Boris Kovalerchuk | Central Washington University (original) (raw)

Uploads

Papers by Boris Kovalerchuk

Research paper thumbnail of Finding your way through blogspace: Using semantics for cross-domain blog analysis

AAAI symposium on computational approaches to analyzing weblogs, Mar 1, 2006

Blogspace is one of the most dynamic areas of today's Internet, and it is increasingly recog... more Blogspace is one of the most dynamic areas of today's Internet, and it is increasingly recognised that blogs are much more than “meaningless chatter”. Many syntaxbased approaches exist to analyse the text and the network structure between blogs. While this is very helpful for purposes such as the detection of discussion bursts concerning uniquely-named topics (eg, a book, product, or person), it is insufficient for understanding blogs discussing new phenomena in different wordings, or for finding and explaining ...

Research paper thumbnail of Comparison and Fusion of Methods and Future Research

Intelligent systems reference library, 2018

In this chapter, we first compare General Line Coordinates with other visualization methods that ... more In this chapter, we first compare General Line Coordinates with other visualization methods that were not analyzed in the previous chapters yet. Then we summarize some comparisons that were presented in other chapters. Next, the hybrid approach that fuses General Line Coordinates with other methods is summarized along with the outline of the future research.

Research paper thumbnail of Visual Knowledge Discovery with General Line Coordinates

arXiv (Cornell University), May 28, 2023

Research paper thumbnail of Parallel Coordinates for Discovery of Interpretable Machine Learning Models

arXiv (Cornell University), May 28, 2023

Research paper thumbnail of Explainable Machine Learning for Categorical and Mixed Data with Lossless Visualization

arXiv (Cornell University), May 28, 2023

Research paper thumbnail of Full High-Dimensional Intelligible Learning In 2-D Lossless Visualization Space

arXiv (Cornell University), May 28, 2023

Research paper thumbnail of Survey of explainable machine learning with visual and granular methods beyond quasi-explanations

arXiv (Cornell University), Sep 21, 2020

Research paper thumbnail of Visual Knowledge Discovery and Machine Learning

Intelligent systems reference library, 2018

Research paper thumbnail of Explainable Mixed Data Representation and Lossless Visualization Toolkit for Knowledge Discovery

Research paper thumbnail of On Foundation of Fuzzy Decision Making

IFAC Proceedings Volumes, Jul 1, 1983

Research paper thumbnail of Retrieval of the maximum upper zero for minimizing the number of attributes in regression analysis

U.S.S.R. computational mathematics and mathematical physics, 1984

Abstract The reduction of certain problems in linear regression analysis to problems of evaluatio... more Abstract The reduction of certain problems in linear regression analysis to problems of evaluation or retrieval of the maximum upper zero is demonstrated. The optimal algorithm B 2 for the retrieval of the maximum upper zero is presented. This algorithm is optimal with respect to computation time and memory requirements in the sense of the Shannon function. A method is given for taking account of the knowledge of experts in order to speed up to the computations involved in the regression analysis problems under consideration.

Research paper thumbnail of Visual and spatial analysis : advances in data mining, reasoning, and problem solving

Research paper thumbnail of Proceedings of the second IASTED International Conference on Computational Intelligence, November 20-22, 2006, San Francisco, California, USA

Research paper thumbnail of Interpretable Machine Learning for Self-Service High-Risk Decision-Making

arXiv (Cornell University), May 9, 2022

Research paper thumbnail of Discovering Visual Features and Shape Perception Capabilities in GLC

Intelligent systems reference library, 2018

The analysis of data visualized with different GLCs in previous chapters shows that multiple visu... more The analysis of data visualized with different GLCs in previous chapters shows that multiple visual features could be estimated for each individual graph. This chapter evaluates efficiency of the human visual system in discovering discriminating features for n-D data classification learning tasks in Closed Contour Paired Coordinates (traditional Stars/Radial Coordinates, and CPC Stars) in comparison with Parallel Coordinates. It is shown that Closed Contour Paired Coordinates are capable representing data in 14-D, 48-D, 96-D, 160-D, 170-D, and 192-D, where humans are capable discovering features and patterns for classification these high-dimensional data. The chapter concludes with a description of the cooperative visualization approach to enhance Knowledge Discovery in solving Data Mining/Machine Learning tasks.

Research paper thumbnail of Interactive Visual Classification, Clustering and Dimension Reduction with GLC-L

Intelligent systems reference library, 2018

The goal of this chapter is to present a new interactive visual machine learning system for solvi... more The goal of this chapter is to present a new interactive visual machine learning system for solving supervised learning classification tasks based on a GLC-L visualization algorithm and associated interactive and automatic algorithms GLC-IL, GLC-AL and GLC-DRL for discovery of linear and non-linear relations and dimension reduction. Classification and dimension reduction tasks from three domains, image processing, computer-aided medical diagnostics and finance (stock market), are used to illustrate this method.

Research paper thumbnail of Bringing a Vector/Image Conflation Tool to the Commercial Market

Research paper thumbnail of Knowledge Discovery and Machine Learning for Investment Strategy with CPC

Intelligent systems reference library, 2018

Knowledge discovery is an important aspect of human cognition. The advantage of the visual approa... more Knowledge discovery is an important aspect of human cognition. The advantage of the visual approach is in the opportunity of solving easier perceptual tasks instead of complex cognitive tasks. H owever for cognitive tasks such as financial investment decision making, this opportunity faces the challenge that financial data are abstract multidimensional and multivariate, i.e., outside of traditional visual perception in 2-D or 3-D world. This chapter presents a visualization-inspired approach to find an investment strategy based on pattern discovery in multidimensional space. It is shown that the new lossless Collocated Paired Coordinates approach is an effective instrument for such inspiration for the investment strategy. It is coming from the two levels of the approach. The first level involves examining the best 4D and 6D coordinate systems to build 2D or 3D visualization spaces. The second level involves learning parameters of attributes in each selected space. A key role of the CPC here is in helping to find the best locations (squares in 2D or cubes in 3D) to open long or short positions, respectively. The main positive result is finding the property in the visualization space that leads to a profitable investment decision for EUR/USD foreign exchange market. The strategy is ready for implementation in algotrading m ode.

Research paper thumbnail of Visual Data Mining in Closed Contour Coordinates

IS&T International Symposium on Electronic Imaging Science and Technology, Feb 14, 2016

Research paper thumbnail of Agents in Quantum and Neural Uncertainty

IGI Global eBooks, Sep 7, 2010

Research paper thumbnail of Finding your way through blogspace: Using semantics for cross-domain blog analysis

AAAI symposium on computational approaches to analyzing weblogs, Mar 1, 2006

Blogspace is one of the most dynamic areas of today's Internet, and it is increasingly recog... more Blogspace is one of the most dynamic areas of today's Internet, and it is increasingly recognised that blogs are much more than “meaningless chatter”. Many syntaxbased approaches exist to analyse the text and the network structure between blogs. While this is very helpful for purposes such as the detection of discussion bursts concerning uniquely-named topics (eg, a book, product, or person), it is insufficient for understanding blogs discussing new phenomena in different wordings, or for finding and explaining ...

Research paper thumbnail of Comparison and Fusion of Methods and Future Research

Intelligent systems reference library, 2018

In this chapter, we first compare General Line Coordinates with other visualization methods that ... more In this chapter, we first compare General Line Coordinates with other visualization methods that were not analyzed in the previous chapters yet. Then we summarize some comparisons that were presented in other chapters. Next, the hybrid approach that fuses General Line Coordinates with other methods is summarized along with the outline of the future research.

Research paper thumbnail of Visual Knowledge Discovery with General Line Coordinates

arXiv (Cornell University), May 28, 2023

Research paper thumbnail of Parallel Coordinates for Discovery of Interpretable Machine Learning Models

arXiv (Cornell University), May 28, 2023

Research paper thumbnail of Explainable Machine Learning for Categorical and Mixed Data with Lossless Visualization

arXiv (Cornell University), May 28, 2023

Research paper thumbnail of Full High-Dimensional Intelligible Learning In 2-D Lossless Visualization Space

arXiv (Cornell University), May 28, 2023

Research paper thumbnail of Survey of explainable machine learning with visual and granular methods beyond quasi-explanations

arXiv (Cornell University), Sep 21, 2020

Research paper thumbnail of Visual Knowledge Discovery and Machine Learning

Intelligent systems reference library, 2018

Research paper thumbnail of Explainable Mixed Data Representation and Lossless Visualization Toolkit for Knowledge Discovery

Research paper thumbnail of On Foundation of Fuzzy Decision Making

IFAC Proceedings Volumes, Jul 1, 1983

Research paper thumbnail of Retrieval of the maximum upper zero for minimizing the number of attributes in regression analysis

U.S.S.R. computational mathematics and mathematical physics, 1984

Abstract The reduction of certain problems in linear regression analysis to problems of evaluatio... more Abstract The reduction of certain problems in linear regression analysis to problems of evaluation or retrieval of the maximum upper zero is demonstrated. The optimal algorithm B 2 for the retrieval of the maximum upper zero is presented. This algorithm is optimal with respect to computation time and memory requirements in the sense of the Shannon function. A method is given for taking account of the knowledge of experts in order to speed up to the computations involved in the regression analysis problems under consideration.

Research paper thumbnail of Visual and spatial analysis : advances in data mining, reasoning, and problem solving

Research paper thumbnail of Proceedings of the second IASTED International Conference on Computational Intelligence, November 20-22, 2006, San Francisco, California, USA

Research paper thumbnail of Interpretable Machine Learning for Self-Service High-Risk Decision-Making

arXiv (Cornell University), May 9, 2022

Research paper thumbnail of Discovering Visual Features and Shape Perception Capabilities in GLC

Intelligent systems reference library, 2018

The analysis of data visualized with different GLCs in previous chapters shows that multiple visu... more The analysis of data visualized with different GLCs in previous chapters shows that multiple visual features could be estimated for each individual graph. This chapter evaluates efficiency of the human visual system in discovering discriminating features for n-D data classification learning tasks in Closed Contour Paired Coordinates (traditional Stars/Radial Coordinates, and CPC Stars) in comparison with Parallel Coordinates. It is shown that Closed Contour Paired Coordinates are capable representing data in 14-D, 48-D, 96-D, 160-D, 170-D, and 192-D, where humans are capable discovering features and patterns for classification these high-dimensional data. The chapter concludes with a description of the cooperative visualization approach to enhance Knowledge Discovery in solving Data Mining/Machine Learning tasks.

Research paper thumbnail of Interactive Visual Classification, Clustering and Dimension Reduction with GLC-L

Intelligent systems reference library, 2018

The goal of this chapter is to present a new interactive visual machine learning system for solvi... more The goal of this chapter is to present a new interactive visual machine learning system for solving supervised learning classification tasks based on a GLC-L visualization algorithm and associated interactive and automatic algorithms GLC-IL, GLC-AL and GLC-DRL for discovery of linear and non-linear relations and dimension reduction. Classification and dimension reduction tasks from three domains, image processing, computer-aided medical diagnostics and finance (stock market), are used to illustrate this method.

Research paper thumbnail of Bringing a Vector/Image Conflation Tool to the Commercial Market

Research paper thumbnail of Knowledge Discovery and Machine Learning for Investment Strategy with CPC

Intelligent systems reference library, 2018

Knowledge discovery is an important aspect of human cognition. The advantage of the visual approa... more Knowledge discovery is an important aspect of human cognition. The advantage of the visual approach is in the opportunity of solving easier perceptual tasks instead of complex cognitive tasks. H owever for cognitive tasks such as financial investment decision making, this opportunity faces the challenge that financial data are abstract multidimensional and multivariate, i.e., outside of traditional visual perception in 2-D or 3-D world. This chapter presents a visualization-inspired approach to find an investment strategy based on pattern discovery in multidimensional space. It is shown that the new lossless Collocated Paired Coordinates approach is an effective instrument for such inspiration for the investment strategy. It is coming from the two levels of the approach. The first level involves examining the best 4D and 6D coordinate systems to build 2D or 3D visualization spaces. The second level involves learning parameters of attributes in each selected space. A key role of the CPC here is in helping to find the best locations (squares in 2D or cubes in 3D) to open long or short positions, respectively. The main positive result is finding the property in the visualization space that leads to a profitable investment decision for EUR/USD foreign exchange market. The strategy is ready for implementation in algotrading m ode.

Research paper thumbnail of Visual Data Mining in Closed Contour Coordinates

IS&T International Symposium on Electronic Imaging Science and Technology, Feb 14, 2016

Research paper thumbnail of Agents in Quantum and Neural Uncertainty

IGI Global eBooks, Sep 7, 2010