Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

When LLM meets Fuzzy-TOPSIS for Personnel Selection through Automated Profile Analysis

Shahria Hoque, Ahmed Akib Jawad Karim, Md. Golam Rabiul Alam, Nirjhar Gope

TL;DR

This paper tackles automated personnel selection by integrating NLP with fuzzy multicriteria decision-making to address subjectivity and uncertainty in candidate evaluation. It introduces a LinkedIn-derived dataset annotated by experts, and a DistilRoBERTa-based scoring pipeline whose attribute scores are aggregated with Fuzzy TOPSIS using triangular fuzzy numbers and centroid defuzzification to rank candidates. The approach achieves high classification accuracy (about 91%) across attributes and strong alignment with human expert rankings, with ranking metrics (MAP ~0.99, MRR ~0.99, NDCG ~0.91–0.93) indicating robust performance. The findings demonstrate a scalable, interpretable workflow for recruitment that reduces human bias and supports real-world decision-making, while outlining future work to expand domains, enhance interpretability, and test robustness.

Abstract

In this highly competitive employment environment, the selection of suitable personnel is essential for organizational success. This study presents an automated personnel selection system that utilizes sophisticated natural language processing (NLP) methods to assess and rank software engineering applicants. A distinctive dataset was created by aggregating LinkedIn profiles that include essential features such as education, work experience, abilities, and self-introduction, further enhanced with expert assessments to function as standards. The research combines large language models (LLMs) with multicriteria decision-making (MCDM) theory to develop the LLM-TOPSIS framework. In this context, we utilized the TOPSIS method enhanced by fuzzy logic (Fuzzy TOPSIS) to address the intrinsic ambiguity and subjectivity in human assessments. We utilized triangular fuzzy numbers (TFNs) to describe criteria weights and scores, thereby addressing the ambiguity frequently encountered in candidate evaluations. For candidate ranking, the DistilRoBERTa model was fine-tuned and integrated with the fuzzy TOPSIS method, achieving rankings closely aligned with human expert evaluations and attaining an accuracy of up to 91% for the Experience attribute and the Overall attribute. The study underlines the potential of NLP-driven frameworks to improve recruitment procedures by boosting scalability, consistency, and minimizing prejudice. Future endeavors will concentrate on augmenting the dataset, enhancing model interpretability, and verifying the system in actual recruitment scenarios to better evaluate its practical applicability. This research highlights the intriguing potential of merging NLP with fuzzy decision-making methods in personnel selection, enabling scalable and unbiased solutions to recruitment difficulties.

When LLM meets Fuzzy-TOPSIS for Personnel Selection through Automated Profile Analysis

TL;DR

This paper tackles automated personnel selection by integrating NLP with fuzzy multicriteria decision-making to address subjectivity and uncertainty in candidate evaluation. It introduces a LinkedIn-derived dataset annotated by experts, and a DistilRoBERTa-based scoring pipeline whose attribute scores are aggregated with Fuzzy TOPSIS using triangular fuzzy numbers and centroid defuzzification to rank candidates. The approach achieves high classification accuracy (about 91%) across attributes and strong alignment with human expert rankings, with ranking metrics (MAP ~0.99, MRR ~0.99, NDCG ~0.91–0.93) indicating robust performance. The findings demonstrate a scalable, interpretable workflow for recruitment that reduces human bias and supports real-world decision-making, while outlining future work to expand domains, enhance interpretability, and test robustness.

Abstract

In this highly competitive employment environment, the selection of suitable personnel is essential for organizational success. This study presents an automated personnel selection system that utilizes sophisticated natural language processing (NLP) methods to assess and rank software engineering applicants. A distinctive dataset was created by aggregating LinkedIn profiles that include essential features such as education, work experience, abilities, and self-introduction, further enhanced with expert assessments to function as standards. The research combines large language models (LLMs) with multicriteria decision-making (MCDM) theory to develop the LLM-TOPSIS framework. In this context, we utilized the TOPSIS method enhanced by fuzzy logic (Fuzzy TOPSIS) to address the intrinsic ambiguity and subjectivity in human assessments. We utilized triangular fuzzy numbers (TFNs) to describe criteria weights and scores, thereby addressing the ambiguity frequently encountered in candidate evaluations. For candidate ranking, the DistilRoBERTa model was fine-tuned and integrated with the fuzzy TOPSIS method, achieving rankings closely aligned with human expert evaluations and attaining an accuracy of up to 91% for the Experience attribute and the Overall attribute. The study underlines the potential of NLP-driven frameworks to improve recruitment procedures by boosting scalability, consistency, and minimizing prejudice. Future endeavors will concentrate on augmenting the dataset, enhancing model interpretability, and verifying the system in actual recruitment scenarios to better evaluate its practical applicability. This research highlights the intriguing potential of merging NLP with fuzzy decision-making methods in personnel selection, enabling scalable and unbiased solutions to recruitment difficulties.
Paper Structure (27 sections, 22 equations, 15 figures, 15 tables, 1 algorithm)

This paper contains 27 sections, 22 equations, 15 figures, 15 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. METHODOLOGY
  4. Dataset Construction and Expert Annotation
  5. Data Description
  6. DistilRoBERTa-Based Scoring with Fuzzy TOPSIS Aggregation for Multi-Criteria Candidate Evaluation
  7. Model Training and Hyperparameter Settings
  8. Architecture of DistilRoBERTa and Scaled Dot-Product Attention Mechanism
  9. Mathematical Representation of the Model Components
  10. Training Objective
  11. Final Output and Prediction
  12. Fuzzy TOPSIS Ranking Algorithm
  13. Performance Metrics
  14. Result Analysis
  15. Performance of DistilRoBERTa on Experience Attribute
  16. ...and 12 more sections

Figures (15)

  • Figure 1: Correlation Heatmap: Encoded Features and Labels
  • Figure 2: Top-level overview of the LLM-Fuzzy TOPSIS Integration for personal selection
  • Figure 3: Top-level overview of the DistilRoBERTa's internals
  • Figure 4: Top-level overview of the Attention head
  • Figure 5: Training and validation loss with accuracy over epochs graph for Experience
  • ...and 10 more figures