• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.39 no.2
• /
• pp.10-20
• /
• 2002

Various blind source separation (BSS) and independent component analysis (ICA) algorithms have been developed. However, comparison study for BSS/ICA algorithms has not been extensively carried out yet. The main objective of this paper is to compare various promising BSS/ICA algorithms in terms of several factors such as robustness to sensor noise, computational complexity, the conditioning of the mixing matrix, the number of sensors, and the number of training patterns. We propose several benchmarks which are useful for the evaluation of the algorithm. This comparison study will be useful for real-world applications, especially EEG/MEG analysis and separation of miked speech signals.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.43 no.1 s.307
• /
• pp.9-16
• /
• 2006

RDA(Relational Discriminant Analysis) is a way of finding classifiers based on the dissimilarity measures among the prototypes extracted from feature vectors instead of the feature vectors themselves. Therefore, the accuracy of the RDA classifier is dependent on the methods of selecting prototypes and measuring proximities. In this paper we propose to utilize PRS(Prototype Reduction Schemes) and Mahalanobis distances to devise a method of increasing classification accuracies. Our experimental results demonstrate that the proposed mechanism increases the classification accuracy compared with the conventional approaches for samples involving real-life data sets as well as artificial data sets.

• Korean Journal of Optics and Photonics
• /
• v.8 no.5
• /
• pp.403-409
• /
• 1997

We proposed a system that can perform pattern recognition based on parrallel optical feature extraction and performed experiments on this. The feature to be extracted are both 6 simple line orientations and two eigenvectors of the covariance matrix of the patterns that cannot be distinguished with the line orientation features alone. Our system consists of a feature-extraction part and a pattern-recognition part. The former that extracts the features in parallel with the multiplexed Vander Lugt filters was implemented optically, while the latter that performs the pattern recognition by the use of the extracted features was implemented in a computer. In the pattern recognition part, two methods are tested;one is to use an artificial neural network, which is trained to recognize the features directly, the other is to count the numbers of specific features simply and then to compare them with the stored reference feature numbers. We report the preliminary experimental results tested for 15 alpabet patterns with only straight line segments.

• Journal of KIISE
• /
• v.45 no.1
• /
• pp.69-75
• /
• 2018

In the streaming data where data samples arrive sequentially in time, it is difficult to apply the dimension reduction method based on batch learning. Therefore an incremental dimension reduction method for the application to streaming data has been studied. In this paper, we propose an incremental linear discriminant analysis method using the least squared error solution. Instead of computing scatter matrices directly, the proposed method incrementally updates the projective direction for dimension reduction by using the information of a new incoming sample. The experimental results demonstrate that the proposed method is more efficient compared with previously proposed incremental dimension reduction methods.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.14 no.1
• /
• pp.378-384
• /
• 2013

In this paper, we propose the fuzzy neural networks based on fuzzy c-means clustering algorithm. Typically, the generation of fuzzy rules have the problem that the number of fuzzy rules exponentially increases when the dimension increases. To solve this problem, the fuzzy rules of the proposed networks are generated by partitioning the input space in the scatter form using FCM clustering algorithm. The premise parameters of the fuzzy rules are determined by membership matrix by means of FCM clustering algorithm. The consequence part of the rules is expressed in the form of polynomial functions and the learning of fuzzy neural networks is realized by adjusting connections of the neurons, and it follows a back-propagation algorithm. The proposed networks are evaluated through the application to nonlinear process.

• Proceedings of the KIEE Conference
• /
• 2001.07d
• /
• pp.2789-2791
• /
• 2001

데이터 마이닝을 통해 우리는 숨겨진 지식, 예상되지 않았던 경향 그리고 새로운 법칙들을 방대한 데이터에서 이끌어내고자 한다. 본 논문에서 우리는 사용자의 구매 패턴을 발견하여 사용자가 원하는 상품을 미리 예측하여 추천하는 알고리즘을 소개하고자 한다. 제안하고 있는 item dependency map은 구매된 상품간의 관계를 수식화 하여 행렬의 형태로 표현한 것이다. Item dependency map의 값은 사용자가 A라는 상품을 구매한 후 B상품을 살 확률이다. 이런 정보를 가지고 있는 item dependency map은 홉필드 네트윅(Hopfield network)에서 연상을 위한 패턴 값으로 적용된다. 홉필드 네트웍은 각 노드사이의 연결가중치에 기억하고자 하는 것들을 연상시킨 뒤 어떤 입력을 통해서 전체 네트워크가 어떤 평형상태에 도달하는 방식으로 작동되는 신경망 중의 하나이다. 홉필드 네트웍의 특징 중의 하나는 부분 정보로부터 전체 정보를 추출할 수 있는 것이다. 이러한 특징을 가지고 사용자들의 일반적인 구매패턴을 일부 정보만 가지고 예측할 수 있다. Item dependency map은 홉필드 네트윅에서 사용자들의 그룹별 패턴을 학습하는데 사용된다. 따라서 item dependency map이 얼마나 사용자 구매패턴에 대한 정보를 가지고 있는지에 따라 그 결과가 결정되는 것이다. 본 논문은 정확한 item dependency map을 계산해 내는 알고리즘을 주로 논의하겠다.

• Communications of Mathematical Education
• /
• v.29 no.3
• /
• pp.465-489
• /
• 2015

The purpose of this study is to suggest implications for the development and revision of future teaching materials for mathematically gifted students by using network text analysis of secondary mathematics materials. Subjects of the analysis were learning goals of 110 teaching materials in a gifted education center in science attached to a university from 2002 to 2014. In analysing the frequency of the texts that appeared in the learning goals, key words were selected. A co-occurrence matrix of the key words was established, and a basic information of network, centrality, centralization, component, and k-core were deducted. For the analysis, KrKwic, KrTitle, and NetMiner4.0 programs were used, respectively. The results of this study were as follows. First, there was a pivot of the network formed with core hubs including 'diversity', 'understanding' 'concept' 'method', 'application', 'connection' 'problem solving', 'basic', 'real life', and 'thinking ability' in the whole network from 2002 to 2014. In addition, knowledge aspects were well reflected in teaching materials based on the centralization analysis. Second, network text analysis based on the three periods of the Mater Plan for the promotion of gifted education was conducted. As a result, a network was built up with 'understanding', and there were strong ties among 'question', 'answer', and 'problem solving' regardless of the periods. On the contrary, the centrality analysis showed that 'communication', 'discovery', and 'proof' only appeared in the first, second, and third period of Master Plan, respectively. Therefore, the results of this study suggest that affective aspects and activities with high cognitive process should be accompanied, and learning goals' mannerism and ahistoricism be prevented in developing and revising teaching materials.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.48 no.10
• /
• pp.773-782
• /
• 2020

The development of joint FE models for deep learning neural network (DLNN)-based hybrid FEA is presented. Material models of bolts and bearings in the front axle of tractor, showing complex behavior induced by various tightening conditions, were replaced with DLNN models. Bolts are modeled as one-dimensional Timoshenko beam elements with six degrees of freedom, and bearings as three-dimensional solid elements. Stress-strain data were extracted from all elements after finite element analysis subjected to various load conditions, and DLNN for bolts and bearing were trained with Tensorflow. The DLNN-based joint models were implemented in the ABAQUS user subroutines where stresses from the next increment are updated and the algorithmic tangent stiffness matrix is calculated. Generalization of the trained DLNN in the FE model was verified by subjecting it to a new loading condition. Finally, the DLNN-based FEA for the front axle of the tractor was conducted and the feasibility was verified by comparing with results of a static structural experiment of the actual tractor.

• Journal of KIISE:Computing Practices and Letters
• /
• v.11 no.6
• /
• pp.503-513
• /
• 2005

With the recognition of the importance of computational approach for protein-protein interaction prediction, many techniques have been developed to computationally predict protein-protein interactions. However, few techniques are actually implemented and announced in service form for general users to readily access and use the techniques. In this paper, we design and implement a protein interaction prediction service system based on the domain combination based protein-protein interaction prediction technique, which is known to show superior accuracy to other conventional computational protein-protein interaction prediction methods. In the prediction accuracy test of the method, high sensitivity($77\%$) and specificity($95\%$) are achieved for test protein pairs containing common domains with teaming sets of proteins in a Yeast. The stability of the method is also manifested through the testing over DIP CORE, HMS-PCI, and TAP data. Performance, openness and flexibility are the major design goals and they are achieved by adopting parallel execution techniques, web Services standards, and layered architecture respectively. In this paper, several representative user interfaces of the system are also introduced with comprehensive usage guides.

• Tunnel and Underground Space
• /
• v.33 no.6
• /
• pp.508-518
• /
• 2023

This paper presents a new approach for the automatic mapping of discontinuities in a tunnel face based on its 3D digital model reconstructed by LiDAR scan or photogrammetry techniques. The main idea revolves around the identification of discontinuity areas in the 3D digital model of a tunnel face by segmenting its 2D projected images using a deep-learning semantic segmentation model called U-Net. The proposed deep learning model integrates various features including the projected RGB image, depth map image, and local surface properties-based images i.e., normal vector and curvature images to effectively segment areas of discontinuity in the images. Subsequently, the segmentation results are projected back onto the 3D model using depth maps and projection matrices to obtain an accurate representation of the location and extent of discontinuities within the 3D space. The performance of the segmentation model is evaluated by comparing the segmented results with their corresponding ground truths, which demonstrates the high accuracy of segmentation results with the intersection-over-union metric of approximately 0.8. Despite still being limited in training data, this method exhibits promising potential to address the limitations of conventional approaches, which only rely on normal vectors and unsupervised machine learning algorithms for grouping points in the 3D model into distinct sets of discontinuities.