DISCUSSION
Rough Set Theory is a mathematical framework that deals with incomplete or uncertain data by analyzing a given dataset's lower and upper approximations. In this work, Rough set theory improved prediction and interpretability of the Generalized linear model through feature selection which ultimately led to detecting key features that influenced the target variable, these detected features were fed to the Generalized Linear model which improved the accuracy and also improved interpretability of the model by removing non important features.
CONCLUSION AND RECOMMENDATIONS
In conclusion, the experiment showed that using the Rough Set theory for interaction detection and feature selection improves the logistic regression model’s accuracy and reduces the dataset’s dimensionality, thus reducing the time required to execute the logistic model. The next part of the experiment will be to model an association rule mining method on the generated reduce and produce decision rules using the technique.