Wadood Alam
18th March 2024
AI 539 Final Project: Credit Score Evaluation
https://github.com/wadoodalam/Credit-Score-Prediction/blob/main/10kData.csv
- Python
- NumPy
- Pandas
- train_test_split
- StratifiedGroupKFold
- Scikit-learn
- accuracy_score
- confusion_matrix
- HistGradientBoostingClassifier
- DummyClassifier
- scipy.stats.mstats: winsorize
- time
- imblearn.under_sampling: RandomUnderSampler
- compute_sample_weight
- re(regex)
- matplotlib.pyplot
- Install the required dependencies using pip
- Ensure Dataset(
train.csv
) is contained in the same directory - Run the program using the command
data_profile.py
- The program will print
Total number of missing values
,Total number of missing values
,Number of numerical features:
- The program will generate 4 csv files, 1 xlsx file, and 5 png files
10kdata.csv
: The cleaned version of dataset containing 10,000 rows and 34 featuresCat_profile.csv
: The data profile for the 4 relevant categorical featurescorrelation_matrix.xlsx
: Correlation matrixprofile.csv
: The data profile for numeric featuresCredit_Mix.png
: Representation of Credit_Mix featureCredit_Score.png
: Representation of Credit_Score featuremissing_values.png
: Representation of the name of the feature with missing value and the number of values missing for each featureMonth.png
: Representation of Month featurePayment_Behaviour.png
: Representation of Payment_Behaviour feature
- Install the required dependencies using pip(if not installed previously)
- Ensure Dataset(
10kdata.csv
) is contained in the same directory - Run the program using the command
train_eval.csv
- The program will output 3 dictionaries for accuracies, 3 dictionaries for runtime, 3 dictionaries for confusion matrices
- The directories for confusion matrices and accuracies follow the following format:
{'Strategy name':[Train-test-split, Stratified Group-wise Cross-Validation],...}
- The runtime dictionary will follow the following format:
Run Time: {'Strategy name': [],...}
- The program will generate 1 csv file called
outliers.csv
to visualize outliers
- train.csv
- 10kdata.csv
- Cat_profile.csv
- correlation_matrix.xlsx
- profile.csv
- Credit_Mix.png
- Credit_Score.png
- missing_values.png
- Month.png
- Payment_Behaviour.png