Object Classification Method Using Dynamic Random Forests and Genetic Optimization
1
조회
0
다운로드
3
인용
0
참고문헌
키워드
초록
In this paper, we proposed the object classification method using genetic and dynamic random forest consisting of optimal combination of unit tree. The random forest can ensure good generalization performance in combination of large amount of trees by assigning the randomization to the training samples and feature selection, etc. allocated to the decision tree as an ensemble classification model which combines with the unit decision tree based on the bagging. However, the random forest is composed of unit trees randomly, so it can show the excellent classification performance only when the sufficient amounts of trees are combined. There is no quantitative measurement method for the number of trees, and there is no choice but to repeat random tree structure continuously. The proposed algorithm is composed of random forest with a combination of optimal tree while maintaining the generalization performance of random forest. To achieve this, the problem of improving the classification performance was assigned to the optimization problem which found the optimal tree combination. For this end, the genetic algorithm methodology was applied. As a result of experiment, we had found out that the proposed algorithm could improve about 3~5% of classification performance in specific cases like common database and self infrared database compare with the existing random forest. In addition, we had shown that the optimal tree combination was decided at 55~60% level from the maximum trees.
Abstract
In this paper, we proposed the object classification method using genetic and dynamic random forest consisting of optimal combination of unit tree. The random forest can ensure good generalization performance in combination of large amount of trees by assigning the randomization to the training samples and feature selection, etc. allocated to the decision tree as an ensemble classification model which combines with the unit decision tree based on the bagging. However, the random forest is composed of unit trees randomly, so it can show the excellent classification performance only when the sufficient amounts of trees are combined. There is no quantitative measurement method for the number of trees, and there is no choice but to repeat random tree structure continuously. The proposed algorithm is composed of random forest with a combination of optimal tree while maintaining the generalization performance of random forest. To achieve this, the problem of improving the classification performance was assigned to the optimization problem which found the optimal tree combination. For this end, the genetic algorithm methodology was applied. As a result of experiment, we had found out that the proposed algorithm could improve about 3~5% of classification performance in specific cases like common database and self infrared database compare with the existing random forest. In addition, we had shown that the optimal tree combination was decided at 55~60% level from the maximum trees.
참고문헌 (25)
- D. E. Rumelhart, "Parallel distributed processing: explorations in the microstructure of cognition" (1986), MIT Press
- C. J. C. Burges, "A tutorial on support vector machines for pattern recognition", Journal of Data Mining and Knowledge Discovery, 2, pp.121~167 (1998)
- 염해영, "소속 함수와 유전자 정보의 신경망을 이용한 유전자 타입의 분류", 전자공학회논문지 - CI, 42(4), pp.251~260 (2005), 대한전자공학회
- 강선경, "인공신경망을 이용한 마커 검출 및 인식의 정확도 개선", 한국컴퓨터정보학회논문지, 13(1), pp.89~98 (2008), 한국컴퓨터정보학회
- 김광성, "불균형 데이터 학습을 위한 지지벡터기계 알고리즘", 한국컴퓨터정보학회논문지, 15(7), pp.11~17 (2010), 한국컴퓨터정보학회
- 김재협, "적외선 영상에서의 시계열 특징 추출을 이용한 Gunnery 분류 기법 연구", 한국컴퓨터정보학회논문지, 19(10), pp.43~53 (2014), 한국컴퓨터정보학회
- Robi Polikar, "Ensemble based systems in decision making", IEEE Circuit and Systems, 6(3), pp.21~45 (2006)
- P. Viola, "Robust real-time face detection", Int. Journal of Computer Vision, 57(2), pp.137~154 (2004)
- L. Breiman, "Random forest", Machine Learning, 45, pp.5~32 (2001)
- 김재협, "아이다부스트(Adaboost)와 원형기반함수를 이용한다중표적 분류 기법", 전자공학회논문지 - CI, 47(3), pp.22~28 (2010), 대한전자공학회
- 정경중, "얼굴정렬과 AdaBoost를 이용한 얼굴 표정 인식", 전자공학회논문지, 51(11), pp.193~202 (2014), 대한전자공학회
- L. Breiman, "Chapman and Hall", ~ (1993)
- R. Banfield, "A comparison of ensemble creation technique", Proc. of Multiple Classifier Systems, 1, pp.223~232 (2004)
- S. Bernard, "Influence of hyperparameters on random forest accuracy", Proc. of Workshop on Multiple Classifier Systems, 1, pp.171~180 (2009)
- S. Bernard, "On the selection of decision trees in random forest", Proc. of Joint Conf. on Neural Networks, pp.302~307 (2009)
- E. Tripoli, "Dynamic construction of random forests: Evaluation using biomedical engineering problems", Proc. of IEEE Int. Conference on Information Technology and Application in Biomedicine, 1, pp.3~5 (2010)
- S. Bernard, "Dynamic random forests", Journal of Pattern Recognition Letters, 33(12), pp.1580~1586 (2012)
- F. Roli, "Methods for designing multiple classifier systems", Proc. of 2nd International Workshop MCS2001, pp.78~87 (2001)
- I. S. Oh, "Hybrid genetic algorithms for feature selection", IEEE Trans. on Pattern Analysis and Machine Intelligence, 26(11), pp.1424~1437 (2004)
- 허기수, "신경망의 노드 가지치기를 위한 유전 알고리즘", 전자공학회논문지 - CI, 46(2), pp.65~74 (2009), 대한전자공학회
- 김철문, "PCA 복원과 HOG 특징 기술자 기반의 효율적인보행자 인식 방법", 전자공학회논문지, 50, pp.162~170 (2013), 대한전자공학회
- R. Maclin, "Boosting Classifiers Regionally", Proc. of the 15th National Conference on Artificial Intelligence, 1, pp.700~705 (1998)
- Venkatadri M., "A multiobjective genetic algorithm for feature selection in data mining", Journal of Computer Science and Information Technology, 1(5), pp.443~448 (2010)
- T. G. Dietterich, "An experimental comparison of three methods for constructing ensembles of decision trees : bagging, boosting, and randomization", Journal of Machine Learning, 40(2), pp.139~157 (2010)