1)Load the provided R Script file (MIS470-W5-CT.R) into RStudio and run its code. The script file builds a logistic regression model of customer churn in a telecommunication company

2) Add R code to the script file to calculate and print out the logistic model accuracy using misClasificError variable

3) Add R code to the script file to display a confusion matrix using the table() function.

4) I already add R code to create and evaluate the performance measures of 3 logistic models using the three (3) most significant predictors (either individual or collectively)

Screenshot 1

Screen Shot 2

Screen Shot 3

III. Summarize your work in one page in which you explain what you did, interpret your results, and reflect on your experience.

Besides your opinions, the summary must include the following things:

-a complete enumeration of the significant factors (they are marked with *, **, *** in the 1st screenshot that was taken in part I); for instance, in my case, there are 11 significant factors;

-interpretation for the misclassification error of the 1st and 2nd model (if you build 3 separate logistic models in part I, 4), you have a total of 4 misclassification errors, so you must write more interpretation). Example: Let's say that you obtain a misclassification error of 0.25. This means that when our model predicts that a customer is going to churn (i.e. Churn=1), in 25% of the cases, the prediction is incorrect.

-interpretation for the confusion maxtrix of the 1st and 2nd model (if you build 3 separate logistic models in part I, 4), you have a total of 4 matrices, so you must write more interpretation). Example: if you check my screenshot from part I, 3), you can see the following numbers: 1414 (0 on row, 0 on column), 134 (0 on row, 1 on column), 290 (1 on row, 0 on column), 270 (1 on row, 1 on column). We can make 4 different combinations of 0s and 1s:

1-1 (testing$Churn=1, fitted.results=1): this is a true positive; in 170 of 2108 cases, the model predicts that a customer will churn and the prediction is correct, i.e. a customer is going to cancel Telco's services.

1-0 (testing$Churn=1, fitted.results=0): this is a false negative (or type II error): in 290 of 2018 cases, the model predicts that a customer won't churn, but the prediction is false/inaccurate, i.e. the customer is going to cancel Telco's services.

0-1 (testing$Churn=0, fitted.results=1): this is a false positive (or type I error); in 134 of 2018 cases, the model predicts that a customer will churn, but the prediction is false/inaccurate, i.e. the customer is going to keep Telco's services.

0-0 (testing$Churn=0, fitted.results=0): this is a true negative; in 1414 of 2018 cases, the model predicts that a customer won't churn and the prediction is correct, i.e. the customer is going to keep Telco's services.

HElp me to interpret the above requirements.