Urszula Stańczyk

Man-machine interactions

Keynote Talks.- Speech Man-Machine Communication.- Stochastic Effects in Signaling Pathways in Cells: Interaction between Visualization and Modeling.- Rough-Granular Computing in Human-Centric Information Processing.- Discovering Affinities between Perceptual Granules.- Human-Computer Interactions.- A Psycholinguistic Model of Man-Machine Interactions Based on Needs of Human Personality.- Adaptable Graphical User Interfaces for Player-Based Applications.- Case-Based Reasoning Model in Process of Emergency Management.- Enterprise Ontology According to Roman Ingarden Formal Ontology.- Hand Shape Recognition for Human-Computer Interaction.- System for Knowledge Mining in Data from Interactions between User and Application.- Computational Techniques in Biosciences.- Analyze of Maldi-TOF Proteomic Spectra with Usage of Mixture of Gaussian Distributions.- Energy Properties of Protein Structures in the Analysis of the Human RAB5A Cellular Activity.- Fuzzy Weighted Averaging of Biomedical Signal Using Bayesian Inference.- Fuzzy Clustering and Gene Ontology Based Decision Rules for Identification and Description of Gene Groups.- Estimation of the Number of Primordial Genes in a Compartment Model of RNA World.- Quasi Dominance Rough Set Approach in Testing for Traces of Natural Selection at Molecular Level.- Decision Support, Rule Inferrence and Representation.- The Way of Rules Representation in Composited Knowledge Bases.- Clustering of Partial Decision Rules.- Decision Trees Constructing over Multiple Data Streams.- Decision Tree Induction Methods for Distributed Environment.- Extensions of Multistage Decision Transition Systems: The Rough Set Perspective.- Emotion Recognition Based on Dynamic Ensemble Feature Selection.- Rough Fuzzy Investigations.- On Construction of Partial Association Rules with Weights.- Fuzzy Rough Entropy Clustering Algorithm Parametrization.- Data Grouping Process in Extended SQL Language Containing Fuzzy Elements.- Rough Sets in Flux: Crispings and Change.- Simplification of Neuro-Fuzzy Models.- Fuzzy Weighted Averaging Using Criterion Function Minimization.- Approximate String Matching by Fuzzy Automata.- Remark on Membership Functions in Neuro-Fuzzy Systems.- Capacity-Based Definite Rough Integral and Its Application.- Advances in Classification Methods.- Classifier Models in Intelligent CAPP Systems.- Classification Algorithms Based on Templates Decision Rules.- Fast Orthogonal Neural Network for Adaptive Fourier Amplitude Spectrum Computation in Classification Problems.- Relative Reduct-Based Selection of Features for ANN Classifier.- Enhanced Ontology Based Profile Comparison Mechanism for Better Recommendation.- Privacy Preserving Classification for Ordered Attributes.- Incorporating Detractors into SVM Classification.- Bayes Multistage Classifier and Boosted C4.5 Algorithm in Acute Abdominal Pain Diagnosis.- Pattern Recognition and Signal Processing.- Skrybot - A System for Automatic Speech Recognition of Polish Language.- Speaker Verification Based on Fuzzy Classifier.- Support Vector Classifier with Linguistic Interpretation of the Kernel Matrix in Speaker Verification.- Application of Discriminant Analysis to Distinction of Musical Instruments on the Basis of Selected Sound Parameters.- Computer Vision, Image Analysis and Virtual Reality.- Spatial Color Distribution Based Indexing and Retrieval Scheme.- Synthesis of Static Medical Images with an Active Shape Model.- New Method for Personalization of Avatar Animation.- Multidimensional Labyrinth - Multidimensional Virtual Reality.- Shape Recognition Using Partitioned Iterated Function Systems.- Computer Vision Support for the Orthodontic Diagnosis.- From Museum Exhibits to 3D Models.- Advances in Algorithmics.- A Method for Automatic Standardization of Text Attributes without Reference Data Sets.- Internal Conflict-Free Projection Sets.- The Comparison of an Adapted Evolutionary Algorithm with the Invasive Weed Optimization Algorithm Based on the Problem of Predetermining the Progress of Distributed Data Merging Process.- Cumulation of Pheromone Values in Web Searching Algorithm.- Mining for Unconnected Frequent Graphs with Direct Subgraph Isomorphism Tests.- Numerical Evaluation of the Random Walk Search Algorithm.- On Two Variants of the Longest Increasing Subsequence Problem.- Computing the Longest Common Transposition-Invariant Subsequence with GPU.- Databases and Data Warehousing.- Usage of the Universal Object Model in Database Schemas Comparison and Integration.- Computational Model for Efficient Processing of Geofield Queries.- Applying Advanced Methods of Query Selectivity Estimation in Oracle DBMS.- How to Efficiently Generate PNR Representation of a Qualitative Geofield.- RBTAT: Red-Black Table Aggregate Tree.- Performing Range Aggregate Queries in Stream Data Warehouse.- LVA-Index: An Efficient Way to Determine Nearest Neighbors.- Embedded Systems Applications.- Basic Component of Computational Intelligence for IRB-1400 Robots.- Factors Having Influence upon Efficiency of an Integrated Wired-Wireless Network.- FFT Based EMG Signals Analysis on FPGAs for Dexterous Hand Prosthesis Control.- The VHDL Implementation of Reconfigurable MIPS Processor.- Time Optimal Target Following by a Mobile Vehicle.- Improving Quality of Satellite Navigation Devices.

Feature Selection for Data and Pattern Recognition

This research book provides the reader with a selection of high-quality texts dedicated to current progress, new developments and research trends in feature selection for data and pattern recognition. Even though it has been the subject of interest for some time, feature selection remains one of actively pursued avenues of investigations due to its importance and bearing upon other problems and tasks. This volume points to a number of advances topically subdivided into four parts: estimation of importance of characteristic features, their relevance, dependencies, weighting and ranking; rough set approach to attribute reduction with focus on relative reducts; construction of rules and their evaluation; and data- and domain-oriented methodologies.

Decision rule length as a basis for evaluation of attribute relevance

Knowledge discovered from data can be represented in a form of decision rules, consisting of required conditions and decisions to which they lead. The quality of rules is usually considered in terms of some quantitative measures such as confidence, support or length. Depending on all these parameters the constructed classifiers can greatly vary in the predictive accuracy and the size of their structure. Both these elements depend strongly on the choice of characteristic features, which can be found by some independent feature selection procedure, but also by applying a wrapper model. In the wrapper model the classifier and its parameters are used to evaluate the importance of attributes. In the paper there are proposed measures of attribute relevance based on rule lengths. The usefulness of the described methodology is shown for rule-based classifiers, obtained through Dominance-Based Rough Set Approach, and a connectionist solution implemented with Artificial Neural Networks, both employed in the task of authorship attribution.

DRSA decision algorithm analysis in stylometric processing of literary texts

When the indiscernibility relation, fundamental to Classical Rough Set Approach, is substituted with dominance relation, it results in Dominance-Based Rough Set Approach to data analysis. It enables support not only for nominal classification tasks, but also when ordinal properties on attribute values can be observed [1], making DRSA methodology well suited for stylometric processing of texts. Stylometry involves handling quantitative features of texts leading to characterisation of authors to the point of recognition of their individual writing styles. As always, selection of attributes is crucial to classification accuracy, as is the construction of a decision algorithm. When minimal cover gives unsatisfactory results, and all rules on examples algorithm returns very high number of rules, usually constraints are imposed by selection of some reduct and limiting the decision algorithm by including within it only rules with certain support. However, reducts are typically numerous and within them some of conditional attributes are used more often than others, which is also true for conditions specified by decision rules. The paper presents observations how the frequency of usage for features reflects on the performance of decision algorithms resulting from selection of rules with conditional attributes exploited most and least often.

Dominance-Based Rough Set Approach Employed in Search of Authorial Invariants

Mechanisms for interpretation and manipulation of data required in decision support systems quite often deal with cases when knowledge is uncertain or incomplete. Dominance-based Rough Set Approach is an example of such methodology, dedicated to analysis of data with ordinal properties, with dominance relation substituting that of indiscernibility of objects as defined by Classical Rough Set Approach. The paper presents application of DRSA procedures to the problem of stylometric analysis of literary texts, which by the notion of its primary concept of authorial invariants enables to identify authors of unattributed or disputed texts.

Ranking of characteristic features in combined wrapper approaches to selection

AbstractThe performance of a classification system of any type can suffer from irrelevant or redundant data, contained in characteristic features that describe objects of the universe. To estimate relevance of attributes and select their subset for a constructed classifier typically either a filter, wrapper, or an embedded approach, is implemented. The paper presents a combined wrapper framework, where in a pre-processing step, a ranking of variables is established by a simple wrapper model employing sequential backward search procedure. Next, another predictor exploits this resulting ordering of features in their reduction. The proposed methodology is illustrated firstly for a binary classification task of authorship attribution from stylometric domain, and then for additional verification for a waveform dataset from UCI machine learning repository.

Indiscernibility Relation for Continuous Attributes: Application in Image Recognition

The paper presents an application of rough sets in a problem defined for the continuous feature space used by hybrid, high speed, pattern recognition system. The feature extraction part of this system is built as a holographic ring-wedge detector based on binary grating. Such feature extractor can be optimized and we apply for this purpose automatic knowledge acquisition and processing. Features from optimized extractor are then classified with the use of probabilistic neural network classifier. The methodology, proposed by one of the authors in earlier works, has been further enhanced here by application of modified indiscernibility relation. Modified version of this relation makes possible natural application of discrete type rough knowledge representation to problems defined in continuous space. We present an application of modified indiscernibility relation in the domain of image recognition.

Rough set-based analysis of characteristic features for ANN classifier

Selection of characteristic features for a classification task is always crucial to high recognition ratio, regardlessly of the particular processing technique applied Most methodologies offer some inherent mechanisms of dimension reduction that lead to expression of available data in more succinct way, however, combining elements of distinctively different approaches to data analysis brings interesting conclusions as to the role of particular features and their influence on the power of the resulting classifier The paper presents research on such fusion of processing techniques, namely employing rough set based analysis of features for ANN classifier within stylometric studies on writing styles.

Relative Reduct-Based Selection of Features for ANN Classifier

Artificial neural networks hold the established position of efficient classifiers used in decision support systems, yet to be efficient an ANN-based classifier requires careful selection of features. The excessive number of conditional attributes is not a guarantee of high classification accuracy, it means gathering and storing more data, and increasing the size of the network. Also the implementation of the trained network can become complex and the classification process takes more time. This line of reasoning leads to conclusion that the number of features should be reduced as far as possible without diminishing the power of the classifier. The paper presents investigations on attribute reduction process performed by exploiting the concept of reducts from the rough set theory and employed within stylometric analysis of literary texts that belongs with automatic categorisation tasks.

Reduct-based analysis of decision algorithms: application in computational stylistics

Computational stylistics focuses on description and quantifiable expression of linguistic styles of written documents that enables author characterisation, comparison, and attribution. It is a case when observation of subtle relationships in data sets is required, with domain knowledge uncertain. Therefore, techniques from the artificial intelligence area, such as Dominance-based Rough Set Approach (DRSA), are well suited to handle the problem. DRSA enables construction of a rule-based classifier consisting of decision rules, selection of which can greatly influence classification accuracy. The paper presents research on application of DRSA classifier in author recognition for literary texts, with considerations on the classifier performance based on an analysis of relative reducts, such subsets of features that maintain classification properties.