8.1 Background
Even though logistic regression predates machine learning, recent excitement over machine learning’s predictive ability has made this technique among the most commonly-employed methodologies in this branch of artificial intelligence. Massive data volume in today’s society, the availability of powerful and affordable computers, and the development of cloud computing infrastructure – along with the precipitous fall in cloud operating costs – have combined to push machine learning into the spotlight.
Consider these results from Google Trends for the 12 months leading up to June 2022, showing common related topics and related queries associated with searches that include ‘logistic regression.’ Most of the world’s frequently asked questions about logistic regression are linked to machine learning. The same pattern can be found by scanning the topics closely related to logistic regression.
Data source: Google Trends 1
You may wonder how machine learning and logistic regression are intrinsically linked? Many machine learning techniques are drawn from statistical techniques like logistic regression2, linear regression, as well as linear algebra and calculus. By using logistic regression and a healthy dose of clean and credible data, a cruise operator can predict if a customer is likely to return, a hotel will know if a customer is likely to cancel a booking, and healthcare researchers can determine if someone is at risk of developing a disease.3 The technique’s common use in the medical field could explain why the online medical repository ‘pubmed’ appears among the most related queries in Google Trends.
1 Data source: Google Trends (https://www.google.com/trends).
2 IBM Cloud Education (15 July 2020) ‘What is machine learning?’ IBM Cloud Learn Hub. https://www.ibm.com/cloud/learn/machine-learning
3 Nusonovici, S. et al (2020) Logistic regression was as good as machine learning for predicting major chronic diseases. Journal of Clinical Epidemiology.122:56-69. doi: 10.1016/j.jclinepi.2020.03.002.