You collect height and weight information of all the 100 participants in the PRML class. But instead of computing PCA on the entire dataset, you arbitrarily split it into 2 datasets with 50 points each. One of your friend says that she suffered an error vector of [-55] with respect to the most important direction of the first dataset and an error vector of [-56] with respect to the most important direction of the second dataset. What is the datapoint corresponding to your friend? Explain with all steps and graphs

icon
Related questions
Question

Alert dont submit AI generated answer.

You collect height and weight information of all the 100 participants in the PRML class.
But instead of computing PCA on the entire dataset, you arbitrarily split it into 2
datasets with 50 points each. One of your friend says that she suffered an error vector
of [-55] with respect to the most important direction of the first dataset and an error
vector of [-56] with respect to the most important direction of the second dataset.
What is the datapoint corresponding to your friend? Explain with all steps and graphs
Transcribed Image Text:You collect height and weight information of all the 100 participants in the PRML class. But instead of computing PCA on the entire dataset, you arbitrarily split it into 2 datasets with 50 points each. One of your friend says that she suffered an error vector of [-55] with respect to the most important direction of the first dataset and an error vector of [-56] with respect to the most important direction of the second dataset. What is the datapoint corresponding to your friend? Explain with all steps and graphs
Expert Solution
steps

Step by step

Solved in 8 steps with 1 images

Blurred answer