0
$\begingroup$

Below are few questions where I unable to find out where I am wrong. I added screen shot of image and explanations of the each options that I am understanding. Questions are purely discussion based and short. Please help me out.


In below question, I checked

A. We need to predict the author gender and it can be either male or female. I think it is classification problem, so it is supervised learning.

B. We have given group of spam emails and need to predict does there sub-types are spam or not. I think it is classification problem, so it is supervised learning.

C. We need to predict data based on height and age. It is a linear regression problem because we create graph height vs age will find out the test case. It is supervised learning.

D. Grouping data is a cluster problem, so it unsupervised learning.

enter image description here


In below question, I had checked C and D options because feature scaling creates our dataset in same range which helps to predict the best theta in less iterations and contour graph will be more cleared and symmetric. Ref : https://medium.com/greyatom/why-how-and-when-to-scale-your-features-4b30ab09db5e

enter image description here


$\endgroup$
1
$\begingroup$

For the 1st question, I think B is not a supervised learning problem. You are already given emails which are spam. You need to analyse them to find if there are subtypes (Clustering.)

For the 2nd question, it is just B. Scaling doesn't make each step less computationally expensive. You do same amount of computation.

| improve this answer | |
$\endgroup$
  • $\begingroup$ Then what will be the second correct answer in question 5. In question 4, we have given dataset but identifying either subtype is spam or not (its either true or false, 0/1 case ). So you dont think predicting the subtype is classification. $\endgroup$ – dahiya_boy May 1 '19 at 17:18
  • $\begingroup$ 4th question, 2nd option asks you to find subtypes of spam. Doesn't ask you to classify between spam and not-spam. And, for the 5th question, I'm not sure about the 3rd option. Do you actually have to select more than one option? $\endgroup$ – Illuminati0x5B May 1 '19 at 17:36
  • $\begingroup$ Thanks, can you pls explain second question too. Where I am wrong. $\endgroup$ – dahiya_boy May 1 '19 at 17:47
  • $\begingroup$ @dahiya_boy we already know that emails are spam. Question says 'examine' this collection and 'discover if there are subtypes.' So, we need to probably cluster those emails and see if they fall into different clusters (sub types). $\endgroup$ – Illuminati0x5B May 1 '19 at 19:36
  • $\begingroup$ Thanks for your explanation, I agreed with your answer. Can you please explain other question (feature scaling). Thanks $\endgroup$ – dahiya_boy May 2 '19 at 3:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.