• Journal of Korean Port Research
• /
• v.7 no.1
• /
• pp.5-12
• /
• 1993

In complex decision making such as ill-defined system, one of the main problem is how to treat ambiguous aspect of the decision making. According to the complexity and ambiguity of the objective systems, many types of evaluation attributes are necessary for the rational decision and the relationship among the attributes become complex and fuzzy. Fuzzy integral is very effective to evalute the complex system with interaction between attributes but how to save the evaluation efforts in the decision making process of grading the membership of the objects or alternative is the problem to be tackled. Because the more object there are to evaluate, the number of decisions to made increase exponentially. Therefore, this paper aimes to propose a new evaluation algorithm based on fuzzy integral which can save the evaluator's efforts in decision making process. The proposed algorithm is constructed as follows : First, compose the fuzzy measure by introducing AHP(Analytical Hierachy Process) & mutual interaction coefficient. Second, generate fuzzy measure value of monotone family set for calculating the fuzzy integral. The effectiveness of the proposed algorithm is investigated through the example and sensitivity of interaction coefficient is illustrated.

• Korean Journal of Artificial Intelligence
• /
• v.10 no.1
• /
• pp.9-14
• /
• 2022

The demand for public bicycles operated by the Seoul Metropolitan Government is increasing every year. The size of the Seoul public bicycle project, which first started with about 5,600 units, increased to 3,7500 units as of September 2021, and the number of members is also increasing every year. However, as the size of the project grows, excessive budget spending and deficit problems are emerging for public bicycle projects, and new bicycles, rental office costs, and bicycle maintenance costs are blamed for the deficit. In this paper, the Azure Machine Learning Studio program and the Boosted Decision Tree Regression technique are used to predict the number of public bicycle rental over environmental factors and time. Predicted results it was confirmed that the demand for public bicycles was high in the season except for winter, and the demand for public bicycles was the highest at 6 p.m. In addition, in this paper compare four additional regression algorithms in addition to the Boosted Decision Tree Regression algorithm to measure algorithm performance. The results showed high accuracy in the order of the First Boosted Decision Tree Regression Algorithm (0.878802), second Decision Forest Regression (0.838232), third Poison Regression (0.62699), and fourth Linear Regression (0.618773). Based on these predictions, it is expected that more public bicycles will be placed at rental stations near public transportation to meet the growing demand for commuting hours and that more bicycles will be placed in rental stations in summer than winter and the life of bicycles can be extended in winter.

• The KIPS Transactions:PartB
• /
• v.16B no.4
• /
• pp.327-331
• /
• 2009

In fuzzy logic, aggregating uncertainties is generally achieved by means of operators such as t-norms and t-conorms. However, existing aggregation operators have some disadvantages as follows : First, they are situation-independent. Thus, they may not be properly applied to dynamic aggregation process. Second, they do not give an intuitional sense to decision making process. To solve these problems, we propose a new $\varepsilon$-AMDA (Aggregation based on the fuzzy Multidimensional Decision Analysis) algorithm to reflect degrees of strength for option i (i = 1, 2, ..., n) in the decision making process. The $\varepsilon$-AMDA algorithm makes adaptive aggregation results between min (the most weakness for an option) and max (the most strength for an option) according to the values of the parameter representing degrees of strength for an option. In this respect, it may be applied to dynamic aggregation process. In addition, it provides a mechanism of the fuzzy multidimensional decision analysis for decision making, and gives an intuitional sense to decision making process. Thus, the proposed method aids the decision maker to get a suitable decision according to the degrees of strength for options (or alternatives).

• Journal of Korean Institute of Industrial Engineers
• /
• v.23 no.1
• /
• pp.163-175
• /
• 1997

This paper presents the algorithm for finding the compensatory solution for fuzzy multiobjective nonlinear programming problem using $\gamma$-operator. The proposed algorithm can be applied to all cases with multiobjective problems since the interactive process with a decision maker is simple, various uncertainties involved on decision making are eliminated and all the objectives are well balanced. On the basis of proposed algorithm, an illustrative numerical example is presented.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1995.10b
• /
• pp.218-223
• /
• 1995

A new design scheme of a binary decision tree is proposed. In this scheme a binary decision tree is constructed by using genetic algorithm and FCM algorithm. At each node optimal or near-optimal feature or feature subset among all the available features is selected based on fitness function in genetic algorithm which is inversely proportional to classification error, balance between cluster, number of feature used. The proposed design scheme is applied to the handwtitten alphabetic characters. Experimental results show the usefulness of the proposed scheme.

• Journal of Institute of Control, Robotics and Systems
• /
• v.6 no.1
• /
• pp.98-103
• /
• 2000

This paper suggests a method to recognize the various defect patterns of a cold mill strip using a binary decision tree automatically constructed by a genetic algorithm(GA). In classifying complex patterns with high similarity like the defect patterns of a cold mill stirp, the selection of an optimal feature set and an appropriate recognizer is important to achieve high recognition rate. In this paper a GA is used to select a subset of the suitable features at each node in the binary decision tree. The feature subset with maximum fitness is chosen and the patterns are classified into two classes using a linear decision function. This process is repeated at each node until all the patterns are classified into individual classes. In this way, the classifier using the binary decision tree is constructed automatically. After constructing the binary decision tree, the final recognizer is accomplished by having neural network learning sits of standard patterns at each node. In this paper, the classifier using the binary decision tree is applied to the recognition of defect patterns of a cold mill strip, and the experimental results are given to demonstrate the usefulness of the proposed scheme.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.1
• /
• pp.43-48
• /
• 2013

Adaboost is widely used for Haar-like feature boosting algorithm in Face Detection. It shows very effective performance on single distribution model. But when detecting front and side face images at same time, Adaboost shows it's limitation on multiple distribution data because it uses linear combination of basic classifier. This paper suggest the HDCT, modified decision tree algorithm for Haar-like features. We still tested the performance of HDCT compared with Adaboost on multiple distributed image recognition.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2003.06a
• /
• pp.22-25
• /
• 2003

It is usually difficult for a single user to have all the essential knowledge on various Rapid Prototyping processes and techniques. It is therefore necessary to capture knowledge and experience of users of expert level into a decision-support system which provides quicker and more interactive way to select proper RP process and/or machine. rather than reading reports on benchmarking studies and comparing tables and graphs. In this paper two algorithms are presented, which may be used in such a decision-support system. together with its applications. The one is an extended PRES(Project Evaluation and Selection) algorithm which applies weighting factors of each attribute. The other is a LCE(Linear Confidence Equation) algorithm which is proposed to apply user's input requirements as well as weighting factors.

• Journal of the Korean Society for Precision Engineering
• /
• v.22 no.11 s.176
• /
• pp.202-209
• /
• 2005

In this paper, benchmark tests of Rapid Prototyping(RP) are presented to evaluate characteristics of various RP Systems and Processes, and several decision-support systems are developed to select RP Machine/Process suitable to user's requirements. Results of the RP benchmark tests are applied to the recently developed RP machines for the purpose of analyzing attributes such as dimensional accuracy, surface roughness, build cost, build time, and etc. Decision-making support systems are also developed, which contain not only new LCE (Linear Confidence Equation) algorithm but also modified PRES and MDS algorithm. Those algorithms are proved to be effective in that reasonably acceptable results are obtained on several cases of different inputs.

• Speech Sciences
• /
• v.10 no.1
• /
• pp.71-84
• /
• 2003

In the acoustic modeling for large vocabulary speech recognition, a sparse data problem caused by a huge number of context-dependent (CD) models usually leads the estimated models to being unreliable. In this paper, we develop a new clustering method based on the C45 decision-tree learning algorithm that effectively encapsulates the CD modeling. The proposed scheme essentially constructs a supervised decision rule and applies over the pre-clustered triphones using the C45 algorithm, which is known to effectively search through the attributes of the training instances and extract the attribute that best separates the given examples. In particular, the data driven method is used as a clustering algorithm while its result is used as the learning target of the C45 algorithm. This scheme has been shown to be effective particularly over the database of low unknown-context ratio in terms of recognition performance. For speaker-independent, task-independent continuous speech recognition task, the proposed method reduced the percent accuracy WER by 3.93% compared to the existing rule-based methods.