• International Journal of Computer Science & Network Security
• /
• v.23 no.11
• /
• pp.21-31
• /
• 2023

In this paper, we focused on the problem of evaluating multi-class classification accuracy and simulation of multiple classifier performance metrics. Multi-class classifiers for sentiment analysis involved many challenges, whereas previous research narrowed to the binary classification model since it provides higher accuracy when dealing with text data. Thus, we take inspiration from the non-linear Support Vector Machine to modify the algorithm by embedding dynamic hyperplanes representing multiple class labels. Then we analyzed the performance of multi-class classifiers using macro-accuracy, micro-accuracy and several other metrics to justify the significance of our algorithm enhancement. Furthermore, we hybridized Enhanced Convolution Neural Network (ECNN) with Dynamic Support Vector Machine (DSVM) to demonstrate the effectiveness and efficiency of the classifier towards multi-class text data. We performed experiments on three hybrid classifiers, which are ECNN with Binary SVM (ECNN-BSVM), and ECNN with linear Multi-Class SVM (ECNN-MCSVM) and our proposed algorithm (ECNNDSVM). Comparative experiments of hybrid algorithms yielded 85.12 % for single metric accuracy; 86.95 % for multiple metrics on average. As for our modified algorithm of the ECNN-DSVM classifier, we reached 98.29 % micro-accuracy results with an f-score value of 98 % at most. For the future direction of this research, we are aiming for hyperplane optimization analysis.

• Tunnel and Underground Space
• /
• v.16 no.4 s.63
• /
• pp.334-346
• /
• 2006

The discrete element code in 2-D, PFC2D, has been used as a tool to simulate various phenomena in rock mechanics and rock engineering. However, the code has an disadvantage that procedure to determine micro-parameters, namely properties of particles and contacts is repetitive and time-consuming. In this study, we analyzed the effect of micro-parameters(for generation of a contact-bonded model) on macro-properties(that were measured numerically by uniaxial compressive test). Based on the analysis, also, the time-saving and reliable method was suggested to determine a complete set of micro-parameters. In order to verify the suggested method, numerical specimens were generated in PFC2D for 10 different rock types at home and abroad. By the two trials for each specimen, in the result, the Young's modulus, Poisson's ratio and uniaxial compressive strength could be reproduced with being in relative error by about 5% to the values obtained by laboratory tests.

• Journal of the Korean Society for Precision Engineering
• /
• v.15 no.5
• /
• pp.167-175
• /
• 1998

A mathematical formulation is presented to model anisotropy from the deformation textures developed in a forming process. In this work, a micro-mechanical-based polycrystalline analysis is implemented into a consistent finite element method for the anisotropic, viscoplastic deformation of polycrystalline metals. As suggested by Taylor, the deformation of each grain in an aggregate is assumed to be same as the macroscopic deformation of an aggregate or a macro-continuum point. Algorithms are developed to represent the plastic anisotropy, such as the anisotropic yield surface and R-value, from the predicted deformation texture. As applications, the evolution of texture in rolling, upsetting and drawing/extrusion processes are simulated and the corresponding changes of mechanical properties such as yield surface and R-value are predicted.

• Korean Journal of Agricultural Science
• /
• v.43 no.2
• /
• pp.163-175
• /
• 2016

For the past decades, advances in computational devices have propelled mathematical modeling to become an effective tool for solving the black box of complex biological systems because of its prominent analytical power and comprehensive insight. Nevertheless, modeling is still limitedly used in the fields of agriculture and food which generally concentrate on producing experimental data rather than processing them. This study, hence, intends to introduce modeling in terms of its procedure types of structure, formulation, analyses, and software, with reviews of current notable studies from micro to macro scales so as to propose the modeling technique as a novel approach in discerning conundrums in agriculture and food systems. We expect this review to provide an eligible source for researchers who are willing to apply modeling techniques into the unexplored fields related to bio-systems that comprehensively include biology, nutrition, agriculture, food, animal science, and ecology.

• Proceedings of the KSME Conference
• /
• 2001.06a
• /
• pp.507-510
• /
• 2001

Brittle failure mechanism has been well known as growth of initial micro-damage, that causes macro crack and failure in the end. Several precise criteria are suggested recently, based on experiments values in a whole load range. Among them, Mohr-Coulomb's criterion is used widely these days, but it has a big error compared with the real failure behavior since it does not show reciprocal actions of stresses. In this study, a new brittle failure criterion is proposed, which includes the effects of brittle damage evolution by taking a brittle damage parameter specifically. Comparisons between the proposed model and the previous ones are also given.

• Transactions of Materials Processing
• /
• v.21 no.5
• /
• pp.319-323
• /
• 2012

In order to predict the internal void closing behavior in open die forging process, multiple scale modeling has been developed and applied. The huge size difference between ingot and inner void makes it almost impossible to simultaneously model the actual loading conditions and the void shape. Multiple scale modeling is designed to integrate macro- and micro- models effectively and efficiently. The void closing behavior was simulated at 39 different locations in a large ingot during upsetting and cogging. The correlation between the closing behavior and variables such as effective plastic strain and maximum compressive strain was studied in order to find an efficient measure for predicting the soundness of the forging.

• Journal of Korea Foundry Society
• /
• v.6 no.2
• /
• pp.116-121
• /
• 1986

The effects of pressure during solidification on macro-and micro-structures have been studied in pure aluminium and Al-Si alloys. The application of pressure during solidifcation accelerated both equiaxed and columnar dendritic-growth due to stimulating of equiaxed survival and faster preferential growth of primary dendrites against the parallel direction of heat flow. Burden-Hunt model was modified to express the significant changes of CET behaviours under pressure. A further point to be noted was that greatly fine eutectic silicon flakes ($0.5\;{\times}\;13{\mu}m$) with the decrease of halo layers ($7{\mu}m$) of aluminium riched phases in the periphery of primary silicon particles were observed when pressure was applied during solidification.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2014.11a
• /
• pp.376-379
• /
• 2014

본 논문에서는 사람의 이동정보인 위치데이터를 바탕으로 위치분석(Location Analysis)을 통해 나타난 개개인의 이동성 모델을 바탕으로 각 개인의 이동성 모델에서 나타난 군집들의 관계를 분석해 개인이 속한 집단의 이동성 모델을 만든다. 집단 이동성 모델은 개인의 모델에서 나타난 군집을 이용하는데, 각 군집을 만드는데 필요한 위치 정보들과 군집의 중심, 군집간의 거리의 값을 계산하여 새로운 통합 군집을 만든다. 새로 만드는 군집은 각 특징에 따라 Micro Cluster, Macro Cluster의 2가지로 분류하였다. 실제 수년간 수집한 2명의 통합 개인 이동성 모델을 바탕으로 집단 이동성 모델을 생성한다. 집단 이동성 모델 생성에는 R Language를 사용하였고 결과 모델을 지도상에 표시할 수 있다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2012.04a
• /
• pp.864-867
• /
• 2012

사람들이 휴대용 위치정보 수집 장비나 혹은 스마트폰을 사용하면서 사람의 이동 정보인 위치정보들을 모으는 일이 가능해 졌다. 이러한 위치정보들을 가지고 본 논문에서는 사람의 이동 모델을 나타내고자 하였다. 이동 정보들은 머물러 있는(Stay)상태와 이동하는(Moving) 상태로 나눌 수 있는데 이러한 상태 중 머물러 있는 상태가 군집화가 되어 연쇄 모델속의 하나의 상태(State)로 나타나 질 수 있다. 물론 이동 정보들을 통해 연쇄모델 속 각 상태간의 전이 확률 또한 계산 할 수 있다. 이러한 일련의 과정을 본 논문에서는 기대치 최대화 기반 군집화 과정을 통해 연속시간 연쇄 모델의 형태로 인간의 이동성을 표현하였다. 또한 이러한 모델에서 대표 군집(macro)과 그 부속 군집(micro)을 표현할 수 있었고 이러한 모습은 대표적인 큰 군집 속의 작은 군집의 형태로 나타나게 된다.

• Coupled systems mechanics
• /
• v.11 no.5
• /
• pp.411-438
• /
• 2022

In this paper, we present the parameter identification for inelastic and multi-scale problems. First, the theoretical background of several fundamental methods used in the upscaling process is reviewed. Several key definitions including random field, Bayesian theorem, Polynomial chaos expansion (PCE), and Gauss-Markov-Kalman filter are briefly summarized. An illustrative example is given to assimilate fracture energy in a simple inelastic problem with linear hardening and softening phases. Second, the parameter identification using the Gauss-Markov-Kalman filter is employed for a multi-scale problem to identify bulk and shear moduli and other material properties in a macro-scale with the data from a micro-scale as quantities of interest (QoI). The problem can also be viewed as upscaling homogenization.