Creating a Data-Driven Taxonomy

• Activity Number: 246 - Data Science
• Type: Contributed
• Date/Time: Wednesday, August 11, 2021 : 10:00 AM to 11:50 AM
• Sponsor: Section on Statistical Computing
• Author(s): Randall Powers* and Wendy Martinez and Terrance Savitsky
• Companies: Bureau of Labor Statistics and Bureau of Labor Statistics and Bureau of Labor Statistics
• Keywords: clustering; R Shiny; text data; unsupervised learning; Monthly Labor Review

Abstract

The Monthly Labor Review (MLR) is published by the U.S. Bureau of Labor Statistics. Issues of the MLR often focus on a particular topic, and most articles are written by BLS staff. The need for a classification system of past MLR articles that can be used to label future articles has been recognized by the agency. To address this problem, we employed various unsupervised learning approaches to cluster MLR articles from 2000 to 2013. In this presentation, we will discuss the processes used to prepare the data set, the cluster approaches used, and the results.