Modeling greenhouse air humidity by means of artificial neural network and principal component analysis

Abstract

A back propagation neural network (BPNN) based on principal component analysis (PCA) was proposed for modeling the internal greenhouse humidity in winter of North China. The environment factors influencing the inside humidity include outside air temperature and humidity, wind speed, solar radiation, inside air temperature, open angle of top vent and side vent, and open ration of sunshade curtain, which were all collected as data samples. Through PCA of these data samples, 4 main factors were extracted, and the relationship between the main factors and the original data was discussed. Taking the principal component values as the input of BPNN, the model showed a good performance. A comparison was made between the performances of the BPNN based on PCA and the stepwise regression method with 20 data samples which had not been used to establish the NN model, and the prediction of stepwise regression method was less accurate than the BPNN based on PCA.