Email-Spam-Detection

We have created 4 models for detection.

In Model 1 we have used a dataset of approx 30000 entries and applied naive bayes algorithm.

In Model 2 we have used a dataset of approx 30000 entries, do some manual changes a created some spam into ham and applied naive bayes algorithm for comparision purpose with model 1.

In Model 3 we have used a dataset of approx 5000 entries and applied naive bayes algorithm.

In Model 4 we have used a dataset of approx 30000 entries and applied stacking to improve the accuracy of model.

After observing all models we see that the stackin model has better accuracy than all the other models