QSPR Study on the Maximum Absorption Wavelength of Indoline Dyes by Using Random Forest Algorithm

This study constructs a quantitative structure-property relationship (QSPR) model based on random forest (RF) for the maximum absorption wavelength (λmax) of indoline dyes, the core photosensitizers for dye-sensitized solar cells (DSSCs). Five key structural descriptors were selected via the filter-wrapper method. The optimal RF-2 model achieved a test set R² of 0.9194, which is significantly superior to the linear multiple linear regression (MLR) model. SHAP analysis identified the BalabanJ index and SlogP_VSA1 as the core influencing factors, providing support for the rapid design of high-performance photosensitizers.