Data Science and Machine Learning (Theory and Projects) A to Z - Optional Estimation: Parametric Distributions

The video tutorial introduces parametric and non-parametric distributions, explaining how parametric distributions are defined by parameters like mean and standard deviation for normal distributions. It contrasts these with non-parametric distributions, which do not follow a specific parametric function. The tutorial also covers kernel density estimation (KDE) as a method for modeling non-parametric distributions, using Python's seaborn library. The video concludes with a preview of upcoming content on parameter estimation methods.

What is a parametric distribution?

• A distribution defined by a set of parameters.

• A distribution without any parameters.

• A distribution that is always discrete.

• A distribution that cannot be described by a function.

Which of the following is a parameter of the normal distribution?

• Probability of success

• Mu

• Lambda

• Number of trials

How does changing the parameters of a distribution affect it?

• It changes the type of distribution.

• It changes the shape of the distribution.

• It has no effect on the distribution.

• It makes the distribution non-parametric.

What is a characteristic of non-parametric distributions?

• They are always continuous.

• They are defined by a set of parameters.

• They follow a specific parametric function.

• They do not follow any parametric function.

What is the purpose of kernel density estimates?

• To convert discrete data into continuous data.

• To estimate parameters of parametric distributions.

• To fit smooth functions on non-parametric distributions.

• To make data follow a normal distribution.

In what situations can real data be modeled by parametric distributions?

• When data is always continuous.

• In several situations with one or more combinations of parametric distributions.

• When data does not fit any known distribution.

• Only when data is discrete.

What will the next video in the series discuss?

• The history of probability distributions.

• Methods to estimate parameters for parametric distributions.

• Applications of kernel density estimates.

• Advanced non-parametric methods.