Abstract
In training machine learning models, loss functions are commonly applied to judge the quality and capability of the models. Traditional loss functions usually neglect the cost-sensitive loss in different intervals, although sensitivity plays an important role for the models. This paper proposes a cost-sensitive loss function based on an interval error evaluation method (IEEM). Using the key points of grade-structured intervals, two methods are proposed to construct the loss function: a piecewise function linking by key points, and a curve function fitting by key points. The proposed function was evaluated against three different loss functions based on a BP neural network. The comparison results show that the proposed loss function based on IEEM made the best prediction of the PM2.5 air quality grade in Guangzhou, China.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsNotes
- 1.
Web site: http://www.stateair.net/web/post/1/3.html.
- 2.
Web site: http://www.cma.gov.cn/2011qxfw/2011qsjgx.
References
Zhu, M.: Comparison Research of SVR Algorithms Based on Several Loss Functions. East China Normal University (2012)
Steinwart, I.: How to compare different loss functions and their risks. Constr. Approx. 26, 225–287 (2007)
Shalev-Shwartz, S., Shamir, O., Sridharan, K.: Learning linear and kernel predictors with the 0–1 loss function. In: International Joint Conference on Artificial Intelligence, pp. 2740–2745 (2011)
Zhao, H., Sinha, A.P., Bansal, G.: An extended tuning method for cost-sensitive regression and forecasting. Decis. Support Syst. 51, 372–383 (2011)
Vapnik, V.N.: The Nature of Statistical Learning Theory. Springer, New York (1995). https://doi.org/10.1007/978-1-4757-3264-1
Cesa-Bianchi, N., Lugosi, G.: Worst-case bounds for the logarithmic loss of predictors. Mach. Learn. 43, 247–264 (2001)
Hu, J., Luo, G., Li, Y., Cheng, W., Wei, X.: An AdaBoost algorithm for multi-class classification based on exponential loss function and its application. Acta Aeronaut. Astronaut. Sin. (2008)
Reich, Y., Barai, S.V.: Evaluating machine learning models for engineering problems. Artif. Intell. Eng. 13, 257–272 (1999)
Cheng, T., Lan, C., Wei, C.P., Chang, H.: Cost-sensitive learning for recurrence prediction of breast cancer (2010)
Huber, P.J.: Robust estimation of a location parameter. Ann. Math. Stat. 35, 73–101 (1964)
Ping, W., Daming, J.: Parameter design based on exponential loss function. J. Changzhou Inst. Technol. 28, 16–20 (2015)
Granger, C.W.J.: Outline of forecast theory using generalized cost functions. Span. Econ. Rev. 1, 161–173 (1999)
Zellner, A.: Bayesian estimation and prediction using asymmetric loss functions. Publ. Am. Stat. Assoc. 81, 446–451 (1986)
Xiaoqing, L., Shihong, C., Danling, T., Yonghui, Y.: Interval Error Evaluation Method (IEEM) and it’s application. Stat. Decis., 84–86 (2016)
Huber, P.J., Ronchetti, E.M.: Robust Statistics, 2nd edn. Wiley, New York (2011)
Karasuyama, M., Takeuchi, I.: Nonlinear regularization path for the modified Huber loss Support Vector Machines. In: International Joint Conference on Neural Networks, pp. 1–8 (2010)
Yamamoto, T., Yamagishi, M., Yamada, I.: Adaptive proximal forward-backward splitting applied to Huber loss function for sparse system identification under impulsive noise. IEICE Tech. Rep. Sig. Process. 111, 19–23 (2012)
Chen, C., Yan, C., Zhao, N., Guo, B., Liu, G.: A robust algorithm of support vector regression with a trimmed Huber loss function in the primal. Soft. Comput. 21, 1–9 (2016)
Chen, C., Li, Y., Yan, C., Dai, H., Liu, G.: A robust algorithm of multiquadric method based on an improved Huber loss function for interpolating remote-sensing-derived elevation data sets. Remote Sens. 7, 3347–3371 (2015)
Cavazza, J., Murino, V.: Active-labelling by adaptive Huber loss regression (2016)
Peker, E., Wiesel, A.: Fitting generalized multivariate Huber loss functions. IEEE Signal Process. Lett. 23, 1647–1651 (2016)
Granger, C.W.J.: Prediction with a generalized cost of error function. J. Oper. Res. Soc. 20, 199–207 (1969)
Coetsee, J., Bekker, A., Millard, S.: Preliminary test and Bayes estimation of a location parameter under BLINEX loss. Commun. Stat. 43, 3641–3660 (2014)
Arashi, M., Tabatabaey, S.M.M.: Estimation in multiple regression model with elliptically contoured errors under MLINEX loss. J. Appl. Probab. Stat. 3, 23–35 (2008)
Domingos, P.: MetaCost: a general method for making classifiers cost-sensitive. In: ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 155–164 (1999)
Bansal, G., Sinha, A.P., Zhao, H.: Tuning data mining methods for cost-sensitive regression: a study in loan charge-off forecasting. J. Manag. Inf. Syst. 25, 315–336 (2008)
Chai, T., Draxler, R.R.: Root mean square error (RMSE) or mean absolute error (MAE)? Geoscientific Model Dev. 7, 1247–1250 (2014)
Chesher, A.: A note on a general definition of the coefficient of determination. Biometrika 78, 691–692 (1991)
Long, L., Lei, M., Jianfeng, H., Dangguo, S., Sanli, Y., Yan, X., Lifang, L.: PM2.5 concentration prediction model of least squares support vector machine based on feature vector. J. Comput. Appl. 34, 2212–2216 (2014)
Ting, D., Jianhui, Z., Yong, H.: AQI levels prediction based on deep neural network with spatial and temporal optimizations. Comput. Eng. Appl. 53, 17–23 (2017)
Song, L.I., Wang, J., Zhang, D.C., Xia, W.: Simulation analysis of prediction optimization model for atmospheric PM2.5 pollution index. Comput. Simul. (2015)
Chen, Y., Wang, L., Zhang, L.: Research on application of BP artificial neural network in prediction of the concentration of PM2.5 in Beijing. J. Comput. Appl. 30, 153–155 (2016)
Zhou, S., Li, W., Qiao, J.: Prediction of PM2.5 concentration based on recurrent fuzzy neural network. In: Control Conference, pp. 3920–3924 (2017)
Acknowledgements
This work was supported by National Natural Science Foundation of China (grant No. 61772146), the Colleges Innovation Project of Guangdong (grant No. 2016KTSCX036) and Guangzhou program of Philosophy and Science Development for 13rd 5-Year Planning (grant No. 2018GZGJ40).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer International Publishing AG, part of Springer Nature
About this paper
Cite this paper
Chen, S., Liu, X., Li, B. (2018). A Cost-Sensitive Loss Function for Machine Learning. In: Liu, C., Zou, L., Li, J. (eds) Database Systems for Advanced Applications. DASFAA 2018. Lecture Notes in Computer Science(), vol 10829. Springer, Cham. https://doi.org/10.1007/978-3-319-91455-8_22
Download citation
DOI: https://doi.org/10.1007/978-3-319-91455-8_22
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-91454-1
Online ISBN: 978-3-319-91455-8
eBook Packages: Computer ScienceComputer Science (R0)