Training method of data processing and recognition model based on machine learning

The invention provides a data processing and recognition model training method based on machine learning, and belongs to the technical field of data processing.The method comprises the steps that firstly, soil information is collected, training samples are marked and used for model training and dimension reduction operation on data, then sample expansion is conducted through an SMOTE sample generation method based on fast clustering, and the data processing and recognition model is obtained; then, a neural network is adopted to perform feature extraction on the data, and a neural network model optimized based on a search operator algorithm is provided to optimize parameters of neurons, so that gradient disappearance and gradient explosion phenomena caused by a traditional neural network parameter optimization method are avoided; finally, a machine learning classification model based on an improved random forest is proposed to classify the hyperspectral data, and in the decision tree training stage, the classi