Air Quality Prediction Based on Wavelet Analysis and Machine Learning

This thesis takes the historical weather time series of Chongqing as experimental samples. Firstly, this uses wavelet transform to organize data, and then divides sample data into training test sets verify accuracy evaluation Naive Bayes Model. Secondly, Model is compared with currently used machine learning models such SVM, XGBoost, bagging, random forest. Finally, results show that has high stability for air quality assessment Chongqing, it can be applied urban ambient quality.