앙상블 학습 알고리즘을 이용한 컨벌루션 신경망의 분류 성능 분석에 관한 연구 = A Study on Classification Performance Analysis of Convolutional Neural Network using Ensemble Learning Algorithm

다국어 초록 (Multilingual Abstract)

In this paper, we compare and analyze the classification performance of deep learning algorithm Convolutional Neural Network(CNN) ac cording to ensemble generation and combining techniques. We used several CNN models(VGG16, VGG19, DenseNet121, DenseNet169, DenseNet201, ResNet18, ResNet34, ResNet50, ResNet101, ResNet152, GoogLeNet) to create 10 ensemble generation combinations and applied 6 combine techniques(average, weighted average, maximum, minimum, median, product) to the optimal combination. Experimental results, DenseNet169-VGG16-GoogLeNet combination in ensemble generation, and the product rule in ensemble combination showed the best performance. Based on this, it was concluded that ensemble in different models of high benchmarking scores is another way to get good results.

참고문헌 (Reference)

1 정윤주, "딥 컨볼루션 신경망을 이용한 자동차번호판 영역 검출 시스템" 한국멀티미디어학회 20 (20): 1166-1174, 2017

2 K. Simonyan, "Very Deep Convolutional Networks for Large-Scale Image Recognition"

3 X. Glorot, "Understanding the Difficulty of Training Deep Feedforward Neural Networks" 249-256, 2010

4 "The BellKor Solution to the Netflix Grand Prize"

5 "Tensorflow Speech Recognition Challenge"

6 M. Abadi, "TensorFlow : Largescale Machine Learning on Heterogeneous Systems"

7 L. Bottou, "Stochastic Gradient Descent Tricks" 7700 : 430-445, 2012

8 "State Farm Distracted Driver Detection"

9 "Santander Product Recommendation"

10 V. Nair, "Rectified Linear Units Improve Restricted Boltzmann Machines" 807-814, 2010

11 J. Bergstra, "Random search for hyper-parameter optimization" 13 (13): 281-305, 2012

12 T. Tieleman, "RMSProp:Divide the Gradient by a Running Average of I ts Recent Magnitude" COURSERA 2012

13 J. Snoek, "Practical Bayesian Optimization of Machine Learning Algorithms" 2 : 2951-2959, 2012

14 "Porto Seguro’s Safe Driver Prediction"

15 I. Sutskever, "On the Importance of Initialization and Momentum in Deep Learning" 28 : 1139-1147, 2013

16 K. Janocha, "On Loss Functions for Deep Neural Networks in Classification"

17 J. Kittler, "On Combining Classifiers" 20 (20): 226-239, 1998

18 M. Lin, "Network in Network"

19 D. Ciresan, "Multi-column Deep Neural Networks for Image Classification" 3642-3649, 2012

20 "Keras"

21 "Kaggle: Your Home for Data Science"

22 K. Alex, "ImageNet Classification with Deep Convolutional Neural Networks" 1097 (1097): 1097-1105, 2012

23 Y. Lecun, "Gradient-based Learning Applied to Document Recognition" 86 (86): 2278-2324, 1998

24 C. Szegedy, "Going Deeper with Convolutions"

25 G. Huang, "Densely Connected Convolutional Networks" 2261-2269, 2017

26 K. He, "Deep Residual Learning for Image Recognition"

27 L. Yann, "Deep Leaning" 521 (521): 436-444, 2015

28 G. Fumera, "A Theoretical and Experimental Analysis of Linear Combiners for Multiple Classifier Systems" 27 (27): 942-956, 2005

29 S. W. Park, "A Study on Classification of Convolutional Neural Network using Ensemble Combining Technique" 757-, 2018