Gearbox oil product state identification method based on integrated learning

The invention discloses a gearbox oil product state recognition method based on integrated learning. The method comprises the three steps of oil data preprocessing, oil product feature optimization and oil product state recognition. According to the oil data preprocessing, a wavelet noise reduction method is adopted for filtering oil data, the quality of the oil data is improved, and interference is removed for subsequent oil product feature optimization; for oil product feature optimization, a principal component analysis method is adopted to carry out dimensionality reduction on the preprocessed oil liquid data, effective oil liquid features are extracted, and then the oil product state recognition efficiency is improved; an Adaboost ensemble learning framework is adopted for oil product state recognition, a plurality of GWO-BP weak classifiers are integrated into a strong classifier GWO-BP-Adaboost with high robustness, and therefore classification and recognition of the gear box oil product state are achi