• Journal of Korean Society for Quality Management
• /
• v.50 no.3
• /
• pp.459-471
• /
• 2022

Purpose: The purpose of this study was to increase prediction accuracy for an anomaly interval identified using an artificial intelligence-based time series anomaly detection technique by establishing a pre-processing process. Methods: Significant variables were extracted by applying feature selection techniques, and anomalies were derived using the TadGAN time series anomaly detection algorithm. After applying machine learning and deep learning methodologies using normal section data (excluding anomaly sections), the explanatory power of the anomaly sections was demonstrated through performance comparison. Results: The results of the machine learning methodology, the performance was the best when SHAP and TadGAN were applied, and the results in the deep learning, the performance was excellent when Chi-square Test and TadGAN were applied. Comparing each performance with the papers applied with a Conventional methodology using the same data, it can be seen that the performance of the MLR was significantly improved to 15%, Random Forest to 24%, XGBoost to 30%, Lasso Regression to 73%, LSTM to 17% and GRU to 19%. Conclusion: Based on the proposed process, when detecting unsupervised learning anomalies of data that are not actually labeled in various fields such as cyber security, financial sector, behavior pattern field, SNS. It is expected to prove the accuracy and explanation of the anomaly detection section and improve the performance of the model.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.1
• /
• pp.133-142
• /
• 2011

When TCP operates in multi-hop wireless networks, it suffers from severe performance degradation due to the different characteristics of wireless networks and wired networks. This is because TCP reacts to wireless packet losses by unnecessarily decreasing its sending rate assuming the losses as congestion losses. Although several loss differentiation algorithms (LDAs) have been proposed to avoid such performance degradation, their detection accuracies are not high as much as we expect. In addition the schemes have a tendency to sacrifice the detection accuracy of congestion losses while they improve the detection accuracy of wireless losses. In this paper, we suggest a new sender-based loss differentiation scheme which enhances the detection accuracy of wireless losses while minimizing the sacrifice of the detection accuracy of congestion losses. Our scheme estimates the rate of queue usage which is highly correlated with the congestion in the network path between a TCP sender and a receiver, and it distinguishes congestion losses from wireless losses by comparing the estimated queue usage with a certain threshold. In the extensive experiments based on a network simulator, QualNet, we measure and compare each detection accuracy of wireless losses and congestion losses, and evaluate the performance enhancement in each scheme. The results show that our scheme has the highest accuracy among the LDAs and it improves the most highly TCP performance in multi-hop wireless networks.

• 한국정보디스플레이학회:학술대회논문집
• /
• 2008.10a
• /
• pp.1358-1361
• /
• 2008

Several different techniques currently exist for measuring display view angle performance. These include conoscopes, goniometric systems, and, most recently introduced to the market, instruments based on Imaging Sphere technology. This paper will compare measurement accuracy and speed of these various methodologies with different FPD types and even 3-D displays.

• The Korean Journal of Emergency Medical Services
• /
• v.11 no.2
• /
• pp.51-66
• /
• 2007

Purpose: With a view to providing basic data to develop cardiopulmonary resuscitation education suitable for elementary students, the cardiopulmonary resuscitation education was conducted to grasp students' knowledge, skills accuracy and the attitude change before and after the education. Methods: Convenience sampling was made on fourth and fifth graders(total-35 students) of S elementary school located in K city, Chungcheongnam-do, and this was a pre-experiment research designed before and after choosing a single group. In terms of methods, specifically we, researchers ; 1) Handed out questionnaires to students directly to make them fill in firsthand and collected the questionnaires. 2) Utilized PPT materials based on 2005 AHA guideline and DVD materials of AHA, to give students theoretical education of cardiopulmonary resuscitation. We used Anne/SkillReporter$^{(R)}$ torso produced by Leardal Inc, and Little Anne to conduct practical education individually. 3) Asked students to give Anne/SkillReporter$^{(R)}$ torso cardiopulmonary resuscitation five times with the ratio of 30 : 2, and then one of researchers filled in the evaluation sheet individually. 4) Evaluated the accuracy of students' ability to perform the resuscitation based on the record of Anne/SkillReporter$^{(R)}$ integrated printer(which was the objective tool to grasp students' skills accuracy). 5) Gave out questionnaires to make students fill them in and then collected them. after completing the practical evaluation. Results: 1) In case of the attitude about cardiopulmonary resuscitation, Students' confidency rose from 19.28%(before the education) to 93.57(after the education)- which is a positive change. 2) As the result of the education, some elementary students scored 11 points (full score-16 points), up from 5 points before the education, in terms of the knowledge about cardiopulmonary resuscitation. The average point also reached 13.14 points(after the education), jump from 8.37(before the education), which was the rise of 29.8%. 3) When it comes to the practical performance, the skills accuracy was 80.93% on average, and the calculation method was as follows: total items were 16, and each item was marked form 0 to 2 points, meaning the full score was 32 points. The minimum score was 19 points and the maximum was 32($M{\pm}SD=25.90{\pm}2.88$), which was calculated based on percentage. 4) Regarding skills accuracy, respiration accuracy(%)($M{\pm}SD=30.20{\pm}27.16$) was higher than pressure accuracy(%) ($M{\pm}SD=15.34{\pm}25.27$). Conclusion: The result showed that students' attitude on cardiopulmonary resuscitation changed positively. and meaningful difference(p = .00) existed in the change of students' knowledge. In terms of skills accuracy. chest compression and airway control showed high accuracy, but the result of Anne/SkillReporter$^{(R)}$ performance showed that the accuracy of chest compression was lower than that of mouth-to-mouth resuscitation.

• Coupled systems mechanics
• /
• v.12 no.6
• /
• pp.539-555
• /
• 2023

Although some prediction models have successfully developed for ultra-high performance concrete (UHPC), they do not provide insights and explicit relations between all constituents and its consistency, and compressive strength. In the present study, based on the experimental results, several mathematical models have been evaluated to predict the consistency and the 28-day compressive strength of UHPC. The models used were Linear, Logarithmic, Inverse, Power, Compound, Quadratic, Cubic, Mixed, Sinusoidal and Cosine equations. The applicability and accuracy of these models were investigated using experimental data, which were collected from literature. The comparisons between the models and the experimental results confirm that the majority of models give acceptable prediction with a high accuracy and trivial error rates, except Linear, Mixed, Sinusoidal and Cosine equations. The assessment of the models using numerical methods revealed that the Quadratic and Inverse equations based models provide the highest predictability of the compressive strength at 28 days and consistency, respectively. Hence, they can be used as a reliable tool in mixture design of the UHPC.

• Proceedings of the KIEE Conference
• /
• 2003.04a
• /
• pp.212-216
• /
• 2003

The performance of active power filter depends on the inverter characteristic, the control method, and the accuracy of reference signal generator. The accuracy of reference generator is the most critical item to determine the performance of active power filter. This paper introduces a novel reference signal generator composed of improved adaptive predictive filter. The performance of proposed reference signal generator was verified by means of simulation with MATLAB. The simulation result confirm that the proposed reference signal generator can be utilized for the active power filter.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2006.05a
• /
• pp.397-401
• /
• 2006

In this paper, we proposed a high precision location estimation algorithm using an enhanced decision scheme for RTLS and analyzed its performance in point of an average estimation error distance at 2D coordinates searching area, $300m\times300m$ and LOS propagation environments. Also the performance was compared with that of conventional TDOA algorithm according to the number of available reader and received sub-blink. From the results, we confirmed that the proposed location estimation algorithm using an enhanced decision scheme was able to improve an estimation accuracy even in boundary region of searching area. Moreover, effectively reduced an error distance in entire searching area so that increased the stability of location estimation in RTLS. Therefore, we verified that the proposed algorithm provided a more higher estimation accuracy and stability than conventional TDOA.

• The Journal of the Acoustical Society of Korea
• /
• v.24 no.4E
• /
• pp.144-151
• /
• 2005

We evaluate the performance of emotion recognition via speech signals when a plain speaker talks to an entertainment robot. For each frame of a speech utterance, we extract the frame-based features: pitch, energy, formant, band energies, mel frequency cepstral coefficients (MFCCs), and velocity/acceleration of pitch and MFCCs. For discriminative classifiers, a fixed-length utterance-based feature vector is computed from the statistics of the frame-based features. Using a speaker-independent database, we evaluate the performance of two promising classifiers: support vector machine (SVM) and hidden Markov model (HMM). For angry/bored/happy/neutral/sad emotion classification, the SVM and HMM classifiers yield $42.3\%\;and\;40.8\%$ accuracy, respectively. We show that the accuracy is significant compared to the performance by foreign human listeners.

• Proceedings of the KIPE Conference
• /
• 2004.07b
• /
• pp.575-578
• /
• 2004

The performance of an active power filter(APF) depends on the inverter characteristic, the control method, and the accuracy of reference signal generator. The accuracy of reference generator is the most critical item to determine the performance of active power filter. This paper introduces a novel reference signal generator composed of improved adaptive predictive filter. The performance of proposed reference signal generator was first verified through a simulation with MATLAB. Furthermore, the application of feasibility was evaluated through experimenting with a single-phase APF prototype based on the proposed reference generator, which was implemented using the TMS320C31 floating-point signal processor. Both simulations and experimental results confirm that our reference signal generator can be used successfully in practical active power filters.

• Structural Engineering and Mechanics
• /
• v.17 no.5
• /
• pp.671-690
• /
• 2004

A meshless method with novel variation of point collocation by finite mixture approximation is developed in this paper, termed the meshless finite mixture (MFM) method. It is based on the finite mixture theorem and consists of two or more existing meshless techniques for exploitation of their respective merits for the numerical solution of partial differential boundary value (PDBV) problems. In this representation, the classical reproducing kernel particle and differential quadrature techniques are mixed in a point collocation framework. The least-square method is used to optimize the value of the weight coefficient to construct the final finite mixture approximation with higher accuracy and numerical stability. In order to validate the developed MFM method, several one- and two-dimensional PDBV problems are studied with different mixed boundary conditions. From the numerical results, it is observed that the optimized MFM weight coefficient can improve significantly the numerical stability and accuracy of the newly developed MFM method for the various PDBV problems.