data: https://utk.instructure.com/courses/97583/files/5834170/download?wrap=1

id: 395246

Use the Q7 dataframe.You want a 'good' model predicting Wins from all variables (without

interactions).Perform the 'all possible' approach, which produces a list of models with

whose AICs are within 4 of the overall lowest AIC of all considered models.

Identify the predictors in the model with the FEWEST predictors (it

could be the case your list has only one model).Redo

the 'all possible' approach considering just those predictors ALONG WITH all two-way

interactions between them to produce yet another list of models (your list might only have one).

Report the AIC of the model at the top of this list, i.e., the one with the lowest AIC.

Note:do not add ANY extra arguments (e.g., nbest or nvmax) to any of the commands.

Split the dataframe Q8 into a training sample (75%) and holdout sample (25%).

Build a predictive logistic regression model predicting Buy from all predictors

(no interactions), choosing as your final model the one suggested by the one standard

deviation rule.Fit the model on the training sample, then report its misclassification

RATE on your holdout sample.

Note:in the TWO places where it is necessary, be sure to set the random number seed to 320.

Build a partition model predicting Buy on the training set you made in #8.

Copy/paste the value of cp corresponding to the tree suggested by the one standard

deviation rule.Again, in the place(s) where it is necessary, be sure to set the

random number seed to 320.

Use the Q10 dataframe.Consider three standard multiple linear regression models predicting

Sales from both predictors and no interaction (i.e., ignoring the time series nature of the data):

one with a linear trend, one with a quadratic trend, and one with a cubic trend.

In the model with lowest of the three AICs, report the STANDARD ERROR of the difference in average sales

between First Quarter and Third Quarter.

Reconsider your model and data from Q10.Now impose an autoregressive model on the residuals.In other words,

consider three additional models predicting Sales from Quarter and Month (no interactions but with whichever

kind of trend was appropriate in Q10) with an autoregressive model for the residuals of lag 1, lag 2, and lag 3.

Note:Quarter is a categorical variable, independent of the trend, so do not use poly() on it.

In the model with lowest of the four AICs, report the STANDARD ERROR of the difference in average sales

between First Quarter and Third Quarter.Note:this requires installation and loading of the library nlme.

So you'll have to do install.packages('nlme') then library(nlme)

Note:while usually standard errors increase when 'correctly' accounting for correlation, yours might decrease.