R Programming Random Forest / Gradient Boosting
Completion requirements
For this assignment you will write an R program to complete the tasks given below. You will hand in two files for this assignment.
- A File with your R program. This file should contain only the code (no output) and must have the typical r extension. No other file extensions will be accepted. The reason is that the assignment be graded based on your R code and not the output file. The output file will be used to verify the code commands. Also, please make sure that all comments, discussion, and conclusions regarding results are also annotated as part of your code.
- A PDF/DOC file with your output code. We are giving you more flexibility regarding how you want to present your output (tables, plots, etc.). You can either use RMD files that combine code, narrative txt, and plots or you can use word document with copy and paste from the R platform you are using. However, please remember that all output (tables, plots, comments, conclusions, etc.) shown in this file has to be generated by the same R code that you submit. This is important! Output shown that is generated using a separate code or output shown that is not supported by the submitted code will not be graded. Screenshots will not be accepted.
- R Data Set: HMEQ_Scrubbed.csv (in the zip file attached).
- The Data Dictionary in the zip file.
Note: The HMEQ_Scrubbed.csv file is a simple scrubbed file from the previous week homework. If you did more advanced scrubbing of data for last week, you may use your own data file instead. You might get better accuracy! If you decide to use your own version of HMEQ_Scrubbed.csv, please hand it in along with the other deliverables.
This assignment is an extension of the Week 4 assignment. The difference is that this assignment will now incorporate Random Forest and Gradient Boosting models.
Step 1: Read in the Data
- Read the data into R
- List the structure of the data (str)
- Execute a summary of the data
- Print the first six records
Step 2: Classification Models
- Using the code discussed in the lecture, split the data into training and testing data sets.
- Create a Decision Tree model using the rpart library to predict the variable TARGET_BAD_FLAG
- Create a Random Forest model using the randomForest library to predict the variable TARGET_BAD_FLAG
- Create a Gradient Boosting model using the gbm library to predict the variable TARGET_BAD_FLAG
- All model parameters such as tree depth are up to you.
- Do not use TARGET_LOSS_AMT to predict TARGET_BAD_FLAG.
- Plot the Decision Tree and list the important variables for the tree.
- List the important variables for the Random Forest and include the variable importance plot.
- List the important variables for the Gradient Boosting model and include the variable importance plot.
- Using the testing data set, create a ROC curves for all models. They must all be on the same plot.
- Display the Area Under the ROC curve (AUC) for all models.
- Rerun with different training and testing data at least three times.
- Determine which model performed best and why you believe this.
- Write a brief summary of which model you would recommend using. Note that this is your opinion. There is no right answer. You might, for example, select a less accurate model because it is faster or easier to interpret.
Step 3: Regression Decision Tree
- Using the code discussed in the lecture, split the data into training and testing data sets.
- Create a Decision Tree model using the rpart library to predict the variable TARGET_LOSS_AMT
- Create a Random Forest model using the randomForest library to predict the variable TARGET_LOSS_AMT
- Create a Gradient Boosting model using the gbm library to predict the variable TARGET_LOSS_AMT
- All model parameters such as tree depth are up to you.
- Do not use TARGET_BAD_FLAG to predict TARGET_LOSS_AMT.
- Plot the Decision Tree and list the important variables for the tree.
- List the important variables for the Random Forest and include the variable importance plot.
- List the important variables for the Gradient Boosting model and include the variable importance plot.
- Using the testing data set, calculate the Root Mean Square Error (RMSE) for all models.
- Rerun with different training and testing data at least three times.
- Determine which model performed best and why you believe this.
- Write a brief summary of which model you would recommend using. Note that this is your opinion. There is no right answer. You might, for example, select a less accurate model because it is faster or easier to interpret.
Step 4: Probability / Severity Model Decision Tree (Push Yourself!)
- Using the code discussed in the lecture, split the data into training and testing data sets.
- Use any model from Step 2 in order to predict the variable TARGET_BAD_FLAG
- Develop three models: Decision Tree, Random Forest, and Gradient Boosting to predict the variable TARGET_LOSS_AMT using only records where TARGET_BAD_FLAG is 1.
- Select one of the models to predict damage.
- List the important variables for both models.
- Using your models, predict the probability of default and the loss given default.
- Multiply the two values together for each record.
- Calculate the RMSE value for the Probability / Severity model.
- Rerun at least three times to be assured that the model is optimal and not over fit or under fit.
- Comment on how this model compares to using the model from Step 3. Which one would your recommend using?
Essential Activities:
- Watch all the training videos
- Execute the example code while watching the training videos.
Notes:
- This assignment is due Sunday at 11:59 PM EST
|
February 4 2026, 3:13 PM |