• Communications for Statistical Applications and Methods
• /
• 제28권2호
• /
• 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.

• International Journal of Internet, Broadcasting and Communication
• /
• 제16권3호
• /
• pp.262-273
• /
• 2024

We provide a detailed analysis of the data processing and model training process for vulnerability classification using Transformer-based language models, especially sentence text-to-text transformers (ST5)-XXL and XLNet. The main purpose of this study is to compare the performance of the two models, identify the strengths and weaknesses of each, and determine the optimal learning rate to increase the efficiency and stability of model training. We performed data preprocessing, constructed and trained models, and evaluated performance based on data sets with various characteristics. We confirmed that the XLNet model showed excellent performance at learning rates of 1e-05 and 1e-04 and had a significantly lower loss value than the ST5-XXL model. This indicates that XLNet is more efficient for learning. Additionally, we confirmed in our study that learning rate has a significant impact on model performance. The results of the study highlight the usefulness of ST5-XXL and XLNet models in the task of classifying security vulnerabilities and highlight the importance of setting an appropriate learning rate. Future research should include more comprehensive analyzes using diverse data sets and additional models.

• 한국대기환경학회지
• /
• 제19권4호
• /
• pp.397-414
• /
• 2003

The performance of one-particle stochastic Lagrangian models for passive tracer dispersion are evaluated against measurements in horizontally-homogeneous neutrally-stratified atmospheric surface layer. State-of-the-technology models as well as classical Langevin models, all in class of well mixed models are numerically implemented for inter-model comparison study. Model results (far-downstream asymptotic behavior and vertical profiles of the time averaged concentrations, concentration fluxes, and concentration fluctuations) are compared with the reported measurements. The results are: 1) the far-downstream asymptotic trends of all models except Reynolds model agree well with Garger and Zhukov's measurements. 2) profiles of the average concentrations and vertical concentration fluxes by all models except Reynolds model show good agreement with Raupach and Legg's experimental data. Reynolds model produces horizontal concentration flux profiles most close to measurements, yet all other models fail severely. 3) With temporally correlated emissions, one-particle models seems to simulate fairly the concentration fluctuations induced by plume meandering, when the statistical random noises are removed from the calculated concentration fluctuations. Analytical expression for the statistical random noise of one-particle model is presented. This study finds no indication that recent models of most delicate theoretical background are superior to the simple Langevin model in accuracy and numerical performance at well.

• 대한기계학회논문집B
• /
• 제22권1호
• /
• pp.70-78
• /
• 1998

The present study has tested semi-empirical loss models for a reliable performance prediction of mixed-flow pumps with four different specific speeds. In order to improve the predictive capabilities, this paper recommends a new internal loss model and a modified parasitic loss model. The prediction method presented here is also compared with that based on two-dimensional cascade theory. Predicted performance curves by the proposed set of loss models agree fairly well with experimental data for a variety of mixed-flow pumps in the normal operating range, but further studies considering 'droop-like' head performance characteristic due to flow reversal in mixed-flow impellers at low flow range near shut-off head are needed.

• 유체기계공업학회:학술대회논문집
• /
• 유체기계공업학회 2000년도 유체기계 연구개발 발표회 논문집
• /
• pp.110-115
• /
• 2000

The mean streamline analysis using the empirical loss correlations has been developed for performance prediction of industrial mixed-flow fan impellers in the present study. New simple, but effective, models for the additional Euler input work characteristic and an internal recirculation loss due to internal flow reversal under the low flowrate conditions are proposed in this paper. Comparison of overall performance predictions with six sets of test data of mixed-flow fans is accomplished to demonstrate the accuracy of the proposed models. Predicted performance curves by the present set of loss models agree fairly well with experimental data for a variety of mixed-flow fan impellers over the entire operating conditions. The prediction method presented herein can be used efficiently in the conceptual design phase of mixed-flow fan impellers.

• 음성과학
• /
• 제10권3호
• /
• pp.225-239
• /
• 2003

In this paper, we develop a keyword spotting system using vocabulary-independent speech recognition technique, and investigate several non-keyword modeling and post-processing methods to improve its performance. In order to model non-keyword speech segments, monophone clustering and Gaussian Mixture Model (GMM) are considered. We employ likelihood ratio scoring method for the post-processing schemes to verify the recognition results, and filler models, anti-subword models and N-best decoding results are considered as an alternative hypothesis for likelihood ratio scoring. We also examine different methods to construct anti-subword models. We evaluate the performance of our system on the automatic telephone exchange service task. The results show that GMM-based non-keyword modeling yields better performance than that using monophone clustering. According to the post-processing experiment, the method using anti-keyword model based on Kullback-Leibler distance and N-best decoding method show better performance than other methods, and we could reduce more than 50% of keyword recognition errors with keyword rejection rate of 5%.

• JSTS:Journal of Semiconductor Technology and Science
• /
• 제10권1호
• /
• pp.28-36
• /
• 2010

Hardware implementation methods for Artificial Neural Network (ANN) have been researched for a long time to achieve high performance. We have proposed a Network on Chip (NoC) for ANN, and this architecture can reduce communication load and increase performance when an implemented ANN is small. In this paper, a multiple NoC models are proposed for ANN, which can implement both a small size ANN and a large size one. The simulation result shows that the proposed multiple NoC models can reduce communication load, increase system performance of connection-per-second (CPS), and reduce system running time compared with the existing hardware ANN. Furthermore, this architecture is reconfigurable and reparable. It can be used to implement different applications of ANN.

• 반도체디스플레이기술학회지
• /
• 제21권3호
• /
• pp.63-68
• /
• 2022

In this study, we produced defect detection models using parallel convolution neural networks. If convolution neural networks are constructed parallel type, the model's detection accuracy will increase and detection time will decrease. We produced parallel-type defect detection models using 4 types of convolutional algorithms. The performance of models was evaluated using evaluation indicators. The model's performance is detection accuracy and detection time. We compared the performance of each parallel model. The detection accuracy of the model using AlexNet is 97 % and the detection time is 0.3 seconds. We confirmed that when AlexNet algorithm is constructed parallel type, the model has the highest performance.

• 한국정보기술학회 영문논문지
• /
• 제10권2호
• /
• pp.139-151
• /
• 2020

Entity name recognition is a part of information extraction that extracts entity names from documents and classifies the types of extracted entity names. Entity name recognition technologies are widely used in natural language processing, such as information retrieval, machine translation, and query response systems. Various deep learning-based models exist to improve entity name recognition performance, but studies that compared and analyzed these models on Korean data are insufficient. In this paper, we compare and analyze the performance of CRF, LSTM-CRF, BiLSTM-CRF, and BERT, which are actively used to identify entity names using Korean data. Also, we compare and evaluate whether embedding models, which are variously used in recent natural language processing tasks, can affect the entity name recognition model's performance improvement. As a result of experiments on patent data and Korean corpus, it was confirmed that the BiLSTM-CRF using FastText method showed the highest performance.

• Journal of Information Processing Systems
• /
• 제19권5호
• /
• pp.673-687
• /
• 2023

Convolutional neural network-based deep learning technology is the most commonly used in image identification, but it requires large-scale data for training. Therefore, application in specific fields in which data acquisition is limited, such as in the military, may be challenging. In particular, the identification of ground weapon systems is a very important mission, and high identification accuracy is required. Accordingly, various studies have been conducted to achieve high performance using small-scale data. Among them, the ensemble method, which achieves excellent performance through the prediction average of the pre-trained models, is the most representative method; however, it requires considerable time and effort to find the optimal combination of ensemble models. In addition, there is a performance limitation in the prediction results obtained by using an ensemble method. Furthermore, it is difficult to obtain the ensemble effect using models with imbalanced classification accuracies. In this paper, we propose a transfer learning-based feature fusion technique for heterogeneous models that extracts and fuses features of pre-trained heterogeneous models and finally, fine-tunes hyperparameters of the fully connected layer to improve the classification accuracy. The experimental results of this study indicate that it is possible to overcome the limitations of the existing ensemble methods by improving the classification accuracy through feature fusion between heterogeneous models based on transfer learning.