Assignment 9

The Data Being Used

• The data we analyzed was Credit Card Data and the attrition rates of customers.

• We had a total of 10,127 observations in this data set.

How the data was split

• The data was split into two sections, attrited customers and existing customers. ## Algorithm used

• The algorithm that we used to sort and analyze our data was the Random Forest algorithm.

The accuracy of the model, tested on the validation data

• Our model predicted that 448 customers were going to leave, however only 421 actually left.

• our model has an accuracy of roughly 94%, with a 6% level of error.

The financial impact of the model, given the following assumptions:

- The customer retention program consists of forgiving the anual fee of $100 (on average)

- The value of a customer is $2,300 per year (on average)

- Customer retention programs have a 30% success rate

• The cost of my program is 448 * 100 = $44,800

• We will save 30% of 421 = 126.3, which has a value of 2300x * 117.6 = $290,490

• The net of this program is $290,498 - $44,800 = $245,698 from 3034

• Per customer is $245,698 / 3034 = $80.98 per customer

Including Plots

rf$importance %>%
  data.frame() %>%
  rownames_to_column(var = "Feature") %>%
  ggplot(aes(x = fct_reorder(Feature, MeanDecreaseAccuracy), y = MeanDecreaseAccuracy)) + 
  geom_pointrange(aes(ymin=0, ymax = MeanDecreaseAccuracy), color = "cadetblue", size = .3) +
  theme_minimal() + 
  coord_flip() + 
  labs(x = "", title = "Variable Importance using MeanDecreasingAccuracy")
  

val$predicted_Attrition_FLag <- predict(rf , val)

Note that the echo = FALSE parameter was added to the code chunk to prevent printing of the R code that generated the plot.