Shapley Values
This is another blog post in the series on model explainability. Here I will provide a brief description of Shapley values in the context of explaining outpu...
Recent posts
Accessing Private GitHub Repositories from Google Colab
According to Google Colab website,
Local Interpretable Model-Agnostic Explanations
Estimating permutation feature importances and plotting relationships between explanatory variables and model outputs by means of partial dependence plots ar...
Permutation Feature Importance
Investigating feature importances for a developed model is a very important step in achieving the goal of interpretable machine learning. This not only allow...
Partial Dependence Plots
The topic of model interpretability has gained a lot of attention recently with the rapid development of highly complex machine learning algorithms for deali...
Twenty-Sided Die Game
In this blog post we are going to discuss a mock quantitative interview question by Jane Street. Suppose you are offered to play the following game: at the s...
Neural Style Transfer
Consider the following image.
Antithetic Variates Method for Variance Reduction in Monte Carlo Simulations
Wikipedia defines Monte Carlo methods as “…a broad class of computational algorithms that rely on repeated random sampling to obtain numerical results.” The ...
Notes on Reinforcement Learning Lectures by David Silver
I have recently finished watching and working through a series of lectures by David Silver on Reinforcement Learning that I found immensely useful. Throughou...
Introduction to Kalman Filter
In this post, we cover the theory behind a discrete-time Kalman filter. Kalman filter is an algorithm that allows us to get a more precise information about ...
Gradient Ascent Algorithm
According to Wikipedia, gradient descent (ascent) is a first-order iterative optimization algorithm for finding a local minimum (maximum) of a differentiable...
Reducing Errors in Data Analysis
Recently a model vetter has pointed out to a mistake that I committed when developing one of the models. The mistake was not of methodological type, rather i...
Training, Validation and Test Datasets
According to different sources, it is advisable that the data that is used to build a model be split into 3 datasets: training, validation and test. This is ...
SMOTE Algorithm
This short blog post relates to addressing a problem of imbalanced datasets. An imbalanced dataset is a dataset where the classes are not approximately equal...
Genetic Algorithms. Introduction.
This is an introductory post about genetic algorithms (GAs) , which are a suite of methods of solving optimization problems. GAs form a subset of more genera...
Abalone. Part 2
This is the second part for the project about constructing a predictive model for the abalone dataset. In this post, we are going to fit 3 different regressi...
Abalone. Part 1
This is the first in a series of posts about constructing a predictive model, based on physical measurements, of age of abalone, where abalone referes to a g...