Eung-Hee Kim

Application of support vector machine for prediction of medication adherence in heart failure patients.

Objectives Heart failure (HF) is a progressive syndrome that marks the end-stage of heart diseases, and it has a high mortality rate and significant cost burden. In particular, non-adherence of medication in HF patients may result in serious consequences such as hospital readmission and death. This study aims to identify predictors of medication adherence in HF patients. In this work, we applied a Support Vector Machine (SVM), a machine-learning method useful for data classification. Methods Data about medication adherence were collected from patients at a university hospital through self-reported questionnaire. The data included 11 variables of 76 patients with HF. Mathematical simulations were conducted in order to develop a SVM model for the identification of variables that would best predict medication adherence. To evaluate the robustness of the estimates made with the SVM models, leave-one-out cross-validation (LOOCV) was conducted on the data set. Results The two models that best classified medication adherence in the HF patients were: one with five predictors (gender, daily frequency of medication, medication knowledge, New York Heart Association [NYHA] functional class, spouse) and the other with seven predictors (age, education, monthly income, ejection fraction, Mini-Mental Status Examination-Korean [MMSE-K], medication knowledge, NYHA functional class). The highest detection accuracy was 77.63%. Conclusions SVM modeling is a promising classification approach for predicting medication adherence in HF patients. This predictive model helps stratify the patients so that evidence-based decisions can be made and patients managed appropriately. Further, this approach should be further explored in other complex diseases using other common variables.

Query-based association rule mining supporting user perspective

Two parameters, namely support and confidence, in association rule mining, are used to arrange association rules in either increasing or decreasing order. These two parameters are assigned values by counting the number of transactions satisfying the rule without considering user perspective. Hence, an association rule, with low values of support and confidence, but meaningful to the user, does not receive the same importance as is perceived by the user. Reflecting user perspective is of paramount importance in light of improving user satisfaction for a given recommendation system. In this paper, we propose a model and an algorithm to extract association rules, meaningful to a user, with an ad-hoc support and confidence by allowing the user to specify the importance of each transaction. In addition, we apply the characteristics of a concept lattice, a core data structure of Formal Concept Analysis (FCA) to reflect subsumption relation of association rules when assigning the priority to each rule. Finally, we describe experiment results to verify the potential and efficiency of the proposed method.

Development of a cervical cancer progress prediction tool for human papillomavirus-positive Koreans: A support vector machine-based approach

Objectives To develop a Web-based tool to draw attention to patients positive for human papillomavirus (HPV) who have a high risk of progression to cervical cancer, in order to increase compliance with follow-up examinations and facilitate good doctor–patient communication. Methods Records were retrospectively analysed from women who were positive for HPV on initial testing (before any treatment). Information concerning age, Papanicolaou (PAP) smear result and presence of 15 high-risk HPV genotypes was used in a support vector machine (SVM) model, to identify the patient features that maximally contributed to progression to high-risk cervical lesions. Results Data from 731 subjects were analysed. The maximum number of correct cancer predictions was seen when four features (PAP, HPV16, HPV52 and HPV35) were used, giving an accuracy of 74.41%. A web-based high-risk cervical lesion prediction application tool was developed using the SVM model results. Conclusions Use of the web-based prediction tool may help to increase patient compliance with physician advice, and may heighten awareness of the significance of regular follow-up HPV examinations for the prevention of cervical cancer, in Korean women predicted to have heightened risk of the disease.

A hypergraph-based storage policy for RDF version management system

Versioning schemes for RDF data can play an important role in the management of Semantic web based information. The storage space and the version retrieval is one of the most significant aspects in version management. However, the existing methods have limitations in that they are not efficient in constructing the version. Therefore, there is a need for new and improved techniques that achieve better performance. In this paper, we proposed a hypergraph-based storage policy for RDF version management, which exploits hyperedges and vertices to maintain the information of version. This scheme enables us to reduce the space overhead and reconstruct a specific version efficiently. In addition, we proposed an optimization technique, which results in considerable space saving. An experimental study with RDF data sets shows the proposed storage policy achieved reduced storage space and efficiently reconstructed the version.

FARM: An FCA-based Association Rule Miner

Abstract Association rule mining is a well-researched and widely applied data mining technique for discovering regularities between items in a dataset. An association rule consists of an antecedent and a consequent with two measures, named support and confidence, which indicate how valuable the rule is. For several decades, intensive studies have been made on efficient association rule mining methods aiming to reduce rule-extraction time and to prevent generation of redundant rules. By incorporating negation and disjunction operators into antecedents, our study offers richer expressive power in describing user interests as antecedents, which in turn translates into more valuable association rules whose consequents match the expressed user interests. This study consists of three components: (1) a conceptual model, called plant , that represents necessary constituents for the proposed extended association rules; (2) three algorithms, called CULTIVATION algorithm family , that demonstrate how the extended association rule is processed within the model; (3) a full-fledged Java-based system, called FARM (FCA-based Association Rule Miner), that is a computerized implementation of our approach. Finally, in order to verify the efficiency and usefulness of our approach, experiments were carried out that compared the approach with extant representative methods.

A FCA-based approach for enriching users' knowledge base

Despite of the various benefits obtainable from Formal Concept Analysis (FCA) in knowledge base construction, FCA-based approaches are not enough to help an expert enrich his knowledge. This is because they provide only the clusters constructed with user-defined knowledge and super-sub relation between the clusters. In this paper, we propose an approach that provides a user with a guideline by suggesting undiscovered knowledge in the form of predicates. This approach firstly generates a set of candidate predicates by analyzing a pre-defined predicate by users. Second, it discards unqualified ones from the set of candidate predicates using a filtering method dealing with two criteria, uniqueness and support. The qualified candidate predicates are suggested, and selected by the user, and finally, his knowledge is enriched by merging the selected predicates with pre-defined ones.

On Development of an Automatic Tool for Extracting Association Rules of a user query using Formal Concept Analysis

ABSTRACT Formal Concept Analysis (FCA) is a widely used methodology for data analysis, which extracts concepts and builds a concept hierarchy from given data. A concept consists of objects and attributes shared by those objects, and a concept hierarchy includes information on super-sub relations among the concepts. In this paper, we propose a method for extracting Implication and Association rules from a concept hierarchy given a query by a user. The method also describes a way for displaying the extracted rules. Based on this method, we implemented an automatic tool, QAG-Wizard. Because the QAG-Wizard not only elicits relation information for the given query, but also displays it in structured form intuitively, we expect that it can be used in the fields of data analysis, data mining and information retrieval for various purposes.Key Words:Formal Concept Analysis, Concept lattice, User Query, Implication, Association Rules 1. 서 론 1) 대용량 데이터를 저장하고 관리할 수 있는 데이터베이스 기술과 정보기술이 발전함에 따라, 개인 및 조직이 보유하고 접근할 수 있는 데이터의 양은 기하급수적으로 증가하고 있다. 그러나, 가용한 데이터 양이 증가함에 따라서, 대용량 데이터를 사용하는 도메인의 특성을 수월하게 파악하기 어렵다는 문제점이 발생하고 있다. 이러한 문제점을 해결하기