• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.36 no.1
• /
• pp.9-18
• /
• 2023

Predicting the compressive strength of high-performance concrete (HPC) is challenging because of the use of additional cementitious materials; thus, the development of improved predictive models is essential. The purpose of this study was to develop an HPC compressive-strength prediction model using an ensemble machine-learning method of combined bagging and stacking techniques. The result is a new ensemble technique that integrates the existing ensemble methods of bagging and stacking to solve the problems of a single machine-learning model and improve the prediction performance of the model. The nonlinear regression, support vector machine, artificial neural network, and Gaussian process regression approaches were used as single machine-learning methods and bagging and stacking techniques as ensemble machine-learning methods. As a result, the model of the proposed method showed improved accuracy results compared with single machine-learning models, an individual bagging technique model, and a stacking technique model. This was confirmed through a comparison of four representative performance indicators, verifying the effectiveness of the method.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.4
• /
• pp.173-178
• /
• 2023

As Korean literature spreads around the world, its position in the overseas publishing market has become important. As demand in the overseas publishing market continues to grow, it is essential to predict future book sales and analyze the characteristics of books that have been highly favored by overseas readers in the past. In this study, we proposed ensemble learning based prediction model and analyzed characteristics of the cumulative sales of more than 5,000 copies classified as good sellers published overseas over the past 5 years. We applied the five ensemble learning models, i.e., XGBoost, Gradient Boosting, Adaboost, LightGBM, and Random Forest, and compared them with other machine learning algorithms, i.e., Support Vector Machine, Logistic Regression, and Deep Learning. Our experimental results showed that the ensemble algorithm outperforms other approaches in troubleshooting imbalanced data. In particular, the LightGBM model obtained an AUC value of 99.86% which is the best prediction performance. Among the features used for prediction, the most important feature is the author's number of overseas publications, and the second important feature is publication in countries with the largest publication market size. The number of evaluation participants is also an important feature. In addition, text mining was performed on the four book reviews that sold the most among good-selling books. Many reviews were interested in stories, characters, and writers and it seems that support for translation is needed as many of the keywords of "translation" appear in low-rated reviews.

• Journal of KIISE:Databases
• /
• v.36 no.4
• /
• pp.247-253
• /
• 2009

Rank learning and processing have gained much attention in the IR and data mining communities for the last decade. While other data mining techniques such as classification and regression have been actively researched to interoperate with RDBMS by using the tightly coupled or loose coupling approaches, ranking has been researched independently without integrating into RDBMS. This paper proposes a tightly coupled integration of the Ranking SVM into MySQL in order to perform the rank learning task efficiently within the RDBMS. We implemented new SQL commands for learning ranking functions and predicting ranking scores. We evaluated our tightly coupled integration of Ranking SVM by comparing it to a loose coupling implementation. The experiment results show that our approach has a performance improvement of $10{\sim}40%$ in the training phase and 60% in the prediction phase.

• Korean Medical Education Review
• /
• v.11 no.2
• /
• pp.25-31
• /
• 2009

The purpose of the study was to investigate the concept and content of teaching portfolios and to consider the use of teaching portfolios in medical education. The concept of teaching portfolios has several implications and has been used in multiple approaches in teaching-learning processes. The ten foreign universities chosen for this study employ teaching portfolios in their professorship and teaching achievement evaluation as a means of deciding promotions or incentivizing employees. However, domestic universities have not yet implemented this system. It is proposed that in order to improve the quality of education programs, teaching portfolios should be used much more frequently than syllabus. Medical school professors should apply what is called "Copernicus's Thinking" to their teaching preparations.

• Journal of Korean Institute of Industrial Engineers
• /
• v.25 no.3
• /
• pp.382-392
• /
• 1999

Industrial robots have increased in both the number and applications in today's material handling systems. However, traditional approaches to robot controling have had limited success in complicated environment, especially for real time applications. One of the main reasons for this is that most traditional methods use a set of kinematic equations to figure out the physical environment of the robot. In this paper, a neural network model to solve robot manipulator's inverse kinematics problem is suggested. It is composed of two Self-Organizing Feature Maps by which the workspace of robot environment and the joint space of robot manipulator is inter-linked to enable the learning of the inverse kinematic relationship between workspace and joint space. The proposed model has been simulated with two robot manipulators, one, consisting of 2 links in 2-dimensional workspace and the other, consisting of 3 links in 2-dimensional workspace, and the performance has been tested by accuracy of the manipulator's positioning and the response time.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.1086-1091
• /
• 2005

This paper focuses on intelligent control of induction motor by hybrid system consisting of GA-PSO. Induction motor has been using in industrial area. However, it is challengeable on how we control effectively. From this point, an optimal solution using GA (Genetic Algorithm) and PSO (Particle Swarm Optimization) is introduced to intelligent control. In this case, it is possible to obtain local solution because chromosomes or individuals which have only a close affinity can convergent. To improve an optimal learning solution of control, This paper deal with applying PSO and Euclidian data distance to mutation procedure on GA's differentiation. Through this approaches, we can have global and local optimal solution together, and the faster and the exact optimal solution without any local solution. Four test functions are used for proof of this suggested algorithm.

• Education of Primary School Mathematics
• /
• v.8 no.1
• /
• pp.13-22
• /
• 2004

There were a lot of challenges to reform mathematics education. These challenges may include reforms of teaching and learning methods, development of mathematics curriculum and textbooks, innovative resources for teaching mathematics. Although there is considerable consensus that meeting these challenges will require that mathematics teachers have deep insights about mathematics, about students as learners of mathematics, and about teaching method, the teachers themselves may have little knowledge of them. The most of the professional development includes elective participation in reeducation course, workshop, and special lectures which designed to transmit a specific set of ideas, techniques, or materials to teachers. But such approaches treat mathematics teaching as routine and technical, and also provide limited opportunities for meaningful interactions within the teaching community. So, this paper suggests that what is needed to develop professional teachers of mathematics is community where teachers work with colleagues rather than working alone.

• The Journal of the Acoustical Society of Korea
• /
• v.16 no.1E
• /
• pp.64-74
• /
• 1997

In many decision making systems, rule-based approaches are used to solve complex problems in the areas of pattern analysis and computer vision. In this paper, we present methods for generating fuzzy IF-THEN rules automatically from training data for pattern classification and high-level computer vision. The rules are generated by construction minimal approximate fuzzy aggregation networks and then training the networks using gradient descent methods. The training data that represent features are treated as linguistic variables that appear in the antecedent clauses of the rules. Methods to generate the corresponding linguistic labels(values) and their membership functions are presented. In addition, an inference procedure is employed to deduce conclusions from information presented to our rule-base. Two experimental results involving synthetic and real are given.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2002.05a
• /
• pp.93-97
• /
• 2002

In this paper, we propose personalized recommendation techniques based on multidimensional scaling (MDS) method for Business to Consumer Electronic Commerce. The multidimensional scaling method is traditionally used in marketing domain for analyzing customers' perceptional differences about brands and products. In this study, using purchase history data, customers in learning dataset are assigned to specific product categories, and after then using MDS a positioning map is generated to map product categories and alternative advertisements. The positioning map will be used to select personalized advertisement in real time situation. In this paper, we suggest the detail design of personalized recommendation method using MDS and compare with other approaches (random approach, collaborative filtering, and TOP3 approach)

• Journal of the Korea Safety Management & Science
• /
• v.7 no.1
• /
• pp.199-210
• /
• 2005

Customer segmentation prediction has attracted a lot of research interests in previous literature, and recent studies have shown that artificial neural networks (ANN) method achieved better performance than traditional statistical ones. However, ANN approaches have suffered from difficulties with generalization, producing models that can overfit the data. This paper employs a relatively new machine learning technique, support vector machines (SVM), to the customer segmentation prediction problem in an attempt to provide a model with better explanatory power. To evaluate the prediction accuracy of SVM, we compare its performance with logistic regression analysis and ANN. The experiment results with real data of insurance company show that SVM superiors to them.