• The Journal of Korean Society for Radiation Therapy
• /
• v.35
• /
• pp.41-51
• /
• 2023

Purpose: This study compares and analyzes the image quality of 3D-CBCT(Cone Beam Computed-Tomography) and Gated CBCT according to baseline changes during SBRT(Stereotactic Body RadioTherapy) in lung cancer patients to find a useful CBCT method for correcting movement due to breathing Materials and methods : Insert a solid tumor material with a diameter of 3 cm into the QUASAR^TM phantom. 4-Dimentional Computed-Tomography(4DCT) images were taken with a speed of the phantom at period 3 sec and a maximum amplitude of 20 mm. Using the contouring menu of the computerized treatment planning system Eclipse^TM Gross Tumor Volume was outlined on solid tumor material. Set-up the same as when acquiring a 4DCT image using Truebeam STx^TM, breathing patterns with baseline changes of 1 mm, 3 mm, and 5 mm were input into the phantom to obtain 3D-CBCT (Spotlight, Full) and Gated-CBCT (Spotlight, Full) images five times repeatedly. The acquired images were compared with the Signal-to-Noise Ratio(SNR), Contrast-to-Noise Ratio(CNR), Tumor Volume Length, and Motion Blurring Ratio(MBR) based on the 4DCT image. Results: The average Signal-to-Noise Ratio, Contrast-to-Noise Ratio, Tumor Volume Length and Motion Blurring Ratio of Spotlight Gated CBCT images were 13.30±0.10%, 7.78±0.16%, 3.55±0.17%, 1.18±0.06%. As a result, Spotlight Gated-CBCT images according to baseline change showed better values than Spotligtht 3D-CBCT images. Also, the average Signal-to-Noise Ratio, Contrast-to-Noise Ratio, Tumor Volume Length and Motion Blurring Ratio of Full Gated CBCT images were 12.80±0.11%, 7.60±0.11%, 3.54±0.16%, 1.18±0.05%. As a result Full GatedCBCT images according to baseline change showed better values than Full 3D-CBCT images. Conclusion : Compared to 3D-CBCT images, Gated-CBCT images had better image quality according to the baseline change, and the effect of Motion Blurring Artifacts caused by breathing was small. Therefore, it is considered useful to image guided using Gated-CBCT when a baseline change occurs due to difficulty in regular breathing during SBRT that exposes high doses in a short period of time

• Journal of Wetlands Research
• /
• v.26 no.2
• /
• pp.147-159
• /
• 2024

Lack of streamflow observations makes model calibration difficult and limits model performance improvement. Satellite-based remote sensing products offer a new alternative as they can be actively utilized to obtain hydrological data. Recently, several studies have shown that artificial intelligence-based solutions are more appropriate than traditional conceptual and physical models. In this study, a data-driven approach combining various recurrent neural networks and decision tree-based algorithms is proposed, and the utilization of satellite remote sensing information for AI training is investigated. The satellite imagery used in this study is from MODIS and SMAP. The proposed approach is validated using publicly available data from 25 watersheds. Inspired by the traditional regionalization approach, a strategy is adopted to learn one data-driven model by integrating data from all basins, and the potential of the proposed approach is evaluated by using a leave-one-out cross-validation regionalization setting to predict streamflow from different basins with one model. The GRU + Light GBM model was found to be a suitable model combination for target basins and showed good streamflow prediction performance in ungauged basins (The average model efficiency coefficient for predicting daily streamflow in 25 ungauged basins is 0.7187) except for the period when streamflow is very small. The influence of satellite remote sensing information was found to be up to 10%, with the additional application of satellite information having a greater impact on streamflow prediction during low or dry seasons than during wet or normal seasons.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.52 no.1
• /
• pp.17-33
• /
• 2024

This study suggests management planning of Bonguisan Neighborhood Park located on the central of Chuncheon city and highly used for citizen's rest and leisure space utilizing its vegetation structure feature. Bonguisan has been the central of the chuncheon since the period of the Three states in Korean history and consistently damaged, especially in present era, an isolation and sererance of its ecosystem has deepen for indiscreet urban development. The percentage of actual vegetation of Boinguisan Neighborhood Park is as follows: Quercus mongolica is 28.5%, Quercus mongolica - Quercus serrata is 2.1%, Pinus densiflora is 15.6%, Pinus densiflora - Quercus mongolica is 15.9%, Betula schmidtii is 1.6%, Robinia pseudoacacia is 5.9%, Pinus koraiensis is 1.6%. Quercus mongolica is distributed on the southwest, northwest, southeast side of region, Pinus densiflora is distributed on the ridge of east and southeast side of region, Betula schmidtii is distributed on the valley of northeast side region and steep slope region which is on the north side of chungwonsa temple. Pinus densiflora community (Comm. I) and Quercus acutissima - Robinia pseudoacacia community (Comm. V) is expected to undergo succession since it's categorized as Quercus spp. and Quercus mongolica community (Comm. II) and Quercus serrata-Quercus mongolica community (Comm. III), Betula schmidtii community (Comm. IV), Pinus koraiensis community (Comm. VI) is expected to maintain. Also for target vegetation and management planning, Vegetation of Bonguisan Neighborhood Park is classified as 1st Natural landscape conservation and improvement type, 2nd Ecological succession type, 3rd Unusual community conservation type, and 4th Recreation and experience type. And we suggested ecological management measure about each management types. For efficient management of Bonguisan Neighborhood Park, it is need to unify management system of it and after designating Pinus densiflora community and Betula schmidtii community which has high ecological preservation value as an ecological landscape protected area and manage it.

• Health Policy and Management
• /
• v.34 no.2
• /
• pp.141-155
• /
• 2024

Background: This study aimed to identified the relationship between the risk of obstructive sleep apnea, subjective health, and health-related quality of life among the middle-aged and elderly population in Korea. Methods: Adults aged 40 or older were extracted from the total 22,559 respondents to the 2019-2020 Korea National Health and Nutrition Examination Survey VIII, and secondary analysis was conducted on a total of 6,659 middle-aged and elderly people with no missing values. Logistic regression analysis and multiple regression analysis were conducted to examine the relationship between obstructive sleep apnea risk factors and subjective health as well as quality of life. Results: The subjective health status decline in the high-risk group compared to the non-risk group for obstructive sleep apnea was statistically significantly higher, with an odds ratio of 1.84 (p<0.001). The health-related quality of life was also statistically significantly lower by 0.02 points (β, -0.02; p<0.001). As a result of subgroup analysis on specific variables, the association between the risk of obstructive sleep apnea and subjective health and health-related quality of life was statistically significant depending on gender, sleep time, presence of depression, household income, and number of household members. Based on the obstructive sleep apnea risk group, women had a higher correlation with low subjective health and lower health-related quality of life scores than men. Sleeping time of more than 8 hours or less than 6 hours was more associated with low subjective health and lower health-related quality of life score than sleeping time of 6-8 hours. Patients with depression were more likely to have low subjective health than those without depression. The lower the household income level and the smaller the number of household members, the higher the association with low subjective health and the lower the health-related quality of life score. Conclusion: It is essential to recognize that the risk of obstructive sleep apnea not only directly affects sleep disorders but also impacts individuals' subjective health and quality of life. Consequently, social support and education should be provided to raise awareness of this issue. Particularly, programs for preventing and managing obstructive sleep apnea should target vulnerable groups such as women, individuals in single-person households, low household income, and those with depression, aiming to improve their subjective health and quality of life.

• The Korean Journal of Nuclear Medicine Technology
• /
• v.15 no.2
• /
• pp.3-12
• /
• 2011

Purpose: With increasing medical use of radiation and radioactive isotopes, there is a need to better manage the risk of radiation exposure. This study aims to grasp and analyze the individual radiation exposure situations of radiation-related workers in a medical facility by specific job, in order to instill awareness of radiation danger and to assist in safety and radiation exposure management for such workers. Materials and Methods: From January 1, 2010 December 31, 2010, medical practitioners working in the radiation is classified as a regular personal radiation dosimetry, and subsequently one year 540 people managed investigation department to target workers, dose sectional area, working period, identify the job function-related tasks for a deep dose, respectively, the annual average radiation dose were analyzed. Frequency analysis methods include ANOVA was performed. Results: Medical radiation workers in the department an annual radiation dose of Nuclear and 4.57 mSv a was highest, dose zone-specific distribution of nuclear medicine and in the 5.01~19.05 mSv in the high dose area distribution showed departmental radiation four of the annual radiation dose of Nuclear and 7.14 mSv showed the highest radiation dose. More work an average annual radiation dose according to the job function related to the synthesis of Cyclotron to 17.47 mSv work showed the highest radiation dose, Gamma camera Cinema Room 7.24 mSv, PET/CT Cinema Room service is 7.60 mSv, 2.04 mSv in order of intervention high, were analyzed. Working period, according to domain-specific average annual dose of radiation dose from 10 to 14 in oral and maxillofacial radiology practitioners as high as 1.01~3.00 mSv average dose showed the Department of Radiology, 1-4 years, 5-9 years, respectively, 1.01 workers~8.00 mSv in the range of the most high-dose region showed the distribution, nuclear medicine, and the 1-4 years, 5-9 years 3.01~19.05 mSv, respectively, workers of the highest dose showed the distribution of the area in the range of 10 to 14 years, Workers at 15-19 3.01~15.00 mSv, respectively in the range of the high-dose region were distributed. Conclusion: These results suggest that medical radiation workers working in Nuclear Medicine radiation safety management of the majority of the current were carried out in the effectiveness, depending on job characteristics has been found that many differences. However, this requires efforts to minimize radiation exposure, and systematic training for them and for reasonable radiation exposure management system is needed.

• Journal of Intelligence and Information Systems
• /
• v.22 no.1
• /
• pp.107-118
• /
• 2016

The advent of 5G mobile communications, which is expected in 2020, will provide many services such as Internet of Things (IoT) and vehicle-to-infra/vehicle/nomadic (V2X) communication. There are many requirements to realizing these services: reduced latency, high data rate and reliability, and real-time service. In particular, a high level of reliability and delay sensitivity with an increased data rate are very important for M2M, IoT, and Factory 4.0. Around the world, 5G standardization organizations have considered these services and grouped them to finally derive the technical requirements and service scenarios. The first scenario is broadcast services that use a high data rate for multiple cases of sporting events or emergencies. The second scenario is as support for e-Health, car reliability, etc.; the third scenario is related to VR games with delay sensitivity and real-time techniques. Recently, these groups have been forming agreements on the requirements for such scenarios and the target level. Various techniques are being studied to satisfy such requirements and are being discussed in the context of software-defined networking (SDN) as the next-generation network architecture. SDN is being used to standardize ONF and basically refers to a structure that separates signals for the control plane from the packets for the data plane. One of the best examples for low latency and high reliability is an intelligent traffic system (ITS) using V2X. Because a car passes a small cell of the 5G network very rapidly, the messages to be delivered in the event of an emergency have to be transported in a very short time. This is a typical example requiring high delay sensitivity. 5G has to support a high reliability and delay sensitivity requirements for V2X in the field of traffic control. For these reasons, V2X is a major application of critical delay. V2X (vehicle-to-infra/vehicle/nomadic) represents all types of communication methods applicable to road and vehicles. It refers to a connected or networked vehicle. V2X can be divided into three kinds of communications. First is the communication between a vehicle and infrastructure (vehicle-to-infrastructure; V2I). Second is the communication between a vehicle and another vehicle (vehicle-to-vehicle; V2V). Third is the communication between a vehicle and mobile equipment (vehicle-to-nomadic devices; V2N). This will be added in the future in various fields. Because the SDN structure is under consideration as the next-generation network architecture, the SDN architecture is significant. However, the centralized architecture of SDN can be considered as an unfavorable structure for delay-sensitive services because a centralized architecture is needed to communicate with many nodes and provide processing power. Therefore, in the case of emergency V2X communications, delay-related control functions require a tree supporting structure. For such a scenario, the architecture of the network processing the vehicle information is a major variable affecting delay. Because it is difficult to meet the desired level of delay sensitivity with a typical fully centralized SDN structure, research on the optimal size of an SDN for processing information is needed. This study examined the SDN architecture considering the V2X emergency delay requirements of a 5G network in the worst-case scenario and performed a system-level simulation on the speed of the car, radius, and cell tier to derive a range of cells for information transfer in SDN network. In the simulation, because 5G provides a sufficiently high data rate, the information for neighboring vehicle support to the car was assumed to be without errors. Furthermore, the 5G small cell was assumed to have a cell radius of 50-100 m, and the maximum speed of the vehicle was considered to be 30-200 km/h in order to examine the network architecture to minimize the delay.

• The Journal of Korean Society for Radiation Therapy
• /
• v.26 no.2
• /
• pp.329-336
• /
• 2014

Purpose : To evaluate the dose on a contralateral breast and the usefulness of shielding according to the distance between the contralateral breast and the side of the beam by breast size when patients who got breast implant receive radiation therapy. Materials and Methods : We equipped 200 cc, 300 cc, 400 cc, and 500 cc breast model on the human phantom (Rando-phantom), acquired CT images (philips 16channel, Netherlands) and established the radiation treatment plan, 180 cGy per day on the left breast (EclipseTM ver10.0.42, Varian Medical Systems, USA) by size. We set up each points, A, B, C, and D on the right(contralateral) breast model for measurement by size and by the distance from the beam and attached MOSFET at each points. The 6 MV, 10 MV and 15 MV X-ray were irradiated to the left(target) breast model and we measured exposure dose of contralateral breast model using MOSFET. Also, at the same condition, we acquired the dose value after shielding using only Pb 2 mm and bolus 3 mm under the Pb 2 mm together. Results : As the breast model is bigger from 200 cc to 500 cc, The surface of the contralateral breast is closer to the beam. As a result, from 200 cc to 500 cc, on 180 cGy basis, the measurement value of the scattered ray inclined by 3.22 ~ 4.17% at A point, 4.06 ~ 6.22% at B point, 0.4~0.5% at C point, and was under 0.4% at D point. As the X-ray energy is higher, from 6 MV to 15 MV, on 180 cGy basis, the measurement value of the scattered ray inclined by 4.06~5% at A point, 2.85~4.94% at B point, 0.74~1.65% at C point, and was under 0.4% at D point. As using Pb 2 mm for shield, scattered ray declined by average 9.74% at A and B point, 2.8% at C point, and is under 1% at D point. As using Pb 2 mm and bolus together for shield, scattered ray declined by average 9.76% at A and B point, 2.2% at C point, and is under 1% at D point. Conclusion : Commonly, in case of patients who got breast implant, there is a distance difference by breast size between the contralateral breast and the side of beam. As the distance is closer to the beam, the scattered ray inclined. At the same size of the breast, as the X-ray energy is higher, the exposure dose by scattered ray tends to incline. As a result, as low as possible energy wihtin the plan dose is good for reducing the exposure dose.

• Journal of Global Scholars of Marketing Science
• /
• v.19 no.4
• /
• pp.44-56
• /
• 2009

Impulse purchasing is defined as an immediate purchase with no pre-shopping intentions. Previous studies of impulse buying have focused primarily on factors linked to marketing mix variables, situational factors, and consumer demographics and traits. In previous studies, marketing mix variables such as product category, product type, and atmospheric factors including advertising, coupons, sales events, promotional stimuli at the point of sale, and media format have been used to evaluate product information. Some authors have also focused on situational factors surrounding the consumer. Factors such as the availability of credit card usage, time available, transportability of the products, and the presence and number of shopping companions were found to have a positive impact on impulse buying and/or impulse tendency. Research has also been conducted to evaluate the effects of individual characteristics such as the age, gender, and educational level of the consumer, as well as perceived crowding, stimulation, and the need for touch, on impulse purchasing. In summary, previous studies have found that all products can be purchased impulsively (Vohs and Faber, 2007), that situational factors affect and/or at least facilitate impulse purchasing behavior, and that various individual traits are closely linked to impulse buying. The recent introduction of new distribution channels such as home shopping channels, discount stores, and Internet stores that are open 24 hours a day increases the probability of impulse purchasing. However, previous literature has focused predominantly on situational and marketing variables and thus studies that consider critical consumer characteristics are still lacking. To fill this gap in the literature, the present study builds on this third tradition of research and focuses on individual trait variables, which have rarely been studied. More specifically, the current study investigates whether impulse buying tendency has a positive impact on impulse buying behavior, and evaluates how consumer characteristics such as the need for cognitive closure (NFCC), psychological wellbeing, and susceptibility to interpersonal influences affect the tendency of consumers towards impulse buying. The survey results reveal that while consumer affective impulsivity has a strong positive impact on impulse buying behavior, cognitive impulsivity has no impact on impulse buying behavior. Furthermore, affective impulse buying tendency is driven by sub-components of NFCC such as decisiveness and discomfort with ambiguity, psychological wellbeing constructs such as environmental control and purpose in life, and by normative and informational influences. In addition, cognitive impulse tendency is driven by sub-components of NFCC such as decisiveness, discomfort with ambiguity, and close-mindedness, and the psychological wellbeing constructs of environmental control, as well as normative and informational influences. The present study has significant theoretical implications. First, affective impulsivity has a strong impact on impulse purchase behavior. Previous studies based on affectivity and flow theories proposed that low to moderate levels of impulsivity are driven by reduced self-control or a failure of self-regulatory mechanisms. The present study confirms the above proposition. Second, the present study also contributes to the literature by confirming that impulse buying tendency can be viewed as a two-dimensional concept with both affective and cognitive dimensions, and illustrates that impulse purchase behavior is explained mainly by affective impulsivity, not by cognitive impulsivity. Third, the current study accommodates new constructs such as psychological wellbeing and NFCC as potential influencing factors in the research model, thereby contributing to the existing literature. Fourth, by incorporating multi-dimensional concepts such as psychological wellbeing and NFCC, more diverse aspects of consumer information processing can be evaluated. Fifth, the current study also extends the existing literature by confirming the two competing routes of normative and informational influences. Normative influence occurs when individuals conform to the expectations of others or to enhance his/her self-image. Whereas informational influence occurs when individuals search for information from knowledgeable others or making inferences based upon observations of the behavior of others. The present study shows that these two competing routes of social influence can be attributed to different sources of influence power. The current study also has many practical implications. First, it suggests that people with affective impulsivity may be primary targets to whom companies should pay closer attention. Cultivating a more amenable and mood-elevating shopping environment will appeal to this segment. Second, the present results demonstrate that NFCC is closely related to the cognitive dimension of impulsivity. These people are driven by careless thoughts, not by feelings or excitement. Rational advertising at the point of purchase will attract these customers. Third, people susceptible to normative influences are another potential target market. Retailers and manufacturers could appeal to this segment by advertising their products and/or services as products that can be used to identify with or conform to the expectations of others in the aspiration group. However, retailers should avoid targeting people susceptible to informational influences as a segment market. These people are engaged in an extensive information search relevant to their purchase, and therefore more elaborate, long-term rational advertising messages, which can be internalized into these consumers' thought processes, will appeal to this segment. The current findings should be interpreted with caution for several reasons. The study used a small convenience sample, and only investigated behavior in two dimensions. Accordingly, future studies should incorporate a sample with more diverse characteristics and measure different aspects of behavior. Future studies should also investigate personality traits closely related to affectivity theories. Trait variables such as sensory curiosity, interpersonal curiosity, and atmospheric responsiveness are interesting areas for future investigation.

• Journal of Intelligence and Information Systems
• /
• v.22 no.3
• /
• pp.143-163
• /
• 2016

The demographics of Internet users are the most basic and important sources for target marketing or personalized advertisements on the digital marketing channels which include email, mobile, and social media. However, it gradually has become difficult to collect the demographics of Internet users because their activities are anonymous in many cases. Although the marketing department is able to get the demographics using online or offline surveys, these approaches are very expensive, long processes, and likely to include false statements. Clickstream data is the recording an Internet user leaves behind while visiting websites. As the user clicks anywhere in the webpage, the activity is logged in semi-structured website log files. Such data allows us to see what pages users visited, how long they stayed there, how often they visited, when they usually visited, which site they prefer, what keywords they used to find the site, whether they purchased any, and so forth. For such a reason, some researchers tried to guess the demographics of Internet users by using their clickstream data. They derived various independent variables likely to be correlated to the demographics. The variables include search keyword, frequency and intensity for time, day and month, variety of websites visited, text information for web pages visited, etc. The demographic attributes to predict are also diverse according to the paper, and cover gender, age, job, location, income, education, marital status, presence of children. A variety of data mining methods, such as LSA, SVM, decision tree, neural network, logistic regression, and k-nearest neighbors, were used for prediction model building. However, this research has not yet identified which data mining method is appropriate to predict each demographic variable. Moreover, it is required to review independent variables studied so far and combine them as needed, and evaluate them for building the best prediction model. The objective of this study is to choose clickstream attributes mostly likely to be correlated to the demographics from the results of previous research, and then to identify which data mining method is fitting to predict each demographic attribute. Among the demographic attributes, this paper focus on predicting gender, age, marital status, residence, and job. And from the results of previous research, 64 clickstream attributes are applied to predict the demographic attributes. The overall process of predictive model building is compose of 4 steps. In the first step, we create user profiles which include 64 clickstream attributes and 5 demographic attributes. The second step performs the dimension reduction of clickstream variables to solve the curse of dimensionality and overfitting problem. We utilize three approaches which are based on decision tree, PCA, and cluster analysis. We build alternative predictive models for each demographic variable in the third step. SVM, neural network, and logistic regression are used for modeling. The last step evaluates the alternative models in view of model accuracy and selects the best model. For the experiments, we used clickstream data which represents 5 demographics and 16,962,705 online activities for 5,000 Internet users. IBM SPSS Modeler 17.0 was used for our prediction process, and the 5-fold cross validation was conducted to enhance the reliability of our experiments. As the experimental results, we can verify that there are a specific data mining method well-suited for each demographic variable. For example, age prediction is best performed when using the decision tree based dimension reduction and neural network whereas the prediction of gender and marital status is the most accurate by applying SVM without dimension reduction. We conclude that the online behaviors of the Internet users, captured from the clickstream data analysis, could be well used to predict their demographics, thereby being utilized to the digital marketing.

• Journal of Intelligence and Information Systems
• /
• v.23 no.1
• /
• pp.95-108
• /
• 2017

Recently, AlphaGo which is Bakuk (Go) artificial intelligence program by Google DeepMind, had a huge victory against Lee Sedol. Many people thought that machines would not be able to win a man in Go games because the number of paths to make a one move is more than the number of atoms in the universe unlike chess, but the result was the opposite to what people predicted. After the match, artificial intelligence technology was focused as a core technology of the fourth industrial revolution and attracted attentions from various application domains. Especially, deep learning technique have been attracted as a core artificial intelligence technology used in the AlphaGo algorithm. The deep learning technique is already being applied to many problems. Especially, it shows good performance in image recognition field. In addition, it shows good performance in high dimensional data area such as voice, image and natural language, which was difficult to get good performance using existing machine learning techniques. However, in contrast, it is difficult to find deep leaning researches on traditional business data and structured data analysis. In this study, we tried to find out whether the deep learning techniques have been studied so far can be used not only for the recognition of high dimensional data but also for the binary classification problem of traditional business data analysis such as customer churn analysis, marketing response prediction, and default prediction. And we compare the performance of the deep learning techniques with that of traditional artificial neural network models. The experimental data in the paper is the telemarketing response data of a bank in Portugal. It has input variables such as age, occupation, loan status, and the number of previous telemarketing and has a binary target variable that records whether the customer intends to open an account or not. In this study, to evaluate the possibility of utilization of deep learning algorithms and techniques in binary classification problem, we compared the performance of various models using CNN, LSTM algorithm and dropout, which are widely used algorithms and techniques in deep learning, with that of MLP models which is a traditional artificial neural network model. However, since all the network design alternatives can not be tested due to the nature of the artificial neural network, the experiment was conducted based on restricted settings on the number of hidden layers, the number of neurons in the hidden layer, the number of output data (filters), and the application conditions of the dropout technique. The F1 Score was used to evaluate the performance of models to show how well the models work to classify the interesting class instead of the overall accuracy. The detail methods for applying each deep learning technique in the experiment is as follows. The CNN algorithm is a method that reads adjacent values from a specific value and recognizes the features, but it does not matter how close the distance of each business data field is because each field is usually independent. In this experiment, we set the filter size of the CNN algorithm as the number of fields to learn the whole characteristics of the data at once, and added a hidden layer to make decision based on the additional features. For the model having two LSTM layers, the input direction of the second layer is put in reversed position with first layer in order to reduce the influence from the position of each field. In the case of the dropout technique, we set the neurons to disappear with a probability of 0.5 for each hidden layer. The experimental results show that the predicted model with the highest F1 score was the CNN model using the dropout technique, and the next best model was the MLP model with two hidden layers using the dropout technique. In this study, we were able to get some findings as the experiment had proceeded. First, models using dropout techniques have a slightly more conservative prediction than those without dropout techniques, and it generally shows better performance in classification. Second, CNN models show better classification performance than MLP models. This is interesting because it has shown good performance in binary classification problems which it rarely have been applied to, as well as in the fields where it's effectiveness has been proven. Third, the LSTM algorithm seems to be unsuitable for binary classification problems because the training time is too long compared to the performance improvement. From these results, we can confirm that some of the deep learning algorithms can be applied to solve business binary classification problems.