• 산업경영시스템학회지
• /
• 제36권2호
• /
• pp.8-16
• /
• 2013

In this paper, we proposed a new method for the determination of strategies that are required in a continuous optimization algorithm based on the fictitious play theory. In order to apply the fictitious play theory to continuous optimization problems, it is necessary to express continuous values of a variable in terms of discrete strategies. In this paper, we proposed a method in which all strategies contain an equal number of selected real values that are sorted in their magnitudes. For comparative analysis of the characteristics and performance of the proposed method of representing strategies with respect to the conventional method, we applied the method to the two types of benchmarking functions: separable and inseparable functions. From the experimental results, we can infer that, in the case of the separable functions, the proposed method not only outperforms but is more stable. In the case of inseparable functions, on the contrary, the performance of the optimization depends on the benchmarking functions. In particular, there is a rather strong correlation between the performance and stability regardless of the benchmarking functions.

• 대한의용생체공학회:의공학회지
• /
• 제32권2호
• /
• pp.134-143
• /
• 2011

The selection of meaningful clinical tests and its reference values from a high-dimensional clinical data with imbalanced class distribution, one class is represented by a large number of examples while the other is represented by only a few, is an important issue for differential diagnosis between similar diseases, but difficult. For this purpose, this study introduces methods based on the concepts of both discernibility matrix and function in rough set theory (RST) with two discretization approaches, equal width and frequency discretization. Here these discretization approaches are used to define the reference values for clinical tests, and the discernibility matrix and function are used to extract a subset of significant clinical tests from the translated nominal attribute values. To show its applicability in the differential diagnosis problem, we have applied it to extract the significant clinical tests and its reference values between normal (N = 351) and abnormal group (N = 101) with either cholecystitis or cholelithiasis disease. In addition, we investigated not only the selected significant clinical tests and the variations of its reference values, but also the average predictive accuracies on four evaluation criteria, i.e., accuracy, sensitivity, specificity, and geometric mean, during l0-fold cross validation. From the experimental results, we confirmed that two discretization approaches based rough set approximation methods with relative frequency give better results than those with absolute frequency, in the evaluation criteria (i.e., average geometric mean). Thus it shows that the prediction model using relative frequency can be used effectively in classification and prediction problems of the clinical data with imbalanced class distribution.

• 한국지능시스템학회논문지
• /
• 제20권2호
• /
• pp.165-172
• /
• 2010

임상 데이터마이닝에서 최적의 특징 집합을 선택하는 것은 주어진 데이터로부터 생성된 모델의 복잡성을 줄일 뿐만 아니라 유용성을 향상시키는 데에 매우 중요하고, 선택된 특징들의 임계값은 질병의 감별진단을 위해 임상 전문가의 결정기준으로 사용된다. 본 논문에서는 데이터의 공간적인 분포, 즉 중첩영역에서 중복 속성값을 포함하는 데이터의 분리성 정도를 평가함으로써 연속형 속성을 가진 데이터에 대한 퍼지 이산화기법을 제안한다. 제안된 방법에서 중복 속성값의 가중치 평균값은 각 특징의 임계값(즉 경계값)을 결정하기 위해서 사용되었고, 러프집합은 전체 특징들 중에서 중요특징들의 집합을 선택하기 위해서 이용하였다. 제안된 방법의 타당성을 검증하기 위해 호흡곤란을 주호소로 내원한 668명의 환자 데이터를 근거로 3가지 이산화방법과 제안된 이산화방법에 대한 실험을 수행하였다. 실험결과, 퍼지분할을 기반으로 한 이산화방법이 하드분할을 기반으로 한 이산화방법에 비해서 평균 분류정확도와 G-mean 성능에서 보다 좋은 결과를 제공함을 확인하였다.