• Journal of the Society of Disaster Information
• /
• v.15 no.3
• /
• pp.440-449
• /
• 2019

Purpose: Fire-fighting organizations are the very first agencies that take actions at a disaster scene, and emergency rescue training is carried out for prompt and systematic response. However, there is a need for a change due to the limitations in emergency rescue trainings such as perfunctory trainings or trainings without considering regional or environmental characteristics. Method: This study is to conduct theoretical review with regard to emergency rescue training and present a measure to improve the emergency rescue training through attitude survey targeting fire-fighting organizations in Gangwon area. Result: Facilities that cause difficulties when doing emergency rescue activity were mostly hazardous material storage and processing facilities. In terms of the level of emergency rescue and response task, most respondents answered that the emergency rescue was insufficient. The respondents answered that the effectiveness of emergency rescue training was helpful, but some responses showed that the training was not helpful because of scenario-based training, seeming training, similar training carried out every year, unrealistic training, and lack of competent authorities' interest and perfunctory participations. Most respondents answered for the appropriateness of emergency rescue training and evaluation that they were satisfied, however, they were not satisfied with the evaluation methods irrelevant to the type of training, evaluation methods requiring unnecessary training scale, and evaluation methods leading perfunctory participations of competent authorities. Lastly, respondents mostly answered that training reflecting various damage situations are necessary regarding the demand on the improvement of emergency rescue training. Conclusion: The improvement measures for emergency rescue training are as follows. First, it is necessary to set and prepare various training contents in accordance with regional characteristics by reviewing major disasters occurred in the region. Second, it is necessary to revise the emergency rescue training guidelines and manuals for appropriate training plan for each fire station, provide education and training for working-level staff members, and establish training in a way that types, tactics, and strategies of emergency rescue training could be utilized practically. Third, it is necessary to prepare a scheme that can lead participation and provide incentive or penalty from the planning stage of training in order to increase the participation of supporting and competent authorities when an actual disaster occurs. Fourth, it is necessary to establish support arrangements and cooperative systems by authority through training by fire stations or zones in preparation for disaster situations that may occur simultaneously. Fifth, it is necessary to put emphasis on the training process rather than the result for emergency rescue training and evaluation, pay attention to the identification of supplement points for each disaster situation and make improvements. Especially, type or form of training should be considered rather than evaluating the execution status of detailed processes, and the evaluation measure that can consider the completeness (proficiency) of training and the status of role performance rather than the scale of training should be prepared. Sixth, type and method of training should be improved in accordance with the characteristics of each fire station by identifying the demand of working-level staff members for an efficient emergency rescue training.

• Journal of the Society of Disaster Information
• /
• v.16 no.4
• /
• pp.701-711
• /
• 2020

Purpose: In this study, the risk of flammability of a liquid mixture was experimentally confirmed because the purpose of this study was to confirm the increase or decrease of the flammability risk in a mixture of two substances (combustible+combustible) and to present the risk of the mixture. Method: Flash point test method and result processing were tested based on KS M 2010-2008, a tag sealing test method used as a flash point test method for crude oil and petroleum products. The manufacturer of the equipment used in this experiment was Japan's TANAKA. The flash point was measured with a test equipment that satisfies the test standards of KS M 2010 with equipment produced by the company, and LP gas was used as the ignition source and water as the cooling water. In addition, when measuring the flash point, the temperature of the cooling water was tested using cooling water of about 2℃. Results: First of all, in the case of flammable + combustible mixtures, there was little change in flash point if the flash point difference between the two substances was not large, and if the flash point difference between the two substances was low, the flash point tended to increase as the number of substances with high flash point increased. However, in the case of toluene and methanol, the flash point of the mixture was lower than that of the material with a lower flash point. Also, in the case of a paint thinner, it was not easy to predict the flash point of the material because it was composed of a mixture, but as a result of experimental measurement, it was measured between -24℃ and 7℃. Conclusion: The results of this study are to determine the risk of mixtures through experimental studies on flammable mixtures for the purpose of securing the effectiveness of the details of the criteria for determining dangerous goods in the existing dangerous goods safety management method and securing the reliability and reproducibility of the determination of dangerous goods Criteria have been presented, and reference data on experimental criteria for flammable liquids that are regulated in firefighting sites can be provided. In addition, if this study accumulates know-how on differences in test methods, it is expected that it can be used as a basis for research on risk assessment of dangerous goods and as a basis for research on dangerous goods determination.

• Journal of Intelligence and Information Systems
• /
• v.27 no.1
• /
• pp.177-190
• /
• 2021

Due to the development of the fourth industrial revolution technology, efforts are being made to improve areas that humans cannot handle by utilizing artificial intelligence techniques such as machine learning. Although on-demand production companies also want to reduce corporate risks such as delays in delivery by predicting total production time for orders, they are having difficulty predicting this because the total production time is all different for each order. The Theory of Constraints (TOC) theory was developed to find the least efficient areas to increase order throughput and reduce order total cost, but failed to provide a forecast of total production time. Order production varies from order to order due to various customer needs, so the total production time of individual orders can be measured postmortem, but it is difficult to predict in advance. The total measured production time of existing orders is also different, which has limitations that cannot be used as standard time. As a result, experienced managers rely on persimmons rather than on the use of the system, while inexperienced managers use simple management indicators (e.g., 60 days total production time for raw materials, 90 days total production time for steel plates, etc.). Too fast work instructions based on imperfections or indicators cause congestion, which leads to productivity degradation, and too late leads to increased production costs or failure to meet delivery dates due to emergency processing. Failure to meet the deadline will result in compensation for delayed compensation or adversely affect business and collection sectors. In this study, to address these problems, an entity that operates an order production system seeks to find a machine learning model that estimates the total production time of new orders. It uses orders, production, and process performance for materials used for machine learning. We compared and analyzed OLS, GLM Gamma, Extra Trees, and Random Forest algorithms as the best algorithms for estimating total production time and present the results.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.33 no.6
• /
• pp.265-274
• /
• 2021

Outlier detection research in ocean data has traditionally been performed using statistical and distance-based machine learning algorithms. Recently, AI-based methods have received a lot of attention and so-called supervised learning methods that require classification information for data are mainly used. This supervised learning method requires a lot of time and costs because classification information (label) must be manually designated for all data required for learning. In this study, an autoencoder based on unsupervised learning was applied as an outlier detection to overcome this problem. For the experiment, two experiments were designed: one is univariate learning, in which only SST data was used among the observation data of Deokjeok Island and the other is multivariate learning, in which SST, air temperature, wind direction, wind speed, air pressure, and humidity were used. Period of data is 25 years from 1996 to 2020, and a pre-processing considering the characteristics of ocean data was applied to the data. An outlier detection of actual SST data was tried with a learned univariate and multivariate autoencoder. We tried to detect outliers in real SST data using trained univariate and multivariate autoencoders. To compare model performance, various outlier detection methods were applied to synthetic data with artificially inserted errors. As a result of quantitatively evaluating the performance of these methods, the multivariate/univariate accuracy was about 96%/91%, respectively, indicating that the multivariate autoencoder had better outlier detection performance. Outlier detection using an unsupervised learning-based autoencoder is expected to be used in various ways in that it can reduce subjective classification errors and cost and time required for data labeling.

• Journal of Intelligence and Information Systems
• /
• v.28 no.2
• /
• pp.191-206
• /
• 2022

Recently, it is a de-facto approach to utilize a pre-trained language model(PLM) to achieve the state-of-the-art performance for various natural language tasks(called downstream tasks) such as sentiment analysis and question answering. However, similar to any other machine learning method, PLM tends to depend on the data distribution seen during the training phase and shows worse performance on the unseen (Out-of-Distribution) domain. Due to the aforementioned reason, there have been many efforts to develop domain-specified PLM for various fields such as medical and legal industries. In this paper, we discuss the training of a finance domain-specified PLM for the Korean language and its applications. Our finance domain-specified PLM, KB-BERT, is trained on a carefully curated financial corpus that includes domain-specific documents such as financial reports. We provide extensive performance evaluation results on three natural language tasks, topic classification, sentiment analysis, and question answering. Compared to the state-of-the-art Korean PLM models such as KoELECTRA and KLUE-RoBERTa, KB-BERT shows comparable performance on general datasets based on common corpora like Wikipedia and news articles. Moreover, KB-BERT outperforms compared models on finance domain datasets that require finance-specific knowledge to solve given problems.

• Journal of The Korean Society of Grassland and Forage Science
• /
• v.41 no.4
• /
• pp.287-294
• /
• 2021

The objective of this study was conducted to calculate the damage of whole crop maize in accordance with abnormal climate using the forage yield prediction model through machine learning. The forage yield prediction model was developed through 8 machine learning by processing after collecting whole crop maize and climate data, and the experimental area was selected as Gyeonggi-do. The forage yield prediction model was developed using the DeepCrossing (R²=0.5442, RMSE=0.1769) technique of the highest accuracy among machine learning techniques. The damage was calculated as the difference between the predicted dry matter yield of normal and abnormal climate. In normal climate, the predicted dry matter yield varies depending on the region, it was found in the range of 15,003~17,517 kg/ha. In abnormal temperature, precipitation, and wind speed, the predicted dry matter yield differed according to region and abnormal climate level, and ranged from 14,947 to 17,571, 14,986 to 17,525, and 14,920 to 17,557 kg/ha, respectively. In abnormal temperature, precipitation, and wind speed, the damage was in the range of -68 to 89 kg/ha, -17 to 17 kg/ha, and -112 to 121 kg/ha, respectively, which could not be judged as damage. In order to accurately calculate the damage of whole crop maize need to increase the number of abnormal climate data used in the forage yield prediction model.

• Korean Journal of Remote Sensing
• /
• v.37 no.3
• /
• pp.449-461
• /
• 2021

Analysis Ready Data (ARD) for optical satellite images represents a pre-processed product by applying spectral characteristics and viewing parameters for each sensor. The atmospheric correction is one of the fundamental and complicated topics, which helps to produce Top-of-Atmosphere (TOA) and Top-of-Canopy (TOC) reflectance from multi-spectral image sets. Most remote sensing software provides algorithms or processing schemes dedicated to those corrections of the Landsat-8 OLI sensors. Furthermore, Google Earth Engine (GEE), provides direct access to Landsat reflectance products, USGS-based ARD (USGS-ARD), on the cloud environment. We implemented the Orfeo ToolBox (OTB) atmospheric correction extension, an open-source remote sensing software for manipulating and analyzing high-resolution satellite images. This is the first tool because OTB has not provided calibration modules for any Landsat sensors. Using this extension software, we conducted the absolute atmospheric correction on the Landsat-8 OLI images of Railroad Valley, United States (RVUS) to validate their reflectance products using reflectance data sets of RVUS in the RadCalNet portal. The results showed that the reflectance products using the OTB extension for Landsat revealed a difference by less than 5% compared to RadCalNet RVUS data. In addition, we performed a comparative analysis with reflectance products obtained from other open-source tools such as a QGIS semi-automatic classification plugin and SAGA, besides USGS-ARD products. The reflectance products by the OTB extension showed a high consistency to those of USGS-ARD within the acceptable level in the measurement data range of the RadCalNet RVUS, compared to those of the other two open-source tools. In this study, the verification of the atmospheric calibration processor in OTB extension was carried out, and it proved the application possibility for other satellite sensors in the Compact Advanced Satellite (CAS)-500 or new optical satellites.

• Journal of Intelligence and Information Systems
• /
• v.28 no.2
• /
• pp.127-146
• /
• 2022

Recently, as word embedding has shown excellent performance in various tasks of deep learning-based natural language processing, researches on the advancement and application of word, sentence, and document embedding are being actively conducted. Among them, cross-language transfer, which enables semantic exchange between different languages, is growing simultaneously with the development of embedding models. Academia's interests in vector alignment are growing with the expectation that it can be applied to various embedding-based analysis. In particular, vector alignment is expected to be applied to mapping between specialized domains and generalized domains. In other words, it is expected that it will be possible to map the vocabulary of specialized fields such as R&D, medicine, and law into the space of the pre-trained language model learned with huge volume of general-purpose documents, or provide a clue for mapping vocabulary between mutually different specialized fields. However, since linear-based vector alignment which has been mainly studied in academia basically assumes statistical linearity, it tends to simplify the vector space. This essentially assumes that different types of vector spaces are geometrically similar, which yields a limitation that it causes inevitable distortion in the alignment process. To overcome this limitation, we propose a deep learning-based vector alignment methodology that effectively learns the nonlinearity of data. The proposed methodology consists of sequential learning of a skip-connected autoencoder and a regression model to align the specialized word embedding expressed in each space to the general embedding space. Finally, through the inference of the two trained models, the specialized vocabulary can be aligned in the general space. To verify the performance of the proposed methodology, an experiment was performed on a total of 77,578 documents in the field of 'health care' among national R&D tasks performed from 2011 to 2020. As a result, it was confirmed that the proposed methodology showed superior performance in terms of cosine similarity compared to the existing linear vector alignment.

• Asia Marketing Journal
• /
• v.15 no.4
• /
• pp.103-129
• /
• 2014

Differentiated positioning becomes increasingly difficult when brand salience weakens. Also, the daily increase in new media use and information load has led to a social climate that regards advertising stimuli as spamming. For these reasons, the focus of advertisement-related communication is shifting from persuading consumers through the direct delivery of information to an emphasis on appealing to their emotions using matching stimuli to enhance persuasion effects. Recently, both academia and industry have increasingly shown an interest in storytelling methods that can generate positive emotional responses and attitude changes by arousing consumers' narrative processing. The purpose of storytelling is to elicit consumers' emotional experience to meet the objectives of advertisement producers. Therefore, the most important requirement for storytelling in advertising is that it evokes consumers' sympathy for the main character in the advertisement. This does not involve advertisements directly persuading consumers, but rather, consumers themselves finding an answer through the advertisement's story. Thus, consumers have an indirect experience regarding the product features and usage through empathy with the advertisement's main character. In this study, we took the results of a precedent study as the starting point, according to which consumers' emotional response can be altered depending on the storytelling methods adopted for storytelling ads. Previous studies have reported that drama-type and vignette-type storytelling methods have a considerably different impact on the emotional responses of advertising audiences, due to their different structural characteristics. Thus, this study aims to verify that emotional response aroused by different types of advertisement storytelling (drama ads vs. vignette ads) can be controlled by the socio-psychological gender difference of advertising audiences and that the interaction effects between the socio-psychological gender differences of the audience and the gender stereotype of emotions to which advertisements appeal can exert an influence on emotional responses to types of storytelling in advertising. To achieve this, an experiment was conducted employing a between-group design consisting of 2 (storytelling type: drama ads vs. vignette ads) × 2 (socio-psychological gender of the audience: masculinity vs. femininity) × 2 (advertising appeal emotion type: male stereotype emotion vs. female stereotype emotion). The experiment revealed that the femininity group displayed a strong and consistent empathy for drama ads regardless of whether the ads appealed to masculine or feminine emotions, whereas the masculinity group displayed a stronger empathy for drama ads appealing to the emotional types matching its own gender as well as for vignette ads. The theoretical contribution of this study is significant in that it sheds light on the controllability of the audiences' emotional responses to advertisement storytelling depending on their socio-psychological gender and gender stereotype of emotions appealed to through advertising. Specifically, its considerable practical contribution consists in easing unnecessary creative constraints by comprehensively analyzing essential advertising strategic factors such as the target consumers' gender and the objective of the advertisement, in contrast to the oversimplified view of previous studies that considered emotional responses to storytelling ads were determined by the different types of production techniques used. This study revealed that emotional response to advertisement storytelling varies depending on the target gender of and emotion type appealed to by the advertisement. This suggests that an understanding of the targeted gender is necessary prior to producing an advertisement and that in deciding on an advertisement storytelling type, strategic attention should be directed to the advertisement's appeal concept or emotion type. Thus, it is safe to use drama-type storytelling that expresses masculine emotions (ex. fun, happy, encouraged) when the advertisement target, like Bacchus, includes both men and women. For brands and advertisements targeting only women (ex. female clothes), it is more effective to use a drama-type storytelling method that expresses feminine emotions (lovely, romantic, sad). The drama method can be still more effective than the vignette when women are the main target and a masculine concept-based creative is to be produced. However, when male consumers are targeted and the brand concept or advertisement concept is focused on feminine emotions (ex. romantic), vignette ads can more effectively induce empathy than drama ads.

• The Journal of the Acoustical Society of Korea
• /
• v.43 no.1
• /
• pp.9-18
• /
• 2024

The importance of active sonar systems is emerging due to the quietness of underwater targets and the increase in ambient noise due to the increase in maritime traffic. However, the low signal-to-noise ratio of the echo signal due to multipath propagation of the signal, various clutter, ambient noise and reverberation makes it difficult to identify underwater targets using active sonar. Attempts have been made to apply data-based methods such as machine learning or deep learning to improve the performance of underwater target recognition systems, but it is difficult to collect enough data for training due to the nature of sonar datasets. Methods based on mathematical modeling have been mainly used to compensate for insufficient active sonar data. However, methodologies based on mathematical modeling have limitations in accurately simulating complex underwater phenomena. Therefore, in this paper, we propose a sonar signal synthesis method based on a deep neural network. In order to apply the neural network model to the field of sonar signal synthesis, the proposed method appropriately corrects the attention-based encoder and decoder to the sonar signal, which is the main module of the Tacotron model mainly used in the field of speech synthesis. It is possible to synthesize a signal more similar to the actual signal by training the proposed model using the dataset collected by arranging a simulated target in an actual marine environment. In order to verify the performance of the proposed method, Perceptual evaluation of audio quality test was conducted and within score difference -2.3 was shown compared to actual signal in a total of four different environments. These results prove that the active sonar signal generated by the proposed method approximates the actual signal.