基于模拟退火半监督学习的信用预测研究

doi:10.3969/j.issn.0253-2778.2018.06.003

中国科学技术大学学报 ›› 2018, Vol. 48 ›› Issue (6): 447-457.DOI: 10.3969/j.issn.0253-2778.2018.06.003

基于模拟退火半监督学习的信用预测研究

张杰，李琳，朱阁

武汉理工大学计算机科学与技术学院，湖北武汉 430070

收稿日期:2017-09-09 修回日期:2018-04-10 接受日期:2018-04-10 出版日期:2018-06-30 发布日期:2018-04-10
通讯作者: 李琳
作者简介:张杰，男1993年生，硕士生.研究方向：机器学习.E-mail: icecream@whut.edu.cn
基金资助:
国家社会科学基金(15BGL048)，武汉理工大学科研基金（2017II39GX)，武汉理工大学研究生优秀学位论文培育项目（2016-YS-068），湖北省科技支撑计划（研发与示范）2015BAA072）资助.

Simulated annealing based semi-supervised support vector machine for credit prediction

ZHANG Jie,LI Lin, ZHU Ge

School of Computer Science & Technology, Wuhan University of Technology,Wuhan 430070， China

Received:2017-09-09 Revised:2018-04-10 Accepted:2018-04-10 Online:2018-06-30 Published:2018-04-10

摘要/Abstract

摘要： 金融机构结合消费者和商业信息来为企业进行信用打分.我国的企业特别是小微企业信用信息少，造成了只有少量企业拥有信用信息，而大量企业没有信用信息的局面.半监督支持向量机可以利用标记数据和未标记数据进行学习，同时可以克服信用数据类别不均衡和样本信息不足等问题.由于半监督支持向量机的参数对算法效果有较大影响，实际参数选取往往根据经验所得.为此提出了一种利用模拟退火（SA）优化基于确定性退火半监督支持向量机（DAS3VM）参数的SAS3VM算法.该算法在少量有标记信用数据的基础上，利用大量无标记信用数据辅助学习，使用模拟退火寻找最优参数.最后在两组企业信用数据集和三组个人信用数据集上进行对比实验，结果表明，半监督学习方法（DAS3VM和SAS3VM）优于监督学习方法，SAS3VM在准确率上比DAS3VM最大提升了13.108%.

关键词: 半监督学习, 确定性退火, 模拟退火, 信用预测

Abstract: In the mid-1990s financial institutions began to combine consumer and business information to create scores for business credits. Enterprises in China, especially small and micro enterprises, have less credit information, resulting in the situation where only a small number of enterprises have credit information, while a large number of enterprises have none. However, semi-supervised support vector machines (S3VM) can learn from labeled data and unlabeled data and solve the problems of imbalanced credit data categories and insufficient sample information. The parameters of S3VM have a great influence on the effect of the algorithm, and the actual parameter selection is often based on experience. An SAS3VM algorithm was proposed to optimize the parameters of deterministic annealing based semi-supervised support vector machine (DAS3VM) with simulated annealing. Based on the small number of labeled credit data, the algorithm takes advantage of the unlabeled credit data to help study and use the simulate annealing to find the optimal parameters. Experiments were conducted on two categories of enterprise credit data and three categories of personal credit data. The results show that semi-supervised learning (DAS3VM and SAS3VM) performs better than supervised learning. The maximum accuracy of SAS3VM has been increased by 13.108% compared with DAS3VM.

Key words: semi-supervised learning, deterministic annealing, simulated annealing, credit prediction

中图分类号:

TP391

张杰，李琳，朱阁. 基于模拟退火半监督学习的信用预测研究[J]. 中国科学技术大学学报, 2018, 48(6): 447-457.

ZHANG Jie,LI Lin, ZHU Ge. Simulated annealing based semi-supervised support vector machine for credit prediction[J]. Journal of University of Science and Technology of China, 2018, 48(6): 447-457.

参考文献

［1］
EISENBEIS R A. Recent developments in the application of credit-scoring technique to the evaliation of commercial loan[J]. IMA Journal of Management Mathematics, 1996, 7(4): 271-290.
[2] BERGER A N, FRAME W S. Small business credit scoring and credit availability[J]. Journal of Small Business Management, 2007, 45(1): 5-22.
[3] BERGER A N, UDELL G F. Small business credit availability and relationship lending: The importance of bank organizational structure[J]. The Economic Journal, 2002, 112(477): F32-F53.
[4] HUANG C L, CHEN M C, WANG C J. Credit scoring with a data mining approach based on support vector machines[J]. Expert Systems With Applications,2007, 33(4): 847-856.
[5] BELLOTTI T, CROOK J. Support vector machines for credit scoring and discovery of significant features[J]. Expert Systems with Applications, 2009, 36(2): 3302-3308.
[6] SHAHSHAHANI B M, LANDGREBE D A. The effect of unlabeled samples in reducing the small sample size problem and mitigating the Hughes phenomenon[J]. IEEE Transactions on Geoscience and Remote Sensing,1994, 32(5): 1087-1095.
[7] NIGAM K, GHANI R. Analyzing the effectiveness and applicability of co-training[J]. Proceedings of the Ninth International Conference on Information and Knowledge Management. New York, NY, USA :ACM, 2000.
[8] BALUJA S. Probabilistic modeling for face orientation discrimination: Learning from labeled and unlabeled data[C]// Proceedings of the 11th International Conference on Neural Information Processing Systems. Cambridge, MA, USA: MIT Press, 1998: 854-860.
[9] BLUM A, CHAWLA S. Learning from labeled and unlabeled data using graph mincuts[C]// Proceedings of the Eighteenth international Conference on Machine Learning. San Francisco, CA, USA : Morgan Kaufmann Publishers Inc., 2001: 19-26.
[10] ZHOU D Y, BOUSQUET O, LAL T N, et al. Learning with local and global consistency[C]// Proceedings of the 16th International Conference on Neural Information Processing Systems. Cambridge, MA, USA: MIT Press, 2003: 321-328.
[11] WANG F, ZHANG C S. Label propagation through linear neighborhoods[J]. IEEE Transactions on Knowledge and Data Engineering,2008, 20(1) : 55-67.
[12] BLUM A, MITCHELL T. Combining labeled and unlabeled data with co-training[C]//Proceedings of the Eleventh Annual Conference on Computational Learning Theory. New York, NY, USA :ACM, 1998: 92-100.
[13] ZHOU Z H, LI M. Tri-training: Exploiting unlabeled data using three classifiers[J]. IEEE Transactions on knowledge and Data Engineering, 2005, 17(11): 1529-1541.
[14] JONES R. Learning to extract entities from labeled and unlabeled text[D]. Pittsburgh，PA，USA：Carnegie Mellon University, 2005.
[15] JOACHIMS T. Transductive inference for text classification using support vector machines[C]// Proceedings of the Sixteenth International Conference on Machine Learning. San Francisco, CA, USA ：Morgan Kaufmann Publishers Inc.，1999: 200-209.
[16] CHAPELLE O, CHI M, ZIEN A. A continuation method for semi-supervised SVMs[C]// Proceedings of the 23rd International Conference on Machine Learning. New York, NY, USA: ACM, 2006:185-192.
[17] SINDHWANI V, KEERTHI S S, CHAPELLE O. Deterministic annealing for semi-supervised kernel machines[C]// Proceedings of the Twenty-Third International Conference on Machine Learning. New York, NY, USA: ACM, 2006: 841- 848.
[18] BELKIN M, NIYOGI P, SINDHWANI V. Manifold regularization: A geometric framework for learning from labeled and unlabeled examples[J]. Journal of Machine Learning Research, 2006,7: 2399-2434.
[19] LI Y F, KWOK J T, ZHOU Z H. Semi-supervised learning using label mean[C]// Proceedings of the 26th Annual International Conference on Machine Learning. New York, NY, USA:ACM, 2009:633-640.
[20] ROSE K. Deterministic annealing for clustering, compression, classifi- cation, regression, and related optimization problems[J]. Proceedings of the IEEE, 1998, 86(11): 2210-2239.
[21] BENNETT K P, DEMIRIZ A. Semi-supervised support vector machin-ES[C]// Proceedings of the 11th International Conference on Neural Information Processing Systems. Cambridge, MA, USA: MIT Press, 1998: 368-374.
[22] CHAPELLE O, ZIEN A. Semi-supervised classification by low density separation[C]// Proceedings of the Tenth International Workshop on Artificial Intelligence and Statistics, 2005:57-64.
[23] BERTSIMAS D, TSITSIKLIS J. Simulated annealing[J]. Statistical Science, 1993, 8(1): 10-15.
[24] KEERTHI S S, DECOSTE D. A modified finite newton method for fast solution of large scale linear SVMs[J]. Journal of Machine Learning Research, 2005, 6: 341-361.
[25] CHAPELLE O, SINDHWANI V, KEERTHI S S. Optimization techniques for semi-supervised support vector machines[J]. Journal of Machine Learning Research, 2008, 9: 203-233.
[26] CHAPELLE O, SCHOLKOPF B, ZIEN A. Semi-supervised learning[J]. IEEE Transactions on Neural Networks, 2009, 20(3): 542-542.
[27] SINDHWANI V, KEERTHI S S. Large scale semi-supervised linear SVMs[C]// Proceedings of the 29th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval. New York, NY, USA: ACM, 2006: 477-484.
[28] MANGASARIAN O L. A finite newton method for classification[J]. Optimization Methods and Software, 2002, 17(5): 913-929.
[29] DEMPSTER A P, LAIRD N M, RUBIN D B. Maximum likelihood from incomplete data via the EM algorithm[J]. Journal of the Royal Statistical Society: Series B (Methodological), 1977, 39(1):1-38.
[30] ZHU X, GHAHRAMANI Z, LAFFERTY J. Semi-supervised learning using Gaussian fields and harmonic functions[C]// Twentieth International Conference on International Conference on Machine Learning. AAAI Press, 2003:912-919.

()
()

[1]	崔文泉，陈伟，程浩洋. 非均衡数据情形的一种协同正则化多视图半监督学习分类器[J]. 中国科学技术大学学报, 2020, 50(5): 596-604.
[2]	刘玮康, 秦晓卫, 卫国. 基于模糊度的半监督自步协同下的微信流业务识别[J]. 中国科学技术大学学报, 2020, 50(1): 29-38.
[3]	汲梦宇，刘利刚. 基于感知的室内场景的多视点渲染优化[J]. 中国科学技术大学学报, 2018, 48(2): 140-147.
[4]	朱威，颜仕星，张磊，李国正. 基于真实世界临床数据的失眠病判别分析[J]. 中国科学技术大学学报, 2016, 46(10): 867-873.

基于模拟退火半监督学习的信用预测研究

Simulated annealing based semi-supervised support vector machine for credit prediction

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 4

编辑推荐

Metrics

本文评价