Ends in
00
days
00
hrs
00
mins
00
secs
SHOP NOW

🚀 25% OFF All Practice Exams, Video Courses, & eBooks – Cyber Week Blowout Deals!

Find answers, ask questions, and connect with our
community around the world.

Home Forums AWS AWS Certified Machine Learning – Specialty question on most accurate model: linear regression vs kNN regressor

  • question on most accurate model: linear regression vs kNN regressor

  • johannes-h

    Member
    March 3, 2021 at 8:59 pm

    I feel that since the question is asking for the most _accurate_ model, the justification for not picking kNN is wrong. Depending on the difficulty of the data, a linear model might underfit heavily, while kNN will probably have at least decent performance. The question does talk about “2 years worth of data”, but this might be just a few hundred measurements in the end – as long as nothing more is specified I wouldn’t automatically assume that there is too much training data for a non-parametric model (and even then you could use clustering and only using cluster means for kNN to reduce the number of data points used in kNN)

  • Carlo-TutorialsDojo

    Member
    March 3, 2021 at 9:45 pm

    Hello johannes,

    Thanks for your feedback. I see what you mean.

    My objective when I wrote this question was to differentiate LR from KNN.

    I think that the question still lacks information (e.g., characteristic of the training data, which one performs faster in giving predictions) to support the correct option. We will modify the scenario so it’ll have a clear distinction between kNN and Linear Learner.

    Let me know your thoughts.

    Regards,

    Carlo

Viewing 1 - 2 of 2 replies

Log in to reply.

Original Post
0 of 0 posts June 2018
Now
Skip to content