Why does dropout prevent Overfitting?

Related questions

1. What is batch size in Lstm?
2. How is Lstm better than RNN?
3. Can Lstm be used for classification?
4. Why does dropout regularization work?
5. Does keras automatically use GPU?
6. What is validation accuracy in keras?
7. What is stateful Lstm?
8. How do you save models in keras?
9. What is dense layer in Lstm?
10. Does batch normalization prevent Overfitting?

Tags

1. #inputs
2. #layer
3. #because
4. #dropped
5. #during
6. #reliance
7. #random
8. #overfitting
9. #prevents
10. #always
11. #these
12. #present
13. #generalization
14. #training
15. #dropout
16. #improving
17. #learns