Fluorescence Emission Wavelength QSPR Application with Linear Blending Method in Machine Learning Algorithms

Machine learning tools have been developed to analyze quantitative structure-activity/property relationship (QSAR/QSPR) modeling research. Better feature selection algorithms in the ensemble methods used advance QSPR/QSAR modeling, helping understand relation between features and target variables reducing computational requirements. Implementing importance allows for a more effective clearer view into features' relative interpret predictions. However, main struggle of is that each model leads different selections interpretation. Therefore, it necessary summarize its corresponding better performance, resulting high prediction accuracy. In this article, we use blending method interpretability terms experimental values fluorescence wavelengths. The blender requires two levels. first level uses multiple classifiers: Random Forest, ExtraTrees, Adaptive Boosting, Gradient Boosting. second linear summarizes information from classifiers. Even though models accurately predict properties activities, are often susceptible so even small changes can drastically impact their efficiency Thus, idea overcome difficulty implement times manipulate sensitivity. Furthermore, predicts data set regression task Decision Tree based (DT-based) QSAR/QSPR model. This paper provides best-optimized when considering specific chemical or biological values. tables figures representing model's accuracy demonstrate result. It shows number predicting Fluorescence Emission Wavelength reduces, training test sets maintained, effectiveness increased.