U.S. flag

An official website of the United States government

Skip Header


Ensemble Modeling Techniques for NAICS Classification in the Economic Census

Written by:
Working Paper Number ADEP-WP-2024-03

Abstract

The Business Establishment Automated Classification of NAICS (BEACON) is a machine learning tool developed by the U.S. Census Bureau to help Economic Census respondents select their establishment’s North American Industry Classification System (NAICS) code. BEACON uses the respondent-provided text, in real time, to predict the respondent’s most likely NAICS code. BEACON utilizes past Economic Census responses in conjunction with other data sources such as NAICS manual descriptions and Internal Revenue Service data  to create a data dictionary for training and testing purposes. Through an ensemble method, BEACON hierarchically predicts a respondent’s NAICS code, first at the 2-digit level and then at the 6-digit level. As a potential means of improving BEACON’s current prediction method, we are exploring the use of model stacking to incorporate predictions from alternative models. This research paper details the ensemble modeling behind BEACON and explores this application of model stacking to improve predictions.

 

Page Last Revised - July 5, 2024
Is this page helpful?
Thumbs Up Image Yes Thumbs Down Image No
NO THANKS
255 characters maximum 255 characters maximum reached
Thank you for your feedback.
Comments or suggestions?

Top

Back to Header