关键词: 绿茶, 气候品质, 评价模型, 机器学习, 湖北省

Abstract: Quality analysis data of green tea within different green tea-plucking periods and observation data of meteorological elements at 15 days before spring tea-plucking in typical tea-producing regions of Hubei Province from 2018 to 2020 was used. Firstly, the key meteorological elements affecting green tea quality and its response relationship were analyzed by using machine learning methods including decision tree and random forest model. Secondly, the single-factor subordinate function model for temperature, sunshine, wind speed and relative humidity was built based on fuzzy mathematics theory. In the meanwhile, a comprehensive index model for climatological quality of green tea was built by using the comprehensive weighted method. Finally, parameters in this model were optimized based on the genetic algorithm. Additionally, grade evaluation standard was also determined. The results showed that the number of samples corresponding and existing one grade difference to the actual grade of samples accounted for 67.2% and 32.8% of the total samples, respectively. The results indicated that this model for green tea could reflect its quality difference under different climate situations.

Key words: green tea, climatological quality, evaluation model, machine learning, Hubei Province