Absorption and emission maxima prediction of fluorescent organic materials using optimized neural network models

Abstract

The aim of the study is to predict the absorption and emission maxima from the given dataset, which consists of 3066 fluorescent organic materials. To fulfill this requirement, five optimized neural network models are employed. Out of them, the wide neural network (WNN) model outperformed the other models on validation as well as test results. The results were obtained on the basis of three evaluation parameters: mean absolute error (MAE), root mean squared error (RMSE), and the coefficient of determination (R2). According to the validation scores, the WNN was the best-predictive model with MAE-18.05 nm, RMSE-28.93 nm, and R2-89.55%, and MAE-29.58 nm, RMSE-42.62 nm, and R2-78.87% for absorption and emission maxima, respectively. On the other hand, on the basis of tested scores, the WNN was the best-predictive model with MAE-19.35 nm, RMSE-29.33 nm, and R2-92.14%, and MAE-29.17 nm, RMSE-41.87 nm, and R2-79.46% for absorption and emission maxima, respectively. The presented automated method does not require an extensive understanding of computer programming to estimate the absorption and emission maxima. The absorption and emission maxima may be predicted, which helps with the design of new fluorescent organic materials and their many uses in electronics, chemistry, materials science, medicine, and other areas.