• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.36 no.2
• /
• pp.74-80
• /
• 2013

In this study, a novel and flexible recommender system was developed, based on product taxonomy and usage patterns of users. The proposed system consists of the following four steps : (i) estimation of the product-preference matrix, (ii) construction of the product-preference matrix, (iii) estimation of the popularity and similarity levels for sought-after products, and (iv) recommendation of a products for the user. The product-preference matrix for each user is estimated through a linear combination of clicks, basket placements, and purchase statuses. Then the preference matrix of a particular genre is constructed by computing the ratios of the number of clicks, basket placements, and purchases of a product with respect to the total. The popularity and similarity levels of a user's clicked product are estimated with an entropy index. Based on this information, collaborative and content-based filtering is used to recommend a product to the user. To assess the effectiveness of the proposed approach, an empirical study was conducted by constructing an experimental e-commerce site. Our results clearly showed that the proposed hybrid method is superior to conventional methods.

• Journal of the Korean Institute of Telematics and Electronics T
• /
• v.36T no.1
• /
• pp.93-102
• /
• 1999

In this paper, we propose a data model and a retrieval method for content-based video query After partitioning a video into frame sets of same length which is called video-window, each video-window can be mapped to a point in a multidimensional space. A video can be represented a trajectory by connection of neighboring video-window in a multidimensional space. The similarity between two video-windows is defined as the euclidean distance of two points in multidimensional space, and the similarity between two video segments of arbitrary length is obtained by comparing corresponding trajectory. A new retrieval method with filtering and refinement step if developed, which return correct results and makes retrieval speed increase by 4.7 times approximately in comparison to a method without filtering and refinement step.

• BMB Reports
• /
• v.39 no.3
• /
• pp.240-246
• /
• 2006

The base compositional correlations that hold among various coding and noncoding regions of the canine genome have been analysed. The distribution pattern of genes, on the basis of $GC_3$ composition, shows a wide range similar to that observed in human. However the occurrence of maximum number of genes was observed in the range of 65-75% of $GC_3$ composition. The correlation between the coding DNA sequences of canine with the different noncoding regions (introns and flanking regions) is found to be significant and in many cases the degree of correlation show similarity to human genome. We found that these correlations are not limited to the GC content alone, but is holding at the level of the frequency of individual bases as well. The present study suggests that canines ideally belong to the predicted 'general mammalian pattern' of genome composition along with human beings.

• Journal of Institute of Control, Robotics and Systems
• /
• v.8 no.4
• /
• pp.333-337
• /
• 2002

A content-based image retrieval method using color, texture, and shape features is proposed in this paper. A region segmentation technique using PIM(Picture Information Measure) entropy is used for similarity indexing. For segmentation, a color image is first transformed to a gray image and it is divided into n$\times$n non-overlapping blocks. Entropy using PIM is obtained from each block. Adequate variance to perform good segmentation of images in the database is obtained heuristically. As variance increases up to some bound, objects within the image can be easily segmented from the background. Therefore, variance is a good indication for adequate image segmentation. For high variance image, the image is segmented into two regions-high and low entropy regions. In high entropy region, hue-saturation-intensity and canny edge histograms are used for image similarity calculation. For image having lower variance is well represented by global texture information. Experiments show that the proposed method displayed similar images at the average of 4th rank for top-10 retrieval case.

• Journal of the Institute of Convergence Signal Processing
• /
• v.3 no.1
• /
• pp.1-7
• /
• 2002

In this paper, when users choose a query image, we implemented a content-based image retrieval system that users can simply choose and extract a object region of query wanted with not only a whole image but various objects in it. Histogram is obtained by improved HSV transformations from query image and then candidate images are retrieved rapidly by a 1st similarity measure with histogram intersection using representative colors of query image. And finally retrieved images are extracted since 2nd similarity measure with banded autocorrelogram is performed so that recall and precision are improved by combining two retrieval methods that can make up for respective weak points. Moreover images in the database are indexed automatically within feature library that makes possible to retrieve images rapidly.

• International Journal of Knowledge Content Development & Technology
• /
• v.9 no.3
• /
• pp.23-41
• /
• 2019

This study combines the concept of degree centrality in complex network with the Term Frequency $^*$ Proportional Document Frequency ($TF^*PDF$) algorithm; the combined method, called PMCN (PDF-Modified similarity and Complex Network), constructs relationship networks among sentences for writing news summaries. The PMCN method is a multi-document summarization extension of the ideas of Bun and Ishizuka (2002), who first published the $TF^*PDF$ algorithm for detecting hot topics. In their $TF^*PDF$ algorithm, Bun and Ishizuka defined the publisher of a news item as its channel. If the PDF weight of a term is higher than the weights of other terms, then the term is hotter than the other terms. However, this study attempts to develop summaries for news items. Because the $TF^*PDF$ algorithm summarizes daily news, PMCN replaces the concept of "channel" with "the date of the news event", and uses the resulting chronicle ordering for a multi-document summarization algorithm, of which the F-measure scores were 0.042 and 0.051 higher than LexRank for the famous d30001t and d30003t tasks, respectively.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.5
• /
• pp.1163-1185
• /
• 2024

With the sharp increase in the volume of literature data, researchers must spend considerable time and energy locating desired papers. A paper recommendation is the means necessary to solve this problem. Unfortunately, the large amount of data combined with sparsity makes personalizing papers challenging. Traditional matrix decomposition models have cold-start issues. Most overlook the importance of information and fail to consider the introduction of noise when using side information, resulting in unsatisfactory recommendations. This study proposes a paper recommendation method (PR-SLSMF) using document-level representation learning with citation-informed transformers (SPECTER) and low-rank and sparse matrix factorization; it uses SPECTER to learn paper content representation. The model calculates the similarity between papers and constructs a weighted heterogeneous information network (HIN), including citation and content similarity information. This method combines the LSMF method with HIN, effectively alleviating data sparsity and cold-start issues and avoiding topic drift. We validated the effectiveness of this method on two real datasets and the necessity of adding side information.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.66 no.8
• /
• pp.1257-1264
• /
• 2017

In video content watching service, when a user requests video content, the content server has to transmit the entire video to the client for watching. This transmission delay increases as the size of video content increases. In order to solve the transmission delay problem, a prefetching technique can be used in which a video content to be watched by a user is predicted and transmitted to a client before the user requests it. In this paper, we propose a prefetching system considering multiple indices for video content. In the proposed method, video content to be prefetched is selected by comprehensively analyzing the order relation index indicating the order of viewing the videos of the users, the similarity index between the video contents, and the popularity index reflecting the viewing frequency of the video content. Experimental results show that the maximum accuracy is achieved when prefetching uses only the order relation index for movie contents.

• Applied Biological Chemistry
• /
• v.29 no.4
• /
• pp.366-371
• /
• 1986

Fatty acid compositions of Panax ginseng leaves (6 year) grown under different light intensity in field and of the detached leaves exposed to high temperature (20 hours) were investigated by gas chromatography. Linoleic, linolenic, palmitic and palmitoleic acid were the major components(80%) of leaf lipid. The higher the growth light intensity, the lower the percentage of unsaturated acids or bonds, indicating metabolic adaptation to high temperature. Pattern similarity of fatty acid composition was little changed until 20% light but significantly different at 30%, suggesting 20% as limitation light intensity. The close similarity of fatty acid composition between the loaves grown under 30% light and the one at harvest rises uncertainty between adaptation to high temperature and senescence. Total fatty acid content decreased with the increase of light intensity. Short term high temperature $(25^{\circ}C\;or\;35^{\circ}C)$ increased total fatty acid content, unsaturated acid percentage and insignificant difference in pattern similarity of composition.

• The KIPS Transactions:PartD
• /
• v.9D no.6
• /
• pp.1009-1016
• /
• 2002

We design and implement a new case-based recommender system using implicit rating information for a digital content site. Our system consists of the User Profile Generation module, the Similarity Evaluation and Recommendation module, and the Personalized Mailing module. In the User Profile Generation Module, we define intra-attribute and inter-attribute weight deriver from own's past interests of a user stored in the access logs to extract individual preferences for a content. A new similarity function is presented in the Similarity Evaluation and Recommendation Module to estimate similarities between new items set and the user profile. The Personalized Mailing Module sends individual recommended mails that are transformed into platform-independent XML document format to users. To verify the efficiency of our system, we have performed experimental comparisons between the proposed model and the collaborative filtering technique by mean absolute error (MAE) and receiver operating characteristic (ROC) values. The results show that the proposed model is more efficient than the traditional collaborative filtering technique.