• Journal of the Korean Society for information Management
• /
• v.25 no.4
• /
• pp.227-243
• /
• 2008

This study investigates an event detection method with the aim of generating an event-focused news summary from a set of news articles on a certain event using a multi-document summarization technique. The event detection method first classifies news articles into the event related topic categories by employing a SVM classifier and then creates event clusters containing news articles on an event by a modified single pass clustering algorithm. The clustering algorithm applies a time penalty function as well as cluster partitioning to enhance the clustering performance. It was found that the event detection method proposed in this study showed a satisfactory performance in terms of both the F-measure and the detection cost.

• The Research Journal of the Costume Culture
• /
• v.17 no.6
• /
• pp.1099-1111
• /
• 2009

The purpose of this study is to classify body shapes of aged women by using 3D body scan data. For the body shape analysis and classification, 3D body scan data of 270 aged women were used, and 16 main measurements consisting of a human body were used to conduct factor analysis, cluster analysis and discriminant analysis. The analysis were performed on all 'the method using the absolute value', 'the method using index of height and weight', and 'the method using index of height', and according to the classification results, the method which categorizes body shapes best in terms of their shapes was adopted. As the factor analysis result using the numerical value of height to categorize the body shapes of the aged women, factor 1 was the thickness and width for the height, factor 2 was the height of the upper part of the body for the height, factor 3 was the height of hips for the height, and factor 4 was the height of belly for the height. When the body shapes were categorized with the deducted factors as variables, they were divided into two types. Type 1 was a short and fat body shape($\blacksquare$ type) and 55.6% of the subjects were of this type. Type 2 was for the body shape whose vertical height, including weight, was long but all kinds of width and thickness were small, that is, tall and thin body shape($\blacksquare$ type), and 44.4% of the aged women were in this case.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.49 no.2
• /
• pp.134-142
• /
• 2012

This paper analyze an LDPC code replication method suitable for cloud computing. First, we determine the number of blocks suitable for cloud computing through analysis of the performance for the file availability and storage overhead. Also we determine the type of LDPC code appropriate for cloud computing through the performance for three types of LDPC codes. Finally we present the graph random generation method and the comparing method of each generated LDPC code's performance by the iterative decoding process. By the simulation, we confirmed the best graph's regularity is left-regular or least left-regular. Also, we confirmed the best graph's total number of edges are minimum value or near the minimum value.

• Journal of Ginseng Research
• /
• v.41 no.3
• /
• pp.290-297
• /
• 2017

Background: The use of X-ray phase-contrast microtomography for the investigation of Chinese medicinal materials is advantageous for its nondestructive, in situ, and three-dimensional quantitative imaging properties. Methods: The X-ray phase-contrast microtomography quantitative imaging method was used to investigate the microstructure of ginseng, and the phase-retrieval method is also employed to process the experimental data. Four different ginseng samples were collected and investigated; these were classified according to their species, production area, and sample growth pattern. Results: The quantitative internal characteristic microstructures of ginseng were extracted successfully. The size and position distributions of the calcium oxalate cluster crystals (COCCs), important secondary metabolites that accumulate in ginseng, are revealed by the three-dimensional quantitative imaging method. The volume and amount of the COCCs in different species of the ginseng are obtained by a quantitative analysis of the three-dimensional microstructures, which shows obvious difference among the four species of ginseng. Conclusion: This study is the first to provide evidence of the distribution characteristics of COCCs to identify four types of ginseng, with regard to species authentication and age identification, by X-ray phase-contrast microtomography quantitative imaging. This method is also expected to reveal important relationships between COCCs and the occurrence of the effective medicinal components of ginseng.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.46 no.11
• /
• pp.952-960
• /
• 2018

Recently, lots of remote sensing satellite require agility to collect more images within the limited time frame. To satisfy this kind of mission requirement, high torque actuator such as CMG is an essential element. In this study, 5Nm class small CMG developed by KARI is introduced to implement for an agile small satellite design. One of the singularity escape CMG steering law, Designated Direction Escape (DDE) method, which is a sort of modified version of Singular Direction Avoidance (SDA) method is summarized for its application on the numerical simulation of agile attitude control system design result. The performance of DDE method is demonstrated properly by escaping well known elliptic internal singularity successfully. 5Nm class small CMG cluster in a pyramid type as well as a roof type configuration is utilized to perform the numerical simulation and to demonstrate its agility design result for a small satellite. Simulation result shows the properness of 5Nm small CMG to a small agile satellite system. Also, the simulation result provides some valuable information that is important to CMG hardware design and manufacturing.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.45 no.5
• /
• pp.881-891
• /
• 2021

This study proposes a simple drape measurement method for the 3D virtualization of garments. The proposed method uses angles or disks of different diameters to evaluate the drape properties easily. We divided 710 fabrics into ten groups based on the drape coefficient, of which 49.6% had drape coefficients of 30 or less. The drape properties were measured to classify the groups into smaller clusters using the angle formed when the center of the fabric was fixed. Accordingly, three clusters were formed for 60° and 100° angles. A method was devised using ten disks of different diameters to classify the remaining two clusters, except the cluster containing only the D10 group (D1-D5 and D5-D9). Three criteria-grade match, a sum of deviation, and standardization of deviation-were used for the classifications. The discriminative ability between groups was high for D1-D5 with disks with 24.0 and 25.5 cm diameters. Furthermore, a disk with a diameter of 16.5 cm was effective for D5-D9. The three-dimensional drape shapes were unique for the ten groups, which can be utilized as fundamental data for 3D virtualization.

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.6
• /
• pp.566-571
• /
• 2009

This paper proposes noise-robust fast speaker adaptation method based on the eigenvoice framework in various noisy environments. The proposed method is focused on de-noising and environment clustering. Since the de-noised adaptation DB still has residual noise in itself, environment clustering divides the noisy adaptation data into similar environments by a clustering method using the cepstral mean of non-speech segments as a feature vector. Then each adaptation data in the same cluster is used to build an environment-clustered speaker adapted (SA) model. After selecting multiple environmentally clustered SA models which are similar to test environment, the speaker adaptation based on an appropriate linear combination of clustered SA models is conducted. According to our experiments, we observe that the proposed method provides error rate reduction of $40{\sim}59%$ over baseline with speaker independent model.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.2
• /
• pp.78-87
• /
• 2006

In this paper. we propose an effective mask estimation scheme for missing-feature reconstruction in order to achieve robust speech recognition under unknown noise environments. In the previous work. colored noise is used for training the mask classifer, which is generated from the entire frequency Partitioned signals. However it gives a limited performance under the restricted number of training database. To reflect the spectral events of more various background noise and improve the performance simultaneously. a new Bayesian classifier for mask estimation is proposed, which works independent of other frequency bands. In the proposed method, we employ the colored noise which is obtained by combining colored noises generated from each frequency band in order to reflect more various noise environments and mitigate the 'sparse' database problem. Combined with the cluster-based missing-feature reconstruction. the performance of the proposed method is evaluated on a task of noisy speech recognition. The results show that the proposed method has improved performance compared to the Previous method under white noise. car noise and background music conditions.

• Journal of Nutrition and Health
• /
• v.53 no.5
• /
• pp.547-561
• /
• 2020

Purpose: This study was conducted to segment the delivery food market and to develop customized products and services. Methods: This study analyzed 636 responses collected from customers who ordered delivery food. Statistical analyses were conducted using the SPSS program (ver. 25.0) for frequency analysis, χ²-test, one-way analysis of variance, factor analysis, and cluster analysis. Results: Four factors were extracted by exploratory factor analysis (safety-orientation, convenience-orientation, taste-orientation, and economy-orientation) to explain the consumers' food-related lifestyles. The results of cluster analysis indicated that the 'low-interest group', 'convenience and economy-oriented group', and 'gourmet and economy-oriented group' should be regarded as the target segments. Characteristic analysis of each cluster showed that lowinterest group had higher rates of married (67.1%) and living with family (85.4%) than other clusters. The convenience and the economy-oriented group had higher rates of living alone (28.9%) than others. The gourmet and the economy-oriented group had a higher percentage of unmarried (62.0%) than the others. In addition, the average age of convenience and economy-oriented group (32.3 years) and gourmet and economy-oriented group (32.5 years) were significantly lower than the safety seeker (40.0 years) (p < 0.001). Difference analysis of the consumption practice according to the cluster, revealed significant differences in the order frequency (p < 0.001), main day to order (p < 0.05), source of information about delivery food (p < 0.001), order method (p < 0.001), and co-consumer (p < 0.01). In addition, the convenience and the economy-oriented group had significantly higher overall satisfaction than the others (p < 0.001). Conclusion: These findings suggest that customer segmentation based on a food-related lifestyle can be used to build a successful marketing strategy. Therefore, restaurant managers and delivery platform operators should consider developing products and services according to the segmentation to maximize customer satisfaction.

• Journal of Korean Society of Environmental Engineers
• /
• v.27 no.8
• /
• pp.870-876
• /
• 2005

The multiple regression models which have two factors(population and commercial area) have been used to forecast the water demand in the future. But, the coefficient of population had a negative value because proper regional classification wasn't performed, and it is not reasonable because the population must be a positive factor. So, the regional classification was performed by principal component and cluster analysis to solve the problem. 6 regional characters were transformed into 4 principal components, and the areas were divided into two groups according to cluster analysis which had 4 principal components. The new regression models were made by each group, and the problem was solved. And, the future water demands were estimated by three scenarios(Active, moderate, and passive one). The increase of water demand ore $89.034\;m^3/day$ in active plat $49,077\;m^3/day$ in moderate plan, and $19,996\;m^3/day$ in passive plan. The water supply ability as scenarios is enough in water treatment plant, however, 2 reservoirs among 4 reservoirs don't have enough retention time in all scenarios.