用于文本相似度计算的新核函数

doi:10.3969/j.issn.1000-436x.2012.12.006

Abstract

Abstract:

To enhance the performance of detecting similar documents,a novel kernel function named S_Wang kernel was constructed.Based on the actual situation of computing text similarity,the S_Wang kernel was newly bu lt with consideration of the Euclidean distance and angle between vectors that represented the text documents to be compared.It was proved that the function could be constructed as a kernel function according to Mercer theorem.Experimental verification of the performance of the kernels in the text document similarity calculation was provided.The results show that the S_Wang kernel is significantly better than the precision and F1 performance of other kernels like Cauchy kernel,Latent Semantic Kernel (LSK) and CLA kernel.S_Wang kernel is suitable for text similarity computation.

Key words: information retrieval, text similarity, kernel function, S_Wang kernel, LSK, Cauchy kernel, CLA kernel

Xiu-hong WANG,Shi-guang JU. Novel kernel function for computing the similarity of text[J]. Journal on Communications, 2012, 33(12): 43-48.

Figures/Tables 5

References 14

[1]	Kernel functions for machine learning applications[EB/OL]. , 2011.
[2]	MARTINS A , FIGUEIREDO M , AGUIAR P . Kernels and similar y measures for text classification[A]. Proceedings of ConfTele’2007[C]. New York,USA, 2007. 1-4.
[3]	SAHAMI M , HEILMAN T D . A Web-based kernel function for measuring the similarity of short text snippets[A]. Proceedings of WWW’06[C]. New York,USA, 2006. 1-10.
[4]	TIAN Y , LI H , CAI Q ,et al. Measuring the similarity of short texts by word similarity and tree kernels[A]. Proceedings of Conf YC-ICT’10[C]. New York,USA, 2010. 69-72.
[5]	HOFMANN T . Learning the Similarity of Documents:an Information-Geometric Approach to Document Retrieval and Categorization[M]. Cambridge,MA: The MIT Press, 2000.
[6]	王秀红, 鞠时光 . 基于混合核函数的分布式信息检索结果融合[J]. 通信学报, 2011,32(4): 112-118. WANG X H , JU S G . Result merging method based on combined kernels for distributed information retrieval[J]. Journal on Communications, 2011,32(4): 112-118.
[7]	王国胜 . 核函数的性质及其构造方法[J]. 计算机科学, 2006,33(6): 171-174,178. WANG G S . Properties and construction methods of kerne in support vector machine[J]. Computer Science, 2006,33(6): 171-174,178.
[8]	任双桥, 魏玺章, 黎湘 ,等. 基于特征可分性的核函数自适应构造[J]. 计算机学报, 2008,31(5): 803-809. REN S Q , WEI X Z , LI X ,et al. Adaptive construction for kernel function based on the feature discriminability[J]. Chinese Journal of Computers, 2008,31(5): 803-809.
[9]	王华忠, 俞金寿 . 核函数方法及其模型选择[J]. 江南大学学报(自然科学版), 2006,5(4): 500-504. WANG H Z , YU J S . Study on the kernel based methods and its model selection[J]. Journal of Southern Yangtze University(Natural Science Edition), 2006,5(4): 500-504.
[10]	吴涛, 贺汉根, 贺明科 . 基于插值的核函数构造[J]. 计算机学报, 2003,26(8): 990-996. WU T , HE H G , HE M K . Interpolation based dernel funct n's construction[J]. Chinese Journal of Computers, 2003,26(8): 990-996.
[11]	SHAWE-TAYLOR J , CRISTIANINI N . Kernel Methods for Pattern Analysis[M]. Lodon: Cambridge University PressPress, 2004.
[12]	CRISTIANINI N , SHAWE-TAYLOR J , LODHI H . Latent semantic kernels[J]. Journal of Intelligent Information Systems, 2002,18(23): 127-152.
[13]	BASAK J . A least square kernel machine with box constraints[A]. 19th International Conference on Pattern Recognition[C]. New York: IEEE Press, 2008. 1-4.
[14]	CHANG C C , LIN C J . LIBSVM:a library for support vector machines[EB/OL]. , 2001.

Metrics

Recommended 0

No Suggested Reading articles found!

文档term	term₁	term₂	term₃	term₄	term₅	term₆	term₇	term₈	term₉
词典（N）	A	B	C	D	F	G	L	M	P
x	1	2	1	0	1	0	0	1	1
z	0	1	1	2	0	1	1	0	0

核函数	指标	top 5	top 10	top 15	top 20	top 25	top 30	top 50	top100
	P	0.290 2	0.271 9	0.250 5	0.240 1	0.226 7	0.207	0.181 1	0.154 1
Cauchy Kernel s =1	R	0.809	0.796 7	0.781 3	0.770 2	0.766	0.753 4	0.747 7	0.726 5
	F₁	0.427 2	0.405 4	0.391 4	0.365 8	0.349 9	0.324 8	0.291 6	0.254 3
	P	0.241 4	0.218 8	0.202 6	0.190 2	0.182 1	0.170 9	0.158 2	0.139 5
LSK	R	0.984 5	0.977 8	0.963 3	0.951 2	0.939	0.921 2	0.901 9	0.881 6
	F₁	0.387 7	0.357 5	0.334 8	0.317	0.305	0.288 3	0.269 2	0.240 9
	P	0.325 3	0.291 2	0.273 4	0.255 9	0.241	0.219 9	0.198 3	0.157 7
CLA kernel	R	0.879 2	0.860 3	0.844 5	0.826 7	0.818 9	0.810 3	0.794 5	0.775 3
	F₁	0.474 8	0.435 6	0.413 2	0.390 8	0.372 4	0.345 9	0.317 4	0.262 1
S_Wang	P	0.356 1	0.326	0.304	0.281	0.264 5	0.241	0.209 3	0.169 8
Kernel	R	0.907 9	0.89	0.867 8	0.849	0.833 3	0.826	0.803 2	0.784 4
s =1	F₁	0.511 6	0.477 2	0.450 3	0.422 2	0.401 5	0.373 1	0.332 1	0.279 2

[1]	Xiaoli DAI, Shifeng LIU, Daqing GONG. Text similarity detection method based on NLP [J]. Journal on Communications, 2021, 42(10): 173-181.
[2]	Yanping SHEN,Kangfeng ZHENG,Chunhua WU,Yixian YANG. Survey of research on application of heuristic algorithm in machine learning [J]. Journal on Communications, 2019, 40(12): 124-137.
[3]	Ling YANG, Liang CHEN, Bin ZHAO, Guolong ZHANG, Yuan LI. Blind equalization algorithm based on complex support vector regression [J]. Journal on Communications, 2019, 40(10): 180-188.
[4]	Zhi-kai XU,Hong-li ZHANG,Xiang-zhan YU,Zhi-gang ZHOU. Combinatorial double auction-based allocation of retrieval tasks in Internet of Things [J]. Journal on Communications, 2015, 36(12): 47-56.
[5]	. Net traffic identifier based on hierarchical clustering [J]. Journal on Communications, 2014, 35(Z1): 9-45.
[6]	Wei DING,Jie XU,Weng-hui ZHUO. Net traffic identifier based on hierarchical clustering [J]. Journal on Communications, 2014, 35(Z1): 41-45.
[7]	. Semi-supervised learning by constructing query-document heterogeneous information network [J]. Journal on Communications, 2014, 35(8): 6-47.
[8]	Yu-feng LIU,Ren-fa LI. Semi-supervised learning by constructing query-document heterogeneous information network [J]. Journal on Communications, 2014, 35(8): 40-47.
[9]	Zhen-long SONG,Gang-yi JIANG,Chao HUANG,Mei YU,Jia-le ZHANG. WLAN indoor positioning algorithm based on skewness-kurtosis testing [J]. Journal on Communications, 2012, 33(5): 99-105.
[10]	Zhuo-yi LV,Ke-bin JIA,Wan-chi WIU. Fast and low-complexity video down-sizing transcoder [J]. Journal on Communications, 2012, 33(1): 160-166.
[11]	Xiu-hong WANG,Shi-guang JU. Result merging method based on combined kernels for distributed information retrieval [J]. Journal on Communications, 2011, 32(4): 112-118.
[12]	Xin ZHOU,Ying WU. Research on signals modulation classification based on SVM and RVM [J]. Journal on Communications, 2010, 31(8A): 94-99.
[13]	Fang ZHAO,Hai-yong LUO,Quan LIN,Yan MA. Node localization algorithm based on kernel function and Markov chains [J]. Journal on Communications, 2010, 31(11): 195-204.
[14]	Huan ZHAO,Ren-fa LI,Jia-qin WANG,Zai-mei ZHANG. Research on an ontology concept similarity calculation approach based on the integrated multilayer information [J]. Journal on Communications, 2009, 30(6): 135-141.
[15]	Jun LUO,Zhi-jian OU. Efficient keyword spotting system for information retrieval [J]. Journal on Communications, 2006, 27(2): 113-118.

Novel kernel function for computing the similarity of text

RichHTML

PDF

Knowledge

Cited

Abstract

Cite this article

share this article

Figures/Tables 5

References 14

Related Articles 15

Metrics

Recommended 0