• Journal of Intelligence and Information Systems
• /
• v.23 no.1
• /
• pp.95-108
• /
• 2017

Recently, AlphaGo which is Bakuk (Go) artificial intelligence program by Google DeepMind, had a huge victory against Lee Sedol. Many people thought that machines would not be able to win a man in Go games because the number of paths to make a one move is more than the number of atoms in the universe unlike chess, but the result was the opposite to what people predicted. After the match, artificial intelligence technology was focused as a core technology of the fourth industrial revolution and attracted attentions from various application domains. Especially, deep learning technique have been attracted as a core artificial intelligence technology used in the AlphaGo algorithm. The deep learning technique is already being applied to many problems. Especially, it shows good performance in image recognition field. In addition, it shows good performance in high dimensional data area such as voice, image and natural language, which was difficult to get good performance using existing machine learning techniques. However, in contrast, it is difficult to find deep leaning researches on traditional business data and structured data analysis. In this study, we tried to find out whether the deep learning techniques have been studied so far can be used not only for the recognition of high dimensional data but also for the binary classification problem of traditional business data analysis such as customer churn analysis, marketing response prediction, and default prediction. And we compare the performance of the deep learning techniques with that of traditional artificial neural network models. The experimental data in the paper is the telemarketing response data of a bank in Portugal. It has input variables such as age, occupation, loan status, and the number of previous telemarketing and has a binary target variable that records whether the customer intends to open an account or not. In this study, to evaluate the possibility of utilization of deep learning algorithms and techniques in binary classification problem, we compared the performance of various models using CNN, LSTM algorithm and dropout, which are widely used algorithms and techniques in deep learning, with that of MLP models which is a traditional artificial neural network model. However, since all the network design alternatives can not be tested due to the nature of the artificial neural network, the experiment was conducted based on restricted settings on the number of hidden layers, the number of neurons in the hidden layer, the number of output data (filters), and the application conditions of the dropout technique. The F1 Score was used to evaluate the performance of models to show how well the models work to classify the interesting class instead of the overall accuracy. The detail methods for applying each deep learning technique in the experiment is as follows. The CNN algorithm is a method that reads adjacent values from a specific value and recognizes the features, but it does not matter how close the distance of each business data field is because each field is usually independent. In this experiment, we set the filter size of the CNN algorithm as the number of fields to learn the whole characteristics of the data at once, and added a hidden layer to make decision based on the additional features. For the model having two LSTM layers, the input direction of the second layer is put in reversed position with first layer in order to reduce the influence from the position of each field. In the case of the dropout technique, we set the neurons to disappear with a probability of 0.5 for each hidden layer. The experimental results show that the predicted model with the highest F1 score was the CNN model using the dropout technique, and the next best model was the MLP model with two hidden layers using the dropout technique. In this study, we were able to get some findings as the experiment had proceeded. First, models using dropout techniques have a slightly more conservative prediction than those without dropout techniques, and it generally shows better performance in classification. Second, CNN models show better classification performance than MLP models. This is interesting because it has shown good performance in binary classification problems which it rarely have been applied to, as well as in the fields where it's effectiveness has been proven. Third, the LSTM algorithm seems to be unsuitable for binary classification problems because the training time is too long compared to the performance improvement. From these results, we can confirm that some of the deep learning algorithms can be applied to solve business binary classification problems.

• Information Systems Review
• /
• v.16 no.3
• /
• pp.161-177
• /
• 2014

Corporate credit rating assessment consists of complicated processes in which various factors describing a company are taken into consideration. Such assessment is known to be very expensive since domain experts should be employed to assess the ratings. As a result, the data-driven corporate credit rating prediction using statistical and artificial intelligence (AI) techniques has received considerable attention from researchers and practitioners. In particular, statistical methods such as multiple discriminant analysis (MDA) and multinomial logistic regression analysis (MLOGIT), and AI methods including case-based reasoning (CBR), artificial neural network (ANN), and multiclass support vector machine (MSVM) have been applied to corporate credit rating.2) Among them, MSVM has recently become popular because of its robustness and high prediction accuracy. In this study, we propose a novel optimized MSVM model, and appy it to corporate credit rating prediction in order to enhance the accuracy. Our model, named 'GAMSVM (Genetic Algorithm-optimized Multiclass Support Vector Machine),' is designed to simultaneously optimize the kernel parameters and the feature subset selection. Prior studies like Lorena and de Carvalho (2008), and Chatterjee (2013) show that proper kernel parameters may improve the performance of MSVMs. Also, the results from the studies such as Shieh and Yang (2008) and Chatterjee (2013) imply that appropriate feature selection may lead to higher prediction accuracy. Based on these prior studies, we propose to apply GAMSVM to corporate credit rating prediction. As a tool for optimizing the kernel parameters and the feature subset selection, we suggest genetic algorithm (GA). GA is known as an efficient and effective search method that attempts to simulate the biological evolution phenomenon. By applying genetic operations such as selection, crossover, and mutation, it is designed to gradually improve the search results. Especially, mutation operator prevents GA from falling into the local optima, thus we can find the globally optimal or near-optimal solution using it. GA has popularly been applied to search optimal parameters or feature subset selections of AI techniques including MSVM. With these reasons, we also adopt GA as an optimization tool. To empirically validate the usefulness of GAMSVM, we applied it to a real-world case of credit rating in Korea. Our application is in bond rating, which is the most frequently studied area of credit rating for specific debt issues or other financial obligations. The experimental dataset was collected from a large credit rating company in South Korea. It contained 39 financial ratios of 1,295 companies in the manufacturing industry, and their credit ratings. Using various statistical methods including the one-way ANOVA and the stepwise MDA, we selected 14 financial ratios as the candidate independent variables. The dependent variable, i.e. credit rating, was labeled as four classes: 1(A1); 2(A2); 3(A3); 4(B and C). 80 percent of total data for each class was used for training, and remaining 20 percent was used for validation. And, to overcome small sample size, we applied five-fold cross validation to our dataset. In order to examine the competitiveness of the proposed model, we also experimented several comparative models including MDA, MLOGIT, CBR, ANN and MSVM. In case of MSVM, we adopted One-Against-One (OAO) and DAGSVM (Directed Acyclic Graph SVM) approaches because they are known to be the most accurate approaches among various MSVM approaches. GAMSVM was implemented using LIBSVM-an open-source software, and Evolver 5.5-a commercial software enables GA. Other comparative models were experimented using various statistical and AI packages such as SPSS for Windows, Neuroshell, and Microsoft Excel VBA (Visual Basic for Applications). Experimental results showed that the proposed model-GAMSVM-outperformed all the competitive models. In addition, the model was found to use less independent variables, but to show higher accuracy. In our experiments, five variables such as X7 (total debt), X9 (sales per employee), X13 (years after founded), X15 (accumulated earning to total asset), and X39 (the index related to the cash flows from operating activity) were found to be the most important factors in predicting the corporate credit ratings. However, the values of the finally selected kernel parameters were found to be almost same among the data subsets. To examine whether the predictive performance of GAMSVM was significantly greater than those of other models, we used the McNemar test. As a result, we found that GAMSVM was better than MDA, MLOGIT, CBR, and ANN at the 1% significance level, and better than OAO and DAGSVM at the 5% significance level.

• Journal of Korean Society of Transportation
• /
• v.23 no.7 s.85
• /
• pp.165-179
• /
• 2005

The object of this study is a development of a stochastic user equilibrium transit assignment algorithm for multiple user classes considering stochastic characteristics and heterogeneous attributes of passengers. The existing transit assignment algorithms have limits to attain realistic results because they assume a characteristic of passengers to be equal. Although one group with transit information and the other group without it have different trip patterns, the past studies could not explain the differences. For overcoming the problems, we use following methods. First, we apply a stochastic transit assignment model to obtain the difference of the perceived travel cost between passengers and apply a multiple user class assignment model to obtain the heterogeneous qualify of groups to get realistic results. Second, we assume that person trips have influence on the travel cost function in the development of model. Third, we use a C-logit model for solving IIA(independence of irrelevant alternatives) problems. According to repetition assigned trips and equivalent path cost have difference by each group and each path. The result comes close to stochastic user equilibrium and converging speed is very fast. The algorithm of this study is expected to make good use of evaluation tools in the transit policies by applying heterogeneous attributes and OD data.

• Computational Structural Engineering
• /
• v.7 no.3
• /
• pp.95-103
• /
• 1994

The safety related components in the nuclear power plant should be designed to withstand the seismic load. Among these components the integrity of reactor internals under earthquake load is important in stand points of safety and economics, because these are classified to Seismic Class I components. So far the modelling methods of reactor internals have been investigated by many authors. In this paper, the dynamic behaviour of reactor internals of Yong Gwang 1&2 nuclear power plants under SSE(Safe Shutdown Earthquake) load is analyzed by using of the simpled Global Beam Model. For this, as a first step, the characteristic analysis of reactor internal components are performed by using of the finite element code ANSYS. And the Global Beam Model for reactor internals which includes beam elements, nonlinear impact springs which have gaps in upper and lower positions, and hydrodynamical couplings which simulate the fluid-filled cylinders of reactor vessel and core barrel structures is established. And for the exciting external force the response spectrum which is applied to reactor support is converted to the time history input. With this excitation and the model the dynamic behaviour of reactor internals is obtained. As the results, the structural integrity of reactor internal components under seismic excitation is verified and the input for the detailed duel assembly series model could be obtained. And the simplicity and effectiveness of Global Beam Model and the economics of the explicit Runge-Kutta-Gills algorithm in impact problem of high frequency interface components are confirmed.

• Journal of Intelligence and Information Systems
• /
• v.22 no.2
• /
• pp.127-142
• /
• 2016

Deep learning model is a kind of neural networks that allows multiple hidden layers. There are various deep learning architectures such as convolutional neural networks, deep belief networks and recurrent neural networks. Those have been applied to fields like computer vision, automatic speech recognition, natural language processing, audio recognition and bioinformatics where they have been shown to produce state-of-the-art results on various tasks. Among those architectures, convolutional neural networks and recurrent neural networks are classified as the supervised learning model. And in recent years, those supervised learning models have gained more popularity than unsupervised learning models such as deep belief networks, because supervised learning models have shown fashionable applications in such fields mentioned above. Deep learning models can be trained with backpropagation algorithm. Backpropagation is an abbreviation for "backward propagation of errors" and a common method of training artificial neural networks used in conjunction with an optimization method such as gradient descent. The method calculates the gradient of an error function with respect to all the weights in the network. The gradient is fed to the optimization method which in turn uses it to update the weights, in an attempt to minimize the error function. Convolutional neural networks use a special architecture which is particularly well-adapted to classify images. Using this architecture makes convolutional networks fast to train. This, in turn, helps us train deep, muti-layer networks, which are very good at classifying images. These days, deep convolutional networks are used in most neural networks for image recognition. Convolutional neural networks use three basic ideas: local receptive fields, shared weights, and pooling. By local receptive fields, we mean that each neuron in the first(or any) hidden layer will be connected to a small region of the input(or previous layer's) neurons. Shared weights mean that we're going to use the same weights and bias for each of the local receptive field. This means that all the neurons in the hidden layer detect exactly the same feature, just at different locations in the input image. In addition to the convolutional layers just described, convolutional neural networks also contain pooling layers. Pooling layers are usually used immediately after convolutional layers. What the pooling layers do is to simplify the information in the output from the convolutional layer. Recent convolutional network architectures have 10 to 20 hidden layers and billions of connections between units. Training deep learning networks has taken weeks several years ago, but thanks to progress in GPU and algorithm enhancement, training time has reduced to several hours. Neural networks with time-varying behavior are known as recurrent neural networks or RNNs. A recurrent neural network is a class of artificial neural network where connections between units form a directed cycle. This creates an internal state of the network which allows it to exhibit dynamic temporal behavior. Unlike feedforward neural networks, RNNs can use their internal memory to process arbitrary sequences of inputs. Early RNN models turned out to be very difficult to train, harder even than deep feedforward networks. The reason is the unstable gradient problem such as vanishing gradient and exploding gradient. The gradient can get smaller and smaller as it is propagated back through layers. This makes learning in early layers extremely slow. The problem actually gets worse in RNNs, since gradients aren't just propagated backward through layers, they're propagated backward through time. If the network runs for a long time, that can make the gradient extremely unstable and hard to learn from. It has been possible to incorporate an idea known as long short-term memory units (LSTMs) into RNNs. LSTMs make it much easier to get good results when training RNNs, and many recent papers make use of LSTMs or related ideas.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.15 no.2
• /
• pp.113-125
• /
• 2012

This research is aimed at evaluating the land surface characterization of KOMPSAT-3A middle infrared (MIR) data. Airborne Hyperspectral Scanner (AHS) data, which has MIR bands with high spatial resolution, were used to assess land surface temperature (LST) retrieval and classification accuracy of MIR bands. Firstly, LST values for daytime and nighttime, which were calculated with AHS thermal infrared (TIR) bands, were compared to digital number of AHS MIR bands. The determination coefficient of AHS band 68 (center wavelength $4.64{\mu}m$) was over 0.74, and was higher than other MIR bands. Secondly, The land cover maps were generated by unsupervised classification methods using the AHS MIR bands. Each class of land cover maps for daytime, such as water, trees, green grass, roads, roofs, was distinguished well. But some classes of land cover maps for nighttime, such as trees versus green grass, roads versus roofs, were not separated. The image classification using the difference images between daytime AHS MIR bands and nighttime AHS MIR bands were conducted to enhance the discrimination ability of land surface for AHS MIR imagery. The classification accuracy of the land cover map for zone 1 and zone 2 was 67.5%, 64.3%, respectively. It was improved by 10% compared to land cover map of daytime AHS MIR bands and night AHS MIR bands. Consequently, new algorithm based on land surface characteristics is required for temperature retrieval of high resolution MIR imagery, and the difference images between daytime and nighttime was considered to enhance the ability of land surface characterization using high resolution MIR data.

• Journal of Life Science
• /
• v.30 no.9
• /
• pp.813-818
• /
• 2020

The archaeal clusters of orthologous genes (arCOG) algorithm, which identifies common genes among archaebacterial genomes, was used to identify conservative genes among 168 archaebacterial strains. The numbers of conserved orthologs were 14, 10, 9, and 8 arCOGs in 168, 167, 166, and 165 strains, respectively. Among 41 conserved arCOGs, 13 were related to function J (translation, ribosomal structure, and biogenesis), and 10 were related to function L (replication, recombination, and repair). Among the 14 conserved arCOGs in all 168 strains, 6 arCOGs of tRNA synthetase comprised the highest proportion. Of the remaining 8 arCOGs, 2 are involved in reactions with ribosomes, 2 for tRNA synthesis, 2 for DNA replication, and 2 for transcription. These results showed the importance of protein expression in archaea. For the classes or orders having 3 or more members, genomic analysis was performed by averaging the distance values of the conservative arCOGs. Classes Archaeoglobi and Thermoplasmata of the phylum Euryarchaeota showed the lowest and the highest average of distance value, respectively. This study can provides data necessary for basic scientific research and the development of antibacterial agents and tumor control.

• Journal of Life Science
• /
• v.25 no.9
• /
• pp.1007-1013
• /
• 2015

A COG (Cluster of Orthologous Groups of proteins) algorithm was applied to detect conserved genes in 711 prokaryotes. Only COG0080 (ribosomal protein L11) was common among all the 711 prokaryotes analyzed and 58 COGs were common in more than 700 prokaryotes. Nine COGs among 58, including COG0197 (endonuclease III) and COG0088 (ribosomal protein L4), were conserved in a form of one gene per one organism. COG0008 represented 1356 genes in 709 of the prokaryotes and this was the highest number of genes among 58 COGs. Twenty-two COGs were conserved in more than 708 prokaryotes. Of these, two were transcription related, four were tRNA synthetases, eight were large ribosomal subunits, seven were small ribosomal subunits, and one was translation elongation factor. Among 58 conserved COGs in more than 700 prokaryotes, 50 (86.2%) were translation related, and four (6.9%) were transcription related, pointing to the importance of protein-synthesis in prokaryotes. Among these 58 COGs, the most conserved COG was COG0060 (isoleucyl tRNA synthetase), and the least conserved was COG0143 (methionyl tRNA synthetase). Archaea and eubacteria were discriminated in the genomic analysis by the average distance and variation in distance of common COGs. The identification of these conserved genes could be useful in basic and applied research, such as antibiotic development and cancer therapeutics.

• Spatial Information Research
• /
• v.18 no.3
• /
• pp.1-11
• /
• 2010

Recently, the construction of three dimensional spatial information of Dam reservoir area is very important part in Dam management work such as sediment survey, but it is difficult to acquire detailed terrain data because totalstation and single beam echo sounder are applied to terrain survey. This study presented method to construct detailed terrain data of Dam reservoir area using LiDAR and multi beam echo sounder. First, LiDAR survey was carried out in land zone and calibration process was applied by ground control point. And also the DEM of land zone was constructed by using algorithm, which eliminated building and vegetation class. As the result of validation of LiDAR DEM using GPS terrain survey, it was possible to construct three dimensional terrain data that was satisfied with the tolerance error of LiDAR, which was the standard error of LiDAR DEM showed as 0.108m. Also multi beam echo sounder was applied to the survey of water zone and it could construct spatial information that was satisfied with bathymetry surveying tolerance error of International Hydrographic Organization by validation with terrain survey data. And LiDAR and multi beam echo sounder data were integrated and it was possible to construct three dimensional spatial imagery information that can be applied to Dam management work such as the estimation of sediment amounts or the monitoring of terrain change by linking with high resolution orthophoto.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.14 no.5
• /
• pp.2393-2399
• /
• 2013

In this paper, we proposed a method to screening of Alzheimer's disease (AD) from Raman spectra of platelet with synthesis of basis spectra using singular value decomposition (SVD). Raman spectra of platelet from AD transgenic mice are preprocessed with denoising, removal background and normalization method. The column vectors of each data matrix consist of Raman spectrum of AD and normal (NR). The matrix is factorized using SVD algorithm and then the basis spectra of AD and NR are determined by 12 column vectors of each matrix. The classification process is completed by select the class that minimized the root-mean-square error between the validation spectrum and the linear synthesized spectrum of the basis spectra. According to the experiments involving 278 Raman spectra, the proposed method gave about 97.6% classification rate, which is better performance about 6.1% than multi-layer perceptron (MLP) with extracted features using principle components analysis (PCA). The results show that the basis spectra using SVD is well suited for the diagnosis of AD by Raman spectra from platelet.