• Proceedings of the Korean Information Science Society Conference
• /
• 2006.06b
• /
• pp.34-36
• /
• 2006

In this paper we present a method of feature extraction for motor imagery single trial EEG classification, where we exploit nonnegative matrix factorization (NMF) to select discriminative features in the time-frequency representation of EEG. Experimental results with motor Imagery EEG data in BCI competition 2003. show that the method indeed finds meaningful EEG features automatically, while some existing methods should undergo cross-validation to find them.

• The Korean Journal of Applied Statistics
• /
• v.25 no.4
• /
• pp.605-612
• /
• 2012

Multivariate data are often generated with multiple outcomes in various fields. Multiple outcomes could be mixed as continuous and discrete. Because of their complexity, the data are often dealt with by separately applying regression analysis to each outcome even though they are associated the each other. This univariate approach results in the low efficiency of estimates for parameters. We study the efficiency gains of the multivariate approaches relative to the univariate approach with the mixed data that include continuous and binary outcomes. All approaches yield consistent estimates for parameters with complete data. By jointly estimating parameters using multivariate methods, it is generally possible to obtain more accurate estimates for parameters than by a univariate approach. The association between continuous and binary outcomes creates a gap in efficiency between multivariate and univariate approaches. We provide a guidance to analyze the mixed data.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.10
• /
• pp.3482-3497
• /
• 2021

Artificial intelligence has emerged as the core of the 4th industrial revolution, and large amounts of data processing, such as big data technology and rapid data analysis, are inevitable. The most fundamental and universal data interpretation technique is an analysis of information through regression, which is also the basis of machine learning. Ridge regression is a technique of regression that decreases sensitivity to unique or outlier information. The time-consuming calculation portion of the matrix computation, however, basically includes the introduction of an inverse matrix. As the size of the matrix expands, the matrix solution method becomes a major challenge. In this paper, a new algorithm is introduced to enhance the speed of ridge regression estimator calculation through series expansion and computation recycle without adopting an inverse matrix in the calculation process or other factorization methods. In addition, the performances of the proposed algorithm and the existing algorithm were compared according to the matrix size. Overall, excellent speed-up of the proposed algorithm with good accuracy was demonstrated.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.9
• /
• pp.53-62
• /
• 2020

Product recommendation services that have been researched recently are only recommended through the customer's product purchase history. In this paper, we propose the digital signage service through customers' behavior pattern analysis that is recommending through not only purchase history, but also behavior pattern that customers take when choosing products. This service analyzes customer behavior patterns and extracts interests about products that are of practical interest. The service is learning extracted interest rate and customers' purchase history through the Wide & Deep model. Based on this learning method, the sparse vector of other products is predicted through the MF(Matrix Factorization). After derive the ranking of predicted product interest rate, this service uses the indoor signage that can interact with customers to expose the suitable advertisements. Through this proposed service, not only online, but also in an offline environment, it would be possible to grasp customers' interest information. Also, it will create a satisfactory purchasing environment by providing suitable advertisements to customers, not advertisements that advertisers randomly expose.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.41 no.8
• /
• pp.851-858
• /
• 2016

This paper presents a new statistical voice activity detection (VAD) based on the probabilistic interpretation of nonnegative matrix factorization (NMF). The objective function of the NMF using Kullback-Leibler divergence coincides with the negative log likelihood function of the data if the distribution of the data given the basis and encoding matrices is modeled as Poisson distributions. Based on this probabilistic NMF, the VAD is constructed using the likelihood ratio test assuming that speech and noise follow Poisson distributions. Experimental results show that the proposed approach outperformed the conventional Gaussian model-based and NMF-based methods at 0-15 dB signal-to-noise ratio simulation conditions.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1998.10a
• /
• pp.63-66
• /
• 1998

In the implementation of the simplex method program, the representation and the maintenance of basis matrix is very important, In the experimental study, we investigates Suhl's idea in the LU factorization and LU update of basis matrix. First, the triangularization of basis matrix is implemented and its efficiency is shown. Second, various technique in the dynamic Markowitz's ordering and threshold pivoting are presented. Third, modified Forrest-Tomlin LU update method exploiting sparsity is presented. Fourth, as a storage scheme of LU factors, Gustavson data structure is explained. Fifth, efficient timing of reinversion is developed. Finally, we show that modified Forrest-Tomlin method with Gustavson data structure is superior more than 30% to the Reid method with linked list data structure.

• East Asian mathematical journal
• /
• v.40 no.3
• /
• pp.307-318
• /
• 2024

This paper explores the application of matrix factorization, specifically CUR decomposition, in the clustering of Korean language documents by topic. It addresses the unique challenges of Natural Language Processing (NLP) in dealing with the Korean language's distinctive features, such as agglutinative words and morphological ambiguity. The study compares the effectiveness of Latent Semantic Analysis (LSA) using CUR decomposition with the classical Singular Value Decomposition (SVD) method in the context of Korean text. Experiments are conducted using Korean Wikipedia documents and newspaper data, providing insight into the accuracy and efficiency of these techniques. The findings demonstrate the potential of CUR decomposition to improve the accuracy of document clustering in Korean, offering a valuable approach to text mining and information retrieval in agglutinative languages.

• Journal of Navigation and Port Research
• /
• v.46 no.1
• /
• pp.42-56
• /
• 2022

Due to the 4th Industrial Revolution announced at the Davos Forum, the World Economic Forum, in 2016, industrial trends around the world are changing rapidly and intelligently. Among them, the digital twin is drawing attention from all industries as a groundbreaking technology that reduces unnecessary costs and trial and error by implementing real objects, systems, and environments in the same way in the real virtual world and using them to perform simulation analysis. In particular, there is a lot of interest in the application of digital twin technology in solving ports safety and efficiency challenges at once. However, there is a lack of in-depth research for the application of digital twin technology in the port, and in particular, there is a lack of research on measurable data for the implementation of the digital twin in ports. The purpose of this study was to increase granularity and connectivity through measurable data investigation for the application of digital twin technology at container ports. Based on the study results, data factors for container port application were classified into crane data, operational data, physical data, and transportation data, and factor composition, correlation with factors, and fitness were confirmed through confirmatory factor analysis.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2007.11a
• /
• pp.217-220
• /
• 2007

In this paper, we present a new anchor shot detection system which is a core step of the preprocessing process for the news video analysis. The proposed system is composed of four modules and operates sequentially: 1) skin color detection module for reducing the candidate face regions; 2) face detection module for finding the key-frames with a facial data; 3) vector representation module for the key-frame images using a non-negative matrix factorization; 4) anchor shot detection module using a support vector data description. According to our computer experiments, the proposed system shows not only the comparable accuracy to the recent other results, but also more faster detection rate than others.

• Proceedings of the KOSOMBE Conference
• /
• v.1997 no.11
• /
• pp.197-200
• /
• 1997

In this paper, we present a 3D reconstruction method of color volume data or a computerized human atlas. Binary volume rendering which takes the advantages of shear-warp factorization and new normal vector calculation method visualizes 3D organs in real time. Various manipulations such as rotation, multiple object rendering, removal, and transparency effect improve the usefulness and comprehensiveness of the computerized atlas.