Mykola Galushka

Case-based tissue classification for monitoring leg ulcer healing

The ability to automatically monitor the wound healing process would reduce the workload of professionals, provide standardization, reduce costs, and improve the quality of care for patients. Here we propose an automatic monitoring system for leg ulcers based on case-based reasoning. We focus on the first stage of the monitoring process in this work, that of tissue classification and examine a number of different feature extraction techniques based on texture and Red, Green, and Blue histograms. Results clearly show a case-based approach to be ideal for this type of task.

Temporal data mining for smart homes

Temporal data mining is a relatively new area of research in computer science. It can provide a large variety of different methods and techniques for handling and analyzing temporal data generated by smart-home environments. Temporal data mining in general fits into a two level architecture, where initially a transformation technique reduces data dimensionality in the first level and indexing techniques provide efficient access to the data in the second level. This infrastructure of temporal data mining provides the basis for high-level data mining operations such as clustering, classification, rule discovery and prediction. These operations can form the basis for developing different smart-home applications, capable of addressing a number of situations occurring within this environment. This paper outlines the main temporal data mining techniques available and provides examples of where they can be applied within a smart home environment.

Grid-enabled data warehousing for molecular engineering

Molecular engineering is concerned with the design and manufacturing of novel chemical compounds and materials. Molecular engineering for drug development is complex, time-consuming, and expensive. To lower costs and improve the overall drug development process, information technology (IT) is increasingly employed in the molecular engineering phase. Key IT components for molecular engineering include public and proprietary databases containing information on molecular structures and properties and computational models capable of calculating global properties of molecular structures based on structural and other descriptors characterizing the compound. Recently data mining and data warehousing have become critical tools in the molecular engineering process. Increasingly, some of the computational resources--such as data and information bases, computational models, compute power to execute these models, specialized data mining algorithms--required to develop a new compound are not available locally, but accessible via the global computing network infrastructure. This paper describes a Grid-based approach to molecular engineering. Focus of this paper is placed on the data warehousing of the OpenMolGRID system.

New protocol for leg ulcer tissue classification from colour images

Measurement of wound healing status is very important for monitoring progress in individual patients. Tissue classification is a vital step in the development of an automatic measurement system for wound healing assessment. We present a new tissue classification protocol using the RGB (Red, Green and Blue) histogram distributions of pixel values from wound color images. These three histogram distributions (extracted features) were used as three two-dimensional (2D) input signals for classification. This protocol has been carried out using the KNN classifier and results show that the proposed protocol provides an extremely competent practical method for the classification of wound tissues.

SOPHIA-TCBR: A knowledge discovery framework for textual case-based reasoning

In this paper, we present a novel textual case-based reasoning system called SOPHIA-TCBR which provides a means of clustering semantically related textual cases where individual clusters are formed through the discovery of narrow themes which then act as attractors for related cases. During this process, SOPHIA-TCBR automatically discovers appropriate case and similarity knowledge. It then is able to organize the cases within each cluster by forming a minimum spanning tree, based on their semantic similarity. SOPHIAs capability as a case-based text classifier is benchmarked against the well known and widely utilised k-Means approach. Results show that SOPHIA either equals or outperforms k-Means based on 2 different case-bases, and as such is an attractive approach for case-based classification. We demonstrate the quality of the knowledge discovery process by showing the high level of topic similarity between adjacent cases within the minimum spanning tree. We show that the formation of the minimum spanning tree makes it possible to identify a kernel region within the cluster, which has a higher level of similarity between cases than the cluster in its entirety, and that this corresponds directly to a higher level of topic homogeneity. We demonstrate that the topic homogeneity increases as the average semantic similarity between cases in the kernel increases. Finally having empirically demonstrated the quality of the knowledge discovery process in SOPHIA, we show how it can be competently applied to case-based retrieval.

A scaleable document clustering approach for large document corpora

In this paper, the scalability and quality of the contextual document clustering (CDC) approach is demonstrated for large data-sets using the whole Reuters Corpus Volume 1 (RCV1) collection. CDC is a form of distributional clustering, which automatically discovers contexts of narrow scope within a document corpus. These contexts act as attractors for clustering documents that are semantically related to each other. Once clustered, the documents are organized into a minimum spanning tree so that the topical similarity of adjacent documents within this structure can be assessed. The pre-defined categories from three different document category sets are used to assess the quality of CDC in terms of its ability to group and structure semantically related documents given the contexts. Quality is evaluated based on two factors, the category overlap between adjacent documents within a cluster, and how well a representative document categorizes all the other documents within a cluster. As the RCV1 collection was collated in a time ordered fashion, it was possible to assess the stability of clusters formed from documents within one time interval when presented with new unseen documents at subsequent time intervals. We demonstrate that CDC is a powerful and scaleable technique with the ability to create stable clusters of high quality. Additionally, to our knowledge this is the first time that a collection as large as RCV1 has been analyzed in its entirety using a static clustering approach.

Efficient Similarity Determination and Case Construction Techniques for Case-Based Reasoning

In this paper, we present three techniques for knowledge discovery in case-based reasoning. The first two techniques D-HS and D-HS+SR are concerned with the discovery of similarity knowledge and operate on an uncompacted case-base while the third technique D-HS+PSR is concerned with the discovery of both similarity and case knowledge and operates on a compacted case-base. All three techniques provide a very efficient and competent means of similarity determination in CBR, which are empirically shown to be up to 25 times faster than k-NN without any loss in competency. D-HS+PSR proposes a novel approach to automatically engineering compact case-bases with a minimal overhead to the system, compared to other approaches such as case deletion/addition. Additionally as the approach provides a means for automatically reducing the number of cases required in the case-base without any loss in problem solving competency it has the greatest implication of the three techniques for reducing the effects of the utility problem in CBR.

SOPHIA: an interactive cluster-based retrieval system for the OHSUMED collection

The ability to perform an exploratory search and retrieval of relevant documents from a large collection of domain-specific documents is an important requirement both in the field of medicine and other areas. In this paper, we present a unsupervised distributional clustering technique called SOPHIA. SOPHIA provides a semantically meaningful visual clustering of the document corpus in conjunction with an intuitive interactive search facility. We assess the effectiveness of SOPHIAs cluster-based information retrieval for the MEDLINE testset collection known as OHSUMED.

Intelligent index selection for case-based reasoning

In this paper, we present an indexing technique for case-based reasoning called D-HS^E, that is shown to be more competent than and twice as efficient as the commonly used R-tree. D-HS^E was designed to addresses periodical competency shortcomings of the related D-HS^M index but unfortunately in doing so some efficiency was seen to be sacrificed. In order to address this problem of competency verses efficiency, we propose an intelligent selection algorithm that automatically analyses the case-base and decides which index (D-HS^M or D-HS^E) should be used to optimize performance. The algorithm is designed to favour competency at the expense of efficiency where a competency gain is deemed highly likely to be achieved by using the less efficient approach. In effect we are proposing a flexible indexing scheme that is aware of changes within its environment and which reacts to these changes to optimize performance.

Efficient real time maintenance of retrieval knowledge in case-based reasoning

In this paper, we investigate two novel indexing schemes called DHS and D-HS+PSR(II) designed for use in case-based reasoning systems. D-HS is based on a matrix of cases indexed by their discretised attribute values. DHS+PSR(II) extends D-HS by combining the matrix with an additional treelike indexing structure to facilitate solution reuse. DHS+PSR(II)s novelty lies in its ability to improve retrieval efficiency over time by reusing previously encountered solution patterns. Its benefits include its accuracy, speed and ability to facilitate efficient real time maintenance of retrieval knowledge as the size of the case-base grows. We present empirical results from an analyses of 20 case-bases and demonstrate the technique to be of similar competency to C4.5 yet much more efficient. Its performance advantages over C4.5 are shown to be especially apparent when tested on case-bases which grow in size over time.