TY - GEN
T1 - Unsupervised Labor Intelligence Systems
T2 - 16th Symposium and Summer School on Service-Oriented Computing, SummerSOC 2022
AU - Cascavilla, Giuseppe
AU - Catolino, Gemma
AU - Palomba, Fabio
AU - Andreou, Andreas S.
AU - Tamburri, Damian A.
AU - Van Den Heuvel, Willem Jan
N1 - Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2022
Y1 - 2022
N2 - In recent years, job advertisements through the web or social media represent an easy way to spread this information. However, social media are often a dangerous showcase of possibly labor exploitation advertisements. This paper aims to determine the potential indicators of labor exploitation for unskilled jobs offered in the Netherlands. Specifically, we exploited topic modeling to extract and handle information from textual data about job advertisements for analyzing deceptive and characterizing features. Finally, we use these features to investigate whether automated machine learning methods can predict the risk of labor exploitation by looking at salary discrepancies. The results suggest that features need to be carefully monitored, e.g., hours. Finally, our results showed encouraging results, i.e., F1-Score 61%, thus meaning that Data Science methods and Artificial Intelligence approaches can be used to detect labor exploitation—starting from job advertisements—based on the discrepancy of delta salary, possibly representing a revolutionary step.
AB - In recent years, job advertisements through the web or social media represent an easy way to spread this information. However, social media are often a dangerous showcase of possibly labor exploitation advertisements. This paper aims to determine the potential indicators of labor exploitation for unskilled jobs offered in the Netherlands. Specifically, we exploited topic modeling to extract and handle information from textual data about job advertisements for analyzing deceptive and characterizing features. Finally, we use these features to investigate whether automated machine learning methods can predict the risk of labor exploitation by looking at salary discrepancies. The results suggest that features need to be carefully monitored, e.g., hours. Finally, our results showed encouraging results, i.e., F1-Score 61%, thus meaning that Data Science methods and Artificial Intelligence approaches can be used to detect labor exploitation—starting from job advertisements—based on the discrepancy of delta salary, possibly representing a revolutionary step.
KW - Artificial Intelligence
KW - Case study
KW - Data science
UR - http://www.scopus.com/inward/record.url?scp=85140740555&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-18304-1_5
DO - 10.1007/978-3-031-18304-1_5
M3 - Conference contribution
AN - SCOPUS:85140740555
SN - 9783031183034
T3 - Communications in Computer and Information Science
SP - 79
EP - 98
BT - Service-Oriented Computing - 16th Symposium and Summer School, SummerSOC 2022, Revised Selected Papers
A2 - Barzen, Johanna
A2 - Leymann, Frank
A2 - Dustdar, Schahram
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 3 July 2022 through 9 July 2022
ER -