随着社会的发展需要,钙钛矿材料因其优异的光电性质在光伏发电领域备受关注。为满足产业发展的需求,本文提出运用机器学习方法对钙钛矿光伏材料进行预测优选,降低成本,节约时间。首先对数据进行预处理,运用随机森林方法建立因素–带隙...随着社会的发展需要,钙钛矿材料因其优异的光电性质在光伏发电领域备受关注。为满足产业发展的需求,本文提出运用机器学习方法对钙钛矿光伏材料进行预测优选,降低成本,节约时间。首先对数据进行预处理,运用随机森林方法建立因素–带隙预测模型,实现初级预测优选;其次运用决策树、BP神经网络模型、随机森林三种方法分别建立关于因素–光电转化率预测模型,结果表明随机森林模型的预测精度最高,实现二级预测优选。With the development needs of society, perovskite materials have attracted much attention in the field of photovoltaic power generation due to their excellent optoelectronic properties. In order to meet the needs of industrial development, this article proposes the use of machine learning methods to predict and optimize perovskite photovoltaic materials, reducing costs and saving time. Firstly, the data is preprocessed and a factor-bandgap prediction model is established using the random forest method to achieve primary prediction optimization;Secondly, three methods including decision tree, BP neural network model, and random forest were used to establish prediction models for factor-photoelectric conversion rate. The results show that the random forest model had the highest prediction accuracy, achieving optimal selection for secondary prediction.展开更多
文摘随着社会的发展需要,钙钛矿材料因其优异的光电性质在光伏发电领域备受关注。为满足产业发展的需求,本文提出运用机器学习方法对钙钛矿光伏材料进行预测优选,降低成本,节约时间。首先对数据进行预处理,运用随机森林方法建立因素–带隙预测模型,实现初级预测优选;其次运用决策树、BP神经网络模型、随机森林三种方法分别建立关于因素–光电转化率预测模型,结果表明随机森林模型的预测精度最高,实现二级预测优选。With the development needs of society, perovskite materials have attracted much attention in the field of photovoltaic power generation due to their excellent optoelectronic properties. In order to meet the needs of industrial development, this article proposes the use of machine learning methods to predict and optimize perovskite photovoltaic materials, reducing costs and saving time. Firstly, the data is preprocessed and a factor-bandgap prediction model is established using the random forest method to achieve primary prediction optimization;Secondly, three methods including decision tree, BP neural network model, and random forest were used to establish prediction models for factor-photoelectric conversion rate. The results show that the random forest model had the highest prediction accuracy, achieving optimal selection for secondary prediction.
