Near infrared spectrum moisture content measurement method based on multi-learner hybrid integration

The invention belongs to the technical field of spectral analysis, and discloses a near infrared spectrum moisture content measurement method based on multi-learner hybrid integration, which comprises the following steps: step 1, collecting near infrared spectrum data and corresponding moisture content, 2, three new training sets are obtained from the collected training sets in a random sampling mode and used for training PLS, SVM and DT base learners respectively, and 3, a near infrared spectrum moisture content prediction model conforming to physical knowledge is established; and 4, inputting the near infrared spectrum measured on line into the established model to obtain a real-time predicted value of the moisture content. According to the invention, the moisture content in the drying process of the fluidized bed can be accurately measured in a complex actual industrial scene. 本发明属于光谱分析技术领域，公开了基于多学习器混合集成的近红外光谱水分含量测量方法，包括以下步骤：步骤1、收集近红外光谱数据和相应的水分含量，获得训练集和验证集：步骤2、从收集的训练集中通过随机采样的方式获得三个新的训练集，分别用于训练PLS、SVM、DT基学习