- Details
- Parent Category: Programming Assignments' Solutions

# We Helped With This R Programming Assignment: Have A Similar One?

Category | Programming |
---|---|

Subject | R | R Studio |

Difficulty | Undergraduate |

Status | Solved |

More Info | Answers To Statistics Homework |

## Short Assignment Requirements

## Assignment Description

**Statistics 108**, Homework 4

*Due: November 6th, 2017, In Class *(turn in paper form)

*You need to show the steps to get the full credits.

This homework is to practice on model diagnostics and data transformation. Total: 90 points.

1. (30 points) *Checking model assumptions – linearity and equal
variance. *For the simple linear regression model, the most essential
assumption is the linearity assumption. Also, the equal variance assumption is
the basics for the least squares method. For seven datasets under
Canvas/Files/Data files (data1.txt – data7.txt), plot the scatter plot, state
whether the two assumptions hold or not and briefly explain.

2. (30 points) *Understanding the Normal Q-Q plot. *We use five
datasets under Canvas/Files/Data files (data8.txt – data12.txt) as illustration
examples to understand the Normal Q-Q plot. For each dataset, first plot the
scatter plot to get an initial idea of the data. Then, fit the model and obtain
the standardized residuals. Plot the histogram of the standardized residuals
and comment on the shape the histogram. Plot the Normal Q-Q plot of the
standardized residuals using the ‘qqnorm()’ function, and add the reference
line through the ‘qqline()’ function. Comment on the Normal Q-Q plot.

3. (30 points) *Data transformation. *When one or more assumptions for
the simple linear regression model do not hold, we may want to seek for
possible transformations on *X *and/or *Y *so that the linear model
is a suitable model for the transformed variables. For three datasets under
Canvas/Files/Data files (data13.txt – data15.txt), first check whether the
three assumptions (linearity, equal variance, and normality) hold based on the
scatter plot and diagnostic plots. Then, try different transformations on *X *and/or *Y *, state the transformation you would take so that the simple linear
regression model is proper on the transformed data. Draw the scatter plot of
the transformed data and the diagnostic plots of the new model. Comment.

1