• Radiation Oncology Journal
• /
• v.16 no.2
• /
• pp.147-157
• /
• 1998

Purpose : The aim of this retrospective study is to assess the necessity of s1aging laparotomy in the management of supradiaphragmatic CS I-II Hodgkin's disease. Prognostic factors and the usefulness of prognostic factor groups were also analyzed. Materials and Methods : From 1985 to 1995, fifty one Patients who were diagnosed as supradiaphragmatic CS I-II Hodgkin's disease at Yonsei Cancer Center in Seoul, Korea were enrolled in this study Age range was 4 to 67 with median age of 30. The number of patients with each CS IA, II A, and IIB were 16, 25, and 10, respectively. Radiotherapy(RT) was delivered using 4 or 6 MV photon beam to a total dose of 19.5 to 55.6Gy (median dose : 45Gy) with a 1.5 to 1.BGy per fraction. Chemotherapy(CT) was given in 2-12 cycles(median : 6 cycles). Thirty one Patients were treated with RT alone, 4 patients with CT alone and 16 patients with combined chemoradiotherapy. RT volumes varied from involved fields(3), subtotal nodal fields(18) or mantle fields(26). Results : Five-year disease-free survival rate(DFS) was $78.0\%$ and overall survival rate(05) was $87.6\%$. Fifty Patients achieved a complete remission after initial treatment and 8 patients were relapsed. Salvage therapy was given to 7 patients, 1 with RT alone, 4 with CT alone, 2 with RT+CT. Only two patients were successfully salvaged. Feminine gender and large media-stinal adenopathy were significant adverse prognostic factors in the univariate analysis for DFS. The significant adverse prognostic factors of OS were B symptom and clinical stage. When patients were analyzed according to European Organization for Research and Treatment of Cancer(EORTC) prognostic factor groups, the DFS in Patients with very favorable, favorable and unfavorable group was 100, 100 and $55.8\%$(p<0.05), and the 05 in each patients' group was 100, 100 and $75.1\%$(p<0.05), respectively. In very favorable and favorable groups, the DFS and 05 were all $100\%$ by RT alone, but in unfavorable group, RT with CT had a lesser relapse rate than RT alone. The subtotal nodal irradiation had better OFS than mantle RT in patients treated with RT. Conclusion : In present study, the DFS and OS in patients who did not undergo s1aging laparotomy were similar with the results in the literatures of which patients were surgically staged. Therefore, we may suggest that staging laparotomy would not influence the outcome of treatments. In univariate analysis, gender, large mediastinal adenopathy. B symptoms and clinical stage were significant prognostic factors for the survival rate. We confirm the usefulness of EORTC prognostic factor groups which may be a good.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.55 no.1
• /
• pp.76-82
• /
• 2010

Experiments were conducted to investigate seed persistence of Chinese milk vetch(CMV) in naturally reseeded rice field in 2007~2009. The seed and pods with seeds were buried in rice field at 0, 5 and 10 cm depths and retrieved from the field at one to three month intervals from July to the following year March and determined change of seed dormancy and viability. In the second experiment, persistence of the CMV seeds in the naturally reseeded rice field at different tillage methods and soil depths were also investigated after rice harvest in autumn. Burial depths and durations affected recovery rate, dormancy and viability of CMV seed. The viability loss was faster and greater in the seed than the pod with seeds and on the soil surface than the 5 or 10 cm burial depths. The recovery rate of CMV seed was decreased starting from one month as seed burial and it was significantly decreased to 52~65% for the seed in September. However, unlike the seed burial, the nearly 100% CMV seeds were recovered for burial as pod with seeds even after four months burial in both 0 and 5 cm depths. However, the recovery rate was sharply declined to below 30% at October in 2007 in both seed and pods with seeds and in the 2008/2009 experiment. the 15~47% of CMV seeds still remained even after October. The CMV had high seed dormancy of 95%, showing only 4~5% germination at the beginning in June but the seed germination increased to 25 to 35% in seed and 55 to 61% in pod with seeds in September due to breakage of hard seed dormancy. The viability loss was faster in the seed than in the pod with seeds regardless of depths of placement in the soil base on decayed seeds. Also the seed placed on the soil surface lost viability faster than the 5~10 burial depths. On the other hand, field observation in the naturally reseeded CMV rice field showed that as many as 917~2,185 CMV seeds $m^2$ were from the 0~15 cm soil depth in the rotary tillage and 250~10,105 CMV seeds in minimum tillage treatmints. The recovered seed germinated 25~33%, 23~43% but still had high percentage of hard seed having 64~72% and 51~77% even after rice harvest in autumn. These results indicate that freshly harvested CMV seeds had high level of primary dormancy and the dormancy was gradually broken in soil with time during rice cultivation periods and appreciable number of CMV seeds remained even 4 month after burial in soil. CMV plant regenerated naturally from the remained seed bank at rice harvest time in autumn. The CMV seedling still emerged even after 2 years of continuous destructive killing of emerged CMV plant by rotary tillage in naturally reseeded CMV plant in rice field, indicating that CMV seeds do persistent as least two years in soil.

• Tuberculosis and Respiratory Diseases
• /
• v.46 no.1
• /
• pp.53-64
• /
• 1999

Background : Massive and untreated hemoptysis is associated with a mortality of greater than 50 percent. Since the bleeding is from a bronchial arterial source in the vast majority of patients, embolization of the bronchial arteries(BAE) has become an accepted treatment in the management of massive hemoptysis because it achieves immediate control of bleeding in 75 to 90 percent of the patients. Methods: Between 1990 and 1996, we treated 146 patients with hemoptysis by bronchial artery embolization. Catheters(4, 5, or 7F) and gelfoam, ivalon, and/or microcoil were used for embolization. Results: Pulmonary tuberculosis and related disorders were the most common underlying disease of hemoptysis(72.6%). Immediate success rate to control bleeding within 24hours was 95%, and recurrence rate was 24.7%. The recurrence rate occured within 6 months after embolization was 63.9%. Initial angiographic findings such as bilaterality, systemic-pulmonary artery shunt, neovascularity, aneurysm were not statistically correlated with rebleeding tendency(P>0.05). Among Initial radiographic findings, only pleural lesions were significantly correlated with rebleeding tendency(P<0.05). At additional bronchial artery angiograpy done due to rebleeding, recanalization of previous embolized arteries were 63.9%, and the presence of new feeding arteries were 16.7%, and 19.4% of patients with rebleeding showed both The complications such as fever, chest pain, headache, nausea and vomiting, arrhythmia, paralylytic ileus, transient sensory loss (lower extremities), hypotension, urination difficulty were noticed at 40 patients(27.4%). Conclusion: We conclude that bronchial artery embolization is relatively safe method achieving immediate control of massive hemoptysis. At initial angiographic findings, we could not find any predictive factors for subsequent rebleeding. It may warrant further study whether patients with pleural disease have definetely increased rebleeding tendency.

• Journal of dental hygiene science
• /
• v.18 no.2
• /
• pp.85-96
• /
• 2018

The purpose of this study is to propose learning objectives in social dental hygiene by analyzing and reviewing learning objectives in oral health administration area of the existing public oral health. This study is a cross-sectional study. The subjects of the study selected with convenience extraction were 15 members of the social dental hygiene subcommittee of the Korean Society of Dental Hygiene Science. Data collection was conducted by self-filling questionnaire. The research tool is from 48 items of A division in the book of learning objectives in the dental hygienist national examination, and this study classified each of them into 'dental hygiene job relevance', 'dental hygiene competency relevance', 'timeliness', and 'value discrimination of educational goal setting' to comprise 192 items. Also, to collect expert opinions, this study conducted Delphi survey on 7 academic experts. Statistical analysis was performed using the IBM SPSS Statistics ver. 23.0 program (IBM Co., Armonk, NY, USA). Recoding was performed according to the degree of relevance of each learning objective and frequency analysis was performed. This study removed 18 items from the whole learning objectives in the dental hygienist national examination in the oral health administration area of public oral health. Fifteen revisions were made and 15 existing learning objectives were maintained. Forty-five learning objectives were proposed as new social dental hygiene learning objectives. The topics of learning objectives are divided into social security and medical assistance, oral health care system, oral health administration, and oral health policy. As a result of this study, it was necessary to construct the learning objectives of social dental hygiene in response to changing situation at the time. The contents of education should be revised in order of revision of learning objectives, development of competency, development of learning materials, and national examination.

• Journal of Intelligence and Information Systems
• /
• v.20 no.2
• /
• pp.73-92
• /
• 2014

An Adaptive Clustering-based Collaborative Filtering Technique was proposed to solve the fundamental problems of collaborative filtering, such as cold-start problems, scalability problems and data sparsity problems. Previous collaborative filtering techniques were carried out according to the recommendations based on the predicted preference of the user to a particular item using a similar item subset and a similar user subset composed based on the preference of users to items. For this reason, if the density of the user preference matrix is low, the reliability of the recommendation system will decrease rapidly. Therefore, the difficulty of creating a similar item subset and similar user subset will be increased. In addition, as the scale of service increases, the time needed to create a similar item subset and similar user subset increases geometrically, and the response time of the recommendation system is then increased. To solve these problems, this paper suggests a collaborative filtering technique that adapts a condition actively to the model and adopts the concepts of a context-based filtering technique. This technique consists of four major methodologies. First, items are made, the users are clustered according their feature vectors, and an inter-cluster preference between each item cluster and user cluster is then assumed. According to this method, the run-time for creating a similar item subset or user subset can be economized, the reliability of a recommendation system can be made higher than that using only the user preference information for creating a similar item subset or similar user subset, and the cold start problem can be partially solved. Second, recommendations are made using the prior composed item and user clusters and inter-cluster preference between each item cluster and user cluster. In this phase, a list of items is made for users by examining the item clusters in the order of the size of the inter-cluster preference of the user cluster, in which the user belongs, and selecting and ranking the items according to the predicted or recorded user preference information. Using this method, the creation of a recommendation model phase bears the highest load of the recommendation system, and it minimizes the load of the recommendation system in run-time. Therefore, the scalability problem and large scale recommendation system can be performed with collaborative filtering, which is highly reliable. Third, the missing user preference information is predicted using the item and user clusters. Using this method, the problem caused by the low density of the user preference matrix can be mitigated. Existing studies on this used an item-based prediction or user-based prediction. In this paper, Hao Ji's idea, which uses both an item-based prediction and user-based prediction, was improved. The reliability of the recommendation service can be improved by combining the predictive values of both techniques by applying the condition of the recommendation model. By predicting the user preference based on the item or user clusters, the time required to predict the user preference can be reduced, and missing user preference in run-time can be predicted. Fourth, the item and user feature vector can be made to learn the following input of the user feedback. This phase applied normalized user feedback to the item and user feature vector. This method can mitigate the problems caused by the use of the concepts of context-based filtering, such as the item and user feature vector based on the user profile and item properties. The problems with using the item and user feature vector are due to the limitation of quantifying the qualitative features of the items and users. Therefore, the elements of the user and item feature vectors are made to match one to one, and if user feedback to a particular item is obtained, it will be applied to the feature vector using the opposite one. Verification of this method was accomplished by comparing the performance with existing hybrid filtering techniques. Two methods were used for verification: MAE(Mean Absolute Error) and response time. Using MAE, this technique was confirmed to improve the reliability of the recommendation system. Using the response time, this technique was found to be suitable for a large scaled recommendation system. This paper suggested an Adaptive Clustering-based Collaborative Filtering Technique with high reliability and low time complexity, but it had some limitations. This technique focused on reducing the time complexity. Hence, an improvement in reliability was not expected. The next topic will be to improve this technique by rule-based filtering.

• Journal of Intelligence and Information Systems
• /
• v.26 no.3
• /
• pp.91-108
• /
• 2020

Korea's famous soccer players are steadily performing well in international leagues, which led to higher interests of Korean fans in the international leagues. Reflecting the growing social phenomenon of rising interests on international leagues by Korean fans, the study examined the overall consumer perception in the consumption of uniform by domestic soccer fans and compared the changes in perception following the transfers of the players. Among others, the paper examined the consumer perception and purchase factors of soccer fans shown in social media, focusing on periods before and after the recruitment of Heung-Min Son to English Premier League's Tottenham Football Club. To this end, the EPL uniform is the collection keyword the paper utilized and collected consumer postings from domestic website and social media via Python 3.7, and analyzed them using Ucinet 6, NodeXL 1.0.1, and SPSS 25.0 programs. The results of this study can be summarized as follows. First, the uniform of the club that consistently topped the league, has been gaining attention as a popular uniform, and the players' performance, and the players' position have been identified as key factors in the purchase and search of professional football uniforms. In the case of the club, the actual ranking and whether the league won are shown to be important factors in the purchase and search of professional soccer uniforms. The club's emblem and the sponsor logo that will be attached to the uniform are also factors of interest to consumers. In addition, in the decision making process of purchase of a uniform by professional soccer fan, uniform's form, marking, authenticity, and sponsors are found to be more important than price, design, size, and logo. The official online store has emerged as a major purchasing channel, followed by gifts for friends or requests from acquaintances when someone travels to the United Kingdom. Second, a classification of key control categories through the convergence of iteration correlation analysis and Clauset-Newman-Moore clustering algorithm shows differences in the classification of individual groups, but groups that include the EPL's club and player keywords are identified as the key topics in relation to professional football uniforms. Third, between 2002 and 2006, the central theme for professional football uniforms was World Cup and English Premier League, but from 2012 to 2015, the focus has shifted to more interest of domestic and international players in the English Premier League. The subject has changed to the uniform itself from this time on. In this context, the paper can confirm that the major issues regarding the uniforms of professional soccer players have changed since Ji-Sung Park's transfer to Manchester United, and Sung-Yong Ki, Chung-Yong Lee, and Heung-Min Son's good performances in these leagues. The paper also identified that the uniforms of the clubs to which the players have transferred to are of interest. Fourth, both male and female consumers are showing increasing interest in Son's league, the English Premier League, which Tottenham FC belongs to. In particular, the increasing interest in Son has shown a tendency to increase interest in football uniforms for female consumers. This study presents a variety of researches on sports consumption and has value as a consumer study by identifying unique consumption patterns. It is meaningful in that the accuracy of the interpretation has been enhanced by using a cluster analysis via convergence of iteration correlation analysis and Clauset-Newman-Moore clustering algorithm to identify the main topics. Based on the results of this study, the clubs will be able to maximize its profits and maintain good relationships with fans by identifying key drivers of consumer awareness and purchasing for professional soccer fans and establishing an effective marketing strategy.

• Korean Journal of Environment and Ecology
• /
• v.33 no.3
• /
• pp.280-291
• /
• 2019

A. antiquum, first observed in Jeju Samdo Island in 1949, was designated as the Natural Monument No. 18 in December 1962 in recognition of its academic value. In Korea, it grows in nature only in Samdo in Jeju Island. Although its natural habitat was greatly damaged and almost destroyed due to firewood, stealing, etc. After the emancipation, it has been maintained by the transplantation and restoration. The site observed by this study has been managed as a restricted area since 2011. Since it has been about 20 years since the restoration of the native site in the 2000s, it is necessary to check the official management history records, such as the origin of transplantation and restoration to monitor the changes in the growth status and to control the habitat. As the results of this study, we have secured the records of cultural property management history, such as the identification of native species and the transplantation and restoration records. We also examined the change of the growth and development of A. antiquum 20 years after the restoration. There are no official records of the individuals transplanted to the restored natural habitat of A. antiquum in the 1970s and 1980s, and there was a controversy about the nativeness of those individuals that were restored and transplanted in 1974 since they were Japanese individuals. The studies of identifying native as the results of this study, we have secured the records of cultural property management history, such as the identification of native species and the transplantation and restoration records. We also examined the change of the growth and development of A. antiquum 20 years after the restoration. There are two sites in natural habitat in Samdo Island. A total of 65 individuals grow in three layers on three stone walls in a site while 29 individuals grow in two columns in the other site. A. antiquum grows in an evergreen broad-leaved forest dominated by Neolitsea sericea, and we did not find any other individuals of naturally growing A. antiquum outside the investigated site. This study checked the distribution of A. antiquum seedlings observed initially after the restoration. There were more than 300 seedling individuals, and we selected three densely populated sites for monitoring. There were 23 A. antiquum seedlings with 4 - 17 leaves per individual and the leaf length of 0.5 - 20 cm in monitoring site 1. There were 88 individuals with 5 - 6 leaves per individual and the leaf length of 1.3 - 10.4 cm in monitoring site 2 while there were 22 individuals with 5 - 9 leaves per individual and the leaf length of 4.5 - 12.1 cm in monitoring site 3. Although the natural habitat of A. antiquum was designated as a restricted public area in 2011, there is a high possibility that the habitat can be damaged because some activities, such as fishing and scuba diving are allowed. Therefore, it is necessary to enforce the law strictly, to provide sufficient education for the preservation of natural treasures, and to present accurate information about cultural assets.

• The Journal of Korean Society for Radiation Therapy
• /
• v.31 no.1
• /
• pp.75-81
• /
• 2019

Purpose: Helmet type bolus for 3D printer is being manufactured because of the disadvantages of Bolus materials when photon beam is used for the treatment of scalp malignancy. However, PLA, which is a used material, has a higher density than a tissue equivalent material and inconveniences occur when the patient wears PLA. In this study, we try to treat malignant scalp tumors by using M3 wax helmet with 3D printer. Methods and materials: For the modeling of the helmet type M3 wax, the head phantom was photographed by CT, which was acquired with a DICOM file. The part for helmet on the scalp was made with Helmet contour. The M3 Wax helmet was made by dissolving paraffin wax, mixing magnesium oxide and calcium carbonate, solidifying it in a PLA 3D helmet, and then eliminated PLA 3D Helmet of the surface. The treatment plan was based on Intensity-Modulated Radiation Therapy (IMRT) of 10 Portals, and the therapeutic dose was 200 cGy, using Analytical Anisotropic Algorithm (AAA) of Eclipse. Then, the dose was verified by using EBT3 film and Mosfet (Metal Oxide Semiconductor Field Effect Transistor: USA), and the IMRT plan was measured 3 times in 3 parts by reproducing the phantom of the head human model under the same condition with the CT simulation room. Results: The Hounsfield unit (HU) of the bolus measured by CT was $52{\pm}37.1$. The dose of TPS was 186.6 cGy, 193.2 cGy and 190.6 cGy at the M3 Wax bolus measurement points of A, B and C, and the dose measured three times at Mostet was $179.66{\pm}2.62cGy$, $184.33{\pm}1.24cGy$ and $195.33{\pm}1.69cGy$. And the error rates were -3.71 %, -4.59 %, and 2.48 %. The dose measured with EBT3 film was $182.00{\pm}1.63cGy$, $193.66{\pm}2.05cGy$ and $196{\pm}2.16cGy$. The error rates were -2.46 %, 0.23 % and 2.83 %. Conclusions: The thickness of the M3 wax bolus was 2 cm, which could help the treatment plan to be established by easily lowering the dose of the brain part. The maximum error rate of the scalp surface dose was measured within 5 % and generally within 3 %, even in the A, B, C measurements of dosimeters of EBT3 film and Mosfet in the treatment dose verification. The making period of M3 wax bolus is shorter, cheaper than that of 3D printer, can be reused and is very useful for the treatment of scalp malignancies as human tissue equivalent material. Therefore, we think that the use of casting type M3 wax bolus, which will complement the making period and cost of high capacity Bolus and Compensator in 3D printer, will increase later.

• Journal of Intelligence and Information Systems
• /
• v.25 no.1
• /
• pp.109-125
• /
• 2019

As interest on intelligent search engines increases, various studies have been conducted to extract and utilize the features related to products intelligencely. In particular, when users search for goods in e-commerce search engines, the 'color' of a product is an important feature that describes the product. Therefore, it is necessary to deal with the synonyms of color terms in order to produce accurate results to user's color-related queries. Previous studies have suggested dictionary-based approach to process synonyms for color features. However, the dictionary-based approach has a limitation that it cannot handle unregistered color-related terms in user queries. In order to overcome the limitation of the conventional methods, this research proposes a model which extracts RGB values from an internet search engine in real time, and outputs similar color names based on designated color information. At first, a color term dictionary was constructed which includes color names and R, G, B values of each color from Korean color standard digital palette program and the Wikipedia color list for the basic color search. The dictionary has been made more robust by adding 138 color names converted from English color names to foreign words in Korean, and with corresponding RGB values. Therefore, the fininal color dictionary includes a total of 671 color names and corresponding RGB values. The method proposed in this research starts by searching for a specific color which a user searched for. Then, the presence of the searched color in the built-in color dictionary is checked. If there exists the color in the dictionary, the RGB values of the color in the dictioanry are used as reference values of the retrieved color. If the searched color does not exist in the dictionary, the top-5 Google image search results of the searched color are crawled and average RGB values are extracted in certain middle area of each image. To extract the RGB values in images, a variety of different ways was attempted since there are limits to simply obtain the average of the RGB values of the center area of images. As a result, clustering RGB values in image's certain area and making average value of the cluster with the highest density as the reference values showed the best performance. Based on the reference RGB values of the searched color, the RGB values of all the colors in the color dictionary constructed aforetime are compared. Then a color list is created with colors within the range of ${\pm}50$ for each R value, G value, and B value. Finally, using the Euclidean distance between the above results and the reference RGB values of the searched color, the color with the highest similarity from up to five colors becomes the final outcome. In order to evaluate the usefulness of the proposed method, we performed an experiment. In the experiment, 300 color names and corresponding color RGB values by the questionnaires were obtained. They are used to compare the RGB values obtained from four different methods including the proposed method. The average euclidean distance of CIE-Lab using our method was about 13.85, which showed a relatively low distance compared to 3088 for the case using synonym dictionary only and 30.38 for the case using the dictionary with Korean synonym website WordNet. The case which didn't use clustering method of the proposed method showed 13.88 of average euclidean distance, which implies the DBSCAN clustering of the proposed method can reduce the Euclidean distance. This research suggests a new color synonym processing method based on RGB values that combines the dictionary method with the real time synonym processing method for new color names. This method enables to get rid of the limit of the dictionary-based approach which is a conventional synonym processing method. This research can contribute to improve the intelligence of e-commerce search systems especially on the color searching feature.

• Journal of Intelligence and Information Systems
• /
• v.27 no.4
• /
• pp.1-22
• /
• 2021

Aspect Based Sentiment Analysis (ABSA), which analyzes sentiment based on aspects that appear in the text, is drawing attention because it can be used in various business industries. ABSA is a study that analyzes sentiment by aspects for multiple aspects that a text has. It is being studied in various forms depending on the purpose, such as analyzing all targets or just aspects and sentiments. Here, the aspect refers to the property of a target, and the target refers to the text that causes the sentiment. For example, for restaurant reviews, you could set the aspect into food taste, food price, quality of service, mood of the restaurant, etc. Also, if there is a review that says, "The pasta was delicious, but the salad was not," the words "steak" and "salad," which are directly mentioned in the sentence, become the "target." So far, in ABSA, most studies have analyzed sentiment only based on aspects or targets. However, even with the same aspects or targets, sentiment analysis may be inaccurate. Instances would be when aspects or sentiment are divided or when sentiment exists without a target. For example, sentences like, "Pizza and the salad were good, but the steak was disappointing." Although the aspect of this sentence is limited to "food," conflicting sentiments coexist. In addition, in the case of sentences such as "Shrimp was delicious, but the price was extravagant," although the target here is "shrimp," there are opposite sentiments coexisting that are dependent on the aspect. Finally, in sentences like "The food arrived too late and is cold now." there is no target (NULL), but it transmits a negative sentiment toward the aspect "service." Like this, failure to consider both aspects and targets - when sentiment or aspect is divided or when sentiment exists without a target - creates a dual dependency problem. To address this problem, this research analyzes sentiment by considering both aspects and targets (Target-Aspect-Sentiment Detection, hereby TASD). This study detected the limitations of existing research in the field of TASD: local contexts are not fully captured, and the number of epochs and batch size dramatically lowers the F1-score. The current model excels in spotting overall context and relations between each word. However, it struggles with phrases in the local context and is relatively slow when learning. Therefore, this study tries to improve the model's performance. To achieve the objective of this research, we additionally used auxiliary loss in aspect-sentiment classification by constructing CNN(Convolutional Neural Network) layers parallel to existing models. If existing models have analyzed aspect-sentiment through BERT encoding, Pooler, and Linear layers, this research added CNN layer-adaptive average pooling to existing models, and learning was progressed by adding additional loss values for aspect-sentiment to existing loss. In other words, when learning, the auxiliary loss, computed through CNN layers, allowed the local context to be captured more fitted. After learning, the model is designed to do aspect-sentiment analysis through the existing method. To evaluate the performance of this model, two datasets, SemEval-2015 task 12 and SemEval-2016 task 5, were used and the f1-score increased compared to the existing models. When the batch was 8 and epoch was 5, the difference was largest between the F1-score of existing models and this study with 29 and 45, respectively. Even when batch and epoch were adjusted, the F1-scores were higher than the existing models. It can be said that even when the batch and epoch numbers were small, they can be learned effectively compared to the existing models. Therefore, it can be useful in situations where resources are limited. Through this study, aspect-based sentiments can be more accurately analyzed. Through various uses in business, such as development or establishing marketing strategies, both consumers and sellers will be able to make efficient decisions. In addition, it is believed that the model can be fully learned and utilized by small businesses, those that do not have much data, given that they use a pre-training model and recorded a relatively high F1-score even with limited resources.