• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제18권2호
• /
• pp.263-283
• /
• 2024

Users' comments after online shopping are critical to product reputation and business improvement. These comments, sometimes known as e-commerce reviews, influence other customers' purchasing decisions. To confront large amounts of e-commerce reviews, automatic analysis based on machine learning and deep learning draws more and more attention. A core task therein is sentiment analysis. However, the e-commerce reviews exhibit the following characteristics: (1) inconsistency between comment content and the star rating; (2) a large number of unlabeled data, i.e., comments without a star rating, and (3) the data imbalance caused by the sparse negative comments. This paper employs Bidirectional Encoder Representation from Transformers (BERT), one of the best natural language processing models, as the base model. According to the above data characteristics, we propose the F_MixBERT framework, to more effectively use inconsistently low-quality and unlabeled data and resolve the problem of data imbalance. In the framework, the proposed MixBERT incorporates the MixMatch approach into BERT's high-dimensional vectors to train the unlabeled and low-quality data with generated pseudo labels. Meanwhile, data imbalance is resolved by Focal loss, which penalizes the contribution of large-scale data and easily-identifiable data to total loss. Comparative experiments demonstrate that the proposed framework outperforms BERT and MixBERT for sentiment analysis of e-commerce comments.

• 한국인쇄학회지
• /
• 제28권2호
• /
• pp.13-30
• /
• 2010

Existing domestic ink market had some problem because of dependence on CPP import, increasing cost, competitive price. For this reason, previous speed of research and developments are very slow. Therefore, develop the new waterborne type of non-chlorine resin, enhance the nation's position with manufacturing outstanding products. Through this thesis, the purpose is mix the polar resin and non-chlorine resin with many proportion and manufacturing better product than existing resin. Pilot production samples have run the test of adhesion, transparency, water resist and resist high pressure car washing. In conclusion, this thesis found that mix acryl Urethane resin and Non-Chlorine resin is the optimum condition.

• Management Science and Financial Engineering
• /
• 제4권2호
• /
• pp.15-42
• /
• 1998

A joint problem of order delivery, setup reduction, and cost-sharing in a two-echelon inventory system in which a vendor supplies multiple products to a group of buyers is studied here. The basic premise is that buyers have independently implemented setup reduction programs to acquire benefits from small order sizes. Doing so, however, causes the buyers' individually optimal order cycles to be differ from that of the vendor. In conjunction with this, two models are considered. In the first model, a multi-buyers single product situation is considered in which the vendor implements a joint supply cycle policy. However, buyers, as the dominant party, insist after implementing the individually optimal setup reduction that the vendor accept their individually optimal order schedules. In the second model. a multi-products, single buyer situation is considered in which the buyer implements a joint order policy. Here, the vendor, as the dominant party, refuses to cooperate fully with the buyer's individually reduced joint order schedule, and designs his own individually optimal setup reduction mix for each product under a given budget constraint. This led to a study of an integrated Setup Reduction/Break-even Pricing Policy for each situation to eliminate mismatches in individually optimal cycle times.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 2004년도 추계학술대회 및 정기총회
• /
• pp.229-244
• /
• 2004

This study addresses the problem of flexible technology acquisition in multi-product market when demands are uncertain. We confine the concept of flexibility to the ability of manufacturing system to produce a number of different types of products, called product-mix flexibility type. And an analytical model in which economies of scope is incorporated explicitly as a feature of flexible technology is presented to find the optimal investment decision to acquire flexible technology and optimal production planning. The characteristics of optimal investment strategy related to capacity and production planning are discussed.

• 대한산업공학회지
• /
• 제31권2호
• /
• pp.135-142
• /
• 2005

This study addresses the problem of flexible technology acquisition in a multi-product market when demands are uncertain. We confine the concept of flexibility to the ability of manufacturing system to produce a number of different types of products, called a product-mix flexibility type. And an analytical model in which the economies of scope are incorporated explicitly as a feature of flexible technology is presented to find the optimal investment decision to acquire flexible technology and optimal production planning. The characteristics of optimal investment strategy related to capacity and production planning are discussed.

• 대한산업공학회지
• /
• 제21권3호
• /
• pp.329-343
• /
• 1995

Traditionally, machining process sequence was influenced and constrained by the design information obtained from CAD data base, i.e., class of operations, geometric shape, tooling, geometric tolerance, etc. However, even though all the constraints from design information are considered, there may exist more than one way to feasibly machine parts. This research is focused on the integrated problem of operations sequencing and machine tools selection in the presence of the product mix and their production volumes. With the transitional costs among machining operations, the operation sequencing problem can be formulated as a well-known Traveling Salesman Problem (TSP). The transitional cost between two operations is expressed as the sum of total machining time of the parts on a machine for the first operation and transportation time of the parts from the first machine to a machine for the second operation. Therefore, the operation sequencing problem formulated as TSP cannot be solved without transitional costs for all operation pairs. When solved separately or serially, their mutual optima cannot be guaranteed. Machining operations sequencing and machine tool selection problems are two core problems in process planning for discretely machined parts. In this paper, the interrelated two problems are integrated and analyzed, zero-one integer programming model for the integrated problem is formulated, and the solution methods are developed using a Tabu Search technique.

• 산업공학
• /
• 제25권2호
• /
• pp.264-275
• /
• 2012

The process of semiconductor(IC Package) manufacturing usually includes lots of complex and sequential processes. Many kinds of equipments are installed with the mixed concept of serial and parallel manufacturing system. The business environments of the semiconductor industry have been changed frequently, because new technologies are developed continuously. It is the main reason of new investment plan and layout consideration. However, it is difficult to change the layout after installation, because the major equipments are expensive and difficult to move. Furthermore, it is usually a multiple-objective problem. Thus, new investment or layout change should be carefully considered when the production environments likewise product mix and production quantity are changed. This paper introduces a simulation case study of a Korean company that produces packaging substrates(especially lead frames) and requires multi-objective decision support. $QUEST^{(R)}$ is used for simulation modelling and AHP(Analytic Hierarchy Process) and DEA(Data Envelopment Analysis) are used for weighting of qualitative performance measures and solving multiple-objective layout problem, respectively.

• International Journal of Advanced Culture Technology
• /
• 제6권3호
• /
• pp.88-97
• /
• 2018

As the trade among the countries increases, domestic agriculture is facing a crisis. Especially domestic rice industry is suffering from the rice market opening problem and the domestic rice industry itself. As a counter measure against this, despite the promotion of the rice processing food industry, the export of rice processed food has decreased by an average annual rate of 12.3% since 2011 due to the lack of customized export strategy, such as low price competitiveness, insufficient quality and brand management. As a measure, we selected indicators to prioritize establishment of export strategy of rice processed food business and used AHP methodology to draw the weight for each indicators. The selection of export strategy prioritization indicator was based on the marketing mix theory, various previous research reviews and expert consulting. Five higher level indicators and nineteen detailed indicators were derived, and a weighted value of them was analyzed using a professional package. As a result, the relative importance of higher level indicators was 45.5% for product differentiation, 18.4% for management competence, 13.0% for export infrastructure, 12.0% for product pricing, and 11.1% for product distribution. The relative importance of the lower level indicators was 17.3% for product safety, 15.3% for product quality, 9.1% for management expertise, and 7.0% for product brand. This result is expected to be used as export strategy indicators of rice processed food export business.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2004년도 심포지엄 논문집 정보 및 제어부문
• /
• pp.98-101
• /
• 2004

Linear Programming(LP) is the term used for defining a wide range of optimization problems in which the objective function to be minimized or maximized is linear in the unknown variables and the constraints are a combination of linear equalities and inequalities. LP problems occur in many real-life economic situations where profits are to be maximized or costs minimized with constraint limits on resources. While the simplex method introduced in a later reference can be used for hand solution of LP problems, computer use becomes necessary even for a small number of variables. Problems involving diet decisions, transportation, production and manufacturing, product mix, engineering limit analysis in design, airline scheduling, and so on are solved using computers. This technique is called Sequential Linear Programming (SLP). This paper describes LP's problems and solves a LP's problems using the neural networks.

• 한국건설순환자원학회논문집
• /
• 제5권3호
• /
• pp.305-312
• /
• 2017

본 연구에서는 국내 화력발전소에서 발생되고 있는 산업부산물을 활용하여 아스팔트 콘크리트용 채움재를 개발 및 혼합물에 적용하여 함량에 따른 특성 및 소성변형 저항성에 미치는 영향을 관찰하였다. 실험결과 석탄회 기반 채움재를 사용할 경우 최적아스팔트함량(OAC)는 기존 채움재 대비 0.1% 낮은 값을 나타내었다. 도출된 최적아스팔트함량에서 채움재 함량에 따른 간접인장강도 및 인장강도비 측정 결과 함량 증가에 따라 높은 성능을 발현하였고, 2.0% 이상의 함량에서 국내 단체표준 품질관리 기준 각 0.8 이상과 0.75 이상을 만족하는 결과를 얻을 수 있었다. 동적안정도는 채움재 2.5%의 함량에서 가장 높은 저항성을 나타내었으며, 함량이 2.5%를 넘어설 때 오히려 성능발현이 저하되는 결과를 나타내었다. 과도한 채움재의 혼입은 상대적으로 아스팔트 함량이 감소되기 때문에 취성파괴를 유발할 수 있으므로 최적의 채움재/아스팔트 의 비율을 찾아 적용하는 방안이 필요할 것으로 판단된다.