BOOTCAMP Credit Risk Modeling

BOOTCAMP CREDIT RISK MODELING

No content

Sln Topics Details
MODULE 1 - APPLICATION OF SCORECARDS
01 VARIABLE EXPLORATION Comprehensive walk through of the variables.
Dropping off irrelevant variables not consistent with business logic.
Creating important Covariate.
02 PORTFOLIO OVERVIEW Product based
Vintage based<
Riskiness based
Accepted and Rejected loans
03 DATA PREPARATION Missing Observations Analysis
Univariate Analysis
Frequency Distribution Analysis
04 SEGMENTATION ANALYSIS Business Segmentation
Statistical Segmentation
05 VARIABLE SELECTION Weight of Evidence
Information Value
06 MODEL DEVELOPMENT Logistic Regression
Create a cut-off score by analysing
The sensitivity and specificity
07 MODEL VALIDATION Discriminatory capacity of the model - Gini, Accuracy Ratio, KS statistic
Stability of the Population -Population Stability Index
Stability of the model components - Variable Deviation Index, Rank Ordering of the model
MODULE 2 - BEHAVIOURAL SCORECARDS
08 DELINQUENCY & BAD FLAGGING Default definition as per Roll Rate analysis
09 RISK ANALYSIS Create derived Variables
10 DATA QUALITY CHECKS Missing Observations Analysis
Univariate Analysis
Frequency Distribution Analysis
11 PERFORMANCE EXCLUSIONS Modelling Exclusions of inactive a/c's.
12 SEGMENTATION Statistical Segmentation & Risk based Segmentation
13 COVARIATES CREATION ANOVA analysis
14 MODEL VALIDATION & MODEL DEVELOPMENT Logistic Regression for Model development
Create a cut-off score by analysing the sensitivity and specificity
Model discriminatory capacity
Model accuracy and Model Stability
MODULE 3 - REJECT INFERENCING
15 METHODS OF REJECT INFERENCING Hard cut-off method
Fuzzy Augmentation
Parcelling Method
MODULE 4 - COLLECTION SCORECARDS
16 METHOD OF MODELLING COLLECTIONS Triangular Matrix approach
Vintage Analysis
Linear Regression
MODULE 5 – BASEL IRB
17 UNDERSTANDING BANKING PRODUCTS Definition of Retail Portfolio
Definition of Commercial Portfolio
Understanding of retail products
18 DATA PREPROCESSING Preliminary data inspection and clean up
Building a Model Data Dictionary
Data reconciliation
Data quality check
19 MODEL DESIGN Incorporate Bad flag as per default definition using Roll rate analysis
Seasoning analysis to identify performance window
Identify snapshots, observation period and performance period and select data points under each of these windows after observation and performance exclusions
Split the data set for training and validation
20 PD MODEL DEVELOPMENT Calculating Mean default rates
Logistic Regression
Decision Trees
Survival Analysis
Segmentation - Business Segmentation & Regulatory Segmentation
21 LGD MODEL DEVELOPMENT LGD Micro structure approach
Probability of Cure
LGD Regression Methods - Tobit Regression & Beta Regression
LGD Machine Learning (ML) Modelling
LGD Survival Analysis
Segmentation - Business Segmentation & Regulatory Segmentation
22 EAD MODEL Full Prepayment Modelling via GLM
Multinomial Regression Competing Risks Modelling
CCF Modelling
Segmentation - Business Segmentation & Regulatory Segmentation
23 MODEL VALIDATION Perform in sample, out of sample and out of time validations and measure performance metrics such as AR, Gini, KS
Measure stability index (Population Stability Index, Rank Ordering)
24 INTERNAL RATING BASED APPROACH Calculating Expected Loss and Unexpected Loss
Capital Requirement as per IRB
MODULE 6 - IFRS 9 ECL Provisioning
25 ONE YEAR PD Generalised Linear Models
Machine Learning (ML) Modelling
Low Default Portfolio, Market- Based, Scarce Data Modelling
26 LIFE TIME PD Lifetime GLM Framework
Survival Modelling
Lifetime Machine Learning (ML) Modelling
Transition Matrix Modelling
27 LIFE TIME PD Lifetime GLM Framework
Survival Modelling
Lifetime Machine Learning (ML) Modelling
Transition Matrix Modelling
28 LGD MODELLING LGD Micro structure approach
Probability of Cure
LGD Regression Methods - Tobit Regression & Beta Regression
LGD Machine Learning (ML) Modelling
LGD Survival Analysis
29 EAD MODELLING Full Prepayment Modelling via GLM
Multinomial Regression Competing Risks Modelling
CCF Modelling
30 PROJECT FULL INDUSTRY LEVEL PROJECT 1 - Dataset containing Corporate Loans, Retail LAP, Retail LAS, Corporate LAS
FULL INDUSTRY LEVEL PROJECT 2 - Corporate Loans Data
ABOUT THE TRAINER

No content

Ans. 1. Anyone with finance background like having studied some level of CFA FRM or actuaries can join this program.

Ans.2. Maths Primers and Python Primers have been included in the program, so no previous experience is expected.

Ans 3. This course is quite long & comprehensive only because we have covered the entire curriculum in 3 parts – theory discussion, visualisations in excel, practical implementation through hands-on session in excel & python

Ans.4. To get certificates you need to complete all topic wise assignments, master project and pass the Final exam.

Ans.5. You can take either 1 year access or lifetime access. Please note that lifetime access is chargeable extra

Ans.6 With this website we have integrated a customized P2T player that will allow you to play encrypted classes. There are no limitations on the number of views. Also the software is compatible with Windows, Mac, Android or iPhone

Ans.7. To interact with the trainer we have a dedicated forum ‘D-forum’. Any questions asked on D-forum are expected to be replied within 24 hours by trainers and team of moderators & experts.

Ans. 8. Presently we are conducting exams in Aug mid and Jan mid. You can choose any of the cohort. In case you are not able to pass the exam in one go, you can re-book at a nominal charge

Ans.9. Every class is supported by One note files, Excel sheets & Python notebooks, Assignments and Quizzes, all these are available in the course section only.

Ans. 10. You get Letter of Recommendation physically delivered to you within 60 days of passing the exam. LOR’s also mention the chosen specialisation with the project details.

No content

Sln Topics Details
MODULE 1 - APPLICATION OF SCORECARDS
01 VARIABLE EXPLORATION Comprehensive walk through of the variables.
Dropping off irrelevant variables not consistent with business logic.
Creating important Covariate.
02 PORTFOLIO OVERVIEW Product based
Vintage based<
Riskiness based
Accepted and Rejected loans
03 DATA PREPARATION Missing Observations Analysis
Univariate Analysis
Frequency Distribution Analysis
04 SEGMENTATION ANALYSIS Business Segmentation
Statistical Segmentation
05 VARIABLE SELECTION Weight of Evidence
Information Value
06 MODEL DEVELOPMENT Logistic Regression
Create a cut-off score by analysing
The sensitivity and specificity
07 MODEL VALIDATION Discriminatory capacity of the model - Gini, Accuracy Ratio, KS statistic
Stability of the Population -Population Stability Index
Stability of the model components - Variable Deviation Index, Rank Ordering of the model
MODULE 2 - BEHAVIOURAL SCORECARDS
08 DELINQUENCY & BAD FLAGGING Default definition as per Roll Rate analysis
09 RISK ANALYSIS Create derived Variables
10 DATA QUALITY CHECKS Missing Observations Analysis
Univariate Analysis
Frequency Distribution Analysis
11 PERFORMANCE EXCLUSIONS Modelling Exclusions of inactive a/c's.
12 SEGMENTATION Statistical Segmentation & Risk based Segmentation
13 COVARIATES CREATION ANOVA analysis
14 MODEL VALIDATION & MODEL DEVELOPMENT Logistic Regression for Model development
Create a cut-off score by analysing the sensitivity and specificity
Model discriminatory capacity
Model accuracy and Model Stability
MODULE 3 - REJECT INFERENCING
15 METHODS OF REJECT INFERENCING Hard cut-off method
Fuzzy Augmentation
Parcelling Method
MODULE 4 - COLLECTION SCORECARDS
16 METHOD OF MODELLING COLLECTIONS Triangular Matrix approach
Vintage Analysis
Linear Regression
MODULE 5 – BASEL IRB
17 UNDERSTANDING BANKING PRODUCTS Definition of Retail Portfolio
Definition of Commercial Portfolio
Understanding of retail products
18 DATA PREPROCESSING Preliminary data inspection and clean up
Building a Model Data Dictionary
Data reconciliation
Data quality check
19 MODEL DESIGN Incorporate Bad flag as per default definition using Roll rate analysis
Seasoning analysis to identify performance window
Identify snapshots, observation period and performance period and select data points under each of these windows after observation and performance exclusions
Split the data set for training and validation
20 PD MODEL DEVELOPMENT Calculating Mean default rates
Logistic Regression
Decision Trees
Survival Analysis
Segmentation - Business Segmentation & Regulatory Segmentation
21 LGD MODEL DEVELOPMENT LGD Micro structure approach
Probability of Cure
LGD Regression Methods - Tobit Regression & Beta Regression
LGD Machine Learning (ML) Modelling
LGD Survival Analysis
Segmentation - Business Segmentation & Regulatory Segmentation
22 EAD MODEL Full Prepayment Modelling via GLM
Multinomial Regression Competing Risks Modelling
CCF Modelling
Segmentation - Business Segmentation & Regulatory Segmentation
23 MODEL VALIDATION Perform in sample, out of sample and out of time validations and measure performance metrics such as AR, Gini, KS
Measure stability index (Population Stability Index, Rank Ordering)
24 INTERNAL RATING BASED APPROACH Calculating Expected Loss and Unexpected Loss
Capital Requirement as per IRB
MODULE 6 - IFRS 9 ECL Provisioning
25 ONE YEAR PD Generalised Linear Models
Machine Learning (ML) Modelling
Low Default Portfolio, Market- Based, Scarce Data Modelling
26 LIFE TIME PD Lifetime GLM Framework
Survival Modelling
Lifetime Machine Learning (ML) Modelling
Transition Matrix Modelling
27 LIFE TIME PD Lifetime GLM Framework
Survival Modelling
Lifetime Machine Learning (ML) Modelling
Transition Matrix Modelling
28 LGD MODELLING LGD Micro structure approach
Probability of Cure
LGD Regression Methods - Tobit Regression & Beta Regression
LGD Machine Learning (ML) Modelling
LGD Survival Analysis
29 EAD MODELLING Full Prepayment Modelling via GLM
Multinomial Regression Competing Risks Modelling
CCF Modelling
30 PROJECT FULL INDUSTRY LEVEL PROJECT 1 - Dataset containing Corporate Loans, Retail LAP, Retail LAS, Corporate LAS
FULL INDUSTRY LEVEL PROJECT 2 - Corporate Loans Data

No content

Ans. 1. Anyone with finance background like having studied some level of CFA FRM or actuaries can join this program.

Ans.2. Maths Primers and Python Primers have been included in the program, so no previous experience is expected.

Ans 3. This course is quite long & comprehensive only because we have covered the entire curriculum in 3 parts – theory discussion, visualisations in excel, practical implementation through hands-on session in excel & python

Ans.4. To get certificates you need to complete all topic wise assignments, master project and pass the Final exam.

Ans.5. You can take either 1 year access or lifetime access. Please note that lifetime access is chargeable extra

Ans.6 With this website we have integrated a customized P2T player that will allow you to play encrypted classes. There are no limitations on the number of views. Also the software is compatible with Windows, Mac, Android or iPhone

Ans.7. To interact with the trainer we have a dedicated forum ‘D-forum’. Any questions asked on D-forum are expected to be replied within 24 hours by trainers and team of moderators & experts.

Ans. 8. Presently we are conducting exams in Aug mid and Jan mid. You can choose any of the cohort. In case you are not able to pass the exam in one go, you can re-book at a nominal charge

Ans.9. Every class is supported by One note files, Excel sheets & Python notebooks, Assignments and Quizzes, all these are available in the course section only.

Ans. 10. You get Letter of Recommendation physically delivered to you within 60 days of passing the exam. LOR’s also mention the chosen specialisation with the project details.