Internet-of-things operation index prediction method based on improved SVR model

The invention discloses an internet-of-things operation index prediction method based on an improved SVR model. The method comprises the following steps: preprocessing climate data and internet-of-things operation index data; feature selection is carried out on the preprocessed climate data, and redundant features are removed; an RF-SVR model is established, parameters of the RF-SVR model are optimized by using an intelligent optimization algorithm PSO, and an RF-PSO-SVR model is obtained; and outputting an internet of things operation index prediction result through the RF-PSO-SVR model. According to the method, the reliability of feature selection is improved, the prediction precision of a subsequent prediction model is improved, and the SVR model after feature selection is optimized by using the PSO algorithm aiming at the situation that accurate prediction of local Internet of Things operation index data cannot be realized only by using the SVR, so that the goodness of fit of the model is improved, the me