抽象的な

Prediction of naringin content based on machine learning methods

Yan Zeng, Xinwen Cheng, Qi Li, Xiao Wang, Yuyun Chen

To increase the accuracy and speed of measurement of Naringin extraction rate, the prediction of Naringin extraction rate is raised based on Weighted Least Square Support Vector Machine (WLSSVM) and improved Artificial Bee Colony (ABC) of the machine learning methods. Taking the ratio of material to solvent, the extracting time, ethanol concentration and extracting temperature which influence Naringin extraction rate as the input of WLSSVM and Naringin extraction rate as output, learn extracting Naringin test data from shaddock peels. The results of simulation indicate that the prediction of improved ABC algorithm and advanced WLSSVM acquires better prediction speed, accuracy and stability and is appropriate for the prediction of Naringin extraction, compared with the methods of LSSVM and ABC-LSSVM.