• Journal of rehabilitation welfare engineering & assistive technology
• /
• v.8 no.4
• /
• pp.305-312
• /
• 2014

In this study, in order to quantitatively confirm walking rehabilitation degree, we analyzed EMG pattern simulated abnormal gait and normal gait by applying a curve fitting. We calculated the suitable high-order function for EMG signal, and classified them into 5 groups by using cluster analysis. Depending on the distance from normal pattern group, we listed the pattern group and then the distribution of each variables were confirmed. The amplitude-decreased pattern was the most similar to the normal pattern, but the reversed pattern showed the lowest similarity. Due to the smaller overlapping range, the distribution of the groups were possible to classify using the value of variable. The standard deviation of each term coefficient was compared to indicate the quantitative rehabilitation extent, and the higher value was confirmed as the pattern is close to the normal pattern. Consequently, the representation of quantitative rehabilitation extent is expected to contribute to the more effective rehabilitation method study.

• Journal of Korean Society on Water Environment
• /
• v.28 no.2
• /
• pp.202-212
• /
• 2012

Due to the deterioration of water supply network and the deficiency of raw water, the water utility of local governments have performed various projects to improve their revenue water ratio. However, it is very difficult to estimate the cost for maintaining the revenue water ratio at higher level after completing the project, because local governments have different conditions affecting the operating and maintenance cost of water supply network. The purpose of this study is to present a procedure to estimate the operating and maintenance cost required to maintain the target revenue water ratio of the water supply network. For this purpose, we estimated the cost used only for operation and maintenance of water supply network of 164 local governments with the aid of K-Mean Clustering Analysis and the data from 40 representative local governments. Then, the regression analysis was performed to find relationship between revenue water ratio and the operating and maintenance cost with two different data sets generated by two classification methods; the first method classifies the local governments by means of k-means clustering, and the other classifies the local governments according to the index standardized by the operating and maintenance cost per unit length of water mains per revenue water ratio. The results shows that the method based on the index standardized by the cost and revenue water ratio of each government produces more reliable results for finding regression equations between revenue water ratio and the operating and maintenance cost only for water supply network. The estimated regression equations for each group can be used to estimate the cost required to keep the target revenue water ratio of the local government.

• Journal of the Ergonomics Society of Korea
• /
• v.33 no.2
• /
• pp.87-96
• /
• 2014

Objective: The present study is intended to objectively classify upper- & lower-body sitting strategies and identify the effects of gender and OPL type on the sitting strategies. Background: A sitting strategy which statistically represents comfortable driving posture can be used as a reference posture of a humanoid in virtual design and evaluation of a driver's seat. Although previous research has classified sitting strategies for driving postures in various occupant package layout (OPL) types, the existing classification methods are not objective and the factors affecting sitting strategies have not been identified. Method: Forty drivers' preferred driving postures in three different OPL types (coupe, sedan, and SUV) were measured by a motion capture system. Next, the measured driving postures were classified by K-means cluster method. Results: Sitting strategies of upper-body were classified as erect (33%), slouched (41%), and reclined (26%) postures, and those of lower-body were classified as knee bent (42%), knee extended (32%), and upper-leg lifted (26%) postures. Significant differences at ${\alpha}$ = 0.05 in the upper-body sitting strategy by gender and lower-body sitting strategy by OPL type were found. Application: Both the classified sitting strategies and the identified factors would be of use in ergonomic seat design and evaluation.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.42 no.1
• /
• pp.1-7
• /
• 2005

Cluster analysis of microarray data has been often used to find biologically relevant Broups of genes based on their expression levels. Since many functionally related genes tend to be co-expressed, by identifying groups of genes with similar expression profiles, the functionalities of unknown genes can be inferred from those of known genes in the same group. In this Paper we address a novel clustering approach, called seed clustering, and investigate its applicability for microarray data analysis. In the seed clustering method, seed genes are first extracted by computational analysis of their expression profiles and then clusters are generated by taking the seed genes as prototype vectors for target clusters. Since it has strong mathematical foundations, the seed clustering method produces the stable and consistent results in a systematic way. Also, our empirical results indicate that the automatically extracted seed genes are well representative of potential clusters hidden in the data, and that its performance is favorable compared to current approaches.

• The Journal of Korean Association of Computer Education
• /
• v.13 no.1
• /
• pp.65-73
• /
• 2010

This study offered lexical knowledge acquisition model of unsupervised learning method in order to overcome limitation of lexical knowledge hand building manual of supervised learning method for research of natural language processing. The offered model obtains the lexical knowledge from the lexical entry which was given by inputting through the process of vectorization, clustering, lexical knowledge acquisition automatically. In the process of obtaining the lexical knowledge acquisition of model, some parts of lexical knowledge dictionary which changes in the number of lexical knowledge and characteristics of lexical knowledge appeared by parameter changes were shown. The experimental results show that is possibility of automatic building of Machine-readable dictionary, because observed to the number of lexical class information cluster collected constant. also building of lexical ditionary including left-morphosyntactic information and right-morphosyntactic information is reflected korean characteristic.

• Geophysics and Geophysical Exploration
• /
• v.14 no.4
• /
• pp.267-273
• /
• 2011

A cluster of micro-earthquakes in the transition zone between the continental and oceanic crust in the East Sea was relocated using the Joint Hypocenter Determination (JHD) method. In order to increase the number of available earthquakes and to take advantage of the high detection capability of the Korea National Seismic Network (KNSN), continuously recorded seismic data were reviewed to identify 56 micro-earthquakes occurring in a 20 km ${\times}$ 20 km region. The initial earthquake hypocenters were determined using a routine single event location method. Single event locations do not reveal any significant structures in the study area. After relocating the earthquake hypocenters using the JHD technique, the earthquakes were clustered and four potential faults responsible for earthquake generation in the subsurface were delineated. They are defined by two sub-vertical and two steeply south-dipping seismicities located next to each other.

• Journal of the Korean Data and Information Science Society
• /
• v.24 no.2
• /
• pp.245-255
• /
• 2013

The object of study is to propose a method for grouping R, F, and M in RFM model. Our model uses 6 levels using standard normal distribution. First level is upper 2.5% and second level next 13.5%, third level next 34%, fourth level next 34%, fifth level next 13.5%, sixth level next 2.5%. Values are symmetric and limits are clear. We compare proposed model with traditional 5 level model and 10 level model using NDSL data of KISTI. Proposed model divides most clearly the distribution of the RFM function for all cases of weights, because it uses the distribution of customers. Comparison studies of our model with grouping using cluster analysis and studies on weights of RFM model are needed.

• Journal of Broadcast Engineering
• /
• v.18 no.4
• /
• pp.643-646
• /
• 2013

We present an improved AdaBoost algorithm to avoid overfitting phenomenon. AdaBoost is widely known as one of the best solutions for object detection. However, AdaBoost tends to be overfitting when a training dataset has noisy samples. To avoid the overfitting phenomenon of AdaBoost, the proposed method divides positive samples into K clusters using k-means algorithm, and then uses only one cluster to minimize the training error at each iteration of weak learning. Through this, excessive partitions of samples are prevented. Also, noisy samples are excluded for the training of weak learners so that the overfitting phenomenon is effectively reduced. In our experiment, the proposed method shows better classification and generalization ability than conventional boosting algorithms with various real world datasets.

• Journal of the Korea Society for Simulation
• /
• v.17 no.3
• /
• pp.27-34
• /
• 2008

Wireless sensor networks are comprised of sensor nodes with resource-constrained hardware. Nodes in the sensor network without adequate protection may be compromised by adversaries. Such compromised nodes are vulnerable to the attacks like false reports injection attacks and false data injection attacks on legitimate reports. In false report injection attacks, an adversary injects false report into the network with the goal of deceiving the sink or the depletion of the finite amount of energy in a battery powered network. In false data injection attacks on legitimate reports, the attacker may inject a false data for every legitimate report. To address such attacks, the probabilistic voting-based filtering scheme (PVFS) has been proposed by Li and Wu. However, each cluster head in PVFS needs additional transmission device. Therefore, this paper proposes a fuzzy logic-based false report detection method (FRD) to mitigate the threat of these attacks. FRD employs the statistical en-route filtering scheme as a basis and improves upon it. We demonstrate that FRD is efficient with respect to the security it provides, and allows a tradeoff between security and energy consumption, as shown in the simulation.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.32 no.3
• /
• pp.349-361
• /
• 2008

The purpose of this study was to develop bodice basic patterns fitted on different body types of adult women. To meet this purpose, first, human bodies were measured using tight fitting technique and classified based on the shapes of upper trunk. The subjects were 214 women $20{\sim}39$ years of age. For the measurement of female upper trunk, tight fitting technique was utilized. The development of figures of upper trunks were obtained from women. These development of figures were then digitized and analysed using the PAD system. A total of 155 measurements were taken from each of the development of figures. then, 32 measurements were selected for the further analysis. As complimentary data, 22 direct body measurements using an anthropometric method and 23 body measurements using a photographic method from the side view pictures of the participants were also obtained. The results and discussions of this study are as follows: Using the body measurements from the development of figures, a factor analysis and a cluster analysis were conducted. As a result, the body types were classified into 5 different types, which differ in terms of bust volumes, shoulder slopes, shoulder tilts, back silhouettes, body axises. The prominent characteristics of each type are as follows: The first type has a large bust volume. The second type has a right figure. The third type has a rounded back silhouette. The fourth type has a back silhouette of scapular coming backward. Finally the fifth type has a shoulder tilted forward.