Michael Perry is an experienced Sparta, NJ, sales and business development executive who guides ClubsGalore’s marketing activities. His endeavors in Sparta include creating integrated campaigns that encompass social media, website content, telemarketing, and e-mail. Michael Perry has an interest in organizational design and economic trends defined through the use of econometrics.
One aspect of this field was recently explored in a Forbes article involving the impact of supervised machine learning (ML) on econometrics. In simple terms, machine learning involves the use of x covariates in predicting (y) outcomes. Current ML methods include systematic approaches such as regression trees, LASSO, and random forest.
Common to these approaches is their use of systems of cross-validation. This involves models being created using one part of available data, which is then tested on another data set. The robustness of the data generated from one model is reinforced by looking at these alternative models.
Supervised machine learning methods employing cross-validation have the advantage of providing actionable results in complex situations, with a multitude of variables. By “supervising” the variables inputted, researchers tailor the results to specific real-world situations.
In contrast, the main currents of social science rely on causal predictions, in which predictions based on trend observation take precedent. These tools may be more accurate in situations where unexpected changes occur and the variables that supervised ML relies on are themselves in flux.