PSO-based band selection method for simultaneous determination of multi-application.docVIP

下载本文档

0
0
约1.09万字
约 10页
2017-05-05 发布于浙江
举报
版权申诉

PSO-based band selection method for simultaneous determination of multi-application.doc

1、本文档共10页，可阅读全部内容。
2、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。
3、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
4、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。
5、该文档为VIP文档，如果想要下载，成为VIP会员后，下载免费。
6、成为VIP后，下载本文档将扣除1次下载权益。下载后，不支持退款、换文档。如有疑问请联系我们。
7、成为VIP后，您将拥有八大权益，权益包括：VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
8、VIP文档为合作方或网友上传，每下载1次，网站将根据用户上传文档的质量评分、类型等，对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档

PSO-based band selection method for simultaneous determination of multi-application

PSO-based band selection method for simultaneous determination of multi-application Study: Shen Tao Qi Zhang Ya LIU Chubb [Abstract] Simultaneous determination of when the interaction occurs between the components should be selected for the band. Proposes an improved particle swarm optimization algorithm is used for band selection. The proposed method for o, p-nitrophenol the concentration predicted in the range of 208 ~ 481 nm to 0.15 mol / L NaOH as the solvent, the configuration set for training the mixture of 27 groups, 27 groups for the prediction set. training set RMS error (RMSE) were .1257, 0.2228 and 0.0846; prediction set RMSE 0.2070,0.1507 and 0.394, respectively, have a good prediction. [Keywords:] Particle swarm optimization, band selection, p-nitrophenol 1 Introduction Simultaneous determination of multi-component system is a chemical analysis of one of the more active areas of research. As multi-component absorption spectra overlap, mutual interference between the components and other reasons, will appear with the full spectrum independent of the concentration or the non linear component between bands. Screening and components related to the wavelength to be analyzed not only can improve the prediction ability and robustness [1 ~ 4], but also simplify the model. wavelength selection methods are the correlation coefficient method [5], stepwise regression, No information variable elimination method [6], and interval partial least squares, based on support vector machines and subspace by band selection [7], and combinatorial optimization-based methods such as simulated annealing, genetic algorithm [8] and so on. Optimization Methods modeling of discrete wavelength points, the spectral data does not take into account the characteristics of continuous relevance, may lead to overfitting, model robustness and poor [9]. popular band selection of fixed position or interval length of the band, it is difficult to achieve the most excellent resul