Which of the following statements about the Machine Learning Process Lifecycle are true
Understanding Machine Learning Projects >>> Which of the following statements about the Machine Learning Process Lifecycle are true >>> Introduction to Applied Machine Learning
1.
Question 1
Which of the following statements about the Machine Learning Process Lifecycle are true? (click all that apply)
1 point
Documentation isn’t important if the QuAM is being built and used internally
A clearly defined process helps with communication
Lifecycle re-sets are easy to anticipate
Testing the QuAM is an important part of the Machine Learning and Model Evaluation phase
Data wrangling is only a small part of an ML project
Looking at whether you can answer your questions based on the data you have is part of the final phase.
The first phase considers exactly what features are going to be used
Iteration is almost always a given; an ML project is rarely straight-forward and linear
Having a clear, specific question for the QuAM to answer is an important part of the first phase
========================================
2.
Question 2
What does it mean to work backwards through the ML Process in oder to define a question?
1 point
Start from the broadest business outcome, identify actions that accomplish that outcome, identify questions that support those actions, refine the question.
Start with the learning data you have, define the question you want your QuAM to answer, choose an algorithm.
Start with the answer you want your QuAM to give, decide on the algorithm, then define operational data.
Start with the required action, identify impact on the business goals, define and refine a question
========================================
3.
Question 3
Why might a business choose to use a QuAM that has lower accuracy than the current system?
1 point
Because the QuAM can do the task much faster than the humans can, and a faster process is highly advantageous for the business
Because the cost of developing the QuAM was high, and the business doesn’t want to waste that time and money
Because your business needs to have machine learning to survive in the modern age
Because the QuAM does the job adequately and provides other benefits
Because the employees hate doing the task, and handing it off to a machine frees them to do other useful things
========================================
4.
Question 4
Why is it important to have a hold-out test set?
1 point
A) It is important to test your QuAM’s ability to generalize before deploying it
B) It’s important to protect the integrity of the test data so that you have confidence in the performance of your QuAM
C) Your QuAM is going to have to answer questions about unseen data, so you need to test it on unseen data
A and B
A, B and C
========================================
5.
Question 5
Which of the following statements are true? (click all that apply)
1 point
A true positive means someone is pregnant
False negatives are always a bigger problem than false positives
Telling a patient that they have cancer when they don’t is a false positive
Telling a patient that they have cancer when they don’t is a false negative
True positives and true negatives are correct predictions