In the last post we explored how to build a particle swarm optimiser from scratch. In this post, let’s explore how we can leverage such an optimiser to perform mean variance optimisation to find optimal portfolios that satisfy certain constraints...

This post is inspired by a really approachable post on particle swarm optimisation by Adrian Tam. We’ll build a basic particle swam optimiser in R and try to visualise the results. Libraries # install.packages(pacman) pacman::p_load(dplyr, ggan...

When building risk scorecards, apart from the variety of performance metrics, analysts also assess something known as risk-ranking i.e. whether or not the observed event rates increase (or decrease) monotonically with increasing (or decreasing) scores...

When developing credit risk scorecards, it is generally a good idea to discretise (bin) numeric variables in a manner that ensures monotonically increasing or decreasing event rates as the variable increases or decreases. While discretising individual variables adds stability to the model, monotonic bins ensure that the model output is ...

While R has variety of options to choose from when it comes to 2D graphics and data visualisation, it is hard to beat ggplot2 in terms of features, functionality and overall visual quality. I wanted to share my take on how to use the package which is, ...