Earticle

Diagnosis of cancer with biomarkers is relatively simple with the use of blood samples, and it can detect cancer at an early stage with expense compared to the other diagnosis methods. We use word embedding to find an alternative biomarker for the early diagnosis of ovarian cancer from the biomedical corpus. Word embedding is a word vector representation with previously proven efficiency in the biomedical domain. First, we derived a low dimensional representation of each biomarker embedding induced from Canonical Correlation Analysis (CCA), which is a powerful and flexible statistical technique for dimensionality reduction. Second, we found a similar pair of biomarkers in the literature by using cosine similarity of biomarker embedding. In order to determine the clinical similarity between the pair of biomarkers, we used the area under the curve (AUC) of the combination of 2 biomarkers used previously. In the experiment, we confirmed that correlation between the high similarity biomarker pair, was highly correlated as the average 0.710 of the actual AUC correlation of the top 10% of the pair.

A multi-tenant Software as a Service (SaaS) application delivers customized run-time application to each tenant using a single application. Multi-tenancy offers several advantages including quick application deployment, reduced application maintenance, effective version control, and much more. However, handling large datacenter resources and providing better Quality of Service (QoS) is a major challenge for SaaS providers. SaaS application have unstable load based on tenant user’s demand. In a typical multitenant application security, reliability and energy efficiency play an important role. However, these aspects are largely ignored while focusing on the performance of the application. In this paper we propose a novel application’s component based multivariable architecture for multi-tenant Software as a Service (SaaS) application. Instances of application’s components are created at component level and dynamically scaled based on the component’s load considering security, reliability and energy based runtime factors applying multi-criteria Zionts–Wallenius optimization method. Our evaluation and discussions show that the proposed multi-instance Towards Multivariable Architecture (TMA) offers enhanced application load distribution and at the same time offers improved application response time and conserves energy in datacenter.

Since the Internet provides a common way for expressing and sharing people’s ideas or minds, corporate marketers can gather data on the web to acquire measurable and actionable insights. As the smartphones are widely spread as hand-held devices, SNSs(Social Network Services) in the smartphones become common media to record people’s daily activities and thoughts, researchers try to gather and read people’s mind on SNS through opinion mining or sentiment analysis. In this study we suggest a framework for clustering brand names and perform a case study of cosmetic products using social big data gathered on the social media - Microblog, Twitter. To cluster the brand names, we calculate the distance of paired brand names based on the total number paired brand names mentioned together. To identify the clusters among these brands names, we projected the brand names onto a 2-dimensional and a 3-dimentional space using MDS(Multi-Dimensional Scaling). After projecting the brand names, we found the clusters of the brand names using k-means clustering and identified the characteristics of each cluster.

This project is to strengthen the competitiveness of ICT devices used in a big data system, which will require the support of information about how to establish infrastructure which will enable to develop and test a large scale of computing devices and a kernel SW. The purpose of this project is to build the research infrastructure for the high-speed big data processing devices and to obtain the kernel SWs and generic technology. The purpose of this study is to implement the feasibility analysis on this project. Therefore this study is using the AHP method which has established the specific analysis processes and can quickly confirm the validity of the project by calculating the weights of the evaluation criteria. As a result of this study, the comprehensive score by AHP analysis is 0.842 and it shows that the implementation of this project is feasible.

This paper researched on the quality of code clone in the software, evaluated the code clone quality of the current versions. Then using Bayesian network to train the existing sample data to get the prediction model of code clone that is able to predict the quality. The prediction results are able to help developers decide which code clone should be reconstructed or efficiently reused. The experiment shows that the method can be used to predict the quality of code clone in software more accurately.

Art-viewers spend time with art. Digital-media-art viewers' engagement is not only time-based but extendable, since depending on the level of aesthetic interaction, people spend more time with the piece, allowing more time to understand the meaning of the work. Interactive digital media art yields an experiential event to art-viewers. Without this interactive event the artwork’s existence becomes questionable. The meaning is output by an art-viewers input; set in a temporal space. A number of studies in HCI, UI, and UX design have dealt with various aspects of time. This paper intends to analyze time as an aesthetic element, and as the reason art-viewers invest their time with interactive digital media art works for artists and designers to understand what they and their prospective viewers are dealing with.

Many security professionals know about Image Metadata, but they do not really know how it can be used to protect from the threats and prove the authenticity of these images and the most significant aspect is the anonymity of perpetrator. The novelty of this paper lies on how to use a new algorithm of Image Metadata to detect internal intrusion and how to protect it in real time inside the cloud providers. In this paper, we propose a novel algorithm to use features and characteristics technologies of Image Metadata, namely, a novel family of Intrusion Detection Systems (IDS) built on Exchangeable Image File Format (EXIF) of Image Metadata. It can assist in eliminating the anonymity advantage of perpetrator which is considered extremely important for the Digital Investigation to detect the attacker in a real time and to prove the authenticity of the images Metadata against attackers.

When using search engine services to search for scholarly articles, obtaining quick and accurate search results from a huge set of scholarly information is always important. However, most of the domestic and foreign search engine services for scholarly articles present a broad range of the results that correspond to the query of the researcher’s name. Such results contribute in lowering the search precision and require users to spend time and effort to verify the results and find the necessary information. Such a problem is called “author ambiguity”, while solving this problem is called “author disambiguation.” An author disambiguation method classifies the authors with the same name into an actual person. By resolving author ambiguity, better search results can be obtained; this increases the recall rate and accuracy when searching for scholarly articles. In order to resolve author ambiguity in this paper, we shall expand the co-author network and identify the author using the co-author network information and basic bibliographic information as the features for machine learning Support Vector Machine. To examine the effectiveness of the proposed method, we test the author disambiguation method by targeting 92,100 IT-related scholarly data generated in Korea. Author disambiguation results through the expansion of co-author network are shown to have an F-1 measure of 94.79%. The result confirms that the author disambiguation method through the implementation of the co-author network is effective.

The data used in this study includes guarantee-related data from January 2007 to September 2014, as well as data of search volume provided by Naver Trend. As a result of the hierarchical regression analysis, it turned out that in step 1, the volume of credit guarantee searching affected guarantee supply while the default amount did not. In addition, it is expected that as the volume of credit guarantee searching increases, guarantee supply will also be expended. In step 2, it turned out that the amount of default normalization affected guarantee supply. It is expected that as the amount of default normalization increases, guarantee supply will also be expanded. Finally, in step 3, it turned out that both the volume of credit guarantee searching and the amount of default normalization affected guarantee supply while the default amount did not. As for the explanation power of the three models, that of step 1 was 8.3%, that of step 2 was 13.7%, and that of step 3 was 19.3%. As the default amount affected guarantee supply, Sobel Test was conducted to measure the mediation effect of the volume of credit guarantee searching and the amount of default normalization. As a result, it turned out that the volume of credit guarantee searching had mediation effect while the amount of default normalization did not. The objective of this study is to examine the effect of guarantee accidents of the Regional Credit Guarantee Foundation, which provides public finance service on guarantee supply. This indicates the necessity of examining the causality through another regression analysis after a Granger Causality test. In terms of demands for guarantee supply, further research may be necessary on referring to the volume of credit guarantee searching.

Distributed computation follows the models of discrete structures in combinatorial forms. In higher-dimensions, the simplex structures of topological spaces as well as homology are employed to model and analyze distributed asynchronous computations. However, the monotone spaces are the general forms of topological spaces and can be effectively employed to analyze distributed computation. This paper proposes an analytical model of distributed computation in monotone spaces. It is illustrated that, the modeling of distributed computation in monotone spaces helps in determining consistent cuts under closure and convergence of computation. Furthermore, a connective mapping between the simplexes and monotone is constructed.

The growth of South Korea’s tourism industry, in turn, has led to a steady growth in the number of international visitors to South Korea (hereinafter referred to as “Korea”). When visiting Korea, foreign visitors are likely to turn to smartphones as the most convenient tool for getting the information they need. A survey by Korea Tourism Organization found that tourists usually search for the information they need on their smartphones. This study was, however, based on the assumption that it is an application (hereinafter “app”) that can provide the most extensive tourist information about Korea. Thus, the study involves analyzing currently available services while centering on the “Visit Korea” app distributed by Korea Tourism Organization, and comparing it with other apps to come up with suggestions for improvement. Integrating Bluetooth 4.0-enabled beacons with mobile apps will provide travelers with useful services, convenience, and a variety of benefits. It is also an easier and more assuring method for administrators when they identify users’ data. As for the problems associated with beacons, incremental improvements are also being made. Combining the “Visit Korea” app with beacon technology will hopefully contribute to the further growth of the Korean tourism industry.

This paper proposes an intelligent recommendation system that has increasing promotion effect. Normally, intelligent recommendation system suggests suitable product to consumer but the Intelligent Recommendation System recommends fit well-known person to company for increasing promotion effect. It uses weight based algorithm and matching model based on well-known person relation graph.

Models transformation is the main key of MDA. The first transformation in MDA is CIM to PIM transformation, the second is PIM to PSM transformation. Most searches deal the transformation from PIM level to PSM level, since there are multiple common points between these two levels. However, the transformation from CIM level to PIM level is rarely addressed in search subjects because they are two distinct levels. Our objective in this paper is to represent an approach that allows controlling transformation from CIM level to PIM level in accordance with the MDA approach. More precisely, we propose a methodology for transforming service-oriented business models, to web-based design models. Despite the importance of service-oriented models and web-based models, the transformation between them is not addressed in MDA researches. Our methodology is based on creating a good CIM models service-oriented, through construction rules, to facilitate transformation towards PIM models web-based. Next, our transformation rules allow a semi-automatic transformation from CIM to PIM. Our approach conforms to MDA recommendations, because it allows considering the business dimension in the CIM level, and it allows modeling this latter level by using SoaML, the OMG standard for modeling services. However, we based on UML 2 to model PIM level, because UML is advocated by MDA in PIM level. Our proposal results a set of web-based design models from service-oriented business models, through semi-automatic transformation in accordance with MDA approach.

A great deal of difficulty in the task of developing software can be traced to the initial phase of its life cycle, when requirements are analyzed and design schemata are constructed. The process involves modeling that part of reality the software system is intended to address. The problem is that some 2014 studies have shown that more than a third of IT projects are cancelled before being completed, the costs of more than half of these projects will be higher than their original estimates, and, on average, barely more than one in six software projects are ever completed on-time and on-budget. The main reasons for these problems include lack of user input and incomplete requirements and specifications. Accordingly, even though previous decades have witnessed a great deal of progress in the area of software development, these difficulties point to the fact that such advancement has not furnished new paradigms that might further advance development in this field. This paper is a contribution in this direction in the area of requirements analysis and design. The paper proposes adopting a new modeling method for use in the software development phase of requirements analysis and design that can contribute to alleviating the problems of lack of user input and incomplete requirements specifications. The basic idea in the proposed method, called Thing-Oriented (TO) modeling, bases requirements analysis and design on “things” that flow. Without loss of generality, TO modeling is contrasted with Object-Oriented modeling. The results point to the feasibility of using TO modeling as a multilevel diagrammatic language with a constrained number of notions.

A data scientist is a relatively new job title and is not yet fully defined or understood. Little research has been conducted on data scientists and there is still incongruity among its practitioners. Yet, the job market for data scientists is already active with high demand. Many companies have created their own definition of a data scientist based on their own needs. The purpose of this research is to explore the definition of a data scientist by examining how it is accepted in these industries and businesses. A content analysis of 1,240 job ads from various companies recruiting data scientists was conducted to identify what types of knowledge and skills were generally demanded. As a result, we found that data scientists were expected to be highly experienced professionals with advanced degrees. The main requisite areas of profession were statistics, modeling, machine learning, and analysis.

Recently, flash memory has been widely used as it has strong points, such as high-speed access, low power consumption, high density and shock resistance. However, due to its hardware characteristics, specifically its erase-before-write feature, it is inefficient for deploying directly B+-tree index which produces a lot of write operations in random order when its nodes are updated. Thus, the normal B+-Tree structure is unfavorable for use in flash memory. Several variants of B+-tree have been proposed for flash memory so far. This paper surveys the state-of-the-art B+-tree index variants on flash memory that have been presented in conference proceedings and journals. It provides a broad overview of existing schemes, algorithms to solve them, and discusses related research issues.