## 02ps-06 Heavytail Guassian Solution

• 0:00 - 0:02
The correct answer here is no.
• 0:02 - 0:06
Suppose we let "x" go to infinity,
• 0:06 - 0:11
then x-Mu2 for any fixed Mu would go to infinity.
• 0:11 - 0:15
So if x with a -infinity would go to zero--
• 0:15 - 0:16
because it is a constant--
• 0:16 - 0:21
Therefore, we know that in the limit of x goes to infinity--
• 0:21 - 0:25
this expression must be zero.
• 0:25 - 0:28
However, in this graph--it stays at alpha, and doesn't go to zero--
• 0:28 - 0:32
So, therefore, there can't be a valid Mu at sigma square.
• 0:32 - 0:35
If a deep improbability, you know that the area in the Gaussian
• 0:35 - 0:37
has to integrate into one,
• 0:37 - 0:41
and this area diverges, it is actually infinite in size,
• 0:41 -
so it's not even a valid execution.
Title:
02ps-06 Heavytail Guassian Solution
Description:

dummy description

more » « less
Video Language:
English
Team:
Udacity
Project:
CS373 - Artificial Intelligence
Duration:
0:44
 Amara Bot added a translation

• Amara Bot