• Journal of Communications and Networks
• /
• 제17권6호
• /
• pp.568-581
• /
• 2015

Predicting locations of users with portable devices such as IP phones, smart-phones, iPads and iPods in public wireless local area networks (WLANs) plays a crucial role in location management and network resource allocation. Many techniques in machine learning and data mining, such as sequential pattern mining and clustering, have been widely used. However, these approaches have two deficiencies. First, because they are based on profiles of individual mobility behaviors, a sequential pattern technique may fail to predict new users or users with movement on novel paths. Second, using similar mobility behaviors in a cluster for predicting the movement of users may cause significant degradation in accuracy owing to indistinguishable regular movement and random movement. In this paper, we propose a novel fusion technique that utilizes mobility rules discovered from multiple similar users by combining clustering and sequential pattern mining. The proposed technique with two algorithms, named the clustering-based-sequential-pattern-mining (CSPM) and sequential-pattern-mining-based-clustering (SPMC), can deal with the lack of information in a personal profile and avoid some noise due to random movements by users. Experimental results show that our approach outperforms existing approaches in terms of efficiency and prediction accuracy.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권1호
• /
• pp.180-194
• /
• 2021

For small image registration, feature-based approaches are likely to fail as feature detectors cannot detect enough feature points from low-resolution images. The classic FFT approach's prediction accuracy is high, but the registration time can be relatively long, about several seconds to register one image pair. To achieve real-time and high-precision rigid registration for small images, we apply deep neural networks for supervised rigid transformation prediction, which directly predicts the transformation parameters. We train deep registration models with rigidly transformed CIFAR-10 images and STL-10 images, and evaluate the generalization ability of deep registration models with transformed CIFAR-10 images, STL-10 images, and randomly generated images. Experimental results show that the deep registration models we propose can achieve comparable accuracy to the classic FFT approach for small CIFAR-10 images (32×32) and our LSTM registration model takes less than 1ms to register one pair of images. For moderate size STL-10 images (96×96), FFT significantly outperforms deep registration models in terms of accuracy but is also considerably slower. Our results suggest that deep registration models have competitive advantages over conventional approaches, at least for small images.

• International Journal of Internet, Broadcasting and Communication
• /
• 제14권3호
• /
• pp.285-291
• /
• 2022

The number of people enrolling in universities is rising due to the simplicity of applying and the benefit of earning a bachelor's degree. However, the on-time graduation rate has declined since plenty of students fail to complete their courses and take longer to get their diplomas. Even though there are various reasons leading to the aforementioned problem, it is crucial to emphasize the cause originating from the management and care of learners. In fact, understanding students' difficult situations and offering timely Number of Test data and advice would help prevent college dropouts or graduate delays. In this study, we present a machine learning-based method for early detection at-risk students, using data obtained from graduates of the Faculty of Information Technology, Dainam University, Vietnam. We experiment with several fundamental machine learning methods before implementing the parameter optimization techniques. In comparison to the other strategies, Random Forest and Grid Search (RF&GS) and Random Forest and Random Search (RF&RS) provided more accurate predictions for identifying at-risk students.

• 품질경영학회지
• /
• 제24권4호
• /
• pp.156-167
• /
• 1996

D-optimality is used often in design augmentation of mixture experiments. Although such alphabetic criteria provide a valuable foundation for generating designs, they often fail to convey the true nature of the design's support of the fitted model in terms of prediction variance over a region of interest. Thus, a graphical method is proposed to evaluate augmented designs in mixture experiments. This method can be used to evaluate the effect of missing observation and outlier in mixture experiments.

• 한국전산유체공학회:학술대회논문집
• /
• 한국전산유체공학회 2005년도 추계 학술대회논문집
• /
• pp.105-110
• /
• 2005

Evaluation of elliptic blending turbulence model (EBM) together with the two-layer model, shear stress transport (SST) model and elliptic relaxation model (V2-F) is performed for a better prediction of thermal stratification in an upper plenum of a liquid metal reactor by applying them to the experiment conducted at JNC. The algebraic flux model is used for treating the turbulent heat flux. There exist much differences between turbulence models in predicting the temporal variation of temperature. The V2-F model and the EBM better predict the steep gradient of temperature at the interface of thermal stratification, and the V2-F model and EBM predict properly the oscillation of temperature. The two-layer model and SST model fail to predict the temporal oscillation of temperature.

• 한국산업정보학회논문지
• /
• 제7권2호
• /
• pp.33-39
• /
• 2002

Accelerated life testing of a car is used to get information quickly on its life distribution. Test cars are no under severe conditions and fail sooner than under usual conditions. A model is fitted to the accelerated failure times and then extrapolated to estimate the life distribution under usual conditions. This paper presents an accelerated test md the reliability growth theory, and applies it to some subsystems of cars during their prototype and pilot testing. The data presented illustrates explicitly the prediction of the reliability growth in the product development cycle. The application of these techniques is a part of the product assurance function that plays an important role in product reliability improvement.

• 한국항공운항학회:학술대회논문집
• /
• 한국항공운항학회 2015년도 추계학술대회
• /
• pp.25-28
• /
• 2015

Aircraft conceptual design usually uses low to medium fidelity analysis to determine the basic configuration of an aircraft. Optimum solution is bounded by at least one of the constraints in most cases. This solution has risk to fail at later stage when analyzed with more sophisticated analysis tools. This research uses pre-constructed database to estimate the analysis prediction errors associated with simplified analysis methods. A possibility based design optimization framework is developed to utilize the newly proposed piecewise-linear fuzzy membership functions that compensate the discrepancies caused by simplified analysis. The proposed approach for aircraft design produces the optimum aircraft configurations that are less likely to fall into infeasible region when analyzed using higher fidelity analysis at later design stages.

• 산업공학
• /
• 제25권2호
• /
• pp.178-186
• /
• 2012

The DRAM module is an important part of servers, workstations and personal computer. Its malfunction causes a lot of damage on customer system. Therefore, customers demand the highest quality products. The company applies DRAM module Outgoing Quality Assurance Inspection(OQA) to secures the highest quality. It is the key process to decides shipment of products through sample inspection method with customer oriented tests. High fraction of defectives entering to OQA causes inevitable high quality cost. This article proposes the application of ensemble learning to classify the lot status to minimize the ratio of wrong decision in OQA, observing a potential in reducing the wrong decision.

• The Plant Pathology Journal
• /
• 제39권1호
• /
• pp.141-148
• /
• 2023

Black shoot blight disease caused by Erwinia pyrifoliae has serious impacts on quality and yield in pear production in Korea; therefore, rapid and accurate methods for its detection are needed. However, traditional detection methods require a great deal of time and fail to achieve absolute quantification. In the present study, we developed a droplet digital polymerase chain reaction (ddPCR) method for the detection and absolute quantification of E. pyrifoliae using a pair of species-specific primers. The detection range was 10³-10⁷ copies/ml (DNA templates) and cfu/ml (cell culture templates). This new method exhibited good linearity and repeatability and was validated by absolute quantification of E. pyrifoliae DNA copies from samples of artificially inoculated immature pear fruits. Here, we present the first study of ddPCR assay for the detection and quantification of E. pyrifoliae. This method has potential applications in epidemiology and for the early prediction of black shoot blight outbreaks.

• 대한통합의학회지
• /
• 제10권3호
• /
• pp.107-117
• /
• 2022

Purpose : The purpose of this study was to establish a model of the predictive factors for success or failure of examinees undertaking the Korean physical therapist licensing examination (KPTLE). Additionally, we assessed the pass/fail cut-off point. Methods : We analyzed the results of 10,881 examinees who undertook the KPTLE, using data provided by the Korea Health Personnel Licensing Examination Institute. The target variable was the test result (pass or fail), and the input variables were: sex, age, test subject, and total score. Frequency analysis, chi-square test, descriptive statistics, independent t-test, correlation analysis, binary logistic regression, and receiver operating characteristic (ROC) curve analyses were performed on the data. Results : Sex and age were not significant predictors of attaining a pass (p>.05). The test subjects with the highest probability of passing were, in order, medical regulation (MR) (Odds ratio (OR)=2.91, p<.001), foundations of physical therapy (FPT) (OR=2.86, p<.001), diagnosis and evaluation for physical therapy (DEPT) (OR=2.74, p<.001), physical therapy intervention (PTI) (OR=2.66, p<.001), and practical examination (PE) (OR=1.24, p<.001). The cut-off points for each subject were: FPT, 32.50; DEPT, 29.50; PTI, 44.50; MR, 14.50; and PE, 50.50. The total score (TS) was 164.50. The sensitivity, specificity, and the classification accuracy of the prediction model was 99 %, 98 %, and 99 %, respectively, indicating high accuracy. Area under the curve (AUC) values for each subject were: FPT, .958; DEPT, .968; PTI, .984; MR, .885; PE, .962; and TS, .998, indicating a high degree of fit. Conclusion : In our study, the predictive factors for passing KPTLE were identified, and the optimal cut-off point was calculated for each subject. Logistic regression was adequate to explain the predictive model. These results will provide universities and examinees with useful information for predicting their success or failure in the KPTLE.