基于BP神经网络的自适应伪最近邻分类

曾勇; 舒欢; 胡江平; 葛月月

doi:10.11999/JEIT160133

基于BP神经网络的自适应伪最近邻分类

doi: 10.11999/JEIT160133

基金项目:

国家自然科学基金(61104104, 61473061)，四川省信号与信息重点实验室基金(SZJJ2009-002)

计量
- 文章访问数: 1092
- HTML全文浏览量: 129
- PDF下载量: 656
- 被引次数: 0
出版历程
- 收稿日期: 2016-01-29
- 修回日期: 2016-06-17
- 刊出日期: 2016-11-19

Adaptive Pseudo Nearest Neighbor Classification Based on BP Neural Network

Funds:

The National Natural Science Foundation of China (61104104, 61473061), The Fund of Sichuan Provincial Key Laboratory of Signal and Information Processing (SZJJ2009-002)

摘要

摘要: 在伪最近邻(PNN)分类算法中，待分类样本点与每一类样本集中各个近邻的距离加权系数都是主观确定的，这就使得算法得不到最优距离加权值。针对这一问题，该文提出一种基于BP神经网络的自适应伪最近邻分类算法。首先通过计算待分类样本点与每一类样本集中各个近邻的距离值，并将其作为BP神经网络的输入。然后根据BP神经网络输入与输出之间的映射来自适应确定相应的距离加权值。最后由BP神经网络的输出值判别样本类别号。实验结果表明，该算法能够自适应地调节距离加权系数，同时还能有效地改善分类准确率。
- 伪最近邻分类 /
- BP神经网络 /
- 自适应
Abstract: Distance-weighted coefficients between unlabeled sample point and its nearest neighbors belonging to same sample set are determined subjectively in the Pseudo Nearest Neighbor (PNN) classification algorithm, which makes it difficult to obtain optimal distance-weighted value. In this paper, an adaptive pseudo neighbor classification algorithm based on BP neural network is proposed. Firstly, the distance-weighted values between unlabeled sample point and its neighbors lying in the same sample set are regarded as the input of BP neural network. Secondly, the corresponding distance-weighted values are adaptively determined according to the mapping between the inputs and outputs of BP neural network. Finally, the classification of unlabeled sample point is judged by the outputs of BP neural network. Experimental results show that the proposed approach adaptively adjusts the distance-weighted coefficients. Moreover, the classification accuracy can be effectively improved.
- Pseudo Nearest Neighbor (PNN) classification /
- BP neural network /
- Adaptive

HTML全文

参考文献(19)

WU Xindong, KUMAR V, QUINLAN J R, et al. Top 10 algorithms in data mining[J]. Knowledge and Information Systems, 2008, 14(1): 1-37. doi: 10.1007/s10115-007-0114-2.

MATEI O, POP P C, and VLEAN H. Optical character recognition in real environments using neural networks and k-nearest neighbor[J]. Applied Intelligence, 2013, 39(4): 739-748. doi: 10.1107/s10489-013-0456-2.

WAN C H, LEE L H, RAJKUMAR R, et al. A hybrid text classi cation approach with low dependency on parameter by integrating k-nearest neighbor and support vector machine[J]. Expert Systems with Applications, 2012, 39(15): 11880-11888. doi: 10.1016/j.eswa.2012.02.068.

CARAWAY N M, MCCREIGHT J L, and RAJAGOPALAN B. Multisite stochastic weather generation using cluster analysis and k-nearest neighbor time series resampling[J]. Journal of Hydrology, 2014, 508: 197-213. doi: 10.1016/ j.jhydrol.2013.10.054.

RAHMAN S A, HUANG Y, CLAASSEN J, et al. Combining Fourier and lagged k-nearest neighbor imputation for biomedical time series data[J]. Journal of Biomedical Informatics, 2015, 58: 198-207. doi: 10.1016/j.jbi.2015. 10.004.

GONZLEZ Mabel, BERGMEIR Christoph, TRIGUERO Isaac, et al. On the stopping criteria for k-Nearest Neighbor in positive unlabeled time series classi cation problems[J]. Information Sciences, 2016, 328: 42-59. doi:10.1016/j.ins. 2015.07.061.

WANG A, AN N, CHEN G, et al. Accelerating wrapper- based feature selection with k-nearest-neighbor[J]. Knowledge-Based Systems, 2015, 83: 81-91. doi: 10.1016/ j.knosys.2015.03.009.

CHEN C H, HUANG W T, Tan T H, et al. Using K-nearest neighbor classification to diagnose abnormal lung sounds[J]. Sensors, 2015, 15(6): 13132-13158. doi: 10.3390/s150613132.

HAN Y, PARK K, HONG J, et al. Distance-constraint k-nearest neighbor searching in mobile sensor networks[J]. Sensors, 2015, 15(8): 18209-18228. doi: 10.3390/s150818209.

TOMAEV N and MLADENIC D. Hubness-aware shared neighbor distances for high-dimensional k-nearest neighbor classification[J]. Knowledge and Information Systems, 2014, 39(1): 89-122. doi: 10.1007/s10115-012-0607-5.

CHOI Sangil, YOUN Ik-hyun, LEMAY Richelle, et al. Biometric gait recognition based on wireless acceleration sensor using k-nearest neighbor classification[C]. 2014 IEEE International Conference on Computing, Networking and Communications (ICNC), Honolulu, HI, 2014: 1091-1095. doi: 10.1109/ICCNC.2014.6785491.

DUDANI S A. The distance-weighted k-nearest-neighbor rule[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1976, 6(4): 325-327. doi: 10. 1109/TSMC. 1976.5408784.

GOU Jianping, XIONG Taisong, and KUANG Yin. A novel weighted voting for k-nearest neighbor rule[J]. Journal of Computers, 2011, 6(5): 833-840. doi: 10.4304/jcp.6.5.833 -840.

GOU Jianping, DU Lan, ZHANG Yuhong, et al. A new distance-weighted k-nearest neighbor classier[J]. Journal of Information Computational Science, 2012, 9(6): 1429-1436.

BAILY T and JAIN A K. A note on distance-weighted k-nearest neighbor rules[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1978, 8(4): 311-313. doi: 10.1109/ TSMC. 1978.4309958.

MORIN R L and RAESIDE B E. A reappraisal of distance-weighted k-nearest-neighbor classification for pattern recognition with missing data[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1981, 11(3): 241-243. doi: 10.1109/TSMC.1981.4308660.

ZENG Yong, YANG Yupu, and ZHAO Liang. Pseudo nearest neighbor rule for pattern classification[J]. Expert Systems with Applications, 2009, 36: 3587-3595. doi: 10.1016/ j.eswa.2008.02.003.

杨凡, 赵建民, 朱信忠. 一种基于BP神经网络的车牌字符分类识别方法[J]. 计算机科学, 2005, 32(8): 192-195.

YANG Fan, ZHAO Jianmin, and ZHU Xinzhong. A new method of license plate characters classified recognition based on BP neural networks[J]. Computer Science, 2005, 32(8): 192-195.

施引文献

资源附件(0)

访问统计