• Smart Media Journal
• /
• v.7 no.4
• /
• pp.17-23
• /
• 2018

We proposed a LSTM-based RNN model that can effectively perform the automatic spacing characteristics. For those long or noisy sentences which are known to be difficult to handle within Neural Network Learning, we defined a proper input data format and decoding data format, and added dropout, bidirectional multi-layer LSTM, layer normalization, and attention mechanism to improve the performance. Despite of the fact that Sejong corpus contains some spacing errors, a noise-robust learning model developed in this study with no overfitting through a dropout method helped training and returned meaningful results of Korean word spacing and its patterns. The experimental results showed that the performance of LSTM sequence-to-sequence model is 0.94 in F1-measure, which is better than the rule-based deep-learning method of GRU-CRF.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.4
• /
• pp.685-692
• /
• 2024

In this paper, we propose a new CNN model to improve the recognition performance of command voice signals. This method obtains a spectrogram image after performing a short-time Fourier transform (STFT) of the input signal and improves command recognition performance through supervised learning using a CNN model. After Fourier transforming the input signal for each short-time section, a spectrogram image is obtained and multi-classification learning is performed using a CNN deep learning model. This effectively classifies commands by converting the time domain voice signal to the frequency domain to express the characteristics well and performing deep learning training using the spectrogram image for the conversion parameters. To verify the performance of the speech recognition system proposed in this study, a simulation program using Tensorflow and Keras libraries was created and a simulation experiment was performed. As a result of the experiment, it was confirmed that an accuracy of 92.5% could be obtained using the proposed deep learning algorithm.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 2017.04a
• /
• pp.135-135
• /
• 2017

Greenhouse have been developed to provide the plants with good environmental conditions for cultivation crop, two major factors of which are the inside air temperature and humidity. The inside temperature are influenced by the heating systems, ventilators and for systems among others, which in turn are geverned by some type of controller. Likewise, humidity environment is the result of complex mass exchanges between the inside air and the several elements of the greenhouse and the outside boundaries. Most of the existing models are based on the energy balance method and heat balance equation for modelling the heat and mass fluxes and generating dynamic elements. However, greenhouse are classified as complex system, and need to make a sophisticated modeling. Furthermore, there is a difficulty in using classical control methods for complex process system due to the process are non linear and multi-output(MIMO) systems. In order to predict the time evolution of conditions in certain greenhouse as a function, we present here to use of recurrent neural networks(RNN) which has been used to implement the direct dynamics of the inside temperature and inside humidity of greenhouse. For the training, we used algorithm of a backpropagation Through Time (BPTT). Because the environmental parameters are shared by all time steps in the network, the gradient at each output depends not only on the calculations of the current time step, but also the previous time steps. The training data was emulated to 13 input variables during March 1 to 7, and the model was tested with database file of March 8. The RMSE of results of the temperature modeling was $0.976^{\circ}C$, and the RMSE of humidity simulation was 4.11%, which will be given to prove the performance of RNN in prediction of the greenhouse environment.

• Journal of the Society of Naval Architects of Korea
• /
• v.51 no.4
• /
• pp.274-282
• /
• 2014

Time series of the dynamic response of a slender marine structure was predicted using quadratic Volterra series. The wave-structure interaction system was identified using the NARX(Nonlinear Autoregressive with Exogenous Input) technique, and the network parameters were determined through the supervised training with the prepared datasets. The dataset used for the network training was obtained by carrying out the nonlinear finite element analysis on the freely standing riser under random ocean waves of white noise. The nonlinearities involved in the analysis were both large deformation of the structure under consideration and the quadratic term of relative velocity between the water particle and structure in Morison formula. The linear and quadratic frequency response functions of the given system were extracted using the multi-tone harmonic probing method and the time series of response of the structure was predicted using the quadratic Volterra series. In order to check the applicability of the method, the response of structure under the realistic ocean wave environment with given significant wave height and modal period was predicted and compared with the nonlinear time domain simulation results. It turned out that the predicted time series of the response of structure with quadratic Volterra series successfully captures the slowly varying response with reasonably good accuracy. It is expected that the method can be used in predicting the response of the slender offshore structure exposed to the Morison type load without relying on the computationally expensive time domain analysis, especially for the screening purpose.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.1
• /
• pp.144-150
• /
• 2017

Recently, deep learning is used for intelligent processing and accuracy improvement of data. It is formed calculation model composed of multi data processing layer that train the data representation through an abstraction of the various levels. A category of deep learning, convolution neural network is utilized in various research fields, which are human pose estimation, face recognition, image classification, speech recognition. When using the deep layer and lots of class, CNN that show a good performance on image classification obtain higher classification rate but occur the overfitting problem, when using a few data. So, we design the training network based on convolution neural network and trained our image data set for object classification in few class problem. The experiment show the higher classification rate of 7.06% in average than the previous networks designed to classify the object in 1000 class problem.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.29 no.3
• /
• pp.531-539
• /
• 2019

This paper proposes the training features for malware family analysis and analyzes the multi-classification performance of ensemble models. We construct training data by extracting API and DLL information from malware executables and use Random Forest and XGBoost algorithms which are based on decision tree. API, API-DLL, and DLL-CM features for malware detection and family classification are proposed by analyzing frequently used API and DLL information from malware and converting high-dimensional features to low-dimensional features. The proposed feature selection method provides the advantages of data dimension reduction and fast learning. In performance comparison, the malware detection rate is 93.0% for Random Forest, the accuracy of malware family dataset is 92.0% for XGBoost, and the false positive rate of malware family dataset including benign is about 3.5% for Random Forest and XGBoost.

• Korean Journal of Labor Studies
• /
• v.23 no.2
• /
• pp.141-167
• /
• 2017

This study carries out an empirical analysis of how workplace innovation affects employment growth. The theoretical model and hypotheses of this study are drawn from the previous research on the employment effects of innovation. I use the data on Workplace Innovation Indicators(2013-14) collected by Labor Foundation. As the regression models for this study, I adopt OLS models whose dependent variable is employment growth rate, and whose main independent variables are the adoption and the intensity(standardized values) of three innovative work practices such as TQM, employee suggestion plans and multi-skill training programs. The results of this study indicate that the adoption of workplace innovation does not have significant effects on employment growth, but that the intensity of workplace innovation has weakly positive effects on employment growth. Besides, the results of this demonstrate show that government-subsided organizational innovation consulting and training hours per capita have positive employment effects, but that wage level and prior employment size have negative ones. Finally the empirical results are outlined, and their limitations and the future direction of research on this topic are discussed.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.181-193
• /
• 2022

Data-driven structural health monitoring (SHM) of civil infrastructure can be used to continuously assess the state of a structure, allowing preemptive safety measures to be carried out. Long-term monitoring of large-scale civil infrastructure often involves data-collection using a network of numerous sensors of various types. Malfunctioning sensors in the network are common, which can disrupt the condition assessment and even lead to false-negative indications of damage. The overwhelming size of the data collected renders manual approaches to ensure data quality intractable. The task of detecting and classifying an anomaly in the raw data is non-trivial. We propose an approach to automate this task, improving upon the previously developed technique of image-based pre-processing on one-dimensional (1D) data by enriching the features of the neural network input data with multiple channels. In particular, feature engineering is employed to convert the measured time histories into a 3-channel image comprised of (i) the time history, (ii) the spectrogram, and (iii) the probability density function representation of the signal. To demonstrate this approach, a CNN model is designed and trained on a dataset consisting of acceleration records of sensors installed on a long-span bridge, with the goal of fault detection and classification. The effect of imbalance in anomaly patterns observed is studied to better account for unseen test cases. The proposed framework achieves high overall accuracy and recall even when tested on an unseen dataset that is much larger than the samples used for training, offering a viable solution for implementation on full-scale structures where limited labeled-training data is available.

• Physical Therapy Korea
• /
• v.31 no.2
• /
• pp.174-181
• /
• 2024

Background: Effective management of clinical assessment tools is critical in stroke and brain injury rehabilitation research. Managing rehabilitation outcome measures (ROMs) scores and training therapists in multicenter randomized clinical trials (RCTs) is challenging. Objects: The aim of this study was to develop a web-based platform, the Korean Rehabilitation Outcome Measurement (KoROM), to address these limitations and improve both therapist training and patient involvement in the rehabilitation process. Methods: The development of the KoROM spanned from June 2021 to July 2022, and included literature and web-based searches to identify relevant ROMs and design a user-friendly platform. Feedback from six physical therapy and informatics experts during pilot testing refined the platform. Results: Several clinical assessment tools categorized under the International Classification of Functioning, Disability, and Health (ICF) model are categorized in the KoROM. The therapist version includes patient management, assessment tool information, and data downloads, while the patient version provides a simplified interface for viewing scores and printing summaries. The master version provides full access to user information and clinical assessment scores. Therapists enter clinical assessment scores into the KoROM and learn ROMs through instructional videos and self-checklists as part of the therapist standardization process. Conclusion: The KoROM is a specialized online platform that improves the management of ROMs, facilitates therapist education, and promotes patient involvement in the rehabilitation process. The KoROM can be used not only in multi-site RCTs, but also in community rehabilitation exercise centers.

• Computers and Concrete
• /
• v.34 no.2
• /
• pp.151-168
• /
• 2024

This research aimed to appraise the effectiveness of four optimization approaches - cuckoo optimization algorithm (COA), multi-verse optimization (MVO), particle swarm optimization (PSO), and teaching-learning-based optimization (TLBO) - that were enhanced with an artificial neural network (ANN) in predicting the bearing capacity of shallow foundations located on cohesionless soils. The study utilized a database of 97 laboratory experiments, with 68 experiments for training data sets and 29 for testing data sets. The ANN algorithms were optimized by adjusting various variables, such as population size and number of neurons in each hidden layer, through trial-and-error techniques. Input parameters used for analysis included width, depth, geometry, unit weight, and angle of shearing resistance. After performing sensitivity analysis, it was determined that the optimized architecture for the ANN structure was 5×5×1. The study found that all four models demonstrated exceptional prediction performance: COA-MLP, MVO-MLP, PSO-MLP, and TLBO-MLP. It is worth noting that the MVO-MLP model exhibited superior accuracy in generating network outputs for predicting measured values compared to the other models. The training data sets showed R² and RMSE values of (0.07184 and 0.9819), (0.04536 and 0.9928), (0.09194 and 0.9702), and (0.04714 and 0.9923) for COA-MLP, MVO-MLP, PSO-MLP, and TLBO-MLP methods respectively. Similarly, the testing data sets produced R² and RMSE values of (0.08126 and 0.07218), (0.07218 and 0.9814), (0.10827 and 0.95764), and (0.09886 and 0.96481) for COA-MLP, MVO-MLP, PSO-MLP, and TLBO-MLP methods respectively.