摘要
无机钙钛矿材料因其优异的光电性质,如较高的光吸收系数和高载流子迁移率,目前在光伏领域受到广泛关注,是具有极大产业化潜力的新型材料。文章采用机器学习方法构建无机钙钛矿材料的形成能高精度预测模型,应用了XGBoost、随机森林、支持向量回归和LightGBM 4种机器学习算法来建立预测模型。在测试集上评估不同模型的预测性能后,结果表明LightGBM模型具有最高的预测精度和效果,支持向量回归和XGBoost模型也显示出较好的预测性能,而随机森林算法的预测效果较差。因此,基于机器学习方法,研究构建了无机钙钛矿材料形成能高精度预测模型。LightGBM算法展示了最优的预测效果,为无机钙钛矿材料的高通量筛选和设计提供了关键技术支持。机器学习技术在材料领域的应用,将大幅提高材料发现和设计的效率,给材料科学发展带来深远影响。
Due to their excellent photoelectric properties,such as high light absorption coefficient and high carrier mobility,inorganic perovskite materials have attracted extensive attention in the field of photovoltaics,and they are new materials with great potential for industrialization.In this study,machine learning methods were used to construct a high-precision prediction model for the formation energy of inorganic perovskite materials,and four machine learning algorithms,XGBoost,random forest,support vector regression and LightGBM,were used to establish the prediction model.After evaluating the prediction performance of different models on the test set,the results show that the LightGBM model has the highest prediction accuracy and effect,and the support vector regression and XGBoost models also show good prediction performance,while the random forest algorithm has poor prediction performance.Therefore,based on machine learning methods,this study constructed a high-precision prediction model for the formation energy of inorganic perovskite materials.The LightGBM algorithm shows the best prediction effect,which provides key technical support for the high-throughput screening and design of inorganic perovskite materials.The application of machine learning technology in the field of materials will greatly improve the efficiency of material discovery and design,and have a profound impact on the development of materials science.
作者
冯顺
Feng Shun(School of Electronic Information,Xijing University,Xi’an 710123,China)
出处
《无线互联科技》
2023年第16期47-51,共5页
Wireless Internet Technology
关键词
无机钙钛矿材料
机器学习
性能预测
形成能
inorganic perovskite materials
machine learning
performance prediction
formation energy
