Enhancing Recruiter Outreach: Predicting Job Ad Quality with Advanced Machine Learning and NLP Models

Abstract

In today’s competitive recruitment landscape, crafting impactful job outreach messages is essential for attracting top talent. This study presents a novel machine learning and NLP-driven framework for predicting recruiter message quality on professional platforms like LinkedIn, aiming to enhance response rates and hiring success. Our approach leverages a multi-label text classification framework that identifies five critical message attributes: call to action, common ground, credibility, incentives, and personalization. Using a labeled dataset of 97,710 messages annotated across these five categories, we benchmark various machine learning and deep learning models, including Decision Trees, Linear SVM, Logistic Regression, Random Forest, LSTM, and customized transformer-based BERT models. The dataset was meticulously curated to address generalization challenges, with 94,010 samples for training and 3,700 samples in a diversified test set. Model performance was assessed using accuracy, with the customized BERT model achieving 95.67%. Our findings underscore the potential of this framework to enhance recruiter outreach strategies, providing actionable insights to refine message quality and improve candidate engagement.

Received: 14 August 2024 | Revised: 29 October 2024 | Accepted: 16 December 2024

Conflicts of Interest

The author declares that he has no conflicts of interest to this work.

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Author Contribution Statement

Shaida Muhammad: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration.