is presented. The dynamical analysis of the respective unimodal maps is performed using symbolic dynamics techniques. The complexity of the correspondent discrete dynamical systems is measured in terms of topological entropy. Different populational dynamics regimes are obtained when the intrinsic growth rates are modified: extinction, bistability, chaotic semistability and essential extinction.
Keywords
Beta(p, 2) densities, Allee effect, symbolic dynamics, topological entropy
Full text is available at IEEE Xplore digital library.
A Framework for Data Representation, Processing, and Dimensionality Reduction with the Best-Rank Tensor Decomposition
Boguslaw Cyganek
Abstract
The paper addresses the problem of efficient multi-dimensional data representation, processing and dimensionality reduction. For this purpose the framework for the best rank-R tensor decomposition is presented. This allows any multi-dimensional data reduction in accordance with chosen ranks. Since computations of tensor decomposition require floating-point operations, we propose special data scaling procedure to allow memory efficient representation in the fixed-point representation. The proposed method is exemplified with processing of the monochrome and color video sequences. The method shows promising results and can be easily applied to other types of multidimensional data.
Keywords
Tensor best rank-R decomposition, dimensionality reduction, data mining
Full text is available at IEEE Xplore digital library.
BICIKELJ: Environmental Data Mining on the Bicycle
Lorand Dali, Dunja Mladenić
Abstract
The paper describes an approach to environmental data mining on a problem of public bicycle system. Environmental infromation including weather information and the bicycle station location is considered, as well as the time of the day and the number of bicycles at each hour at each station. Data mining methods are applied to predict the number of available bicycles at a certain time at a given station, to describe situations of empty and full stations and, to estimate the most common paths and usage patterns. The experiemntal evaluation of the proposed approach on real-world data gives promissing results, with machine learning mehtods achieving significantely lower error on predicting the number of bicycles compared to a baseline method.
Keywords
bicycle sharing, data mining, traffic models
Full text is available at IEEE Xplore digital library.
Analysis of Heavy Metals Concentration in Wastewater along Highways in Croatia
Dijana Grd, Jasminka Dobsa, Vesna Simunic-Meznaric, Teuta Tompic
Abstract
In this paper we have analysed concetrations of heavy metals (lead, copper, nickel, zink, mercury, cadmium, and chromium) in wastewater along highways in Croatia. We have used standard statistical methods: analysis of variance, Kruskal-Wallis test and principal analysis. Analysis of variance and Kruskal-Wallis test were used to detect factors that influence the concentration of lead, copper, nickel, and zink in wastewater. We have investigated the influence of the highway section, the side of a highway, and the influence of the season of the year. Principal components were used to identify groups of elements with similar characteristics in wastewater.
Keywords
Wastewater, heavy metals, highway, ANOVA, Kruskal-Wallis test, principal components analysis
Full text is available at IEEE Xplore digital library.
Web Connectivity of Higher Education Institutions within a Country: A Comparison of Croatia and Ireland
Tomislav Jakopec, Cathal Hoare, Adrian O’Riordan
Abstract
This study’s aim was to carry out a comparative analysis of hyperlink connectivity between higher education and research sites within two European countries. This is analogous to citation analysis but Website interlinking is often less formal and largely unregulated. A categorization of both Ireland’s and Croatia’s educational and research domains and sub-domains served as input to a customized Web crawler to compute relevant connectivity measures. Trends and differences in both countries’ interconnections were observed.
Keywords
Webometrics, hyperlinking, Web crawling, Websites
Full text is available at IEEE Xplore digital library.
Knowledge Visualization in Biometric Face Recognition on Two-dimensional Images
Koruga Petra, Baca Miroslav, Fotak Tomislav
Abstract
Biometric face recognition is one of the fields of computer vision. The part of biometric person identification has been extensivly researched. This resulted in a large number of algorithms used for person identification. This paper gives an overview and classification of those algorithms. Because of large number of algorithms, knowledge visualization is used for easier understanding of structure and connections between those algorithms.
Keywords
knowledge, visualization, biometrics, face recognition, algorithms
Full text is available at IEEE Xplore digital library.
Improved Bisector Pruning for Uncertain Data Mining
Ivica Lukić, Mirko Kohler, Ninoslav Slavek
Abstract
Uncertain data mining is well studied and very challenging task. This paper is concentrated on clustering uncertain objects with location uncertainty. Uncertain locations are described by probability density function (PDF). Number of uncertain objects can be very large and obtaining quality result within reasonable time is a challenging task. Basic clustering method is UK-means, in which all expected distances (ED) from objects to clusters are calculated. Thus UK-means is inefficient. To avoid ED calculations various pruning methods are proposed. The pruning methods are significantly more effective than UK-means method. In this paper, Improved Bisector pruning method is proposed as an improvement of clustering process.
Keywords
Clustering, data mining, expected distance, pruning, uncertain data
Full text is available at IEEE Xplore digital library.
Two-Way Collaborative Filtering on Semantically Enhanced Movie Ratings
Hasan Ogul, Emrah Ekmekciler
Abstract
A key step in recommendation systems is to estimate if a user would likely enjoy an item who has not considered yet. In this study, a new framework is defined to predict user ratings on new items from previously given ratings by other users. The systems has two major steps: (1) Enhancing available data based on semantic content to get a full item-user matrix, and (2) Predicting the unknown rating using an integrated feature set of "other ratings given by the same user" and "other ratings given to the same item". This allows the classifier to consider both user similarities and item similarities simultaneously. The system is shown to outperform existing methods in terms of prediction accuracy on a benchmark movie dataset.
Keywords
Recommendation system, contentboosted collaborative filtering, movie rating, data mining
Full text is available at IEEE Xplore digital library.
Educational Benefit Factors of Convergence of Academic Libraries’ E-services and Learning Management System
Anita Papić, Boris Badurina
Abstract
Concerning the needs of today’s university oriented to virtual university development, aim of this paper is to explore convergence of academic libraries' e-services and learning management system in regards to quality assurance in higher education. A special attention in this research is given to students. Namely, research results based on exploring students’ attitudes towards libraries' e-services integrated in the subject contents within learning management system revealed four educational benefit factors of convergence of academic libraries' e-services and learning management systems which are presented in this paper.
Keywords
Academic libraries, factor analysis, learning management system, students, virtual university
Full text is available at IEEE Xplore digital library.
Two Factor Authentication using EEG Augmented Passwords
Ivan Švogor, Tonimir Kišasondi
Abstract
The current research with EEG devices in the user authentication context has some deficiencies that address expensive equipment, the requirement of laboratory conditions and applicability. In this paper we address this issue by using widely available and inexpensive EEG device to verify its capability for authentication. As a part of this research, we developed two phase authentication that enables users to enhance their password with the mental state by breaking the password into smaller elements, marry them with mental state, and generate one time pad for a secure session.
Keywords
EEG, BCI, password augmentation, security, brain waves, one time pass, two phase
Full text is available at IEEE Xplore digital library.