Description:
Used Imputation and Adaboosting ( Adaptive Boosting) technique in SAS Enterprise miner to analyse the Dataset.
Analyzed the dataset first without using imputation i.e with missing values.
Used Imputation technique to replace all the missing value with mean
Used Gradient boosting node to generate a set of decision trees to form a single predictive model.
Key Results:
Without Imputation: With Imputation:
Without Gradient Boosting: With Gradient Boosting:
In our dataset, we will not recommend using gradient boosting since even though the true negative count increased, the true positive count decreased substantially increasing the misclassification rate.
Collaborators: Ishan Dindorkar, Rohan Ashok Patil, Valay Raval and Sridevi Allela.