• Annual Conference of KIPS
• /
• 2023.05a
• /
• pp.647-649
• /
• 2023

Sentiment analysis as part of natural language processing (NLP) has received much attention following the demand to understand people's opinions. Aspect-based sentiment analysis (ABSA) is a fine-grained subtask from sentiment analysis that aims to classify sentiment at the aspect level. Throughout the years, researchers have formulated ABSA into various tasks for different scenarios. Unlike the early works, the current ABSA utilizes many elements to improve performance and provide more details to produce informative results. These ABSA formulations have provided greater challenges for researchers. However, it is difficult to explore ABSA's works due to the many different formulations, terms, and results. In this paper, we conduct a comparative analysis of recent studies on ABSA. We mention some key elements, problem formulations, and datasets currently utilized by most ABSA communities. Also, we conduct a short review of the latest papers to find the current state-of-the-art model. From our observations, we found that span-level representation is an important feature in solving the ABSA problem, while multi-task learning and generative approach look promising. Finally, we review some open challenges and further directions for ABSA research in the future.

• Journal of the Society of Naval Architects of Korea
• /
• v.32 no.3
• /
• pp.62-71
• /
• 1995

In the numerical analysis of incompressible unsteady Navier-stokes equation, large time is required for solving the pressure Poisson equation of the elliptic type at each time step. In this paper, a numerical analysis by the direct method is carried out to solve the pressure Poisson equation and the computing time is analyzed as mesh size increases. The pressure Poisson equation can be transformed to the boundary value problem by the Green theorem. The computing time for the convolution type of the domain integral can be reduced by using F.F.T. and the computing time in the direct method depends entirely on obtaining the solution of the boundary value problem. The numerical analysis on the known solutions is carried out and compared for the verification of the direct method. And the numerical analysis on the body boundary and domain decomposition problem are carried out with the computing time less than O($n^{3}$) in the (n.n) mesh.

• Structural Engineering and Mechanics
• /
• v.53 no.2
• /
• pp.205-226
• /
• 2015

Finite element method (FEM) is an effective quantitative method to solve complex engineering problems. The basic idea of FEM for a complex problem is to be able to find a solution by reducing the problem made simple. If mathematical tools are inadequate to obtain precise result, even approximate result, FEM is the only method that can be used for structural analyses. In FEM, the domain is divided into a large number of simple, small and interconnected sub-regions called finite elements. FEM has been used commonly for linear and nonlinear analyses of different types of structures to give us accurate results of plane stress and plane strain problems in civil engineering area. In this paper, FEM is used to investigate stress analysis of a shear wall which is subjected to concentrated loads and fundamental principles of stress analysis of the shear wall are presented by using matrix displacement method in this paper. This study is consisting of two parts. In the first part, the shear wall is discretized with constant strain triangular finite elements and stiffness matrix and load vector which is attained from external effects are calculated for each of finite elements using matrix displacement method. As to second part of the study, finite element analysis of the shear wall is made by ANSYS software program. Results obtained in the second part are presented with tables and graphics, also results of each part is compared with each other, so the performance of the matrix displacement method is demonstrated. The solutions obtained by using the proposed method show excellent agreements with the results of ANSYS. The results show that this method is effective and preferable for the stress analysis of shell structures. Further studies should be carried out to be able to prove the efficiency of the matrix displacement method on the solution of plane stress problems using different types of structures.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.42 no.5
• /
• pp.659-669
• /
• 2022

Recently, parking problems have repeatedly occurred not only in the residential areas of the old town, but in detached residential areas, neighborhood living facilities, and commercial areas of the planned New Town. This study suggests a strategy to minimize parking problems before constructing a New Town by reviewing these parking problems at the new city district planning stage. Parking problems can be divided into supply-demand and non-supply-based, and the solution strategy is reviewed in terms of supply and operation. The procedure for applying the solution strategy is proposed according to the effectiveness and ease of application.First, this paper suggests the PAZ (Parking Analysis Zone) as the basic analysis unit. Second, the supply-based parking problem in the concerned area based on the land use plan of Hanam Gyosan is reviewed. Last, solutions to a parking problem for each PAZ are presented.

• Journal of Families and Better Life
• /
• v.25 no.5
• /
• pp.31-46
• /
• 2007

The purpose of this study was to determine the differences in internet use motivations and problem behaviors among vocational high school students according to internet addiction and family system type. Five hundred vocational high school students from Daegu took part in the survey. The questionnaire survey consisted of an internet addiction self-assessment scale, a family system type scale, an internet use motivation scale, and a problem behavior scale. The data was analysed using MANOVA and Cronbach a, and the $Scheff\acute{e}$ test was used for post-hoc analysis. The major findings of this study were as follows: First, the study found significant differences in adolescent internet use motivation according to internet addiction. Adolescents in the high dangerous user group showed higher communicative, experiential, and addictive motivations than adolescents in potentially dangerous and normal user groups. Also, there were no significant differences in informative motivation according to internet addiction. Regardless of internet addiction, adolescents tend to use the internet to search for information and to solve problems. Second, there were significant differences in adolescents' problem behaviors according to internet addiction. More specifically, adolescents who were in the high dangerous user group displayed more problem behaviors toward their body, family, school, and society than adolescents in the potentially dangerous and the normal user groups. Third, an adolescent's family system type significantly influenced his/her internet use motivation. Adolescents from an extreme family showed higher communicative and addictive motivations than adolescents from middle and balanced families. Fourth, there were significant differences in the problem behaviors of adolescents according to family system types. Adolescents from an extreme family showed more problem behaviors toward their body, family, school, and society than adolescents from middle and balanced families.

• Journal of Gifted/Talented Education
• /
• v.23 no.2
• /
• pp.289-310
• /
• 2013

The purpose of the study is to investigate science process skills and suggest several considerations about developing scientific inquiries for secondary science gifted students. To do this, we analyzed scientific inquiries of science gifted programs and evaluated them on the quantity of problem perception, problem finding and inquiry planning that are regarded as high level science process skills, then revised each inquiry to include those high level skills. The result was that the first, there were differences in frequencies and types of science process skills among those inquiries. The second, there were very few problem perception and problem finding and were not many inquiry planning. The third, some of the revised inquiries showed those high level skills. From this, we would like to suggest we should construct scientific inquiries of science gifted program out of many and various themes. And there should be more high level science process skills such as problem perception, problem finding, and inquiry planning. For this, scientific inquiry developers should have intentions to involve such science process skills which is appropriate for science gifted student.

• Journal of Elementary Mathematics Education in Korea
• /
• v.20 no.4
• /
• pp.563-581
• /
• 2016

Studies on meta-affect in problem solving tried to build similar structures among affective elements as the structure of cognition and meta-cognition. But it's still need to be more systematic as meta-cognition. This study defines meta-affect as the connection of cognitive elements and affective elements which always include at least one affective element. We logically categorized types of meta-affect in problem solving, and then observed and analyzed the real cases for each type of meta-affect based on the logical categories. We found the operating mechanism of meta-affect in mathematical problem solving. In particular, we found the characteristics of meta function which operates in the process of problem solving. Finally, this study contributes in efficient analysis of meta-affect in problem solving and educational implications of meta-affect in teaching and learning in problem solving.

• Education of Primary School Mathematics
• /
• v.19 no.3
• /
• pp.177-191
• /
• 2016

Students can calculate the addition and subtraction problem using informal knowledge before receiving the formal instruction. Recently, the value that a computation lesson focus on the understanding and developing the various strategies is highlighted by curriculum developers as well as in reports. Ideally, a educational setting and classroom culture reflected students' learning and problem solving strategies. So, this paper analyzed the similarity and difference with respect to the numeric sentence and word problem in the addition and subtraction. The subjects for the study were 100 third-grade Korean students and 68 third-grade American students. Researcher developed the questionnaire in the addition and subtraction and used it for the survey. The following results have been drawn from this study. The computational ability of Korean students was higher than that of American students in both the numeric sentence and word problem. And it was revealed the differences of the strategies which were used problem solving process. Korean students tended to use algorithms and numbers' characters and relations, but American students tended to use the drawings and algorithms with drawings.

• Journal of Elementary Mathematics Education in Korea
• /
• v.1 no.1
• /
• pp.67-86
• /
• 1997

The purposes of this paper are to show problem-solving strategies and their typical problems to suggest specific ways to teach strategies to promote problem-solving abilities. (1) Problem-solving strategies can be divided into general strategies and specific strategies. General strategies refer to procedural teaching-learning activities based on Polya's 4 step problem-solving. Specific strategies refer to Lenchner's 12 problem solving strategies and their characteristics which are helpful to the substantial solution of specific problems. (2) Concerning to problem-solving strategies teaching, the followings are suggested. First, the sequence of strategy teaching should be from easy to difficult ones, from short to long ones. Second problems for strategy training should be simple and good enough to serve as examples of the strategies. Repetition with similar problems are needed. Third, analysis and comparison of various strategies, and extension and adaptation of the strategies to complicate problems are needed. Fourth, procedures of strategies teaching are the follows: Have students make their own strategies focused on the solution process; Have students solve the problems with expectation of the solving methods; Have students compare and reflect on their solving methods; And assess problem - solving processes.

• Journal of The Korean Association of Information Education
• /
• v.21 no.4
• /
• pp.425-435
• /
• 2017

The purpose of this study is to explore the differences between Korea and Japan in terms of the effects of student- and school-level ICT-related factors on computer-based problem solving ability in PISA 2012 achievement results. PISA 2012's computer-based problem solving assessment included 5,033 students from 156 schools in Korea and 6,351 students from 191 schools in Japan. A 2-level hierarchical linear model (HLM) was employed to analyze data. Both of Korean and Japanese students, students with the earlier use of computer and the Internet, higher socioeconomic background and higher openness to problem reported better problem solving scores than their counterparts. Also, accessibility to ICT equipment provided by home or school was found to be a negative effect on problem solving in both countries. Differences in the effects of purposes and attitudes for ICT use, teacher-student relationship, and creative out-of-school activities on problem solving ability were found between two countries.