Exploring Aromatase Inhibitors to Breast Cancer with Machine Learning Approach Based Quantitative Structure-Activity Relationships Model Technique

The conversion of androgen depends in large part on the overexpression of aromatase in many females. Many therapeutic and pharmaceutical drugs, as well as inhibitors, have been developed, but some of them either don't have the strength to control diseases caused by the aromatase enzyme or they encourage unwelcome mammary epithelial cells inside the breast. The prospect and potential for creating medicines with unique pharmacological and therapeutic capabilities using the Multi-Machine Learning Classifier Based-Quantitative Structure-Activity Relationship (MMLCB-QSAR) process model were looked into. In order to forecast the chemical classes with the highest potential for Aromatase, The model uses chemical descriptors and supervised machine learning to inhibit utilizing real multidimensional Aromatase inhibitors that were collected from a curated database. The multi-dimensional dataset is mainly explored and visualized using the t-SNE algorithm. Despite having excellent predictive abilities, the Bootstrap aggregation or Bagging Classifier (RandomForest Classifier) algorithm excels. More specifically, regularization and hyperparameter tuning are responsible for outstanding model results in statistics and more accurate prediction using 10-fold cross-validation. Model parameters like the Matthews correlation coefficient score (0.9397), the accuracy Score (0.973%), and the log loss probability (0.1208).