• Proceedings of the Earthquake Engineering Society of Korea Conference
• /
• 2003.09a
• /
• pp.13-19
• /
• 2003

Discriminant method using seismic signal was studied for discrimination of surface explosion. By means of the seismic spectral characteristics, multi-variate discriminant analysis was performed. Four single discriminant techniques - Pg/Lg, Lg1/Lg2, Pg1/Pg2, and Rg/Lg - based on seismic source theory were applied to explosion and earthquake training data sets. The Pg/Lg discriminant technique was most effective among the four techniques. Nevertheless, it could not perfectly discriminate the samples of the training data sets. In this study, a compound linear discriminant analysis was defined by using common characteristics of the training data sets for the single discriminants. The compound linear discriminant analysis was used for the single discriminant as an independent variable. From this analysis, all the samples of the training data sets were correctly discriminated, and the probability of misclassification was lowered to 0.7％.

• Bulletin of the Korean Chemical Society
• /
• v.32 no.7
• /
• pp.2397-2404
• /
• 2011

Variable selection k nearest neighbor QSAR modeling approach was applied to a data set of 80 3-arylisoquinolines exhibiting cytotoxicity against human lung tumor cell line (A-549). All compounds were characterized with molecular topology descriptors calculated with the MolconnZ program. Seven compounds were randomly selected from the original dataset and used as an external validation set. The remaining subset of 73 compounds was divided into multiple training (56 to 61 compounds) and test (17 to 12 compounds) sets using a chemical diversity sampling method developed in this group. Highly predictive models characterized by the leave-one out cross-validated $R^2$ ($q^2$) values greater than 0.8 for the training sets and $R^2$ values greater than 0.7 for the test sets have been obtained. The robustness of models was confirmed by the Y-randomization test: all models built using training sets with randomly shuffled activities were characterized by low $q^2{\leq}0.26$ and $R^2{\leq}0.22$ for training and test sets, respectively. Twelve best models (with the highest values of both $q^2$ and $R^2$) predicted the activities of the external validation set of seven compounds with $R^2$ ranging from 0.71 to 0.93.

• International Journal of Internet, Broadcasting and Communication
• /
• v.14 no.4
• /
• pp.222-227
• /
• 2022

The purpose of this study was to investigate changes in heart rate according to recovery methods after circuit weight training exercise. Fourteen men in their twenties were selected as subjects, and three sets of circuit weight training were performed by cycling six sports, and two recovery conditions (dynamic and static) were performed immediately after exercise. Changes in heart rate did not have an interactive effect according to recovery method and time, and both conditions showed significant changes between sets 1 and 2, and between sets 3 and after recovery. In this study, the high heart rate of 2 sets and 3 sets was seen as a result of exercise stimulation, and the low heart rate of 1 set was thought to be due to the decrease in vagus nerve activity rather than the role of catecholamines. On the other hand, the heart rate after 20 minutes of exercise did not show any difference according to the recovery method, which could mean that the recovery process due to the aquatic environment can act more strongly than the process of dynamic recovery and static recovery. It is thought that the characteristics affected the sensory and circulation of the body, and thus the change of the afferent signal and the level of metabolic products generated in the active muscle.

• The Journal of Korean Physical Therapy
• /
• v.27 no.6
• /
• pp.407-412
• /
• 2015

Purpose: This study aimed to evaluate changes in the balance ability of patients whose head positions were altered due to stroke. Subjects were divided into three groups to determine the effects of the training on dynamic balance and gait. Methods: Forty-two stroke patients were enrolled. The Visual Feedback Training (VFT) group performed four sets of exercises per training session using a Sensoneck device, while the Active Range of Motion (ART) group performed eight sets per training session after receiving education from an experienced therapist. The Visual Feedback with Active Range of Motion (VAT) group performed four sets of active range of motion and two sets of visual-feedback training per session using a Sensoneck device. The training sessions were conducted three days a week for eight weeks. Results: The comparison of changes in dynamic balance ability showed that a significant difference in the total distance of the body center was found in the VFT group (p<0.05) and Significant differences were found according to the training period (p<0.05). The comparison of the 10 m walk test showed that the main effect test, treatment period and interactions between group had statistically significant differences between the three groups (p<0.05). Conclusion: Head-adjustment training using visual feedback can improve the balance ability and gait of stroke patients. These results show that coordination training between the eyes and head with visual feedback exercises can be used as a treatment approach to affect postural control through various activities involving the central nervous system.

• ETRI Journal
• /
• v.44 no.2
• /
• pp.327-338
• /
• 2022

Classification models for human action recognition require robust features and large training sets for good generalization. However, data augmentation methods are employed for imbalanced training sets to achieve higher accuracy. These samples generated using data augmentation only reflect existing samples within the training set, their feature representations are less diverse and hence, contribute to less precise classification. This paper presents new data augmentation and action representation approaches to grow training sets. The proposed approach is based on two fundamental concepts: virtual video generation for augmentation and representation of the action videos through robust features. Virtual videos are generated from the motion history templates of action videos, which are convolved using a convolutional neural network, to generate deep features. Furthermore, by observing an objective function of the genetic algorithm, the spatiotemporal features of different samples are combined, to generate the representations of the virtual videos and then classified through an extreme learning machine classifier on MuHAVi-Uncut, iXMAS, and IAVID-1 datasets.

• 한국체육학회지인문사회과학편
• /
• v.51 no.5
• /
• pp.639-647
• /
• 2012

The purpose of this study was to examine the effects of varied resistance training intensities and rest intervals between training sets on integral electromyography (iEMG), repetition rate, and total work. All subjects, 14 college students, were tested one repetition maximum (1RM). Then, all subjects were weekly tested with 9 practice procedures, composed of diverse intensities (60, 75, 90% of 1RM) and rest intervals (1, 3, 5 min). As results show, to maintain the same load and target repetition maximum for an untrained person, muscular power training (90% of 1RM), muscular hypertrophy training (75% of 1RM), and muscular endurance training (60% of 1RM) should be applied with 5 min or longer rest interval periods for 3 training sets. In addition, 2 training sets with 3 min rest intervals and a set with an 1 min rest interval were capable by the subjects. Thus, at least 3 min or longer rest intervals should be applied to maintain multiple training sets. In case for muscular endurance training, which requires shorter rest intervals, the intensity of exercise should be adjusted to 60% of 1RM or less. In conclusion, depending on diverse purposes of resistance training such as improving muscular power, muscular hypertrophy, or muscular endurance, appropriate exercise intensity and rest intervals should be applied.

• Computers and Concrete
• /
• v.17 no.4
• /
• pp.489-498
• /
• 2016

In this paper, the flexural strength ($f_{fs}$) and splitting tensile strength ($f_{sts}$) of concrete containing different proportions of fly ash have been modeled by using gene expression programming (GEP). Two GEP models called GEP-I and GEP-II are constituted to predict the $f_{fs}$ and $f_{sts}$ values, respectively. In these models, the age of specimen, cement, water, sand, aggregate, superplasticizer and fly ash are used as independent input parameters. GEP-I model is constructed by 292 experimental data and trisected into 170, 86 and 36 data for training, testing and validating sets, respectively. Similarly, GEP-II model is constructed by 278 experimental data and trisected into 142, 70 and 66 data for training, testing and validating sets, respectively. The experimental data used in the validating set of these models are independent from the training and testing sets. The results of the statistical parameters obtained from the models indicate that the proposed empirical models have good prediction and generalization capability.

• Korean Journal of Remote Sensing
• /
• v.24 no.5
• /
• pp.473-481
• /
• 2008

This paper presents an approach to Back-propagation and Radial Basis Function neural network method with various training set for automatic vehicle detection from aerial images. The initial extraction of candidate object is based on Mean-shift algorithm with symmetric property of a vehicle structure. By fusing the density and the symmetry, the method can remove the ambiguous objects and reduce the cost of processing in the next stage. To extract features from the detected object, we describe the object as a log-polar shape histogram using edge strengths of object and represent the orientation and distance from its center. The spatial histogram is used for calculating the momentum of object and compensating the direction of object. BPNN and RBFNN are applied to verify the object as a vehicle using a variety of non-car training sets. The proposed algorithm shows the results which are according to the training data. By comparing the training sets, advantages and disadvantages of them have been discussed.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2019.01a
• /
• pp.123-124
• /
• 2019

This paper proposes an improved deep learning method based on small data sets for animal image classification. Firstly, we use a CNN to build a training model for small data sets, and use data augmentation to expand the data samples of the training set. Secondly, using the pre-trained network on large-scale datasets, such as VGG16, the bottleneck features in the small dataset are extracted and to be stored in two NumPy files as new training datasets and test datasets. Finally, training a fully connected network with the new datasets. In this paper, we use Kaggle famous Dogs vs Cats dataset as the experimental dataset, which is a two-category classification dataset.

• Journal of Radiation Protection and Research
• /
• v.46 no.4
• /
• pp.204-212
• /
• 2021

Background: Identification of radioisotopes for plastic scintillation detectors is challenging because their spectra have poor energy resolutions and lack photo peaks. To overcome this weakness, many researchers have conducted radioisotope identification studies using machine learning algorithms; however, the effect of data normalization on radioisotope identification has not been addressed yet. Furthermore, studies on machine learning-based radioisotope identifiers for plastic scintillation detectors are limited. Materials and Methods: In this study, machine learning-based radioisotope identifiers were implemented, and their performances according to data normalization methods were compared. Eight classes of radioisotopes consisting of combinations of ²²Na, ⁶⁰Co, and ¹³⁷Cs, and the background, were defined. The training set was generated by the random sampling technique based on probabilistic density functions acquired by experiments and simulations, and test set was acquired by experiments. Support vector machine (SVM), artificial neural network (ANN), and convolutional neural network (CNN) were implemented as radioisotope identifiers with six data normalization methods, and trained using the generated training set. Results and Discussion: The implemented identifiers were evaluated by test sets acquired by experiments with and without gain shifts to confirm the robustness of the identifiers against the gain shift effect. Among the three machine learning-based radioisotope identifiers, prediction accuracy followed the order SVM > ANN > CNN, while the training time followed the order SVM > ANN > CNN. Conclusion: The prediction accuracy for the combined test sets was highest with the SVM. The CNN exhibited a minimum variation in prediction accuracy for each class, even though it had the lowest prediction accuracy for the combined test sets among three identifiers. The SVM exhibited the highest prediction accuracy for the combined test sets, and its training time was the shortest among three identifiers.