• Journal of radiological science and technology
• /
• v.46 no.4
• /
• pp.277-285
• /
• 2023

In this study, pneumonia identification networks with the small number of layers were constructed by using chest X-ray images. The networks had similar trainable-parameters, and the performance of the trained models was quantitatively evaluated with the modification of the network architectures. A total of 6 networks were constructed: convolutional neural network (CNN), VGGNet, GoogleNet, residual network with identity blocks, ResNet with bottleneck blocks and ResNet with identity and bottleneck blocks. Trainable parameters for the 6 networks were set in a range of 273,921-294,817 by adjusting the output channels of convolution layers. The network training was implemented with binary cross entropy (BCE) loss function, sigmoid activation function, adaptive moment estimation (Adam) optimizer and 100 epochs. The performance of the trained models was evaluated in terms of training time, accuracy, precision, recall, specificity and F1-score. The results showed that the trained models with the small number of layers precisely detect pneumonia from chest X-ray images. In particular, the overall quantitative performance of the trained models based on the ResNets was above 0.9, and the performance levels were similar or superior to those based on the CNN, VGGNet and GoogleNet. Also, the residual blocks affected the performance of the trained models based on the ResNets. Therefore, in this study, we demonstrated that the object detection networks with the small number of layers are suitable for detecting pneumonia using chest X-ray images. And, the trained models based on the ResNets can be optimized by applying appropriate residual-blocks.

• Journal of Broadcast Engineering
• /
• v.28 no.3
• /
• pp.267-274
• /
• 2023

There is a problem that the size of the dataset is insufficient due to the limitation of the collection of the medical image public data, so there is a possibility that the existing studies are overfitted to the public dataset. In this paper, we compare the performance of eight (Unet, X-Net, HarDNet, SegNet, PSPNet, SwinUnet, 3D-ResU-Net, UNETR) medical image semantic segmentation models to revalidate the superiority of existing models. Anatomical Tracings of Lesions After Stroke (ATLAS) V1.2, a public dataset for stroke diagnosis, is used to compare the performance of the models and the performance of the models in ATLAS V2.0. Experimental results show that most models have similar performance in V1.2 and V2.0, but X-net and 3D-ResU-Net have higher performance in V1.2 datasets. These results can be interpreted that the models may be overfitted to V1.2.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1996.04a
• /
• pp.741-744
• /
• 1996

Artifical Neural Network(ANN) models were used for forecasting interest rate as a new methodology, which has proven itself successful in financial domain. This research intended to construct ANN models which can maximize the performance of prediction, regarding Corporate Bond Yield (CBY) as interest rate. Synergistic Market Analysis (SMA) was applied to the construction of models [Freedman et al.]. In this aspect, while the models which consist of only time series data for corporate bond yield were devloped, the other models generated through conjunction and reorganization of fundamental variables and market variables were developed. Every model was constructed to predict 1,6, and 12 months after and we obtained 9 ANN models for interest rate forecasting. Multi-layer perceptron networks using backpropagation algorithm showed good performance in the prediction for 1 and 6 months after.

• Journal of the Korean GEO-environmental Society
• /
• v.16 no.5
• /
• pp.13-25
• /
• 2015

This paper proposes an extended model evaluation method that considers not only the model performance but also the model structure and parameter uncertainties in hydrologic modeling. A simple reservoir model (SFM) and distributed kinematic wave models (KWMSS1 and KWMSS2 using topography from 250-m, 500-m, and 1-km digital elevation models) were developed and assessed by three evaluative criteria for model performance, model structural stability, and parameter identifiability. All the models provided acceptable performance in terms of a global response, but the simpler SFM and KWMSS1 could not accurately represent the local behaviors of hydrographs. Moreover, SFM and KWMSS1 were structurally unstable; their performance was sensitive to the applied objective functions. On the other hand, the most sophisticated model, KWMSS2, performed well, satisfying both global and local behaviors. KMSS2 also showed good structural stability, reproducing hydrographs regardless of the applied objective functions; however, superior parameter identifiability was not guaranteed. A number of parameter sets could result in indistinguishable hydrographs. This result indicates that while making hydrologic models complex increases its performance accuracy and reduces its structural uncertainty, the model is likely to suffer from parameter uncertainty.

• KIEAE Journal
• /
• v.10 no.1
• /
• pp.65-72
• /
• 2010

The use of natural light has the potential for improving both the energy efficiency and indoor environmental quality in buildings. A light-pipe system can introduce daylight to spaces that would otherwise not be able to benefit from the advantages of daylight penetration. For the light-pipe system to be widely used in Korea, it is important to quantify its daylighting performance with due consideration regarding the effects imposed by the local climate conditions. This paper presents the evaluation results of existing semi-empirical models to predict daylighting performance of a light-pipe system. The evaluation of the existing models was based on the monitoring data obtained from a underground parking lot in which the light-pipe system was installed. Comparisons were made between the predicted and the monitored data obtained from the study. The results indicated that semi-empirical models which was developed using the experimental data obtained under the Korean climatic conditions had a good prediction performance. We also quantified the effects caused by sky conditions, solar altitudes, room dimensions, and the aspect ratio of a light-pipe system on both the daylighting performance of the light-pipe system and the indoor illuminance distributions of the space using the semi-empirical model. Finally, this paper provides the design guideline of the light-pipe system for its application to an underground parking lot space.

• Korean Journal of Materials Research
• /
• v.33 no.7
• /
• pp.273-278
• /
• 2023

In this study, based on the saturation magnetic flux density experimental values (B_s) of 622 Fe-based bulk metallic glasses (BMGs), regression models were applied to predict B_s using artificial neural networks (ANN), and prediction performance was evaluated. Model performance evaluation was investigated by using the F1 score together with the coefficient of determination (R² score), which is mainly used in regression models. The coefficient of determination can be used as a performance indicator, since it shows the predicted results of the saturation magnetic flux density of full material datasets in a balanced way. However, the BMG alloy contains iron and requires a high saturation magnetic flux density to have excellent applicability as a soft magnetic material, and in this study F1 score was used as a performance indicator to better predict B_s above the threshold value of B_s (1.4 T). After obtaining two ANN models optimized for the R² and F1 score conditions, respectively, their prediction performance was compared for the test data. As a case study to evaluate the prediction performance, new Fe-based BMG datasets that were not included in the training and test datasets were predicted using the two ANN models. The results showed that the model with an excellent F1 score achieved a more accurate prediction for a material with a high saturation magnetic flux density.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.32 no.1
• /
• pp.61-75
• /
• 2007

The purpose of this paper is to develop a new group classification method based on genetic algorithm and to com-pare its prediction performance with those of existing methods in the area of bond rating. To serve this purpose, we conduct various experiments with pilot and general models. Specifically, we first conduct experiments employing two pilot models : the one searching for the cluster center of each group and the other one searching for both the cluster center and the attribute weights in order to maximize classification accuracy. The results from the pilot experiments show that the performance of the latter in terms of classification accuracy ratio is higher than that of the former which provides the rationale of searching for both the cluster center of each group and the attribute weights to improve classification accuracy. With this lesson in mind, we design two generalized models employing genetic algorithm : the one is to maximize the classification accuracy and the other one is to minimize the total misclassification cost. We compare the performance of these two models with those of existing statistical and artificial intelligent models such as MDA, ANN, and Decision Tree, and conclude that the genetic algorithm-based group classification method that we propose in this paper significantly outperforms the other methods in respect of classification accuracy ratio as well as misclassification cost.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.41 no.4
• /
• pp.95-112
• /
• 2016

The purpose of this study is to develop a quantitative model to evaluate the performance of the servitized business model. We aim to quantitatively analyze the decisions of participants in the servitized supply chains, and provide methods to maximize the performance. We consider servitized supply chains consisting of a manufacturer and a service provider, that can be integrated, separated or coordinated based on the relationship between the manufacturer and the service provider. The decision models in each case are developed, and performance and profitability are analyzed. Utilizing the decision models in different cases, we compare the performances of different business models of the servitization. Since our models can be applied to analyze a wide range of the servitization business models, we expect this study can contribute to promote servitization in manufacturing companies by providing methods to evaluate the profitability of the servitization business model.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.44 no.3
• /
• pp.33-38
• /
• 2021

Recently, transfer learning techniques with a base convolutional neural network (CNN) model have widely gained acceptance in early detection and classification of crop diseases to increase agricultural productivity with reducing disease spread. The transfer learning techniques based classifiers generally achieve over 90% of classification accuracy for crop diseases using dataset of crop leaf images (e.g., PlantVillage dataset), but they have ability to classify only the pre-trained diseases. This paper provides with an evaluation scheme on selecting an effective base CNN model for crop disease transfer learning with regard to the accuracy of trained target crops as well as of untrained target crops. First, we present transfer learning models called CDC (crop disease classification) architecture including widely used base (pre-trained) CNN models. We evaluate each performance of seven base CNN models for four untrained crops. The results of performance evaluation show that the DenseNet201 is one of the best base CNN models.

• Communications for Statistical Applications and Methods
• /
• v.28 no.2
• /
• pp.189-204
• /
• 2021

In a regression analysis, a single best model is usually selected among several candidate models. However, it is often useful to combine several candidate models to achieve better performance, especially, in the prediction viewpoint. Model combining methods such as stacking and Bayesian model averaging (BMA) have been suggested from the perspective of averaging candidate models. When the candidate models include a true model, it is expected that BMA generally gives better performance than stacking. On the other hand, when candidate models do not include the true model, it is known that stacking outperforms BMA. Since stacking and BMA approaches have different properties, it is difficult to determine which method is more appropriate under other situations. In particular, it is not easy to find research papers that compare stacking and BMA when regression model assumptions are violated. Therefore, in the paper, we compare the performance among model averaging methods as well as a single best model in the linear regression analysis when standard linear regression assumptions are violated. Simulations were conducted to compare model averaging methods with the linear regression when data include outliers and data do not include them. We also compared them when data include errors from a non-normal distribution. The model averaging methods were applied to the water pollution data, which have a strong multicollinearity among variables. Simulation studies showed that the stacking method tends to give better performance than BMA or standard linear regression analysis (including the stepwise selection method) in the sense of risks (see (3.1)) or prediction error (see (3.2)) when typical linear regression assumptions are violated.