What is var()? Knowing this, what would be the best criteria or method to select

Question

What is var()? Knowing this, what would be the best criteria or method to select a model that will reduce errors?

What are the advantages/disadvantages between parametric and non-parametric modeling? Given certain data set, what approach do you recommend?

Which type of prediction is aimed at predicting a discrete random variable?

Explain in detail non-parametric and parametric modeling and explain their differences and when you use each one.

Given a scenario on probability, what distribution would be used to solve the scenario?

Explanation / Answer

Answer to question# 1)

Var() is the variance of the error term associated with the regression model

it tells us about the fluctuations in the value of the error terms. for some data values the error for the regrssion model will be large, sometimes it is small. This fluctuation in the error is measured by V()

.

The aim for the best regression model is to reduce these fluctuations in the regression model. Less the errors and its fluctuations better will be the model in predicting the values. Hence the rule of "least squares" is used to fit the best regression model to the data.

Related Questions

Navigate

• Browse (All)
• Browse W
• Subjects

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.