DC3K 2014 Abstracts


Short Papers
Paper Nr: 2
Title:

Knowledge Management Concepts and Processes in Healthcare - Research Plan for Doctoral Thesis

Authors:

Helvi Nyerwanire, Erja Mustonen-Ollila, Antti Valpas and Jukka Heikkonen

Abstract: This study presents a research plan for a doctoral thesis about Knowledge Management in Healthcare. It outlines objectives, research problems, state of the art, methodology and expected outcome. The study introduces current knowledge management concepts, the research questions, and a conceptual framework of knowledge management processes. It also outlines data collection methods and data analyzing methods. In this study both qualitative analyzing methods with the grounded theory approach and quantitative data analysis with novel intelligent computing and analyzing methods are applied. This doctoral study is planned to take a total of five (5) years (January 2012- January 2017) in which the output will be five (5) conference articles, and one journal article. Furthermore, a relevant introductory section of the thesis will be written in this period. One conference article has been accepted in 2012, another conference article was submitted for a review in 2014, and one journal article in under work.

Paper Nr: 3
Title:

A Model-Driven Approach to Create and Maintain an Executable Transferal Management Platform

Authors:

Emanuele Laurenzi

Abstract: My work falls within the eHealth application domain and it is embedded into the just started research project Patient Radar. The Patient Radar project wants to facilitate intersectoral collaboration within the inpatient sector, also called “transferal management”, i.e. between acute hospitals and rehabilitation clinics in Switzerland. My research aims at supporting and optimizing such a collaboration by setting up a framework which adopts a model-driven approach to enable the creation and maintenance of a transferal management platform. The model-driven approach makes the platform highly configurable to accommodate new clinical pathways and be easily extendable to include additional functions to meet future needs. All domain-specific aspects are described declaratively in application models. Hence, domain experts will be able to create/use/manage application models with no required programming skills. To provide an executable platform, models are first specified in a description logic and then their elements are mapped to corresponding elements in an application framework. In this way, we will ensure that executable code can be derived from all application models. Additionally, the transferal management platform includes reference models from which domain experts can easily create and adapt application models.

Paper Nr: 4
Title:

Knowledge Transfer in Regulatory Analytical Sciences through the Implementation of Communities of Practice

Authors:

Joanna Jesionkowska, Brigitte Denis and Philip Taylor

Abstract: Today, organisations, work groups, teams, and individuals work together in new ways. Inter-organisational collaboration is increasingly important. Communities of practice provide a new model for connecting people in the spirit of learning, knowledge sharing, and collaboration as well as individual, group, and organizational development. This gives new possibilities which could be used in scientific environment. Promoting affiliation between scientists is relatively easy, but creating larger organizational structures is much more difficult, due to traditions of scientific independence, difficulties of sharing implicit knowledge, and formal organizational barriers. Hence there is a need to research possibilities given by CoPs in the area of sciences.

Paper Nr: 9
Title:

Semantic Enrichment of Relevant Feature Selection Methods for Data Mining in Oncology

Authors:

Adriana Da Silva Jacinto, Ricardo Da Silva Santos and José Maria Parente De Oliveira

Abstract: This project presents a proposal of capturing of the semantic importance of each feature by computational manner. The proposal enriches the traditional methods of feature selection by using of Natural Language Processing, the NCI ontology, WordNet and medical documents. A prototype of this approach was implemented and tested with five data sets related to cancer patients. The results show that the use of semantic improves the pre – processing by selecting of the most relevant semantic features.

Paper Nr: 12
Title:

An Approach to Develop Flexible Systems with Organizational-interoperability Requirements

Authors:

Diego Fuentealba, Kecheng Liu and Weizi Li

Abstract: This paper reports the initial state of a PhD project with an aim of discussing the definition of flexibility and the interoperability of IS in order to identify their key factors. These factors are significant to propose an assessment model, which can address the design. After a literature review, the method of this stage is the analysis of each definition, using a semiotic framework (Stamper, 1973) to identify the impacts of each definition between the organization and the IS.

Paper Nr: 13
Title:

Ontology based Knowledge Extraction with Application to Finance

Authors:

Özgür Bağlıoğlu and Mesut Çeviker

Abstract: Public and private enterprise finance performance is reflected and affected by unorganized, unstructured data such as news, reports (IMF, OECD and other periodical reports) as well as structured statistical data extracted by Statistical Institutes and other organizations. The role of raw data in influencing performance and decision making is not negligible. In this context, this paper presents knowledge extraction methodology for precise and fast decision making in finance by using ontological tools. For this purpose, we firstly design finance ontology and collect datasets. The aim of this ontology is to support the knowledge management in the finance domain and to increase the productivity through evidence base, comprising raw finance data to be retrieved from various operational sources. We then propose to populate the ontology by using past project properties and project progress reports. After population of data, we plan to develop and use a semantic search engine to gather meaningful data i.e. knowledge. The semantic search engine will assist decision makers to make better decisions. The output of this work will be also used as an input for decision making and scenario based future prediction for finance as this study is a part of a larger project called “ontology based decision support system”.

Paper Nr: 14
Title:

Ensemble Method for Prediction of Prostate Cancer from RNA-Seq Data

Authors:

Yongjun Piao, Nak Hyun Choi, Meijing Li, Minghao Piao and Keun Ho Ryu

Abstract: The main idea of our research is to develop an ensemble machine learning algorithm to accurately classify prostate cancer using RNA-Seq data. To date, many studies have focused on predicting prostate cancer using microarray data. Recently, RNA-Seq is rapidly being used for cancer studies as an alternative for microarray. Thus, new machine learning algorithms are needed to analyze RNA-Seq data which have different characteristic compared with microarray. Currently the PhD research has been running for one year and has focused on analyzing existing state-of-art normalization methods, gene expression data analysis, and ensemble methods. Besides that, we designed an ensemble feature selection algorithm to select relevant genes from the gene expression data. Moreover, we have developed a 'digital' gene expression data simulator for evaluating the performance of proposed algorithms. The next step will be to construct an accurate ensemble prediction model to diagnosis of prostate cancer. Finally, the model will be fine-tuned based on the feedback from the medical doctors.

Paper Nr: 15
Title:

The Comprehensive Modelling of BPMN Business Processes and Business Rules using SBVR Profile

Authors:

Egle Mickeviciute and Rimantas Butleris

Abstract: In order to have a full and comprehensive business process model we have to consider using two different modelling approaches that define business process models in two different ways. Modelling of business processes and business rules must be complementary each other, that means they should be used together in one environment. The goal of this paper is to present created approach for business process and business vocabulary and rules combination in one CASE tool keeping linkage between elements from two different modelling approaches as a result of current research. In order to achieve the goal we present transformation rules that allow us to transform business process model to business vocabulary and rules.

Paper Nr: 16
Title:

Integration Method of Business Vocabularies and Business Rules Specifications (Models)

Authors:

Edvinas Sinkevicius and Rimantas Butleris

Abstract: Information system development starts from defining business vocabulary and rules. There are some cases when several business vocabularies and rules from the same domain must be used. Therefore there is a need to merge those business vocabularies and rules to make one and use it for development of a system. The research problem is that there is a need to use merged information from several business vocabularies and rules. Merging them could cause to occur conflicts between the elements from different sources. To our knowledge, the problem yet is not solved nor in the scientific literature nor in ptactical applications.

Paper Nr: 17
Title:

Rule Generation for Scenario based Decision Support System on Public Finance Domain

Authors:

Mesut Çeviker and Özgür Bağlıoğlu

Abstract: This study is a part of a larger project called “Ontology Based Decision Support System”. In this document, we report methodology of the Rule Generation (RG) that is planned to be taken from the knowledge queried from ontology based Knowledge Extraction System (KES). Rule generation aims producing rules for a rule based system, which will be used for future prediction of an organization or an organizational unit. The term “scenario based” implies that the system will do future prediction for possible scenarios of next movements like different budget scheduling scenarios. Future prediction will be limited to the prediction of parameters that the organization is willing to know, such as the parameters related to the objectives and the goals on their strategic plan. In literature, rule generation problems are addressed by variety of different learners; so what we plan is using a learners system with many learners possibly with different types. The system will be valuable for merging an ontology based KES and DSS with future prediction capability. In addition, this will be the first composite system (having mentioned KES+DES) for public finance domain.

Paper Nr: 18
Title:

A Multi-domain Hybrid Recommender Systems Based on a Dynamic Contextual Ontological User Profile

Authors:

Aleksandra Karpus and Krzysztof Goczyla

Abstract: The aim of research presented here is creation of a multi-domain hybrid recommender system based on a dynamic contextual ontological user profile. Recently, we have built a contextual ontology for representing user preferences. The next step that need to be undertaken is validation of correctness and completeness of this idea of representing a user profile. We consider three context parameters: location, time and user mood. The validity of these parameters, and hence, their impact on user preferences, has been confirmed by the results of a survey among the potential users of recommendation systems. The research on knowledge aquisition and new recommendation algorithms is still in the early stages.

Paper Nr: 19
Title:

Ontology Supported Case based Reasoning and Semantic Similarity Comparison of Documents to Decision-making of the Court Judgment

Authors:

Egon Sewald Junior

Abstract: The duty of maintaining the rule of law applied by the judiciary board, watching the constitution and its laws, judging conflicts of interest and maintaining the social order, must be provided to meet the citizens and businesses needs properly. One of the main points to be considered is the fight against the slowness of the judiciary; in other words, achieve reasonable quickness in the processes. The electronic process reduced the judicial proceeding time, allowing to diminish “dead time”, for example, assembly the process files, page numbering, and mainly with the physical transit of procedural parts. This work starts from researching the organizational context, identifying the intensive knowledge activities and define the Knowledge System pattern to support the decision-making process within the state courts. It has brought up engineering applications expertise applied to the judiciary and from this literature review, to generate a pattern to the knowledge system model. For the organizational system pattern "Court of the State of Amazonas" (TJ/AM) the CommomKADS methodology was applied to raise the organizational context as well as the concepts to be designed to solve problems and exploit opportunities, as well as the design definitions of the artefact, in other words, the Knowledge System. From the application of this methodology, it was established organization models, where it can define problems and opportunities, which may or may not be related to knowledge intensive activities, so identifying applied knowledge assets. The task model describes the processes of the TJ/AM, identifying the use of knowledge and a review about its correct application. Based on these models, the context is presented and it is defined the concept of the knowledge system, by the use of knowledge systems models, which describes its application during the tasks, as well as the communication models that define the interaction between the agents. Thenceforth, it is defined the software design, outlining an artefact that I should develop to support the decision of judges.We will develop semi-automatic system based on sentiment analysis and intentionality to build base cases from court documents. The expected scientific contribution to the development of the work is a model to support the sentencing of the Judge. This model should include a comparison of legal cases through semantic and contextual comparisons. This model will assist the judge in sentencing task, without this, requires a large number of human experts providing the basis for cases to be compared. The generated system must be validated by a comparison of the performance of judges, by comparing times and the perception of judges.