Construction of Machine Learning Models to Predict the Maximum Absorption Wavelength Considering the Solute and Solvent and Inverse Analysis of the Models.

Fluorescent organic molecules are used in a wide range of fields, such as organic light-emitting diodes, paints, and fluorescent imaging, and they are indispensable in our daily lives. However, the development of fluorescent organic molecules is difficult because fluorescence is only exhibited by specific molecules. In addition, the optical properties of the same fluorescent organic molecule can change by changing the solvent. Therefore, to design new fluorescent organic molecules, it is necessary to predict their optical properties by considering the interaction between the solute and solvent. In this study, mathematical models of the maximum absorption wavelength obtained from the interaction between the solute and solvent were constructed by machine learning, and the solute-solvent combination with a specific maximum absorption wavelength can be predicted. In addition to the solute-solvent combinations in the data set used to construct the model, we also predicted the maximum absorption wavelengths of new solute-solvent combinations and new solute data sets.