• The Journal of the Acoustical Society of Korea
• /
• v.18 no.8
• /
• pp.3-8
• /
• 1999

Previous variable vocabulary speech recognition systems with context-independent acoustic modeling, could not represent the effect of neighboring phonemes. To solve this problem, we use allophone-based context-dependent acoustic model. This paper describes the method to improve acoustic model of the system effectively. Acoustic model is improved by using allophone clustering technique that uses entropy as a similarity measure and the optimal allophone model is generated by changing the number of allophones. We evaluate performance of the improved system by using Phonetically Optimized Words(POW) DB and PC commands(PC) DB. As a result, the allophone model composed of six hundreds allophones improved the recognition rate by 13% from the original context independent model m POW test DB.

• Journal of the Korea Society of Computer and Information
• /
• v.17 no.2
• /
• pp.49-57
• /
• 2012

Conventional fuzzy c-means (FCM) algorithms have achieved a good clustering performance. However, they do not fully utilize the spatial information in the image and this results in lower clustering performance for images that have low contrast, vague boundaries, and noises. To overcome this issue, we propose an enhanced spatial fuzzy c-means (ESFCM) algorithm that takes into account the influence of neighboring pixels on the center pixel by assigning weights to the neighbors in a $3{\times}3$ square window. To evaluate between the proposed ESFCM and various FCM based segmentation algorithms, we utilized clustering validity functions such as partition coefficient ($V_{pc}$), partition entropy ($V_{pe}$), and Xie-Bdni function ($V_{xb}$). Experimental results show that the proposed ESFCM outperforms other FCM based algorithms in terms of clustering validity functions.

• Journal of the Korea Society of Computer and Information
• /
• v.23 no.9
• /
• pp.81-86
• /
• 2018

Software defect severity is very important in projects with limited historical data or new projects. But general software defect prediction is very difficult to collect the label information of the training set and cross-project defect prediction must have a lot of data. In this paper, an unclassified data set with defect severity is clustered according to the distribution ratio. And defect severity-based prediction model is proposed by way of labeling. Proposed model is applied CLAMI in JM1, PC4 with the least ambiguity of defect severity-based NASA dataset. And it is evaluated the value of ACC compared to original data. In this study experiment result, proposed model is improved JM1 0.15 (15%), PC4 0.12(12%) than existing defect severity-based prediction models.

• 한국전산유체공학회:학술대회논문집
• /
• 2006.05a
• /
• pp.361-363
• /
• 2006

• Structural Engineering and Mechanics
• /
• v.82 no.3
• /
• pp.271-282
• /
• 2022

The active learning surrogate model based on adaptive sampling strategy is increasingly popular in reliability analysis. However, most of the existing sampling strategies adopt the trial and error method to determine the size of the Monte Carlo (MC) candidate sample pool which satisfies the requirement of variation coefficient of failure probability. It will lead to a reduction in the calculation efficiency of reliability analysis. To avoid this defect, a new method for determining the optimal size of the MC candidate sample pool is proposed, and a new structural reliability analysis method combining polynomial chaos-based Kriging model (PC-Kriging) with adaptive sampling region is also proposed (PCK-ASR). Firstly, based on the lower limit of the confidence interval, a new method for estimating the optimal size of the MC candidate sample pool is proposed. Secondly, based on the upper limit of the confidence interval, an adaptive sampling region strategy similar to the radial centralized sampling method is developed. Then, the k-means++ clustering technique and the learning function LIF are used to complete the adaptive design of experiments (DoE). Finally, the effectiveness and accuracy of the PCK-ASR method are verified by three numerical examples and one practical engineering example.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.6
• /
• pp.383-394
• /
• 2013

As a way of augmenting constrained resources of mobile devices such as CPU and memory, many works on mobile cloud computing (MCC), where mobile devices utilize remote resources of cloud services or PCs, have been proposed. Typically, in MCC, many nodes with different operating systems and platform and diverse mobile applications or services are located, and a central manager autonomously performs several management tasks to maintain a consistent level of MCC overall quality. However, as there are a larger number of nodes, mobile applications, and services subscribed by the mobile applications and their interactions are extremely increased, a traditional management method of MCC reveals a fundamental problem of degrading its overall performance due to overloaded management tasks to the central manager, i.e. a bottle neck phenomenon. Therefore, in this paper, we propose a clustering-based optimization method to solve performance-related problems on large-scaled MCC and to stabilize its overall quality. With our proposed method, we can ensure to minimize the management overloads and stabilize the quality of MCC in an active and autonomous way.

• Journal of KIISE:Information Networking
• /
• v.31 no.6
• /
• pp.582-594
• /
• 2004

In school PC labs or other educational purpose PC labs with a few dozens of PCs, computers are configured in a distributed architecture so that they are set up, maintained and upgraded separately. As an alternative to the distributed architecture, we can consider a thin-client computing environment. In a thin-client computing environment, client side devices provide mainly I/O functions with user friendly GUI and multimedia processing support whereas remote servers called terminal server provide computing power. In order to support many clients in the environment, a cluster of terminal servers can be configured. In this architecture, it is difficult due to the characteristics of terminal session persistence and different pattern of computing usage of users so that the utilization of terminal server resources becomes low. To overcome this disadvantage, we propose an adaptive terminal cluster where terminal servers ,ire partitioned into groups and a terminal server in a light-loaded group can be dynamically reassigned to a heavy-loaded group at run time. The proposed adaptive scheme is compared with a generic terminal service cluster and a group based non-adaptive terminal server cluster. Experimental results show the effectiveness of the proposed scheme.

• Journal of Food Hygiene and Safety
• /
• v.35 no.5
• /
• pp.430-437
• /
• 2020

This is an approach study on the sensory properties (taste and odor) of 15 types of Korean conventional soups and stews using electronic nose and tongue. The relative sensor intensity for the taste components of the samples using electronic tongue was demonstrated. By SRS (sourness) sensor, sogogi-baechuguk (beef and cabbage soup) had the highest rate of 9.0. The STS (saltiness) sensor showed the highest score of 8.2 for ojingeoguk (squid soup). For the UMS (umami) sensor, which identifies savoriness, the sogogi-baechuguk was the highest at 10.1. The SWS (sweetness) sensors showed relatively little difference, with sigeumchi-doenjangguk (spinach and bean paste soup) at the highest of 7.3. According to the BRS sensor, which tests for bitterness, the siraegi-doenjangguk (dried radish green and bean paste soup) was the highest at 7.8. By principal component analysis (PCA), we observed variances of 56.21% in principal component 1 (PC1) and 25.23% in PC2. For each flavor component, we observed -0.95 and -0.20 for factor loading of PC1 and PC2 for SRS sensors, 0.96 and 0.14 for STS sensors, and -0.94 and 0.22 for PC1 and PC2 for UMS sensors, and PC1 and 0.22 for PC1 and PC2 loading for SWS sensors. The similarity between the samples identified by clustering analysis was largely identified by 4 clusters. A total of 25 kinds of volatile compounds in 15 samples were identified, and the ones showing the highest relative content in all samples were identified as ethanol and 2-methylthiophhene. The main ingredient analysis confirmed variances of 28.54% in PC1 and 20.80% in PC2 as a result of the pattern for volatile compounds in 15 samples. In the cluster analysis, it was found to be largely classified into 3 clusters. The data in this study can be used for a sensory property database of conventional Korean soups and stews using electronic sensors.

• Spatial Information Research
• /
• v.11 no.3
• /
• pp.213-226
• /
• 2003

With the increasing demand of processing massive geographic data, conventional GISs based on the single processor architecture appear to be problematic. Especially, performing complex GIS operations on the massive geographic data is very time consuming and even impossible. This is due to the processor speed development does not keep up with the data volume to be processed. In the field of GIS, this PC clustering is one of the emerging technology for handling massive geographic data effectively. In this study, a MPI(Message Passing Interface)-based parallel processing approach was conducted to implement the existing raster GIS operations that typically requires massive geographic data sets in order to improve the processing capabilities and performance. Specially for this research, four types of raster CIS operations that Tomlin(1990) has introduced for systematic analysis of raster GIS operation. A data decomposition method was designed and implemented for selected raster GIS operations.

• International Journal of Oral Biology
• /
• v.43 no.2
• /
• pp.101-111
• /
• 2018

Understanding the classification of malocclusion is a crucial issue in Orthodontics. It can also help us to diagnose, treat, and understand malocclusion to establish a standard for definite class of patients. Principal component analysis (PCA) and k-means algorithms have been emerging as data analytic methods for cephalometric measurements, due to their intuitive concepts and application potentials. This study analyzed the macro- and meso-scale classification structure and feature basis vectors of 1020 (415 male, 605 female; mean age, 25 years) orthodontic patients using statistical preprocessing, PCA, random matrix theory (RMT) and k-means algorithms. RMT results show that 7 principal components (PCs) are significant standard in the extraction of features. Using k-means algorithms, 3 and 6 clusters were identified and the axes of PC1~3 were determined to be significant for patient classification. Macro-scale classification denotes skeletal Class I, II, III and PC1 means anteroposterior discrepancy of the maxilla and mandible and mandibular position. PC2 and PC3 means vertical pattern and maxillary position respectively; they played significant roles in the meso-scale classification. In conclusion, the typical patient profile (TPP) of each class showed that the data-based classification corresponds with the clinical classification of orthodontic patients. This data-based study can provide insight into the development of new diagnostic classifications.