• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2005.04a
• /
• pp.363-366
• /
• 2005

유비쿼터스 컴퓨팅에서 사용자에게 필요한 서비스를 지능적으로 제공하기 위해서는 컨텍스트 정보의 효과적인 필터링이 필요하다. 현재까지 사용되고 있는 필터링 기술은 온라인상에서 사용되는 사용자 정보를 기준으로 서비스를 제공하고 있다. 하지만 휴대용 유$\cdot$무선기기에서 컨텍스트 인식에 기반을 둔 서비스를 제공하기 위해서는 복잡한 필터링과정과 큰 저장 공간이 요구된다. 따라서 본 논문에서는 사용자 주변에 널려 있는 센서를 통해 입력된 컨텍스트 정보들을 효율적으로 필터링하여 사용자에게 필요한 서비스만을 제공하도록 하였다. 이를 위해서 기존의 P2P 모바일 에이전트에서 사용되는 협력적 필터링 기술에 Naive Bayesian 알고리즘을 혼합한 컨텍스트 협력적 필터링 알고리즘을 제안한다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.24 no.8
• /
• pp.970-977
• /
• 2020

The collaborative filtering recommendation technique has been the most widely used since the beginning of e-commerce companies introducing the recommendation system. As the online purchase of products or contents became an ordinary thing, however, recommendation simply applying purchasers' ratings led to the problem of low accuracy in recommendation. To improve the accuracy of recommendation, in this paper suggests the method of collaborative filtering that analyses product reviews and uses them as a weighted value. The proposed method refines product reviews with text mining to extract features and conducts sentiment analysis to draw a sentiment score. In order to recommend better items to user, sentiment weight is used to calculate the predicted values. The experiment results show that higher accuracy can be gained in the proposed method than the traditional collaborative filtering.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2008.05a
• /
• pp.145-147
• /
• 2008

현재까지 유해한 컨텐츠(Contents)를 차단하기 위한 활발한 연구가 있었으나, 사람의 음성(speech)정보를 이용한 필터링(filtering) 기법에 대한 연구는 활발히 이루어지지 않은 측면이 있다. 본 논문은 동영상 데이터를 가지고 있는 여러 데이터 중에서 음성 정보의 분석을 통하여 일반 컨텐츠와 성인 컨텐츠를 분류하기 위함이다. 본 논문은 음성 정보 중에서 음성 정보의 특징을 가장 잘 다루는 피치 검출을 통한 정보의 분석을 통한 성인 컨텐츠의 필터링에 그 목적이 있다. 현재까지 진행되고 있는 필터링(filtering)방법에 대한 수행 결과보다 개선된 성능을 보이고자 한다. 즉, 음성 정보의 특징 정보를 이용한 성인 컨텐츠(Adult Contents)분류 기법을 활용하는 것으로 성인 컨텐츠(Adult Contents)에서 두드러지는 특징을 보이는 사운드 패턴을 분석하는 것이다.

• Journal of Intelligence and Information Systems
• /
• v.21 no.1
• /
• pp.119-142
• /
• 2015

With the explosive growth in the volume of information, Internet users are experiencing considerable difficulties in obtaining necessary information online. Against this backdrop, ever-greater importance is being placed on a recommender system that provides information catered to user preferences and tastes in an attempt to address issues associated with information overload. To this end, a number of techniques have been proposed, including content-based filtering (CBF), demographic filtering (DF) and collaborative filtering (CF). Among them, CBF and DF require external information and thus cannot be applied to a variety of domains. CF, on the other hand, is widely used since it is relatively free from the domain constraint. The CF technique is broadly classified into memory-based CF, model-based CF and hybrid CF. Model-based CF addresses the drawbacks of CF by considering the Bayesian model, clustering model or dependency network model. This filtering technique not only improves the sparsity and scalability issues but also boosts predictive performance. However, it involves expensive model-building and results in a tradeoff between performance and scalability. Such tradeoff is attributed to reduced coverage, which is a type of sparsity issues. In addition, expensive model-building may lead to performance instability since changes in the domain environment cannot be immediately incorporated into the model due to high costs involved. Cumulative changes in the domain environment that have failed to be reflected eventually undermine system performance. This study incorporates the Markov model of transition probabilities and the concept of fuzzy clustering with CBCF to propose predictive clustering-based CF (PCCF) that solves the issues of reduced coverage and of unstable performance. The method improves performance instability by tracking the changes in user preferences and bridging the gap between the static model and dynamic users. Furthermore, the issue of reduced coverage also improves by expanding the coverage based on transition probabilities and clustering probabilities. The proposed method consists of four processes. First, user preferences are normalized in preference clustering. Second, changes in user preferences are detected from review score entries during preference transition detection. Third, user propensities are normalized using patterns of changes (propensities) in user preferences in propensity clustering. Lastly, the preference prediction model is developed to predict user preferences for items during preference prediction. The proposed method has been validated by testing the robustness of performance instability and scalability-performance tradeoff. The initial test compared and analyzed the performance of individual recommender systems each enabled by IBCF, CBCF, ICFEC and PCCF under an environment where data sparsity had been minimized. The following test adjusted the optimal number of clusters in CBCF, ICFEC and PCCF for a comparative analysis of subsequent changes in the system performance. The test results revealed that the suggested method produced insignificant improvement in performance in comparison with the existing techniques. In addition, it failed to achieve significant improvement in the standard deviation that indicates the degree of data fluctuation. Notwithstanding, it resulted in marked improvement over the existing techniques in terms of range that indicates the level of performance fluctuation. The level of performance fluctuation before and after the model generation improved by 51.31% in the initial test. Then in the following test, there has been 36.05% improvement in the level of performance fluctuation driven by the changes in the number of clusters. This signifies that the proposed method, despite the slight performance improvement, clearly offers better performance stability compared to the existing techniques. Further research on this study will be directed toward enhancing the recommendation performance that failed to demonstrate significant improvement over the existing techniques. The future research will consider the introduction of a high-dimensional parameter-free clustering algorithm or deep learning-based model in order to improve performance in recommendations.

• Journal of Korea Spatial Information System Society
• /
• v.12 no.1
• /
• pp.18-27
• /
• 2010

In u-GIS environment, GeoSensor environment requires that dynamic data captured from various sensors and static information in terms of features in 2D or 3D are fused together. GeoSensors, the core of this environment, are distributed over a wide area sporadically, and are collected in any size constantly. As a result, storage space could be exceeded because of restricted memory in DSMS. To solve this kind of problems, a lot of related studies are being researched actively. There are typically 3 different methods - Random Load Shedding, Semantic Load Shedding, and Sampling. Random Load Shedding chooses and deletes data in random. Semantic Load Shedding prioritizes data, then deletes it first which has lower priority. Sampling uses statistical operation, computes sampling rate, and sheds load. However, they are not high accuracy because traditional ones do not consider spatial characteristics. In this paper 'Pre-Filtering based Post Load Shedding' are suggested to improve the accuracy of spatial query and to restrict load shedding in DSMS. This method, at first, limits unnecessarily increased loads in stream queue with 'Pre-Filtering'. And then, it processes 'Post-Load Shedding', considering data and spatial status to guarantee the accuracy of result. The suggested method effectively reduces the number of the performance of load shedding, and improves the accuracy of spatial query.

• The KIPS Transactions:PartD
• /
• v.18D no.6
• /
• pp.415-422
• /
• 2011

Finding similar strings is an important operation because textual data can have errors, duplications, and inconsistencies by nature. Many algorithms have been developed for string approximate searches and most of them make use of inverted lists to find similar strings. These algorithms basically perform merge operations on inverted lists. In this paper, we develop a bitmap representation of an inverted list and propose an efficient search algorithm that can skip unnecessary inverted lists without searching using bitmap filters. Experimental results show that the proposed technique consistently improve the performance of the search.

• The Journal of the Korea Contents Association
• /
• v.7 no.4
• /
• pp.66-73
• /
• 2007

Internet is a part of our common life and tremendous information is cumulated. In these trends, the personalization becomes a very important technology which could find exact information to present users. Previous personalized services use content based filtering which is able to recommend by analyzing the content and collaborative filtering which is able to recommend contents according to preference of users group. But, collaborative filtering needs the evaluation of some amount of data. Also, It cannot reflect all data of users because it recommends items based on data of some users who have similar inclination. Therefore, we need a new recommendation method which can recommend prefer items without preference data of users. In this paper, we proposed the apparel coordination system using personalized preference on the semantic web. This paper provides the results which this system can reduce the searching time and advance the customer satisfaction measurement according to user's feedback to system.

• Journal of KIISE:Databases
• /
• v.33 no.5
• /
• pp.437-452
• /
• 2006

In this paper, we propose an extended XPath specification which includes a special matching character '%' used in the LIKE operation of SQL in order to solve the difficulty of writing some queries to filter element contents well, using the previous XPath specification. We also present a novel technique for filtering a collection of document-centric XMLs, called Pfilter, which is able to exploit the extended XPath specification. Owing to sharing the common prefix characters of the operands in value-based predicates, the Pfilter improves the performance in processing those. We show several performance studies, comparing Pfilter with Yfilter in respect to efficiency and scalability as using multi-query processing time (MQPT), and reporting the results with respect to inserting, deleting, and processing of value-based predicates. In conclusion, our approach provides a core algorithm for evaluating the contains() function of XPath queries in previous XML filtering researches, and a foundation for building XML-based distributed information systems.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2011.05a
• /
• pp.73-76
• /
• 2011

본 논문에서는 세라믹 비파괴 검사를 이용하여 획득한 소재 영상에서 기존의 결함 검출 방법보다 결함 검출의 정확도를 개선하기 위한 개선된 결함 검출 방법을 제안한다. 제안된 결함 검출 방법은 명암 대비를 강조하기 위해 최소 명암도와 최대 명암도를 이용한 Ends-in Search Stretching 기법을 적용하여 비파괴 영상의 명암 대비를 강조한다. Stretching 기법이 적용된 영상에 $7{\times}7$ Sobel 마스크를 적용하여 비파괴 영상의 경계 영역을 추출하고, 영상의 잡음을 제거하기 위해 비등방성 필터링을 적용하여 영상을 보정한다. 보정된 영상에서 임계치 이진화 기법을 적용하여 경계 영역의 기울기를 계산하고, 계산된 기울기를 이용하여 비파괴 영상의 영역을 세분화한다. 세분화된 영역을 구분하기 위해 Grassfire Labeling 기법을 적용한다. Grassfire Labeling 기법이 적용된 영상을 Ends-in Search Stretching 기법이 적용된 비파괴 영상에 적용한 후에 대각선 이진화 기법을 적용한다. 이진화된 영상에서 형태학적 정보를 이용하여 잡음을 제거하고 결함을 검출한다. 본 논문에서 제안한 방법을 획득한 세라믹 소재 영상을 대상으로 실험한 결과, 기존의 결함 검출 방법보다 더 효과적으로 소재의 결함을 추출할 수 있는 것을 확인할 수 있었다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.10a
• /
• pp.33-35
• /
• 2000

최근 멀티미디어 객체를 위한 효율적인 색인 기술에 대한 많은 연구가 활발히 진행되고 있으나 이러한 색인 구조들은 단일 사용자만을 지원하는데 그치고 있는 실정이며 상용 DBMS에 통합되어 실제 응용되는 사례는 드물다. 이에 본 논문에서는 필터링에 기반한 고차원 색인구조를 위한 효율적인 동시성 제어 기법을 제안하고, 지속성 객체 시스템인 SHORE 하부 저장 구조에 통합한다. 제안하는 동시성 제어 기법은 쓰레드의 개수를 증가시켜 삽입과 검색 측면에서 실험한 결과 약 30%의 응답시간 감소를 보였다.