ECON4280/6280 : Final Project

Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

ECON4280/6280 : Final Project

Due by 11:59PM on Dec 11, 2022

To minimize loss, credit card companies need decision rules regarding approvals of credit card applications. An applicant’s demographic and socio-economic profiles are considered by credit card companies before a decision is taken regarding his/her application. This empirical project gives you the opportunity to implement various statistical learning methods discussed in the class on a credit card approval dataset. After you run the analysis in R, write a report of about 1500-2000 words (together with important plots , important statistics) summarizing your finding. Please attach your R codes in the appendix.Your report should be no more than 20 pages in total (inluding R code appendix).

The Credit Approval dataset is available in the archives of machine learning repository of University of California, Irvine (http://archive.ics.uci.edu/ml/datasets/credit+approval). It consists of 690 rows , repre- senting 690 individuals applying for a credit card, and 16 variables in total. The first 15 variables represent various attributes of the individual like gender, age, marital status, years employed etc.  The 16th variable is the one of interest: credit approved(or just approved). It contains the outcome of the application, either positive(represented by +) meaning approved or negative (represented by -) meaning rejected.  The expla- nation of the variables can be found in reference 1. Note that you don’t need to know the exact meaning of say, for example b or a in A1 variable.  You just need to know that A1 represents Gender, one of a and b refers to male. The same logic applies to other variables.

Keep  in  mind that you may need to pre-process the  data  (for example, data cleaning, data in- tegration, data transformation , data reduction, missing values imputation among other tasks), carry out exploratory data analysis, and then build various models to fit the training data, compute the test error and explain the results. Finally pick the best model in terms of lowest test error and try to justify your choice of model.

Some reference

1. Deepesh Khaneja,“Credit Approval Analysis using R”,  Technical Report,  November 2017

2. Ms.D.Jayanthi,“Credit Approval Data Analysis Using Classification and Regression Models”, Interna- tional Journal of Research and Analytical Reviews, vol5 (3), 2018

3. Fu, Z and Z, Liu,“Classifier Comparisons On Credit Approval Prediction”, Final project of Course CS229 in 2014 at Stanford University.