Question

I need new and unique answers, please. (Use your own words, don't copy and paste), Please Use your keyboard (Don't use handwriting) Thank you..

Q1:What is K-Fold Cross Validation? Draw an example for k=5.

Q2: Explain the difference in neural network architectures when neural network models are used as classifiers and when they are used as optimizers. Give an example of each ANN architecture type.    

Answer 1

Solution:

Note:

We are authorized to answer one question at a time, since you have not mentioned which question you are looking for, so we are answering the first one, please re-post your question separately for the remaining question.

Answer 1:

K-Fold Cross Validation:

• Cross validation is a method to get ensure that the developed model is reliable and robust.
• In the k cross fold validation, the data is divided into the k number of subsets.
• Here, k can be any number of elements or tuple or it can be said as the number of iterations.
• In k fold cross validation, huge quantity of data is trained, and a fragment of data is tested against of it.
• For set, 1/k=1/5, 1/5(20%) of the data for testing and remaining 80% of data for training purpose.

Consider the five subsets for the given condition:

       I, II, III, IV, V

drawing example for k=5 :

In all the 5 iterations, some average error estimates is carried out at each stage (denoted as E1, E2, E3, E4 AND E5 in the above diagram). The sum of error estimates can be computed by adding all the error estimates (E1, E2, E3, E4 AND E5) together. The formula for calculating error estimates is given below:

Advantage:

All the data present in the test case must be should be among the provided data set.

Disadvantage:

• It is computationally expensive.
• It can be used for complex models and is expensive.