• Journal of Information Technology Applications and Management
• /
• v.15 no.1
• /
• pp.271-282
• /
• 2008

Database administrators are demanded to acquire much knowledges and take great efforts for keeping consistent performance in system. Various principles, methods, and tools have been proposed in many studies and commercial products in order to alleviate such burdens on database administrators, and it has resulted to the automation of DBMS which reduces the intervention of database administrator. This paper suggests a resource selection method that estimates the status of the database system based on the workload characteristics and that recommends tuneable resources. Our method tries to simplify selection information on DBMS status using data-mining techniques, enhance the accuracy of the selection model, and recommend tuneable resource. For evaluating the performance of our method, instances are collected in TPC-C and TPC-W workloads, and accuracy are calculated using 10 cross validation method, comparisons are made between our scheme and the method which uses only the classification procedure without any simplification of informations. It is shown that our method has over 90% accuracy and can perform tuneable resource selection.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.6
• /
• pp.319-325
• /
• 2023

In this study, we proposed a method for hyperparameter optimization in the building and training of a deep learning model designed to process point cloud data collected by a millimeter-wave radar system. The primary aim of this study is to facilitate the deployment of a baseline model in resource-constrained IoT devices. We evaluated a RadHAR baseline deep learning model trained on a public dataset composed of point clouds representing five distinct human activities. Additionally, we introduced a coarse-to-fine hyperparameter optimization procedure, showing substantial potential to enhance model efficiency without compromising predictive performance. Experimental results show the feasibility of significantly reducing model size without adversely impacting performance. Specifically, the optimized model demonstrated a 3.3% improvement in classification accuracy despite a 16.8% reduction in number of parameters compared th the baseline model. In conclusion, this research offers valuable insights for the development of deep learning models for resource-constrained IoT devices, underscoring the potential of hyperparameter optimization and model size reduction strategies. This work contributes to enhancing the practicality and usability of deep learning models in real-world environments, where high levels of accuracy and efficiency in data processing and classification tasks are required.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.35 no.9B
• /
• pp.1287-1294
• /
• 2010

The traffic classification is a preliminary and essential step for stable network service provision and efficient network resource management. While a number of classification methods have been introduced in literature, the payload signature-based classification method shows the highest performance in terms of accuracy, completeness, and practicality. However, the payload signature-based method has a significant drawback in high-speed network environment that the processing speed is much slower than other classification method such as header-based and statistical methods. In this paper, We describes various design options to improve the processing speed of traffic classification in design of a payload signature based classification system and describes our selections on the development of our traffic classification system. Also the feasibility of our selection was proved through experimental evaluation on our campus traffic trace.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.38B no.7
• /
• pp.519-525
• /
• 2013

The traffic classification is a preliminary and essential step for stable network service provision and efficient network resource management. However, the payload signature-based method has a significant drawback in high-speed network environment that the processing speed is much slower than other method such as header-based and statistical methods. In this paper, We propose the server IP, Port cache-based traffic classification method using application traffic locality to improve the processing speed of traffic classification. The suggested method achieved about 10 folds improvement in processing speed and 10% improvement in completeness over the payload-based classification system.

• Journal of the Korea Society of Computer and Information
• /
• v.13 no.1
• /
• pp.1-10
• /
• 2008

In order to high-performance data Processing, effective resource selection is needed since grid resources are composed of heterogeneous networks and OS systems in the grid environment. In this paper. we classify grid resources with data properties and user requirements for resource selection using a decision tree method. Our resource selection method can provide suitable resource selection methodology using classification with a decision tree to grid users. This paper evaluates our grid system performance with throughput. utilization, job loss, and average of turn-around time and shows experiment results of our resource selection model in comparison with those of existing resource selection models such as Condor-G and Nimrod-G. These experiment results showed that our resource selection model provides a vision of efficient grid resource selection methodology.

• Proceedings of the KIEE Conference
• /
• 2004.05a
• /
• pp.62-64
• /
• 2004

P2P Systems have grown dramatically in recent years. Now many P2P systems have developed and been confronted by P2P technical challenges. We should consider how to efficiently locate desired resources. In this paper we integrated the existing pure P2P and hybrid P2P model. We try to keep roles of super peer in hybrid and concurrently use pure P2P model for searching resource. In order to improve the existing search mechanism, we present contents-based classification mechanism. Proposed system have the following features. This can forward only query to best peer using RI. Second, it is self-organization. A peer can reconfigure network that it can communicate directly with based on best peer. Third, peers can cluster each other through contents-based classification.

• The KIPS Transactions:PartD
• /
• v.9D no.5
• /
• pp.939-944
• /
• 2002

In this paper, we propose and apply new classification method to the remotely sensed image acquired from airborne multi-spectral scanner. This is a neuro-fuzzy image classifier derived from the generic model of a 3-layer fuzzy perceptron. We implement a classification software system with the proposed method for land cover image classification. Comparisons with the proposed and maximum-likelihood classifiers are also presented. The results show that the neuro-fuzzy classification method classifies more accurately than the maximum likelihood method. In comparing the maximum-likelihood classification map with the neuro-fuzzy classification map, it is apparent that there is more different as amount as 7.96% in the overall accuracy. Most of the differences are in the "Building" and "Pine tree", for which the neuro-fuzzy classifier was considerably more accurate. However, the "Bare soil" is classified more correctly with the maximum-likelihood classifier rather than the neuro-fuzzy classifier.

• Journal of the Korean Data and Information Science Society
• /
• v.16 no.2
• /
• pp.313-325
• /
• 2005

In Korea, there exists a large corpus of handwritten historical documents that serve as a valuable resource. Most of them are hand-written by the King's chroniclers and secretaries. Recently, the historical archives of Lee dynasty have been digitalized. Since it is extremely difficult to utilize conventional OCR system, most of the processes have been performed manually. In this paper, we propose OCR-based digitalization system using Mahalanobis distance-based rejection and interface for eye inspection about historical Hanja documents. Compared with our previous work, experimental results show that the proposed system can help enhancing the overall efficiency of the process.

• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.2
• /
• pp.81-86
• /
• 2014

In this paper, we present an effective sound classification system for recognizing the real-time context of a smartphone user. Our system avoids unnecessary consumption of limited computational resource by filtering both silence and white noise out of input sound data in the pre-processing step. It also improves the classification performance on low energy-level sounds by amplifying them as pre-processing. Moreover, for efficient learning and application of HMM classification models, our system executes the dimension reduction and discretization on the feature vectors through k-means clustering. We collected a large amount of 8 different type sound data from daily life in a university research building and then conducted experiments using them. Through these experiments, our system showed high classification performance.

• Journal of Power System Engineering
• /
• v.19 no.4
• /
• pp.76-81
• /
• 2015

The production of electricity from wind energy is noticed as economic power generation system in the natural resource. Lately, since lots of wind turbine have installed globally, the accidents have increased gradually. In this paper, we classified domestic information for 10 years, information of new energy and industrial technology development organization(NEDO) for 4years and caithness windfarms information forum(CWIF) for 15 years according to part and cause of wind turbine accident. We found that the main causes of accidents are storm, lightening and carelessness. The results of classifying and analyzing the informations, should be used to take measures on the accident prevention of wind turbine.