• Journal of Korea Society of Industrial Information Systems
• /
• v.29 no.3
• /
• pp.1-12
• /
• 2024

In this paper, we fine-tuned KoAlpaca, a large-scale Korean language model, and implemented a blog text generation system utilizing it. Blogs on social media platforms are widely used as a marketing tool for businesses. We constructed training data of positive reviews through emotion analysis and refinement of collected sponsored blog texts and applied QLoRA for the lightweight training of KoAlpaca. QLoRA is a fine-tuning approach that significantly reduces the memory usage required for training, with experiments in an environment with a parameter size of 12.8B showing up to a 58.8% decrease in memory usage compared to LoRA. To evaluate the generative performance of the fine-tuned model, texts generated from 100 inputs not included in the training data produced on average more than twice the number of words compared to the pre-trained model, with texts of positive sentiment also appearing more than twice as often. In a survey conducted for qualitative evaluation of generative performance, responses indicated that the fine-tuned model's generated outputs were more relevant to the given topics on average 77.5% of the time. This demonstrates that the positive review generation language model for sponsored content in this paper can enhance the efficiency of time management for content creation and ensure consistent marketing effects. However, to reduce the generation of content that deviates from the category of positive reviews due to elements of the pre-trained model, we plan to proceed with fine-tuning using the augmentation of training data.

• International Commerce and Information Review
• /
• v.19 no.1
• /
• pp.67-98
• /
• 2017

This study intended to investigate to what extent country image of America and previous visit experience affect the cross-border shopping intention. In particular, the present study used a country image measurement brought from another research study the factors of which are economy-technology image, social-cultural image, and citizen image. A total of 155 respondents participated in the survey targeting Korean citizen for the present study. Single regression, multiple regression, and independent t-test were conducted for data analysis. The result of the single regression indicated that country image is a critical predictor of cross-border shopping intention. The Multiple regression revealed that among three factors composing country image, social-cultural image plays the most significant and economy-technology image plays the second-most significant role in influencing cross-border shopping intention. However, it was found that citizen image does not play a substantial role for some reason. Moreover, the result of t-test showed that those who have a prior visit experience to America are more likely to buy products online from America than those who don't have prior visit experience. More detailed findings and implications will be discussed in the manuscript.

• The Journal of the Korea Contents Association
• /
• v.21 no.4
• /
• pp.38-49
• /
• 2021

This study analyzed the trend of the term 'false and exaggerated advertisement' in 5,141 newspaper articles from 1990 to 2019 using text mining methodology. First of all, we identified the most frequent keywords of false and exaggerated advertisements through frequency analysis for all newspaper articles, and understood the context between the extracted keywords. Next, to examine how false and exaggerated advertisements have changed, the frequency analysis was performed by separating articles by 10 years, and the tendency of the keyword that became an issue was identified by comparing the number of academic papers on the subject of the highest keywords of each year. Finally, we identified trends in false and exaggerated advertisements based on the detailed keywords in the topic using the topic modeling. In our results, it was confirmed that the topic that became an issue at a specific time was extracted as the frequent keywords, and the keyword trends by period changed in connection with social and environmental factors. This study is meaningful in helping consumers spend wisely by cultivating background knowledge about unfair advertising. Furthermore, it is expected that the core keyword extraction will provide the true purpose of advertising and deliver its implications to companies and related employees who commit misconduct.

• Information Systems Review
• /
• v.23 no.1
• /
• pp.45-68
• /
• 2021

As the influence of online reviews on consumer decision-making increases, concerns about review manipulation are also increasing. Fake reviews or review manipulations are emerging as an important problem by posting untrue reviews in order to increase sales volume, causing the consumer's reverse choice, and acting at a high cost to the society as a whole. Most of the related prior studies have focused on predicting review manipulation through data mining methods, and research from a consumer perspective is insufficient. However, since the possibility of manipulation of reviews perceived by consumers can affect the usefulness of reviews, it can provide important implications for online word-of-mouth management regardless of whether it is false or not. Therefore, in this study, we analyzed whether there is a difference between the review evaluated by the consumer as being manipulated and the general review, and verified whether the manipulated review negatively affects the review usefulness. For empirical analysis, 34,711 online book reviews on the LibraryThing website were analyzed using multilevel logistic regression analysis and Poisson regression analysis. As a result of the analysis, it was found that there were differences in product level, reviewer level, and review level factors between reviews that consumers perceived as being manipulated and reviews that were not. In addition, manipulated reviews have been shown to negatively affect review usefulness.

• The Journal of the Korea Contents Association
• /
• v.11 no.4
• /
• pp.361-377
• /
• 2011

This study has three primary purposes, firstly to identify how leader's personal characters influence to his/her transformational leadership, secondly to find how transformational leadership influences to innovative behavior, finally to explore how organizational cultures moderate between transformational leadership and innovative behavior. The first part of the study, based on literature study on transformational leadership, provides insight into what are antecedents, moderators and dependent variable in transformational leadership. Firstly, leader's personal characters are selected as antecedent variables such as extroversion and self-efficacy. Secondly, innovative behavior is introduced as a dependent variable. Thirdly, two types of organizational culture are considered as moderators between leader's personal character and leadership In this study, a comprehensive research model and hypothesis were empirically tested based on data from three types of questionnaires involving 663 employees in Korean organizations. In order to test the hypotheses, we have used Structural Equations Model (SEM) from AMOS7.0. In this analysis, we have employed raw data as it is instead of correlation matrix or covariance matrix. We have tested hypotheses by examining the significance of each path of the model, and gone through the process of testing the goodness of fit of the model itself. The results of statistical analysis show the following. Firstly, one of leader's personalities, self-efficacy has positive effect on his/her transformational leadership, but extroversion does not have positive effect. Secondly, transformational leadership has positive effect on innovative behavior. Finally, there was not any cultural moderating effects between transformational leadership and innovative behavior.

• The Journal of Society for e-Business Studies
• /
• v.17 no.2
• /
• pp.129-147
• /
• 2012

Collaborative filtering (CF), one of the most successful recommendation techniques, has been used in a number of different applications such as recommending web pages, movies, music, articles and products. One of the critical issues in CF is why recommendation performances are different depending on application domains. However, prior literatures have focused on only data characteristics to explain the origin of the difference. Scant attentions have been paid to provide systematic explanation on the issue. To fill this research gap, this study attempts to systematically explain why recommendation performances are different using structural indexes of social network. For this purpose, we developed hypotheses regarding the relationships between structural indexes of social network and recommendation performance of collaboration filtering, and empirically tested them. Results of this study showed that density and inconclusiveness positively affected recommendation performance while clustering coefficient negatively affected it. This study can be used as stepping stone for understanding collaborative filtering recommendation performance. Furthermore, it might be helpful for managers to decide whether they adopt recommendation systems.

• Proceedings of the CALSEC Conference
• /
• 1999.11a
• /
• pp.529-534
• /
• 1999

전자상거래는 구매자와 판매자 모두에게 많은 이점을 제공할 수 있어 최근 이에 관한 연구들이 많이 진행되고 있다. 특히 중소제조업체의 경우, 전자상거래라는 경영환경의 변화는 새로운 기회로 다가오고 있어, 상대적으로 기술력이 취약한 중소제조업체의 전자상거래를 지원하기 위한 요소 기술들의 개발 필요성이 점차 부각되고 있다. 이에 본 연구에서는 중소 금형제조업체의 판매과정을 사이버 공간에서 수행할 수 있는 전자상거래 기술을 개발하였다. 일반적으로 변화하는 경영환경에서는 생산과 관련된 계획과 통제가 보다 더 신속하고 정확하게 이루어져야 한다. 즉 전자상거래 환경에서의 제조업체는 구매자가 요구한 제품의 생산과 납기일을 맞추어 줄 수 있는지의 여부를 실시간으로 응답할 수 있어야 한다. 나아가서 인터넷을 통해 접수된 주문들은 해당 제조업체의 생산능력을 초과할 수 있는데 이 때에는 접수된 주문들 중에서 자사의 이익을 극대화할 수 있는 주문집합을 선별하여 접수여부를 결정해야 한다. 이와 같이 전자상거래 환경하에서의 제조업체는 생산과 관련된 정보를 신속하게 전달 받아 주문접수여부에 관한 의사결정을 올바르게 수행하는 것이 중요한데 본 연구에서는 중소 금형제조업체의 일정계획 및 주문처리를 위한 일정계획 기반의 선정 에이전트의 구조와 방법론을 제시하였다. 지금까지 일정계획에 관한 연구들은 대부분 납기일의 만족과 비용의 최소화 측면을 위주로 다루었다. 그러나 본 연구에서의 문제는 비용의 최소화보다는 납기일을 준수하면서 가장 많은 이익을 가져다 줄 수 있는 최적주문집합을 선정하는 문제를 다루고있다.자료를 수집하고, 통계분석 패키지를 이용하여 자료를 분석하였다. 방식을 결합한 하이브리드 형태이다.인터넷으로 주문처리하고, 신속 안전한 배달을 기대한다. 더불어 고객은 현재 자신의 물건이 배달되는 경로를 알고싶어 한다. 웹을 통해 물건을 주문한 고객이 자신이 물건의 배달 상황을 웹에서 모니터링 한다면 기업은 고객으로 공간적인 제약으로 인한 불신을 불식시키는 신뢰감을 주게 된다. 이러한 고객서비스 향상과 물류비용 절감은 사이버 쇼핑몰이 전국 어디서나 우리의 안방에서 자연스럽게 점할 수 있는 상황을 만들 것이다.SP가 도입되어, 설계업무를 지원하기위한 기본적인 시스템 구조를 구상하게 된다. 이와 함께 IT Model을 구성하게 되는데, 객체지향적 접근 방법으로 Model을 생성하고 UML(Unified Modeling Language)을 Tool로 사용한다. 단계 4)는 Software Engineering 관점으로 접근한다. 이는 최종산물이라고 볼 수 있는 설계업무 지원 시스템을 Design하는 과정으로, 시스템에 사용될 데이터를 Design하는 과정과, 데이터를 기반으로 한 기능을 Design하는 과정으로 나눈다. 이를 통해 생성된 Model에 따라 최종적으로 Coding을 통하여 실제 시스템을 구축하게 된다.the making. program and policy decision making, The objectives of the study are to develop the methodology of modeling the socioeconomic evaluation, and b

• Journal of Convergence for Information Technology
• /
• v.11 no.5
• /
• pp.111-117
• /
• 2021

With the rapid development of advanced science and technology including the 4th industrial revolution, the future battlefield environment is evolving at a rapid pace. In order to actively respond to issues such as reduction of military resources and shortening of service period, and to emphasize the realization of human-centered values, the Ministry of National Defense is re-establishing the role of the Army in accordance with the defense reform and is promoting the Warrior Platform, a next-generation individual combat system. In this paper, we intend to present the optimal warrior platform architecture suitable for the Korean Army by realizing the concept of future ground operations and analyzing overseas cases. We analyze the essential abilities required of individual combatants and the abilities required for each unit type, and specifically presents a plan for integration and linkage of warrior platform equipment. We also propose an efficient business promotion direction by presenting the data flow and power connection diagram between the devices that need integration and interworking.

• Journal of Convergence for Information Technology
• /
• v.11 no.5
• /
• pp.30-37
• /
• 2021

Various devices are connected to the Internet, and attacks using the Internet are occurring. Among such attacks, there are attacks that use malicious URLs to make users access to wrong phishing sites or distribute malicious viruses. Therefore, how to detect such malicious URL attacks is one of the important security issues. Among recent deep learning technologies, neural networks are showing good performance in image recognition, speech recognition, and pattern recognition. This neural network can be applied to research that analyzes and detects patterns of malicious URL characteristics. In this paper, performance analysis according to various parameters was performed on a method of detecting malicious URLs using neural networks. In this paper, malicious URL detection performance was analyzed while changing the activation function, learning rate, and neural network structure. The experimental data was crawled by Alexa top 1 million and Whois to build the data, and the machine learning library used TensorFlow. As a result of the experiment, when the number of layers is 4, the learning rate is 0.005, and the number of nodes in each layer is 100, the accuracy of 97.8% and the f1 score of 92.94% are obtained.

• Journal of Korea Entertainment Industry Association
• /
• v.14 no.8
• /
• pp.513-523
• /
• 2020

This study has derived the patents of the technology that have been filed and registered so far to investigate the trends of virtual reality(VR) experience contents technology, and analyzed them focusing on core patent technologies. The patents of Korea, USA, Japan, Europe and PCT, which were released until June 2020, were targeted, and patent search was conducted using WISDOMAIN search DB. The keywords for patent search were related to experience technology using VR, and a total of 1,013 data were obtained after creating a search formula by combining the derived keywords. Among them, a total of 65 data were extracted from the result of selecting valid patents, and a political analysis was conducted on them. Looking at the overall application trend, most of Korean patent applications accounted for, and noise patents are system-related devices to implement VR technology. The United States and Europe are focused on developing augmented reality(AR) technology, the study found. The technology of VR experience has increased rapidly since 2017, and the technology growth stage is the period from the beginning to the growth stage. As a result of examining the valid patents related to VR experience, technology was searched in various fields such as rural tour, exhibition, education, and performance, and patents for contents writing and general virtual experience related technology were also searched. If we predict the possibility of development of VR industry in the future, it is necessary to respond to preemption of intellectual property rights by proceeding technology development and patent application for more diverse fields.