• The KIPS Transactions:PartB
• /
• v.10B no.3
• /
• pp.273-280
• /
• 2003

Recommender System can help users to find products to Purchase. A representative method for recommender systems is collaborative filtering (CF). It predict products that user may like based on a group of similar users. User information is based on user's ratings for products and similarities of users are measured by ratings. As user is increasing tremendously, the performance of the pure collaborative filtering is lowed because of high dimensionality and scarcity of data. We consider the effect of dimension deduction in collaborative filtering to cope with scarcity of data experimentally. We suggest that SVD improves the performance of collaborative filtering in comparison with pure collaborative filtering.

• Journal of Satellite, Information and Communications
• /
• v.6 no.2
• /
• pp.117-124
• /
• 2011

In this paper, we present more effective throughput enhancement technique to improve the data rate and reliability by using the multiple antenna technique. The conventional spatial diversity scheme is limited in according with the interference from each antenna channel status, and the orthogonality of spreading codes and subcarriers are destroyed due to the frequency selectivity. Proposed system is considered MC-CDMA system with 4 transmit antennas and 1 receive antenna. Proposed system based on SVD with the MS-RLS MMSE subcarrier combining method in order to achieve better performance with low computational complexity. Via computer simulation, we confirm that the proposed system is able to improve the BER performance by suppressing the interference of other antenna signals.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.29 no.5
• /
• pp.71-79
• /
• 2015

Recently, new Low Voltage DC (LVDC) power distribution systems have been constantly researched as uses of DC in end-user equipment are increased. As in conventional AC distribution system, High Impedance Fault (HIF) which may cause a failure of protective relay can occur in LVDC distribution system as well. It, however, is hard to be detected since change in magnitude of current due to the fault is too small to detect the fault by the protective relay using overcurrent element. In order to solve the problem, this paper presents an algorithm for detecting HIF using accumulated energy in LVDC distribution system. Wavelet Singular Value Decomposition (WSVD) is used to extract abnormal high frequency components from fault current and accumulated energy of high frequency components is considered as the element to detect the fault. LVDC distribution system including AC/DC and DC/DC converter is modeled to verify the proposed algorithm using ElectroMagnetic Transient Program (EMTP) software. Simulation results considering various conditions show that the proposed algorithm can be utilized to effectively detect HIF.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.8A
• /
• pp.829-835
• /
• 2002

When finite length image signal is decomposed into M-band synthesized using M-band orthogonal wavelet transform, the boundary signal of image are not reconstructed perfectly. for boundary signals to be reconstructed perfectly, different type filter bank or technique is applied to them when the dimension of analysed is proposed. It can be designed using the singular value decomposition of boundary perfect reconstruction matrix which is obtained from paraunitary matrix which is obtained from paraunitary matrix. And it is also discussed to design the boundary filter bank for improving the coding performance when it is applied to subband coder. The proposed boundary filter bank shows 7% gains in PSNR compared with reflected method.

• The Korean Journal of Applied Statistics
• /
• v.28 no.1
• /
• pp.49-60
• /
• 2015

Various kinds of ordination methods such as correspondence analysis and canonical correspondence analysis are used in community ecology to visualize relationships among species, sites, and environmental variables. Ter Braak (1986), Jackson and Somers (1991), Parmer (1993), compared the ordination methods using eigenvalue and distance graph. However, these methods did not show the relationship between population and biplot because they are only based on surveyed data. In this paper, a method that measures the extent to show population information to biplot was introduced to compare ordination methods objectively.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.23 no.7
• /
• pp.617-623
• /
• 2013

Accurate prediction of the radiated noise is important to reduce the noise of the washing machine. It is also necessary to predict the excitation force accurately because excitation force can induce noise. In order to predict the excitation force acting on the washing machine, this paper conducts source identification method by use of phase reference spectrum. In this method, the transfer function between the cabinet and the motor through FEM and the measured response from the surface of the cabinet is used. The analysis of the radiation noise from the identified exciting force has been investigated. The comparison between the predicted SPL and the measured SPL at 1m apart from the front side of the washing machine showed good tendency.

• Journal of the korean society of oceanography
• /
• v.31 no.2
• /
• pp.64-74
• /
• 1996

Numerical experiments of OAT (Ocean Acoustic Tomography) are carried out in the East Sea of Korea where the canonical ocean has been perturbed by a mesoscale warm eddy and a thermal front. In order to estimate the horizontal and vertical structure of water temperature of the perturbed ocean, the experimental area is divided into 16 cells with 8 pairs of sources and receivers for a horizontal slice and the water column is divided into 8 layers for a vertical slice. The inversely estimated temperature field by using SVD (Singular Value Decomposition) method reveals the eddy and frontal structure clearly. The rms errors of the two horizontal slices are less than $0.4^{\circ}C$ and $1.7^{\circ}C$ at 400 m and 200 m depths, respectively, while the error in the vertical slice is less than $1.0^{\circ}C.$ For better estimation of temperature by OAT method, particularly for the East Sea, a range-dependent ray model should be used to solve the forward problem. At the same time, improvement in computing the refracted ray path between vertical layers is required to obtain more accurate travel time information. The results of the present experiment give rise to a possibility of application of OAT in remote sensing of the ocean thermal structure.

• Journal of the Optical Society of Korea
• /
• v.17 no.6
• /
• pp.525-532
• /
• 2013

Contrary to the academic interests of other existing studies elsewhere, this study deals with how the alignment algorithms such as sensitivity or Differential Wavefront Sampling (DWS) can be better used under effects from field, compensator positioning and environmental errors unavoidable from the shop-floor alignment work. First, the influences of aforementioned errors to the alignment state estimation was investigated with the algorithms. The environmental error was then found to be the dominant factor influencing the alignment state prediction accuracy. Having understood such relationship between the distorted system wavefront caused by the error sources and the alignment state prediction, we used it for simulated and experimental alignment runs for Infrared Optical System (IROS). The difference between trial alignment runs and experiment was quite close, independent of alignment methods; 6 nm rms for sensitivity method and 13 nm rms for DWS. This demonstrates the practical usefulness and importance of the prior error analysis using the alignment algorithms before the actual alignment runs begin. The error analysis methodology, its application to the actual alignment of IROS and their results are described together with their implications.

• Journal of Information Processing Systems
• /
• v.17 no.4
• /
• pp.707-720
• /
• 2021

To improve the accuracy of the recommendation system, multi-criteria recommendation systems have been widely researched. However, it is highly complicated to extract the preferred features of users and items from the data. To this end, subjective indicators, which indicate a user's priorities for personalized recommendations, should be derived. In this study, we propose a method for generating recommendation candidates by predicting multi-criteria ratings from reviews and using them to derive user priorities. Using a deep learning model based on convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM), multi-criteria prediction ratings were derived from reviews. These ratings were then aggregated to form a linear regression model to predict the overall rating. This model not only predicts the overall rating but also uses the training weights from the layers of the model as the user's priority. Based on this, a new score matrix for recommendation is derived by calculating the similarity between the user and the item according to the criteria, and an item suitable for the user is proposed. The experiment was conducted by collecting the actual "TripAdvisor" dataset. For performance evaluation, the proposed method was compared with a general recommendation system based on singular value decomposition. The results of the experiments demonstrate the high performance of the proposed method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.7
• /
• pp.3620-3637
• /
• 2019

Since the birth of Synthetic Aperture Radar (SAR), it has been widely used in the military field and so on. However, the existence of speckle noise makes a good deal inconvenience for the subsequent image processing. The continuous development of sparse representation (SR) opens a new field for the speckle suppressing of SAR image. Although the SR de-noising may be effective, the over-smooth phenomenon still has bad influence on the integrity of the image information. In this paper, one novel SAR image de-noising method based on residual image fusion and sparse representation is proposed. Firstly we can get the similar block groups by the non-local similar block matching method (NLS-BM). Then SR de-noising based on the adaptive K-means singular value decomposition (K-SVD) is adopted to obtain the initial de-noised image and residual image. The residual image is processed by Shearlet transform (ST), and the corresponding de-noising methods are applied on it. Finally, in ST domain the low-frequency and high-frequency components of the initial de-noised and residual image are fused respectively by relevant fusion rules. The final de-noised image can be recovered by inverse ST. Experimental results show the proposed method can not only suppress the speckle effectively, but also save more details and other useful information of the original SAR image, which could provide more authentic and credible records for the follow-up image processing.