Transparent Peer Review By Scholar9

Paper Title

“Explainable AI-Based Multilingual Fake Job Detection Using Transformer Models for Indian Recruitment Platform”

Description / Abstract

The rapid expansion of online recruitment platforms has significantly transformed the employment process by connecting employers and job seekers through digital technologies. However, the growing number of fake job advertisements has become a serious cybersecurity issue, leading to financial fraud, identity theft, phishing attacks, and misinformation. The multilingual nature of Indian recruitment platforms further complicates fake job detection because advertisements are published in different regional languages and code-mixed text. Existing machine learning methods often struggle to classify multilingual job postings accurately and provide limited explanation for their predictions.This research proposes an Explainable Artificial Intelligence (XAI)based multilingual fake job detection framework using transformer-based language models such as Multilingual BERT (mBERT) and XLM-RoBERTa. The proposed framework performs data collection, preprocessing, language identification, tokenization, transformer-based classification, and explainability analysis. Explainability techniques including SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Modelagnostic Explanations) are incorporated to improve transparency by highlighting the important features influencing each prediction. Experimental evaluation demonstrates that the XLM-ROBERTA model achieves superior classification performance compared with conventional machine learning and deep learning approaches. The proposed framework improves recruitment security, enhances user confidence, and supports trustworthy online recruitment systems for multilingual environments.

Sumit Shekhar

Reviewer 4.8

Vishesh Narendra Pamadi

Sumit Shekhar Reviewer

Review Request Accepted

Sumit Shekhar Reviewer

Approved Rating

Relevance and Originality

Methodology

Validity & Reliability

Clarity and Structure

Results and Analysis

Comment

Relevance and Originality

The manuscript addresses an important cybersecurity problem involving fraudulent job advertisements in multilingual Indian recruitment platforms. Combining multilingual transformers with SHAP and LIME is relevant and potentially useful. However, the specific novelty should be stated more clearly in relation to existing transformer based fake job detection research.

Methodology

The proposed workflow is logical, covering data collection, preprocessing, language identification, transformer classification, and explainability. However, details concerning dataset sources, annotation, data splitting, model parameters, and training procedures are insufficient for reproducibility. The preprocessing choices also require further justification for transformer based models.

Validity and Reliability

The reported XLM RoBERTa results are strong, but the manuscript does not provide statistical validation, repeated experiments, confusion matrices, or language specific performance. Since the dataset contains more genuine than fake advertisements, additional evaluation measures are necessary to establish reliability.

Clarity and Structure

The manuscript has a logical structure, but several sections repeat general explanations of transformers and explainability. Some tables and technical terms require formatting and language correction. Greater emphasis on the actual experimental contribution would improve presentation.

Results and Analysis

XLM RoBERTa achieves the strongest reported performance, but the analysis is mainly descriptive. Language wise results, error analysis, explanation examples, and comparison with recent studies would strengthen the interpretation and support the practical claims.

IJ Publication Publisher

We sincerely appreciate the reviewer’s careful assessment and valuable professional observations. The detailed comments provide useful guidance for strengthening the manuscript and improving its overall scholarly quality. The review will be valuable to the editorial evaluation process.

Publisher

IJ Publication

All Reviewers

Sumit Shekhar

Reviewer

Vishesh Narendra Pamadi

Reviewer

Das Pakanti Yadav

Reviewer

Raja Kumar Kolli

Reviewer

Nimeshkumar Patel