基于最小二乘支持向量机的森林火灾预测研究
Prediction of Forest Fires Based on Least Squares Support Vector Machine

作者: 李恩来 :云南财经大学统计与数学学院,云南 昆明;

关键词: 森林火灾预测机器学习最小二乘支持向量机Forest Fires Prediction Machine Learning Least Squares Support Vector Machine

摘要:
森林火灾是一个主要的环境问题,造成经济损失和生态破坏而且危及生命。如何预测、防治或减少森林火灾的危害成为诸多学科领域共同关注的科学任务。传统的做法是使用卫星,红外线扫描仪和局部传感器。但是由于卫星定位的延迟和扫描仪高昂的设备成本和维护成本,这些方案不能用来解决所有的情况。然而,研究表明气象因素对森林火灾有重要的影响。因此,有不少的学者建立森林火灾预测系统并将气象数据纳入量化指标体系。随机计算机的迅速发展,不少的学者将机器学习的方法运用到森林火灾等级预测模型中,但是其预测效果并不十分理想。本文提出基于机器学习中支持向量机方法的改进方法-最小二乘支持向量机,由于最小二乘支持向量机对处理样本容量较小的数据具有较高的准确度而且耗时较短。本文选用UCI数据库中的森林火灾数据进行预测处理,选用高斯函数(径向基函数)作为最小二乘支持向量机的核函数,根据一对一的多分类算法设计出最小二乘支持向量机的多分类器,使用粒子群算法选择最优参数。最后与支持向量机、BP神经网络、决策树等方法进行对比。

Abstract: Forest fire is a major environmental problem, resulting in economic loss and ecological damage, and endangering life. How to predict, prevent or reduce the damage of forest fire has become a scientific task of many disciplines. The traditional approach is to use a satellite, an infrared scanner, and a local sensor. However, due to the delay of the satellite positioning and the high cost of the scanner’s equipment and maintenance costs, these solutions can not be used to solve all the situation. However, the study shows that the meteorological factors have an important influence on forest fire. Therefore, many scholars have established system for forest fire prediction and the meteorological data into the quantitative index system. With the rapid development of random computer, many scholars have applied the method of machine learning to forest fire grade prediction model, but the effect is not very ideal. This paper presents an improved method of support vector machine method based on machine learning, because the least squares support vector machine is with a higher accuracy and shorter time consuming to process small sample size of the data. In this paper, we select the UCI database of forest fire forecast data processing, select Gaussian function (radial basis function) as the kernel function of least squares support vector machine, according to one of multiple classification algorithm design of least squares support vector machine classifier, using particle swarm optimization algorithm to choose the optimal parameters. Finally, it is compared with the support vector machine, BP neural network, decision tree and so on.

文章引用: 李恩来 (2016) 基于最小二乘支持向量机的森林火灾预测研究。 数据挖掘, 6, 15-27. doi: 10.12677/HJDM.2016.61003

参考文献

[1] 曲智林, 胡海清. 基于气象因子的森林火灾面积预测模型[J]. 应用生态学报, 2007, 18(12): 2705-2709.

[2] 牛若芸, 翟盘茂, 孙明华. 森林火险气象指数及其构建方法回顾[J]. 气象, 2006, 32(12): 3-9.

[3] Arrue, B., Ollero, A. and Matinez de Dios, J.R. (2000) An Intelligent System for False Alarm Reduction in Infrared Forest-Fire Detection. IEEE Intelligent Systems and their Applications, 15, 64-73.
http://dx.doi.org/10.1109/5254.846287

[4] Bi, J. and Bennett, K. (2003) Regression Error Characteristic Curves. Proceedings of 20th International Conference on Machine Learning (ICML), Washington DC, 21-24 August 2003, 43-50.

[5] Breiman, L. (2001) Random Forests. Machine Learning, 45, 5-32.
http://dx.doi.org/10.1023/A:1010933404324

[6] Han, J.W. (2001) Data Mining Concepts and Techniques. Morgan Kaufmann, San Francisco.

[7] Fraser, R.H. and Li, Z. (2002) Estimating Fire-Related Parameters in Boreal Forest Using SPOT VEGETATION. Remote Sensing of Environment, 82, 95-110.
http://dx.doi.org/10.1016/S0034-4257(02)00027-5

[8] Cherkassy, V. and Ma, Y. (2004) Practical Selection of SVM Parameters and Noise Estimation for SVM Regression. Neural Networks, 17, 113-126.
http://dx.doi.org/10.1016/S0893-6080(03)00169-2

[9] Han, J.W. and Kamber, M. 数据挖掘: 概念与技术[M]. 第3版. 北京: 机械工业出版社, 2012.

[10] 黄家荣, 刘倩, 高光芹, 刘小丽, 魏春生, 段群迷, 程作峰. 森林火灾成灾面积的人工神经网络BP模型预测[J]. 河南农业大学学报, 2007, 41(3): 273-275.

[11] 杨景标, 马晓茜. 基于人工神经网络预测预测广东省森林火灾的发生[J]. 林业科学, 2005, 41(4): 127-132.

[12] 许志卿, 苏喜友, 张顾. 基于支持向量机方法的森林火险预测研究[J]. 中国农学通报, 2012, 28(13): 126-131.

[13] Vapnik, V.N. (1995) The Nature of Statistical Learning Theory. Springer-Verlag, New York.

[14] Vapnik, V.N. (1998) Statistical Learning Theory. John Wiley, New York.

[15] 朱家元, 陈开陶, 张恒喜. 最小二乘支持向量机算法研究[J]. 计算机科学, 2003, 30(7): 157-159.

[16] Cristianini, N. and Shawe-Taylor, J. (2000) An Introduction to Support Vector Machines and Other Kernel-Based Learning Methods. The Press Syndicate of Cambridge University, Cambridge.
http://dx.doi.org/10.1017/CBO9780511801389

[17] Bredensteniner, E.J. and Bennett, K.P. (1999) Multicategory Classification by Support Vector Machines. Computational Optimization and Applications, 12, 53-79.
http://dx.doi.org/10.1023/A:1008663629662

[18] Vapnik, V. (2005) Universal Learning Technology: Support Vector Machines. NEC Journal of Advanced Technology, 2, 137-144.

[19] Eberhart, R.C. and Kennedy, J. (1995) A New Optimizer Using Particles Swarm Theory. Proceedings of the 6th International Symposium on Micro Machine and Human Science, Nagoya, 4-6 October 1995, 39-43.
http://dx.doi.org/10.1109/mhs.1995.494215

分享
Top