# News and announcements

## Credit Analytics 2.3 Release

Written for CreditAnalytics by Lakshmi Krishnamurthy on 2014-01-24

· Basis Spline Library Extensions:

o Segment/

o Segment Local Curvature + Length Penalty Setup (aka pseudo splines)

o Best-Fit Penalizing Splines with Custom Penalty Order and arbitrary (but “well-behaved”) basis spline set

o Penalty evaluated Regression Splines

o Local Hermite Smoothing Schemes – Akima, Bessel, Hagan-West, Harmonic, Huynh-Le Floch, Hyman, Kruger, and Preuss schemes.

· B Spline Functionality:

o Raw/Processed Basis Hat Functions Implementation

o Synthetic Monic Basis B Spline generation with shape control

o B Spline Sequence build-out using multic segment basis function aggregation

o Custom closed form cubic KLK Hyperbolic Tension

o Incorporation of the B Spline basis onto the segment/

· Spline Based Discount Curve Build-Out:

o Shape Preserving Discount Curve Build with and without turn list adjustment

o Smoothing Discount Curve Build Pass with and without turn list adjustment

o Transition Spline Based Discount Curve Construction

o Estimation of the in-situ discount curve input quote Jacobian

o Implementation of the discount curve build-out using standard schemes such as DENSE, DUALDENSE, and CUSTOMDENSE

· Spline Based Forward Curve Build-Out:

o Shape Preserving Forward Curve Build

o Smoothing Forward Curve Build Pass

o Linearized Forward Basis Calibrator Constraint setup for fix-float and float-float

o Labeled correlated discount factor/forward rate merge sub-stretch setup

o Manifest Measure/

· Canned Product Metric Calculation:

o Day-over-day discount curve build-out for 20 years for EM/G10

o 1D/1M/3M/6M Carry PnL

o 1D/1M/3M/6M Curve Roll Down PnL

o 1D Curve Shift PnL

o Daily Forward Rate Matrix

## Credit Analytics 2.2 Release

Written for CreditAnalytics by Lakshmi Krishnamurthy on 2013-08-15

· CreditAnalytics Integration with Non-linear fixed-point searcher: Integration of the curve builder functionality with non-linear fixed-point searcher – with multiple search algorithms. Incorporation of linear searches as well.

· Rich Set of Bloomberg Samples: Full replication of the standard BBG screens – YAS, SWPM, and CDSW – along with the measure details and cash flows. Also added targeted RV measures and multi-leg swap samples.

· Product/Curve Jacobian Generation: Curve/Product Jacobian generation, both independent, and as part of the calibration, using adjoint algorithmic differentiation. Jacobians are available across the full set of curve construction/

· Serverization of CreditAnalytics: Build-out of CreditAnalytics Stub and CreditAnalytics Proxy to act as the distributor of the CreditAnalytics functionality. Incorporation of bit-wise serialization and de-serialization across all product, computed output, curves, quotes, and parameters.

· CreditAnalytics Integration with the Basis Spline Library: Integration of the design components, calibration formulation, implementation framework, and sufficiency evaluation/Jacobian generation in conjunction with the Basis Spline Library. Also exposing the full variety of discount curve construction techniques available using the basis spline library

## Credit Analytics 2.1 Release

Written for CreditAnalytics by Lakshmi Krishnamurthy on 2013-03-12

· Fast, Multi-layer, interpolating curve building: Discount/Credit Curve build out using

highly efficient and robust curve calibration techniques. Also customize the build out

based off of different node interpolation techniques, curve variate parameters, and

boundary conditions.

· Fast calibration of CDS/bond measures: Make the calibration of the bond/CDS

calibration much more faster and robust. For bonds, the new calibration is applied for

yield, Z Spread, and implied credit spread. For CDS, the new calibration is applied to

flat spreads.

· Calculation of Curve Self-Jacobian: Calculation of the curve Jacobian of the

characterizing curve variate to itself. This curve variate could be discount factor, zero

rate, or forward rate. Process also involves evaluating the corresponding canonical

measure cross-Jacobian.

· Calculation of product measure Jacobian: Fast computation of the rates/credit product

measure Jacobian to the curve variate factors. In particular, we compute the a)

product quoted measure Jacobian, and the b) product PV Jacobian. Products

considered all the standard rates and credit products (cash/EDF/

their variants).

· Monte-Carlo based product algorithmic differentiation: Formulation and

implementation of a) path-wise self-Jacobian estimator, b) path-wise parameter

Jacobian estimator, and c) path-wise product payoff derivative estimator. Also

computes product path-wise pay-off derivative estimator, along with the

implementation for specific products.

## Credit Analytics 1.6

Written for CreditAnalytics by Lakshmi Krishnamurthy on 2012-08-16

CreditAnalytics provides the functionality behind creation, calibration, and implementation of the curve, the parameter, and the product interfaces defined in CreditProduct. It also implements a curve/parameter

## Credit Analytics 1.5 Release

Written for CreditAnalytics by Lakshmi Krishnamurthy on 2012-05-29

CreditAnalytics is a full featured fixed income credit analytics library, developed with a special focus towards the needs of the credit products community (CDS, CDX, CDO, and bonds of all types and variants).

What's new in Credit Analytics 1.5:

Regressor Framework: Implementation of the regressor set, tolerance check, curve scenario regressors, regression framework suite, and the eventual regression output.

Discount Curve Regression: Regressing Base Curve Creation, scenario Curve creation, and calculation of spot/effective implied rates and discount factors.

Credit Curve Regression: Regressing Base Curve Creation, scenario Curve creation, and calculation of spot/effective implied hazard rates, recoveries, and survival.

FX Curve Regression: Creation of the basis and forward curves, conversion from one to another, and implying of the basis nodes and the enhanced discount curve on the domestic/foreign discount curves.

Zero Curve Regression: Creation of the zero curve from the product cash flow nodes, implying of the zero rates and zero discount factors at the relevant nodes.