• 한국컴퓨터정보학회논문지
• /
• 제23권12호
• /
• pp.219-226
• /
• 2018

Collaborative filtering has been most popular approach to recommend items in online recommender systems. However, collaborative filtering is known to suffer from data sparsity problem. As a simple way to overcome this problem in literature, Jaccard index has been adopted to combine with the existing similarity measures. We analyze performance of such combination in various data environments. We also find optimal weights of factors in the combination using a genetic algorithm to formulate a similarity measure. Furthermore, optimal weights are searched for each user independently, in order to reflect each user's different rating behavior. Performance of the resulting personalized similarity measure is examined using two datasets with different data characteristics. It presents overall superiority to previous measures in terms of recommendation and prediction qualities regardless of the characteristics of the data environment.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제7권5호
• /
• pp.1252-1271
• /
• 2013

Automatic image annotation has become an increasingly important research topic owing to its key role in image retrieval. Simultaneously, it is highly challenging when facing to large-scale dataset with large variance. Practical approaches generally rely on similarity measures defined over images and multi-label prediction methods. More specifically, those approaches usually 1) leverage similarity measures predefined or learned by optimizing for ranking or annotation, which might be not adaptive enough to datasets; and 2) predict labels separately without taking the correlation of labels into account. In this paper, we propose a method for image annotation through collaborative similarity metric learning from dataset and modeling the label correlation of the dataset. The similarity metric is learned by simultaneously optimizing the 1) image ranking using structural SVM (SSVM), and 2) image annotation using correlated label propagation, with respect to the similarity metric. The learned similarity metric, fully exploiting the available information of datasets, would improve the two collaborative components, ranking and annotation, and sequentially the retrieval system itself. We evaluated the proposed method on Corel5k, Corel30k and EspGame databases. The results for annotation and retrieval show the competitive performance of the proposed method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제7권3호
• /
• pp.522-538
• /
• 2013

Content based image retrieval has become an increasingly important research topic for its wide application. It is highly challenging when facing to large-scale database with large variance. The retrieval systems rely on a key component, the predefined or learned similarity measures over images. We note that, the similarity measures can be potential improved if the data distribution information is exploited using a more sophisticated way. In this paper, we propose a similarity measure learning approach for image retrieval. The similarity measure, so called Fisher kernel, is derived from the probabilistic distribution of images and is the function over observed data, hidden variable and model parameters, where the hidden variables encode high level information which are powerful in discrimination and are failed to be exploited in previous methods. We further propose a discriminative learning method for the similarity measure, i.e., encouraging the learned similarity to take a large value for a pair of images with the same label and to take a small value for a pair of images with distinct labels. The learned similarity measure, fully exploiting the data distribution, is well adapted to dataset and would improve the retrieval system. We evaluate the proposed method on Corel-1000, Corel5k, Caltech101 and MIRFlickr 25,000 databases. The results show the competitive performance of the proposed method.

• 한국지능시스템학회논문지
• /
• 제8권6호
• /
• pp.119-121
• /
• 1998

퍼지 시스템의 퍼지 하이퍼그래프에 의해서 표현되었다고 할때, 퍼지 집합을 나타내는 퍼지 에지사이의 유사도를 측정할 필요가 있다. 또한 원소들 사이의 유사도를 측정할 필요가 있다. 본 논문은 이런 필요성에 따라서 퍼지 유사도를 측정하는 척도를 제안한다. 하나는 퍼지 집합 사이의 유사도를 측정하고, 또 하나는 원소 사이의 퍼지 유사도를 측정해 준다. 이 척도는 퍼지집합과 원소 개개의 유사성을 중시하고 시스템 분석 분야에서 이용될 수 있다.

• Communications for Statistical Applications and Methods
• /
• 제11권2호
• /
• pp.369-380
• /
• 2004

Clustering algorithms with different similarity measures are commonly used to find an optimal clustering or close to original clustering. The recovery level of using Euclidean distance and distances transformed from correlation coefficients is evaluated and compared using Rand's (1971) C statistic. The C values present how the resultant clustering is close to the original clustering. In simulation study, the recovery level is improved by applying the correlation coefficients between objects. Using the data set from Spellman et al. (1998), the recovery levels with different similarity measures are also presented. In general, the recovery level of true clusters was increased by using the correlation coefficients.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2022년도 춘계학술발표대회
• /
• pp.366-369
• /
• 2022

The glut of information aggravated the process of data analysis and other procedures including data mining. Many algorithms were devised in Big Data and Data Mining to solve such an intricate problem. In this paper, we conducted research about the comparison of several similarity measures and community detection algorithms in collaborative filtering for movie recommendation systems. Movielense data set was used to do an empirical experiment. We applied three different similarity measures: Cosine, Euclidean, and Pearson. Moreover, betweenness and eigenvector centrality were used to detect communities from the network. As a result, we elucidated which algorithm is more suitable than its counterpart in terms of recommendation accuracy.

• The Journal of the Acoustical Society of Korea
• /
• 제26권2E호
• /
• pp.39-43
• /
• 2007

Music summarization refers to a technique which automatically extracts the most important and representative segments in music content. In this paper, we propose and evaluate a technique which provides the repeated part in music content as music summary. For extracting a repeated segment in music content, the proposed algorithm uses the weighted sum of similarity measures based on multi-level vector quantization for fixed-length summary or optimal-length summary. For similarity measures, count-based similarity measure and distance-based similarity measure are proposed. The number of the same codeword and the Mahalanobis distance of features which have same codeword at the same position in segments are used for count-based and distance-based similarity measure, respectively. Fixed-length music summary is evaluated by measuring the overlapping ratio between hand-made repeated parts and automatically generated ones. Optimal-length music summary is evaluated by calculating how much automatically generated music summary includes repeated parts of the music content. From experiments we observed that optimal-length summary could capture the repeated parts in music content more effectively in terms of summary length than fixed-length summary.

• 정보관리학회지
• /
• 제24권4호
• /
• pp.267-283
• /
• 2007

이 연구에서는 분포 유사도를 문헌 클러스터링에 적용하여 전통적인 코사인 유사도 공식을 대체할 수 있는 가능성을 모색해보았다. 대표적인 분포 유사도인 KL 다이버전스 공식을 변형한 Jansen-Shannon 다이버전스, 대칭적 스큐 다이버전스, 최소스큐 다이버전스의 세 가지 공식을 문헌 벡터에 적용하는 방안을 고안하였다. 분포 유사도를 적용한 문헌 클러스터링 성능을 검증하기 위해서 세 실험 집단을 대상으로 두 가지 실험을 준비하여 실행하였다. 첫 번째 문헌클러스터링실험에서는 최소스큐다이버전스가 코사인 유사도 뿐만 아니라 다른 다이버전스공식의 성능도 확연히 앞서는 뛰어난 성능을 보였다. 두번째 실험에서는 피어슨 상관계수를 이용하여1차 유사도 행렬로부터2차 분포 유사도를 산출하여 문헌 클러스터링을 수행하였다. 실험결과는 2차 분포 유사도가 전반적으로더 좋은 문헌 클러스터링성능을 보이는 것으로 나타났다. 문헌클러스터링에서 처리 시간과 분류 성능을 함께 고려한다면 이 연구에서 제안한 최소 스큐 다이버전스 공식을 사용하고, 분류 성능만 고려할 경우에는 2차 분포 유사도 방식을 사용하는 것이 바람직하다고 판단된다.

• 한국컴퓨터정보학회논문지
• /
• 제24권12호
• /
• pp.93-100
• /
• 2019

유사도 계산은 메모리 기반 협력필터링 시스템의 성능에 매우 중요하다. 이 시스템들은 사용자 평가치들을 이용하여 온라인 상업 사이트에서 고객들에게 상품을 추천한다. 더욱 적합한 추천을 위해 현 사용자와 가장 유사한 사용자들을 선정하여 참조한다. 기존 문헌에는 많은 유사도 척도들이 개발되었는데, 이들은 대개 데이터 희소성이나 완전 시작 문제를 내포하고 있다. 본 논문에서는 기존 척도들과는 달리 사용자 평가치들로부터 선호 정보를 최대한 추출함으로써 희소한 데이터 조건에서도 더욱 신뢰할 수 있는 유사도값을 산출하고자 한다. 사용자 평가치 뿐만 아니라 데이터셋이 제공하는 영화장르 정보를 이용하는 새로운 유사도 척도를 제시한다. 본 척도와 기존의 관련된 척도들의 성능 실험을 하였고, 그 결과, 제안 척도는 주요 성능 평가기준 상으로 더욱 우수하거나 유사한 성능 결과를 보임을 확인하였다.

• 대한기계학회논문집B
• /
• 제23권8호
• /
• pp.1010-1021
• /
• 1999

Camouflage measures in military purpose utilize the apparent temperature difference between target and background, so it is essential to develop a thermal analysis program for apparent temperature predictions and to apply some camouflage measures to real military targets for camouflage purpose. In this study, a thermal analysis program including conduction, convection and radiation is developed and the validity of radiation heat transfer terms is examined. The results show that longwave radiation along with solar radiation should be included in order to predict the apparent temperature as well as the physical temperature precisely. Longwave emissivity variation as an effective camouflage measures is applied to a real M2 tank. From the simulation results, it is found that an effective surface treatment, such as painting of a less emissive material or camouflage clothing, may provide a temperature similarity or a spatial similarity, resulting in an effective camouflage.