• 대한화장품학회:학술대회논문집
• /
• 대한화장품학회 2003년도 IFSCC Conference Proceeding Book I
• /
• pp.733-740
• /
• 2003

To find a new substance with superior melanogenesis inhibitory activity, the bioactivities of alkyl-3,4-dihydroxy-5-substituted benzoate (A) and N-alkyl-3,4-dihydroxy-5-substituted benzamide (B) derivatives as substrate of tyrosinase were measured in mouse melanoma cells. And the bioactivities analyzed using comparative molecular similarity indices analysis (CoMSIA). From the CoMSIA model, when cross-validation value (q$^2$) is 0.713 at four components, the pearson correlation coefficient ($r^2$) is 0.900. Unknown compounds were predicted, using QSAR analyzed results from the CoMSIA methods. Excellent agreement was obtained between the measured and the predicted bioactivities of unknown compounds. As the results of prediction from CoMSIA, we could conclude that the bioactivities were increased from pl$_{50}$=3.18-4.80 to above 5.17 by creation of 6-methylheptyl, n-pentylphenyl and 2-hydroxypentylphenyl group etc,.,.

• 농약과학회지
• /
• 제6권3호
• /
• pp.209-217
• /
• 2002

Metalaxyl 살균제 저항성(RPC)과 감수성(SPC) 고추역병균주(Phytaphthora capsici)들에 대한 2-N-benzyl-5-phenoxy-3-isothiazolone 유도체들의 살균활성을 비교분자 유사성 지수분석(CoMSIA)과 홀로그램 구조-활성 관계(HQSAR) 방법으로 분석하였다. 두 균주의 살균활성에 대한 PLS 계산결과, 교차 확인값($q^2$)과 Pearson 상관계수($r^2$) (CoMSIA: RPC; $q^2=0.675,\;r^2=0.942$, SPC; $q^2=0.350,\;r^2=0.876$ 및 HQSAR: RPC; $q^2=0.519,\;r^2=0.869$, SPC; $q^2=0.483,\;r^2=0.990$)를 비교한 바, 두 방법 모두 양호한 분석 결과를 나타내었다. 그리고 CoMSIA 등고도로부터 특히, RPC에 대한 선택적인 살균활성 요소는 phenoxy-기의 meta, para(C1-C6) 위치에 소수성이 작고 입체적으로 크지 않은 H-결합 받게가 치환 될 경우이었으며 CoMSIA 보다는 HQSAR 방법이 높은 예측성을 나타내었다.

• Journal of Microbiology and Biotechnology
• /
• 제20권1호
• /
• pp.169-178
• /
• 2010

The microbial community structure in Thailand soils contaminated with low and high levels of arsenic was determined by denaturing gradient gel electrophoresis. Band pattern analysis indicated that the bacterial community was not significantly different in the two soils. Phylogenetic analysis obtained by excising and sequencing six bands indicated that the soils were dominated by Arthrobacter koreensis and $\beta$-Proteobacteria. Two hundred and sixty-two bacterial isolates were obtained from arsenic-contaminated soils. The majority of the As-resistant isolates were Gramnegative bacteria. MIC studies indicated that all of the tested bacteria had greater resistance to arsenate than arsenite. Some strains were capable of growing in medium containing up to 1,500 mg/l arsenite and arsenate. Correlations analysis of resistance patterns of arsenite resistance indicated that the isolated bacteria could be categorized into 13 groups, with a maximum similarity value of 100%. All strains were also evaluated for resistance to eight antibiotics. The antibiotic resistance patterns divided the strains into 100 unique groups, indicating that the strains were very diverse. Isolates from each antibiotic resistance group were characterized in more detail by using the repetitive extragenic palindromic-PCR (rep-PCR) DNA fingerprinting technique with ERIC primers. The PCR products were analyzed by agarose gel electrophoresis. The genetic relatedness of 100 bacterial fingerprints, determined by using the Pearson product-moment similarity coefficient, showed that the isolates could be divided into four clusters, with similarity values ranging from 5-99%. Although many isolates were genetically diverse, others were clonal in nature. Additionally, the arsenic-resistant isolates were examined for the presence of arsenic resistance (ars) genes by using PCR, and 30% of the isolates were found to carry an arsenate reductase encoded by the arsC gene.

• 한국정보과학회논문지:컴퓨팅의 실제 및 레터
• /
• 제9권5호
• /
• pp.521-529
• /
• 2003

본 논문에서는 추천시스템에서 많이 활용되는 협업 여과 방법의 문제점으로 지적되고 있는 희소성(sparsity)으로 인한 유사도의 부정확한 문제를 개선하기 위하여, 인구 통계 정보를 이용한 기법을 제안하였다. 두 사용자간의 유사도는 같은 항목에 동시에 평가된 점수를 기반으로 결정되며, 두 사용자가 동시에 평가하지 않은 항목은 유사도 계산에서 제외된다. 제안된 기법은 이러한 평가 점수 부족으로 인하여 유사도 계산이 정확치 못한 단점을 보완하기 위하여, 인구 통계 정보를 이용한 가상 평가 점수를 부가하여 유사도 계산을 개선, 예측의 정확도를 향상시킨 방식으로 기존의 피어슨 상관관계를 이용한 협업여과 방식의 확장이다. 실험은 Grouplens의 영화 평가 자료를 활용하였고, 평균절대오차(MAE)와 반응자 작용특성(ROC)값을 이용하여 제안 기법과 피어슨 상관관계를 이용한 협업 여과 방식을 비교하였다. 제안한 기법이 피어슨 상관관계를 이용한 협업 여과 추천 방식에 비하여 평균절대오차는 9%, 반응자 작용 특성의 민감도는 13% 향상되었음을 확인하였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권8호
• /
• pp.2948-2966
• /
• 2014

To promote recommendation services through prediction quality, some privacy-preserving collaborative filtering solutions are proposed to make e-commerce parties collaborate on partitioned data. It is almost probable that two parties hold ratings for the same users and items simultaneously; however, existing two-party privacy-preserving collaborative filtering solutions do not cover such overlaps. Since rating values and rated items are confidential, overlapping ratings make privacy-preservation more challenging. This study examines how to estimate predictions privately based on partitioned data with overlapped entries between two e-commerce companies. We consider both user-based and item-based collaborative filtering approaches and propose novel privacy-preserving collaborative filtering schemes in this sense. We also evaluate our schemes using real movie dataset, and the empirical outcomes show that the parties can promote collaborative services using our schemes.

• 아동학회지
• /
• 제13권2호
• /
• pp.251-272
• /
• 1992

The purpose of this study is to investigate the relationships among the beliefs of parents and children, children's cognitive and emotional behavior based on cognitive and interreactionary approach models. The Subjects were 138 children (68 eight-year-olds and 70 eleven-year-olds) and their parents. Instruments used in this study were the modified Family Belief Interview Schedule(Alessandri & Wozniak, 1987), the Standard Achievement Test, and Harter's Perceived Competence Scale. Data analysis was by Pearson's r product moment correlation, two-way ANOVA, Fisher-Z test and Student-Newman-Keuls post-hoc test. The major findings are as follows: (1) The beliefs of parents and children has a significant influence children's perception of competence. (2) The congruence of parents' and children's beliefs was hightest in "assumed similarity". (3) Mother's positive view of their children and congruence of mother's and father's beliefs were correlated with children's academic achievement. Parents' positive beliefs and congruence of beliefs were also correlated with children's self-perception of competence.

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 2001년도 봄 학술발표논문집 Vol.28 No.1 (B)
• /
• pp.103-105
• /
• 2001

인터넷 사용 인구의 폭증으로 인터넷 사이트가 경쟁적으로 유용한 각종 정보를 사용자들에게 제공하여 보다 많은 수의 회원을 확보하기 위해 노력하고 있지만 여러 사이트를 동시에 사용하고 있는 대부분의 인터넷 사용자들에게는 각 사이트에서 날아드는 정보를 매번 일일이 검색해야 하는 일이 여간 번거롭지 않을 뿐만 아니라 이런 무분별하고 획일적인 정보 서비스는 오히려 사용자들의 인터넷 사용을 불편하게 하며 더욱이 그 내용이 관심 밖의 것이 경우 네트워크의 효율적인 사용을 저해하는 정보공해에 지나지 않게 된다. 추천엔진은 기본으로 끊임없이 유입되는 다량의 정보 중에서 필요한 것을 추천해 주는 것이다. 이에 본 논문에서는 사용자들에게 필요한 정보만을 효율적으로 전달 해주기 위해서 먼저 개인화된 정보의 전달을 위해 사용자의취향을 파악하여 선택 가능성이 높은 항목을 예측할 수 있어야 한다. 그리고 사용자와 가까운 K 명의 사용자들을 효율적으로 검색하기 위해서 K-최근접 이웃 방식을 사용하고 인덱싱을 사용할 수 있는 세가지 벡터 유사도를 기존의 피어슨 상관계수(Pearson Correlation)와 비교하여 제안한다. 이를 통해 정보의 효율적인 제공방법, 즉 일반적인 검색으로 인한 정보의 제공이 아닌 일반 사용자들의 추천에 의해 정보를 제공하는 K-최근접 이웃 추천 엔진을 세가지 벡터 유사도를 이용해서 분석한다.

• 한국컴퓨터정보학회논문지
• /
• 제22권7호
• /
• pp.125-131
• /
• 2017

It is not easy for the user to find the information that is appropriate for the user among the suddenly increasing information in recent years. One of the ways to help individuals make decisions in such a lot of information is the recommendation system. Although there are many recommendation methods for such recommendation systems, a representative method is collaborative filtering. In this paper, we design and implement the movie recommendation system on user-based collaborative filtering of apache mahout. In addition, Pearson correlation coefficient is used as a method of measuring the similarity between users. We evaluate Precision and Recall using the MovieLens 100k dataset for performance evaluation.

• 한국컴퓨터정보학회논문지
• /
• 제23권4호
• /
• pp.89-95
• /
• 2018

It is not easy for the user to find the information that is appropriate for the user among the suddenly increasing information in recent years. One of the ways to help individuals make decisions in such a lot of information is the recommendation system. Although there are many recommendation methods for such recommendation systems, a representative method is collaborative filtering. In this paper, we design and implement the movie recommendation system on user-based collaborative filtering of apache mahout based on mongoDB. In addition, Pearson correlation coefficient is used as a method of measuring the similarity between users. We evaluate Precision and Recall using the MovieLens 100k dataset for performance evaluation.

• 한국IT서비스학회지
• /
• 제7권4호
• /
• pp.199-208
• /
• 2008

Tagging represents one of the Web 2.0 technology, and has an appropriate mechanism for the classification of dynamically changing Web informations. This technique is capable of searching the Web informations using the user specified tags, but still it has a limitation of providing only the limited informations to the tags. Therefore, in order to search the related informations easily, we need to extend this technique further to search not only the desired informations through the designated tags and also the related informations. In this paper, we first have designed and developed an algorithm that can get a desired tag cluster, which is capable of collecting the searched tags along with the related tags. We first performed a test to compare the difference between the user collected tag data through RSS and the reduced data. The second test focused on the accuracy of extracted related tags that depends on the similarity functions, such as the Pearson Correlation and Euclidean. Finally, we showed the final results visually using the graph algorithm.