Tracing origins: Comparing a heuristic, machine learning, and a large language model for migrant identification

Migration has become a defining feature of contemporary societies and digital economies. Migration significantly reshapes the global digital economy, yet the lack of explicit country-of-origin data in digital records hinders large-scale quantitative research. This study addresses this methodological gap by systematically evaluating and comparing three computational approaches for inferring migrant origins from social networking data: a rule-based heuristic, supervised machine learning, and a large language model. Drawing on a validated dataset of LinkedIn profiles enriched with external geographic and linguistic sources, we assess performance across 80 countries of origin. Our results demonstrate that a transparent, theoretically grounded rule-based heuristic outperforms both machine learning and large language model approaches. We find that inference robustness improves substantially when integrating multiple profile attributes—namely, education, native language, and work history—rather than relying solely on name-based cues. To demonstrate the practical scalability of the heuristic, we applied it to a dataset of 5479 IT-related startup founders in Germany, with results indicating comparatively higher migrant participation in IT-related startups than in the wider German startup landscape. Beyond its methodological contributions, this study conceptualizes country of origin as a vital construct for information systems research, offering a validated framework for studying diversity, global knowledge flows, and digital entrepreneurship. • Conceptualizes country of origin as a construct in information systems research • Compares country-of-origin inference across three computational methods • Shows the rule-based heuristic outperforms machine learning and a large language model • Reveals multi-feature profiles improve inference beyond name-based cues • Applies approach using LinkedIn data from IT-related startup founders in Germany

Authors

Institutions

Publication Details

Journal
Technological Forecasting and Social Change
Published
2026-04-24
DOI
https://doi.org/10.1016/j.techfore.2026.124666
Primary Topic
Migration, Ethnicity, and Economy
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Tracing origins: Comparing a heuristic, machine learning, and a large language model for migrant identification

Jonathan Hoke, David Bendig, Leoni Onken
Technological Forecasting and Social Change
Migration, Ethnicity, and Economy
article

Tracing origins: Comparing a heuristic, machine learning, and a large language model for migrant identification

Jonathan Hoke, David Bendig, Leoni Onken
article en

Abstract

Migration has become a defining feature of contemporary societies and digital economies. Migration significantly reshapes the global digital economy, yet the lack of explicit country-of-origin data in digital records hinders large-scale quantitative research. This study addresses this methodological gap by systematically evaluating and comparing three computational approaches for inferring migrant origins from social networking data: a rule-based heuristic, supervised machine learning, and a large language model. Drawing on a validated dataset of LinkedIn profiles enriched with external geographic and linguistic sources, we assess performance across 80 countries of origin. Our results demonstrate that a transparent, theoretically grounded rule-based heuristic outperforms both machine learning and large language model approaches. We find that inference robustness improves substantially when integrating multiple profile attributes—namely, education, native language, and work history—rather than relying solely on name-based cues. To demonstrate the practical scalability of the heuristic, we applied it to a dataset of 5479 IT-related startup founders in Germany, with results indicating comparatively higher migrant participation in IT-related startups than in the wider German startup landscape. Beyond its methodological contributions, this study conceptualizes country of origin as a vital construct for information systems research, offering a validated framework for studying diversity, global knowledge flows, and digital entrepreneurship. • Conceptualizes country of origin as a construct in information systems research • Compares country-of-origin inference across three computational methods • Shows the rule-based heuristic outperforms machine learning and a large language model • Reveals multi-feature profiles improve inference beyond name-based cues • Applies approach using LinkedIn data from IT-related startup founders in Germany

Technological Forecasting and Social ChangeVol. 229
University of Münster (DE)
Deutsche Forschungsgemeinschaft
Decent work and economic growth
Openalex Percentile: Top 3%
Migration, Ethnicity, and Economy
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Tracing origins: Comparing a heuristic, machine learning, and a large language model for migrant identification — Jonathan Hoke, David Bendig, et al. · Technological Forecasting and Social Change (2026) | TGRS Research Map | TGRS