A hierarchical classifier based on human blood plasma fluorescence for non-invasive colorectal cancer screening
作者：
Highlights：
• A hierarchical classification method based on plasma fluorescence to identify colorectal cancer.
• A unified approach that expands CRC identification to other non-malignant findings.
• SVM-RFE is used to deal with high dimensionality and select discriminatory features.
• Superior results achieved for CRC classification (sensitivity of 0.86 and specificity of 0.95).
• The classification of non-malignant conditions achieved sensitivity of 0.60 and specificity of 0.79.
摘要
•A hierarchical classification method based on plasma fluorescence to identify colorectal cancer.•A unified approach that expands CRC identification to other non-malignant findings.•SVM-RFE is used to deal with high dimensionality and select discriminatory features.•Superior results achieved for CRC classification (sensitivity of 0.86 and specificity of 0.95).•The classification of non-malignant conditions achieved sensitivity of 0.60 and specificity of 0.79.
论文关键词：Colorectal cancer,Hierarchical classification,Fluorescence spectroscopy,Support vector machine,Human blood plasma,SVM-RFE
论文评审过程：Received 7 June 2017, Revised 15 August 2017, Accepted 12 September 2017, Available online 20 September 2017, Version of Record 26 October 2017.
论文官网地址：https://doi.org/10.1016/j.artmed.2017.09.004
