Dimitar Kazakov
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Dimitar Kazakov has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Daniel Kudenko
University of York
2 shared publications Eduardo Alonso
2 shared publications- Amer Alzaidi
University of York
1 shared publication - Julian Sedding
Department of Computer Science University of York Heslington YO10 5DD York United Kingdom
1 shared publication - Mark Bartlett
University of York
1 shared publication Matthew Butler
Artificial Intelligence Group in the Department of Computer Science, The University of York, Deramore Lane, Heslington, York, United Kingdom
1 shared publication
Latest Publications
- 18 December 2014
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 27 (4) , 1162-1166
SAX Discretization Does Not Guarantee Equiprobable Symbols
In time series analysis research, there is a strong interest in discrete representations of real valued data streams. One approach still considered state-of-the-art is the Symbolic Aggregate Approximation (SAX) algorithm. The interest of this paper concerns the SAX assumption of data being highly Gaussian and the use of the standard normal curve to choose partitions to discretize the data. The SAX approach chooses partitions on the standard normal curve that would produce an equal probability for each symbol. This procedure is generally valid as a time series is normalized to have μ = 0 and σ = 1. However, there exists a caveat to this assumption of equi-probability due to the intermediate step of Piecewise Aggregate Approximation (PAA). We show in this paper that when PAA is applied, the distribution of the data is altered, resulting in a shrinking standard deviation that is proportional to the number of points used to create a segment of the PAA representation and the degree of auto-correlation within the series. Data that exhibits statistically significant auto-correlation is less affected by this shrinking distribution. As the standard deviation of the data contracts, the mean remains the same, however the distribution is no longer standard normal and therefore the partitions based on the standard normal curve are no longer valid for the assumption of equal probability.
- 18 December 2014
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 27 (4) , 1162-1166
SAX Discretization Does Not Guarantee Equiprobable Symbols
In time series analysis research, there is a strong interest in discrete representations of real valued data streams. One approach still considered state-of-the-art is the Symbolic Aggregate Approximation (SAX) algorithm. The interest of this paper concerns the SAX assumption of data being highly Gaussian and the use of the standard normal curve to choose partitions to discretize the data. The SAX approach chooses partitions on the standard normal curve that would produce an equal probability for each symbol. This procedure is generally valid as a time series is normalized to have μ = 0 and σ = 1. However, there exists a caveat to this assumption of equi-probability due to the intermediate step of Piecewise Aggregate Approximation (PAA). We show in this paper that when PAA is applied, the distribution of the data is altered, resulting in a shrinking standard deviation that is proportional to the number of points used to create a segment of the PAA representation and the degree of auto-correlation within the series. Data that exhibits statistically significant auto-correlation is less affected by this shrinking distribution. As the standard deviation of the data contracts, the mean remains the same, however the distribution is no longer standard normal and therefore the partitions based on the standard normal curve are no longer valid for the assumption of equal probability.
- 1 January 2013
- Published by John Benjamins Publishing Company in Interaction Studies
- Vol. 14 (1) , 107-135
Evolutionary pressures promoting complexity in navigation and communication
This article presents results from simulations studying the hypothesis that mechanisms for landmark-based navigation could have served as preadaptations for compositional language. It is argued that sharing directions would significantly have helped bridge the gap between general and language-specific cognitive faculties. A number of different levels of navigational and communicative abilities are considered, resulting in a range of possible evolutionary paths. The selective pressures for, resp. against, increased complexity in either faculty are then evaluated for a range of environments.The study aims specifically to identify whether there is a viable evolutionary path leading to compositional language, and if so, under what circumstances.The results show that environmental conditions can render a step towards more complex communication either desirable or harmul, and suggest that very specific initial conditions and changes in the environment, resp. the ecological niche occupied, would have been needed to select for compositional language. Subject to these conditions, a (proto)language using order, but no hierarchical structure could evolve. This represents a middle ground, which brings closer hypotheses about syntax that have so far appeared difficult to reconcile.
- 1 January 2013
- Published by John Benjamins Publishing Company in Interaction Studies
- Vol. 14 (1) , 107-135
Evolutionary pressures promoting complexity in navigation and communication
This article presents results from simulations studying the hypothesis that mechanisms for landmark-based navigation could have served as preadaptations for compositional language. It is argued that sharing directions would significantly have helped bridge the gap between general and language-specific cognitive faculties. A number of different levels of navigational and communicative abilities are considered, resulting in a range of possible evolutionary paths. The selective pressures for, resp. against, increased complexity in either faculty are then evaluated for a range of environments.The study aims specifically to identify whether there is a viable evolutionary path leading to compositional language, and if so, under what circumstances.The results show that environmental conditions can render a step towards more complex communication either desirable or harmul, and suggest that very specific initial conditions and changes in the environment, resp. the ecological niche occupied, would have been needed to select for compositional language. Subject to these conditions, a (proto)language using order, but no hierarchical structure could evolve. This represents a middle ground, which brings closer hypotheses about syntax that have so far appeared difficult to reconcile.
- Open Access
Creating a level playing field for all symbols in a discretization
In time series analysis research there is a strong interest in discrete representations of real valued data streams. One approach that emerged over a decade ago and is still considered state-of-the-art is the Symbolic Aggregate Approximation algorithm. This discretization algorithm was the first symbolic approach that mapped a real-valued time series to a symbolic representation that was guaranteed to lower-bound Euclidean distance. The interest of this paper concerns the SAX assumption of data being highly Gaussian and the use of the standard normal curve to choose partitions to discretize the data. Though not necessarily, but generally, and certainly in its canonical form, the SAX approach chooses partitions on the standard normal curve that would produce an equal probability for each symbol in a finite alphabet to occur. This procedure is generally valid as a time series is normalized before the rest of the SAX algorithm is applied. However there exists a caveat to this assumption of equi-probability due to the intermediate step of Piecewise Aggregate Approximation (PAA). What we will show in this paper is that when PAA is applied the distribution of the data is indeed altered, resulting in a shrinking standard deviation that is proportional to the number of points used to create a segment of the PAA representation and the degree of auto-correlation within the series. Data that exhibits statistically significant auto-correlation is less affected by this shrinking distribution. As the standard deviation of the data contracts, the mean remains the same, however the distribution is no longer standard normal and therefore the partitions based on the standard normal curve are no longer valid for the assumption of equal probability.
Open AccessCreating a level playing field for all symbols in a discretization
In time series analysis research there is a strong interest in discrete representations of real valued data streams. One approach that emerged over a decade ago and is still considered state-of-the-art is the Symbolic Aggregate Approximation algorithm. This discretization algorithm was the first symbolic approach that mapped a real-valued time series to a symbolic representation that was guaranteed to lower-bound Euclidean distance. The interest of this paper concerns the SAX assumption of data being highly Gaussian and the use of the standard normal curve to choose partitions to discretize the data. Though not necessarily, but generally, and certainly in its canonical form, the SAX approach chooses partitions on the standard normal curve that would produce an equal probability for each symbol in a finite alphabet to occur. This procedure is generally valid as a time series is normalized before the rest of the SAX algorithm is applied. However there exists a caveat to this assumption of equi-probability due to the intermediate step of Piecewise Aggregate Approximation (PAA). What we will show in this paper is that when PAA is applied the distribution of the data is indeed altered, resulting in a shrinking standard deviation that is proportional to the number of points used to create a segment of the PAA representation and the degree of auto-correlation within the series. Data that exhibits statistically significant auto-correlation is less affected by this shrinking distribution. As the standard deviation of the data contracts, the mean remains the same, however the distribution is no longer standard normal and therefore the partitions based on the standard normal curve are no longer valid for the assumption of equal probability.
- 10 May 2012
- Published by Springer Nature in Cognitive Computation
- Vol. 4 (3) , 347-353
The Self-Cognisant Robot
- 10 May 2012
- Published by Springer Nature in Cognitive Computation
- Vol. 4 (3) , 347-353
The Self-Cognisant Robot
- 1 January 2012
- Published by Springer Nature
- p. 315-322
Earn More, Stay Legal: Novel Multi-agent Support for Islamic Banking
- 1 January 2012
- Published by Springer Nature
- p. 315-322
Earn More, Stay Legal: Novel Multi-agent Support for Islamic Banking
- 18 January 2011
- Published by IGI Global Scientific Publishing
Machine Learning for Agents and Multi-Agent Systems
Machine Learning for Agents and Multi-Agent Systems: 10.4018/978-1-59904-941-0.ch023: In order to be truly autonomous, agents need the ability to learn from and adapt to the environment and other agents. This chapter introduces key concepts of
- 18 January 2011
- Published by IGI Global Scientific Publishing
Machine Learning for Agents and Multi-Agent Systems
Machine Learning for Agents and Multi-Agent Systems: 10.4018/978-1-59904-941-0.ch023: In order to be truly autonomous, agents need the ability to learn from and adapt to the environment and other agents. This chapter introduces key concepts of
- 18 January 2011
- Published by IGI Global Scientific Publishing
- p. 1-26
Machine Learning for Agents and Multi-Agent Systems
- 18 January 2011
- Published by IGI Global Scientific Publishing
- p. 1-26
Machine Learning for Agents and Multi-Agent Systems
- 1 September 2007
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 196-201
Handling Datasets in a Multi-Relational Environment: Cluster Dispersion vs Cluster Purity
Clustering multiple-instances in a multi-relational environment requires data transformations (e.g. data aggregation) from datasets stored in multiple tables into a single table. Unfortunately, most relational databases are limited to a few basic methods of aggregation (e.g. max, min, sum, count, ave) to aggregate continuous and categorical values. Therefore, data transformation is limited only to aggregation of continuous and categorical values. In this paper, to get the best number of clusters, we propose a genetic semi-supervised clustering technique as a means of aggregating data stored in multiple tables. This algorithm is suitable for classification of datasets with a high degree of one-to-many associations, in which a single record has multiple instances that are associated with it. The clustering algorithm can be used in two ways. One is the unsupervised clustering, where the user may control the result of clustering by optimizing the value of cluster dispersion. The other is a semi-supervised clustering, where the user uses an unsupervised clustering method optimized with a genetic algorithm incorporating a measure of classification accuracy used in decision tree algorithm, the GINI index. In this paper, we examine both methods to dynamically cluster multiple instances, as a means of aggregating them, and illustrate the effectiveness of the semi-supervised genetic algorithm-based clustering technique.
- 1 September 2007
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 196-201
Handling Datasets in a Multi-Relational Environment: Cluster Dispersion vs Cluster Purity
Clustering multiple-instances in a multi-relational environment requires data transformations (e.g. data aggregation) from datasets stored in multiple tables into a single table. Unfortunately, most relational databases are limited to a few basic methods of aggregation (e.g. max, min, sum, count, ave) to aggregate continuous and categorical values. Therefore, data transformation is limited only to aggregation of continuous and categorical values. In this paper, to get the best number of clusters, we propose a genetic semi-supervised clustering technique as a means of aggregating data stored in multiple tables. This algorithm is suitable for classification of datasets with a high degree of one-to-many associations, in which a single record has multiple instances that are associated with it. The clustering algorithm can be used in two ways. One is the unsupervised clustering, where the user may control the result of clustering by optimizing the value of cluster dispersion. The other is a semi-supervised clustering, where the user uses an unsupervised clustering method optimized with a genetic algorithm incorporating a measure of classification accuracy used in decision tree algorithm, the GINI index. In this paper, we examine both methods to dynamically cluster multiple instances, as a means of aggregating them, and illustrate the effectiveness of the semi-supervised genetic algorithm-based clustering technique.
- Open Access
- 1 January 2004
- Published by Association for Computational Linguistics (ACL)
- p. 104-113
WordNet-based text document clustering
Text document clustering can greatly simplify browsing large collections of documents by reorganizing them into a smaller number of manageable clusters. Algorithms to solve this task exist; however, the algorithms are only as good as the data they work on. Problems include ambiguity and synonymy, the former allowing for erroneous groupings and the latter causing similarities between documents to go unnoticed. In this research, naïve, syntax-based disambiguation is attempted by assigning each word a part-of-speech tag and by enriching the 'bag-ofwords' data representation often used for document clustering with synonyms and hypernyms from WordNet.
- 1 January 2004
- Published by Association for Computational Linguistics (ACL)
- p. 104-113
Open AccessWordNet-based text document clustering
Text document clustering can greatly simplify browsing large collections of documents by reorganizing them into a smaller number of manageable clusters. Algorithms to solve this task exist; however, the algorithms are only as good as the data they work on. Problems include ambiguity and synonymy, the former allowing for erroneous groupings and the latter causing similarities between documents to go unnoticed. In this research, naïve, syntax-based disambiguation is attempted by assigning each word a part-of-speech tag and by enriching the 'bag-ofwords' data representation often used for document clustering with synonyms and hypernyms from WordNet.