Normal dist. doubt

#1

sir you mentioned in the lecture that we will use it for generative modelling?what is that? i am not able to understand the wiki defn.

#2

Suppose you want to generate some data(lets say height and weight) of 1000 monkeys. But don’t have access to monkeys and can’t measure it directly. So you will make an assumption that this data is distributed normally according to some mean and covariance.

Let say avg height of monkey is 75 cm and average wt is 20 kg and some reasonable covariance matrix we assume.

Then you can generate data using the Multivariate Normal Distribution function we saw in the video. (Some Algorithms are based on Generative Modelling which we will study later)
The data generated is not the real data but an assumption about how the real data might look like. So in a sense we are trying to MODEL a real distribution of moneky features.

1 Like