• Journal of the Korea Society of Computer and Information
• /
• v.9 no.4 s.32
• /
• pp.33-39
• /
• 2004

In this paper. we design the learning algorithm of LVQ which is used Forward Counter Propagation Networks to improve classification performance of LVQ networks. The weights of Forward Counter Propagation Networks which is between input layer and cluster layer can be learned to determine initial reference vectors by using SOM algorithm and to learn reference vectors by using LVQ algorithm. Finally. pattern vectors is classified into subclasses by neurons which is being in the cluster layer, and the weights of Forward Counter Propagation Networks which is between cluster layer and output layer is learned to classify the classified subclass, which is enclosed a class. Also. kr the number of classes is determined, the number of neurons which is being in the input layer, cluster layer and output layer can be determined. To prove the performance of the proposed learning algorithm. the simulation is performed by using training vectors and test vectors that ate Fisher's Iris data, and classification performance of the proposed learning method is compared with ones of the conventional LVQ, and it was a confirmation that the proposed learning method is more successful classification than the conventional classification.

• Korean Journal of Mathematics
• /
• v.26 no.2
• /
• pp.337-348
• /
• 2018

MNIST dataset is a database containing images of handwritten digits, with each image labeled by an integer from 0 to 9. It is used to benchmark the performance of machine learning algorithms. Neural networks for MNIST are regarded as the starting point of the studying machine learning algorithms. However it is not easy to start the actual programming. In this expository article, we will give a step-by-step instruction to build neural networks for MNIST dataset using MATLAB.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.7
• /
• pp.399-412
• /
• 2006

A clustering based wireless internet proxy server with cooperative caching has a problem of minimizing overall performance because some servers become overloaded if client request pattern is Hot-Spot or uneven. We propose a self-learning based adaptive clustering scheme to solve the poor performance problems of the existing clustering in case of Hot-Spot or uneven client request pattern. In the proposed scheme, requests are dynamically redistributed to the other servers if some servers supposed to handle the requests become overloaded. This is done by a self-learning based method based dynamic weight adjustment algorithm so that it can be applied to a situation with even various request pattern or a cluster of hosts with different performance. We performed experiments in a clustering environment with 16 PCs and a load balancer. Experimental results show the 54.62% performance improvement of the proposed schemes compared to the existing schemes.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.4
• /
• pp.537-542
• /
• 2021

Smart TVs provide a variety of services and information compared to existing TVs based on the Internet. In order to provide more personalized services or information, it is necessary to analyze users' viewing patterns and provide customized services or information based on them. The proposed system receives the user's TV viewing pattern, analyzes it, and recommends a TV program or movie as customized information to the user. For this, the system was constructed with a preprocessor and a deep learning model. The preprocessor refines the name of the TV program watched by the user, the date the TV program was watched, and the watched time. Then, the multi-attribute LSTM model trains the refined data and performs prediction.The proposed system is a system that provides customized information to users, and is believed to be a leading technology in digital convergence that combines existing IoT technology and deep learning technology.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.7
• /
• pp.83-92
• /
• 2022

In this paper, we designed an algorithm that can control the state of PM by learning the chain failure pattern of PHM based air conditioning facility. It is an inevitable spread of PM due to the downtime caused by the failure of the air conditioning facility. The algorithm developed by us is to establish a PM management system through PHM, and it is an algorithm that maintains a constant stabilization state through learning the stop/operation pattern of the air conditioner and manages PM based on this. As a result of the simulating at a subway station for the performance qualification of the algorithm, it was verified that the concentration of PM reduces by 30% on average. In the case of stations with many passengers using the subway, the concentration of PM exceeded the Ministry of Environment Standards(100 ㎍/m³), but it was verified that the concentration of PM was improved at all stations where the simulation was conducted. In the future research is to expand the system to comprehensively manage not only PM but also pollutants such as CO₂, CO, and NO₂ in subway stations.

• Journal of KIISE:Software and Applications
• /
• v.34 no.1
• /
• pp.11-17
• /
• 2007

One of the popular methods used for pattern classification is the MBR (Memory-Based Reasoning) algorithm. Since it simply computes distances between a test pattern and training patterns or hyperplanes stored in memory, and then assigns the class of the nearest training pattern, it cannot explain how the classification result is obtained. In order to overcome this problem, we propose an incremental teaming algorithm based on RPA (Recursive Partition Averaging) to extract IF-THEN rules that describe regularities inherent in training patterns. But rules generated by RPA eventually show an overfitting phenomenon, because they depend too strongly on the details of given training patterns. Also RPA produces more number of rules than necessary, due to over-partitioning of the pattern space. Consequently, we present the IREA (Incremental Rule Extraction Algorithm) that overcomes overfitting problem by removing useless conditions from rules and reduces the number of rules at the same time. We verify the performance of proposed algorithm using benchmark data sets from UCI Machine Learning Repository.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.22 no.3
• /
• pp.319-327
• /
• 2012

In this paper, we present a sign language recognition model which does not use any wearable devices for object tracking. The system design issues and implementation issues such as data representation, feature extraction and pattern classification methods are discussed. The proposed data representation method for sign language patterns is robust for spatio-temporal variances of feature points. We present a feature extraction technique which can improve the computation speed by reducing the amount of feature data. A neural network model which is capable of incremental learning is described and the behaviors and learning algorithm of the model are introduced. We have defined a measure which reflects the relevance between the feature values and the pattern classes. The measure makes it possible to select more effective features without any degradation of performance. Through the experiments using six types of sign language patterns, the proposed model is evaluated empirically.

• Journal of the Korean Institute of Educational Facilities
• /
• v.27 no.3
• /
• pp.3-14
• /
• 2020

The purpose of this study is to discuss the planning direction of educational spaces to support children's healthy and creative learning based on bio_philic theory. This study analyzed the characteristics of the application of biophilic patterns in children's education space through case analysis. The conclusion of this study is summarized as follows. As a result of the analysis of children's classroom space, the pattern of 'A(Visual connection with nature), F(Dynamic & Diffuse Light), K(Prospect)' shows high application rate, but the pattern of 'C(Non-Rhythmic Sensory Stimuli), G(Connection with Natural Systems), I(Material Connection with Nature)' shows low application rate. In particular, there is a lack of connection with patterns such as hearing, smell, touch, taste stimulation and water experience, and curiosity through exploration of nature about 'B(Non-visual connection with nature), E(Presence of Water), N(Risk/Peril)' changes in nature and ecosystem. In the corridor and rest space, the pattern of 'A(Visual connection with nature), D(Thermal & Airflow Variability), F(Dynamic & Diffuse Light), G(Connection with Natural Systems), K(Prospect)' shows high application rate, but 'B(Non-visual connection with nature)' shows low application rate. In addition, the application of patterns related to the stimulation of curiosity through direct exploration of nature and the exploration of the patterns of 'E(Presence of Water), N(Risk/Peril)' is insufficient. Therefore, in the case of classroom spaces, the active use of nature as it is should be considered within the scope that does not cause visual confusion, and it should provide an area that can be experienced through the five senses. And corridors and rest spaces should be designed to introduce more active natural elements as spaces to recover stress caused by learning. In other words, the characteristics of children's education facilities need to be connected between classroom space, corridor, rest space and external space. This study is meaningful in that it analyzes and derives the application characteristics of 'biophilic design' which affects the 'Attention Restoration' of children's educational spaces through foreign cases.

• KIPS Transactions on Computer and Communication Systems
• /
• v.4 no.6
• /
• pp.185-196
• /
• 2015

Energy aware server clusters aim to reduce power consumption at maximum while keeping QoS(Quality of Service) compared to energy non-aware server clusters. They adjust the power mode of each server in a fixed or variable time interval to let only the minimum number of servers needed to handle current user requests ON. Previous studies on energy aware server cluster put efforts to reduce power consumption further or to keep QoS, but they do not consider energy efficiency well. In this paper, we propose an energy efficient cluster management based on autonomous learning for energy aware server clusters. Using parameters optimized through autonomous learning, our method adjusts server power mode to achieve maximum performance with respect to power consumption. Our method repeats the following procedure for adjusting the power modes of servers. Firstly, according to the current load and traffic pattern, it classifies current workload pattern type in a predetermined way. Secondly, it searches learning table to check whether learning has been performed for the classified workload pattern type in the past. If yes, it uses the already-stored parameters. Otherwise, it performs learning for the classified workload pattern type to find the best parameters in terms of energy efficiency and stores the optimized parameters. Thirdly, it adjusts server power mode with the parameters. We implemented the proposed method and performed experiments with a cluster of 16 servers using three different kinds of load patterns. Experimental results show that the proposed method is better than the existing methods in terms of energy efficiency: the numbers of good response per unit power consumed in the proposed method are 99.8%, 107.5% and 141.8% of those in the existing static method, 102.0%, 107.0% and 106.8% of those in the existing prediction method for banking load pattern, real load pattern, and virtual load pattern, respectively.

• Journal of Information Processing Systems
• /
• v.13 no.4
• /
• pp.643-652
• /
• 2017