基于ARIMA-SVM模型的博物馆经书库TVOC浓度预测
Prediction of TVOC concentration in museum scripture libraries based on ARIMA-SVM model
摘要：
为满足文物预防性保护需求，分别用ARIMA和ARIMA-SVM模型对某博物馆经书库TVOC浓度进行了预测研究。结果表明：ARIMA-SVM模型的精度高，能够较好地预测TVOC浓度序列趋势；基于ARIMA-SVM组合预测方法的平均绝对误差（MAF)、平均绝对百分比误差(MAPE)和均方根误差(RMSE)分别为0.001 5×10-6、0.000 5和0.005 5×10-6，印证了该模型预测博物馆TVOC浓度的可行性，可以为经书库环境调控提供科学依据。
Abstract：
In order to meet the need of preventive protection of cultural relics, the TVOC concentration of a museums scripture library is predicted and studied by the ARIMA model and the ARIMA-SVM model, respectively. The prediction results show that the ARIMA-SVM model has high accuracy and can better predict the trend of TVOC concentration series. The MAE, MAPE and RMSE based on the ARIMA-SVM combined forecasting method are 0.001 5×10-6, 0.000 5 and 0.005 5×10-6, respectively. This confirms the feasibility of the ARIMA-SVM model in the prediction of the TVOC concentration of the museum, which can provide a scientific basis for the environmental regulation of the scripture library.
Keywords:museum; scripture library; preventive protection; TVOC concentration; ARIMA-SVM combined model; time series forecasting; model evaluation