My Data Science Internship Experience at Statistics Canada

• August 26, 2019
• Pascal Schmidt

In this blog post, I’ll be talking about my data science internship experience at Statistics Canada, in Ottawa, and what I have learned throughout my 4 months. First off, I am very happy that I went to Ottawa during the summer. A lot of people warned me about the winter months and advised me to return to Vancouver in September… Read More

Newton’s Method and Quasi-Newton’s Method (BFGS) for Linear Regression

• August 7, 2019
• Pascal Schmidt

In this blog post, we will be going over two more optimization techniques, Newton’s method and Quasi-Newton’s Method (BFGS), to find the minimum of the objective function of linear regression. What we will be covering: An introduction to Newton’s method and when to apply this optimization technique. Newton’s method for linear regression with a step length of 1 and backtracking… Read More

A Practical Data Wrangling Example of purrr’s imap Function

• July 28, 2019
• Pascal Schmidt

Lately, I have been looking at purrr‘s imap function and was wondering in which cases it is useful. I have always wanted to use it but never found the right problem to try it out until last week at work. Today I will be sharing a short example about purrr‘s pmap function to provide some inspiration. Let’s start by providing… Read More

Conjugate Gradient Descent for Linear Regression

• July 15, 2019
• Pascal Schmidt

Optimization techniques are constantly used in machine learning to minimize some function. In this blog post, we will be using two optimization techniques used in machine learning. Namely, conjugate gradient descent and the preconditioned conjugate gradient descent to find the beta values that are minimizing the objective linear regression function. In the last blog post, we found the beta values… Read More

Gradient Descent and Line Search Methods for Linear Regression

• July 8, 2019
• Pascal Schmidt

In this tutorial, we will be using the gradient descent optimization algorithm to find the beta values that are minimizing the objective function of a multiple linear regression. In the last blog post, we discussed the intuition behind gradient descent and showed how to minimize . Check it out for some more information about gradient descent and line search methods…. Read More

An Introduction to Gradient Descent and Line Search Methods

• July 1, 2019
• Pascal Schmidt

The gradient descent algorithm is an optimization technique that can be used to minimize objective function values. This algorithm can be used in machine learning for example to find the optimal beta coefficients that are minimizing the objective function of a linear regression. In this blog post, we are going over the gradient descent algorithm and some line search methods… Read More

Speeding up For Loops in R With Vectorization, Rcpp, and C++ Loops

• June 15, 2019
• Pascal Schmidt

Currently I am working at Statistics Canada with administrative data. Therefore, the data sets are a lot larger than at my previous job at the BC Cancer Agency. Hence, I often run into trouble when doing data manipulation by using loops. With ~2,000,000 observations, loops in R can become ridiculously slow and a real pain. In this tutorial, I am… Read More

Rowwise Data Wrangling Example With purrr, pmap, and Others in R

• May 20, 2019
• Pascal Schmidt

Last time, we talked about row-wise operations with purrr and pmap() after a colleague of mine got me thinking about row-wise operations in R. In this post, I will be going over a small example data set which outlines the problem we wanted to solve. First, the data set: library(tidyverse) dat <- data.frame( “SCORE_2010” = c(0, 0, 1, 0), “SCORE_2011″… Read More