• Journal of Intelligence and Information Systems
• /
• v.26 no.1
• /
• pp.23-45
• /
• 2020

Big data is creating in a wide variety of fields such as medical care, manufacturing, logistics, sales site, SNS, and the dataset characteristics are also diverse. In order to secure the competitiveness of companies, it is necessary to improve decision-making capacity using a classification algorithm. However, most of them do not have sufficient knowledge on what kind of classification algorithm is appropriate for a specific problem area. In other words, determining which classification algorithm is appropriate depending on the characteristics of the dataset was has been a task that required expertise and effort. This is because the relationship between the characteristics of datasets (called meta-features) and the performance of classification algorithms has not been fully understood. Moreover, there has been little research on meta-features reflecting the characteristics of multi-class. Therefore, the purpose of this study is to empirically analyze whether meta-features of multi-class datasets have a significant effect on the performance of classification algorithms. In this study, meta-features of multi-class datasets were identified into two factors, (the data structure and the data complexity,) and seven representative meta-features were selected. Among those, we included the Herfindahl-Hirschman Index (HHI), originally a market concentration measurement index, in the meta-features to replace IR(Imbalanced Ratio). Also, we developed a new index called Reverse ReLU Silhouette Score into the meta-feature set. Among the UCI Machine Learning Repository data, six representative datasets (Balance Scale, PageBlocks, Car Evaluation, User Knowledge-Modeling, Wine Quality(red), Contraceptive Method Choice) were selected. The class of each dataset was classified by using the classification algorithms (KNN, Logistic Regression, Nave Bayes, Random Forest, and SVM) selected in the study. For each dataset, we applied 10-fold cross validation method. 10% to 100% oversampling method is applied for each fold and meta-features of the dataset is measured. The meta-features selected are HHI, Number of Classes, Number of Features, Entropy, Reverse ReLU Silhouette Score, Nonlinearity of Linear Classifier, Hub Score. F1-score was selected as the dependent variable. As a result, the results of this study showed that the six meta-features including Reverse ReLU Silhouette Score and HHI proposed in this study have a significant effect on the classification performance. (1) The meta-features HHI proposed in this study was significant in the classification performance. (2) The number of variables has a significant effect on the classification performance, unlike the number of classes, but it has a positive effect. (3) The number of classes has a negative effect on the performance of classification. (4) Entropy has a significant effect on the performance of classification. (5) The Reverse ReLU Silhouette Score also significantly affects the classification performance at a significant level of 0.01. (6) The nonlinearity of linear classifiers has a significant negative effect on classification performance. In addition, the results of the analysis by the classification algorithms were also consistent. In the regression analysis by classification algorithm, Naïve Bayes algorithm does not have a significant effect on the number of variables unlike other classification algorithms. This study has two theoretical contributions: (1) two new meta-features (HHI, Reverse ReLU Silhouette score) was proved to be significant. (2) The effects of data characteristics on the performance of classification were investigated using meta-features. The practical contribution points (1) can be utilized in the development of classification algorithm recommendation system according to the characteristics of datasets. (2) Many data scientists are often testing by adjusting the parameters of the algorithm to find the optimal algorithm for the situation because the characteristics of the data are different. In this process, excessive waste of resources occurs due to hardware, cost, time, and manpower. This study is expected to be useful for machine learning, data mining researchers, practitioners, and machine learning-based system developers. The composition of this study consists of introduction, related research, research model, experiment, conclusion and discussion.

• The Journal of the Acoustical Society of Korea
• /
• v.34 no.6
• /
• pp.423-431
• /
• 2015

KIOST-HYU joint acoustics experiment was performed on the western shallow water off the Taean peninsula in the Yellow Sea in May 2013. In this paper, mid-frequency (6~16 kHz) bottom loss data measured in a grazing angle range of $17{\sim}60^{\circ}$ are presented and compared to the predictions obtained using a Rayleigh reflection model. The sediment structure of the experimental site was characterized by multi-layered sediment and the components of the surficial sediment consisted of various types of particles with a mean grain size of $5.9{\phi}$. The model predictions obtained using the mean grain size were not in agreement with the measured bottom loss, and those obtained using the grain size of $4{\phi}$, which was estimated by an inversion process, showed a best fit to the measurements. It would be because the standard deviation of the gain-size distribution of surficial sediment is $4.3{\phi}$, which is much larger than those of other areas around the experimental site. Finally, the model predictions obtained using the geoacoustic parameters estimated from the inversion process for the surficial sediment layer and those corresponding to the mean grain size of $1.3{\phi}$ for lower layer are reasonably agreement with the measured bottom loss data.

• Journal of the Korea Concrete Institute
• /
• v.19 no.3
• /
• pp.359-366
• /
• 2007

Recently, analysis researches on durability are focused on chloride attack and carbonation due to increased social and engineering significance. Generally, chloride penetration and carbonation occur simultaneously except for in submerged condition and chloride behavior in carbonated concrete is evaluated to be different from that in normal concrete. Furthermore, if unavoidable crack occurs in concrete, it influences not only single attack but also coupled deterioration more severely. This is a study on analysis technique with system dynamics for chloride penetration in concrete structures exposed to coupled chloride attack and carbonation through chloride diffusion, permeation, and carbonation reaction. For the purpose, a modeling for chloride behavior considering diffusion and permeation is performed through previous models for early-aged concrete such as MCHHM (multi component hydration heat model) and MPSFM (micro pore structure formation). Then model for combined deterioration is developed considering changed characteristics such as pore distribution, saturation and dissociation of bound chloride content under carbonation. The developed model is verified through comparison with previous experimental data. Additionally, simulation for combined deterioration in cracked concrete is carried out through utilizing previously developed models for chloride penetration and carbonation in cracked concrete. From the simulated results, CCTZ (chloride-carbonation transition zone) for evaluating combined deterioration is proposed. It is numerically verified that concrete with slag has better resistance to combined deterioration than concrete with OPC in sound and cracked concrete.

• Journal of the Earthquake Engineering Society of Korea
• /
• v.13 no.1
• /
• pp.35-43
• /
• 2009

Two types of piloti-type high-rise RC building structures having irregularity in the lower two stories were selected as prototypes, and nonlinear time history analysis was performed using OpenSees to verify the analysis technique and to investigate the seismic capacity of those buildings. One of the buildings studied had a symmetrical moment-resisting frame (BF), while the other had an infilled shear wall in only one of the exterior frames (ESW). A fiber model, consisting of concrete and reinforcing bar represented from the stress-strain relationship, was adapted and used to simulate the nonlinearity of members, and MVLEM (Multi Vertical Linear Element Model) was used to simulate the behavior of the wall. The analytical results simulate the behavior of piloti-type high-rise RC building structures well, including the stiffness and yield force of piloti stories, the rocking behavior of the upper structure and the variation of the axial stiffness of the column due to variation in loading condition. However, MVLEM has a limitation in simulating the abrupt increasing lateral stiffness of a wall, due to the torsional mode behavior of the building. The design force obtained from a nonlinear time history analysis was shown to be about $20{\sim}30%$ smaller than that obtained in the experiment. For this reason, further research is required to match the analytical results with real structures, in order to use nonlinear time history analysis in designing a piloti-type high-rise RC building.

• Korean Journal of Remote Sensing
• /
• v.38 no.6_2
• /
• pp.1709-1722
• /
• 2022

The rise in temperature induced by global warming caused in El Nino and La Nina, and abnormally changed the temperature of seawater. Rainfall concentrates in some locations due to abnormal variations in seawater temperature, causing frequent abnormal floods. It is important to rapidly detect flooded regions to recover and prevent human and property damage caused by floods. This is possible with synthetic aperture radar. This study aims to generate a model that directly derives flood-damaged areas by using modified U-NET and TerraSAR-X images based on Multi Kernel to reduce the effect of speckle noise through various characteristic map extraction and using two images before and after flooding as input data. To that purpose, two synthetic aperture radar (SAR) images were preprocessed to generate the model's input data, which was then applied to the modified U-NET structure to train the flood detection deep learning model. Through this method, the flood area could be detected at a high level with an average F1 score value of 0.966. This result is expected to contribute to the rapid recovery of flood-stricken areas and the derivation of flood-prevention measures.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.3
• /
• pp.359-366
• /
• 2022

In this paper, we present the Korean menu-ordering Sentence Text-to-Speech (TTS) system using conformer-based FastSpeech2. Conformer is the convolution-augmented transformer, which was originally proposed in Speech Recognition. Combining two different structures, the Conformer extracts better local and global features. It comprises two half Feed Forward module at the front and the end, sandwiching the Multi-Head Self-Attention module and Convolution module. We introduce the Conformer in Korean TTS, as we know it works well in Korean Speech Recognition. For comparison between transformer-based TTS model and Conformer-based one, we train FastSpeech2 and Conformer-based FastSpeech2. We collected a phoneme-balanced data set and used this for training our models. This corpus comprises not only general conversation, but also menu-ordering conversation consisting mainly of loanwords. This data set is the solution to the current Korean TTS model's degradation in loanwords. As a result of generating a synthesized sound using ParallelWave Gan, the Conformer-based FastSpeech2 achieved superior performance of MOS 4.04. We confirm that the model performance improved when the same structure was changed from transformer to Conformer in the Korean TTS.

• The Journal of the Korea Contents Association
• /
• v.8 no.12
• /
• pp.1-9
• /
• 2008

Fuzzy Cognitive Map, one of ways to model, describe and infer reasoning relations, is widely used in the field of reasoning knowledge engineering. Despite of the natural and easy understanding of decision and smooth explanation of relation between front and rear, reasoning relation is organized with mathematical haziness and complex algorithm and rarely has an interactive user interface. This paper suggests an interactive Fuzzy Cognitive Map(FCM) construction support system. It builds a FCM increasingly concerning multiple experts' knowledge. Futhermore, it supports user-supportive environment by dynamically displaying the structure of Fuzzy Cognitive Map which is constructed by the interaction between experts and the system.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.61 no.2
• /
• pp.312-320
• /
• 2012

In this study, we introduce a new design methodology of a granular-oriented self-organizing polynomial neural networks (GoSOPNNs) that is based on multi-layer perceptron with Context-based Polynomial Neurons (CPNs) or Polynomial Neurons (PNs). In contrast to the typical architectures encountered in polynomial neural networks (PNN), our main objective is to develop a methodological design strategy of GoSOPNNs as follows : (a) The 1st layer of the proposed network consists of Context-based Polynomial Neuron (CPN). In here, CPN is fully reflective of the structure encountered in numeric data which are granulated with the aid of Context-based Fuzzy C-Means (C-FCM) clustering method. The context-based clustering supporting the design of information granules is completed in the space of the input data while the build of the clusters is guided by a collection of some predefined fuzzy sets (so-called contexts) defined in the output space. (b) The proposed design procedure being applied at each layer of GoSOPNN leads to the selection of preferred nodes of the network (CPNs or PNs) whose local characteristics (such as the number of contexts, the number of clusters, a collection of the specific subset of input variables, and the order of the polynomial) can be easily adjusted. These options contribute to the flexibility as well as simplicity and compactness of the resulting architecture of the network. For the evaluation of performance of the proposed GoSOPNN network, we describe a detailed characteristic of the proposed model using a well-known learning machine data(Automobile Miles Per Gallon Data, Boston Housing Data, Medical Image System Data).

• Journal of the Microelectronics and Packaging Society
• /
• v.24 no.4
• /
• pp.19-22
• /
• 2017

In this study, wavelength-free $1{\times}N$ optical splitter using a tapered polymer waveguide was studied to realize the transmission network of high-speed information communication network. Based on the evaluation of mode converting characteristics of splitter having two tapered multi-mode interference structures, an optimized structure of the $1{\times}N$ splitter was proposed for wide-range of input wavelength. 2D-BPM analysis $1{\times}8$ model showed that insertion loss of the proposed splitter is less than 10 dB for wavelength of input source from 1260 nm to 1650 nm.

• Electrical & Electronic Materials
• /
• v.16 no.9
• /
• pp.64.2-65
• /
• 2003

A 4 nm layer of ZrOx (targeted x-2) was deposited on an interfacial layer(IL) of native oxide (SiO, t∼1.2 nm) surface on 200 mm Si wafers by a manufacturable atomic layer chemical vapor deposition technique at 30$0^{\circ}C$. Some as-deposited layers were subjected to a post-deposition, rapid thermal annealing at $700^{\circ}C$ for 5 min in flowing oxygen at atmospheric pressure. The experimental x-ray diffraction, x-ray photoelectron spectroscopy, high-resolution transmission electron microscopy, and high-resolution parallel electron energy loss spectroscopy results showed that a multiphase and heterogeneous structure evolved, which we call the Zr-O/IL/Si stack. The as-deposited Zr-O layer was amorphous $ZrO_2$-rich Zr silicate containing about 15% by volume of embedded $ZrO_2$ nanocrystals, which transformed to a glass nanoceramic (with over 90% by volume of predominantly tetragonal-$ZrO_2$(t-$ZrO_2$) and monoclinic-$ZrO_2$(m-$ZrO_2$) nanocrystals) upon annealing. The formation of disordered amorphous regions within some of the nanocrystals, as well as crystalline regions with defects, probably gave rise to lattice strains and deformations. The interfacial layer (IL) was partitioned into an upper Si $o_2$-rich Zr silicate and the lower $SiO_{x}$. The latter was sub-toichiometric and the average oxidation state increased from Si0.86$^{＋}$ in $SiO_{0.43}$ (as-deposited) to Si1.32$^{＋}$ in $SiO_{0.66}$ (annealed). This high oxygen deficiency in $SiO_{x}$ indicative of the low mobility of oxidizing specie in the Zr-O layer. The stacks were characterized for their dielectric properties in the Pt/{Zr-O/IL}/Si metal oxide-semiconductor capacitor(MOSCAP) configuration. The measured equivalent oxide thickness (EOT) was not consistent with the calculated EOT using a bilayer model of $ZrO_2$ and $SiO_2$, and the capacitance in accumulation (and therefore, EOT and kZr-O) was frequency dispersive, trends well documented in literature. This behavior is qualitatively explained in terms of the multi-layer nanostructure and nanochemistry that evolves.ves.ves.