Question

please chooose all corrct answers to the question( it could be more than one corrct answers):-

Q1) What are the best practices for building out good unit tests?

- You do not need an assert statement for each edge case.

- Write a test function to check for implementation given all parameters.

- Have multiple assert statements to test for all edge cases.

- Think about all edge cases before you write the function.

- Run a test function to see if your function works in all edge cases.

Q2) Which one of the following statements is INCORRECT?

- In regression models, input and output variables have numerical values.

- When fitting the line, we are trying to maximize the distance of each point to the line.

- We can make predictions if input values are given and the equation of the line is known.

- Although it is called a linear regression model, any curve can be used as long as it can fit the data well.

Q3) "The following descriptions best describe what: 1. Value that has to be assigned manually. 2. The K value in K-nearest-neighbore is an example of this. 3. Value is set before the training."

- Model Parameter

- Model hyperparameter

- Significance

- Centroid

Q4) Which of the following describes code that is CLEAN?

- Simple

- Readable

- Vague

- Concise

- Repetitive

Q5)What are some useful metrics to evaluate a GAN’s performance?

- Precision

- Discriminator and generator loss

- Recall

- Softmax

Answer 1

1) What are the best practices for building out good unit tests?

A: - Write a test function to check for implementation given all parameters.

   - Have multiple assert statements to test for all edge cases.

   - Think about all edge cases before you write the function.

   - Run a test function to see if your function works in all edge cases.

2) Which one of the following statements is INCORRECT?

A: When fitting the line, we are trying to maximize the distance of each point to the line.

   Explanation : Because generally we try to minimize to distance between each point in regression.

               And all other options are true i.e. in regression we have only numerical input and output, if input values are given and the equation of the line is known, any curve can be used as long as it can fit the data well.

3) "The following descriptions best describe what: 1. Value that has to be assigned manually. 2. The K value in K-nearest-neighbore is an example of this. 3. Value is set before the training."

A: Model HyperParameter

Explanation : K in KNN is a hyperparameter that you, as a designer, must pick in order to get the best possible fit for the data set. Intuitively, you can think of K as controlling the shape of the decision boundary we talked about earlier.

4) Which of the following describes code that is CLEAN?

A: Simple , Readable , Concise (giving lot of information in few lines)

   Explanation : Because we don’t want our code to be long and complex which repetitive and vogue (unclear)

5) What are some useful metrics to evaluate a GAN’s performance?

A: Precision , Discriminator and generator loss , Recall

   Explanation : Softmax is not a metric to evaluate GAN performance whereas Annealed Importance Sampling , FID , PRECISION and RECALL , Geometry score , discriminator are metrics to evaluate GAN performance.