PDtoolkit1.2.0 package

Collection of Tools for PD Rating Model Development and Validation

univariate

Univariate analysis

auc.model

Area under curve (AUC)

bivariate

Bivariate analysis

boots.vld

Bootstrap model validation

cat.bin

Categorical risk factor binning

cat.slice

Slice categorical variable

confusion.matrix

Confusion matrix

constrained.logit

Constrained logistic regression

create.partitions

Create partitions (aka nested dummy variables)

cutoff.palette

Palette of cutoff values that minimize and maximize metrics from the c...

decision.tree

Custom decision tree algorithm

dp.testing

Testing the discriminatory power of PD rating model

embedded.blocks

Embedded blocks regression

encode.woe

Encode WoE

ensemble.blocks

Ensemble blocks regression

evrs

Modelling the Economic Value of Credit Rating System

fairness.vld

Model fairness validation

heterogeneity

Testing heterogeneity of the PD rating model

hhi

Herfindahl-Hirschman Index (HHI)

homogeneity

Testing homogeneity of the PD rating model

imp.outliers

Imputation methods for outliers

imp.sc

Imputation methods for special cases

interaction.transformer

Extract risk factors interaction from decision tree

kfold.idx

Indices for K-fold validation

kfold.vld

K-fold model cross-validation

normal.test

Multi-period predictive power test

num.slice

Slice numeric variable

nzv

Near-zero variance

power

Power of statistical tests for predictive ability testing

pp.testing

Testing the predictive power of PD rating model

predict.cdt

Predict method for custom decision tree

psi

Population Stability Index (PSI)

replace.woe

Replace modalities of risk factor with weights of evidence (WoE) value

rf.clustering

Risk factor clustering

rf.interaction.transformer

Extract interactions from random forest

rs.calibration

Calibration of the rating scale

scaled.score

Scaling the probabilities

segment.vld

Model segment validation

smote

Synthetic Minority Oversampling Technique (SMOTE)

staged.blocks

Staged blocks regression

stepFWD

Customized stepwise regression with p-value and trend check

stepFWDr

Customized stepwise regression with p-value and trend check on raw ris...

stepMIV

Stepwise logistic regression based on marginal information value (MIV)

stepRPC

Stepwise logistic regression based on risk profile concept

stepRPCr

Stepwise regression based on risk profile concept and raw risk factors

ush.bin

U-shape binning algorithm

ush.test

Testing for U-shape relation

woe.tbl

Weights of evidence (WoE) table

The goal of this package is to cover the most common steps in probability of default (PD) rating model development and validation. The main procedures available are those that refer to univariate, bivariate, multivariate analysis, calibration and validation. Along with accompanied 'monobin' and 'monobinShiny' packages, 'PDtoolkit' provides functions which are suitable for different data transformation and modeling tasks such as: imputations, monotonic binning of numeric risk factors, binning of categorical risk factors, weights of evidence (WoE) and information value (IV) calculations, WoE coding (replacement of risk factors modalities with WoE values), risk factor clustering, area under curve (AUC) calculation and others. Additionally, package provides set of validation functions for testing homogeneity, heterogeneity, discriminatory and predictive power of the model.