Models

This module defines the BaselineModel, an abstract base class for constructing Quantitative Structure-Activity Relationship (QSAR) models within a machine learning framework. The class is designed to streamline the integration of different machine learning algorithms into QSAR studies by providing a uniform interface and methodology for model construction, training, prediction, and hyperparameter optimization.

Child classes inheriting from BaselineModel are required to implement the model-specific logic within the defined abstract methods. This structure promotes a clear separation between the generic model workflow and the specific implementation details of different QSAR modeling techniques.

The module also includes several concrete implementations of the BaselineModel class, each corresponding to a different machine learning algorithm. These classes provide a starting point for constructing QSAR models using popular machine learning libraries that use regression algorithms and ensemble methods.

• Base Model
• CatBoost Model
• ElasticNet Model
• Lasso Model
• RandomForest Model
• Ridge Model
• XGBoost Model