Complete Scholarly Publications Timeline

High-Performance Fraud Detection Using Tree-Based Ensembles with SMOTE: An Optimized RF-XGBoost Framework

3rd IEEE Technology & Innovation Conference (TIC-2026), Spectrum International University College, Kuala Lumpur, Malaysia, June 5–6, 2026 · IEEE Conference ID #68483 · co-authored

Uses Random Forest and XGBoost with SMOTE oversampling to detect credit card fraud in highly imbalanced datasets, reporting near-perfect accuracy.

Fail-safe Autonomy: Detecting and Preventing Silent Failures in Clinical Agentic AI

3rd IEEE Technology & Innovation Conference (TIC-2026), organized by Spectrum International University College, Kuala Lumpur, Malaysia, June 5–6, 2026 · supported by the World Research Union · IEEE Conference ID #68483 · co-authored

Proposes an autonomous AI workflow to catch “silent failures” in clinical decision-support systems before they cause patient harm, validated on real clinical notes.

Certificate of Award signed by Valliappan Raju (Conference Chair, GISMA University, Germany) and Ts Dr Jamal Zaidi (Deputy Vice-Chancellor, Spectrum International University College, Malaysia).

Cross-Modal Extrinsic Calibration of LiDAR and RGB Cameras (re-presented)

FET'26, University of Bahrain, April 22–23, 2026 (online)

Re-presentation of the 2021 JISEM paper (item 1) at an international engineering & technology forum.

Optimal Deep Belief Network for Online Banking Fraud Detection on Digital Financial Services

IEEE conference paper, Jan 2026 · co-authored · published & indexed in IEEE Xplore

Proposes a refined UNet + deep belief network pipeline for detecting online banking fraud, tested on public fraud datasets.

Soft Computing Models for Fraudulent Transactions Prediction Using Multi-Objective Optimization

IEEE conference paper, Jan 2026 · co-authored

Combines deep learning (dual-attention UNet) with multi-objective optimization to detect fraudulent e-commerce transactions with high accuracy.

Cross-Modal Learning Between LiDAR and RGB for Smart Fleet Systems

Working paper · dated Nov 2025

Surveys LiDAR–camera sensor fusion methods (PointPainting, BEVFusion, TransFusion) and proposes a fleet-specific deployment framework robust to sensor dropouts.

Artificial Intelligence in ERP Inventory Management: Transforming Supply Chain Agility and Efficiency

IJSET-associated

Case-study-driven paper on how AI improves demand forecasting and stock optimization in ERP inventory modules, using Zara, Amazon, and healthcare examples.

“AI Cancers”: Systemic Limitations Threatening the Integrity of Artificial Intelligence

IJSET · presented at ICSEMT 2025, Sept 29–30, 2025

Uses a “cancer” metaphor to categorize systemic AI flaws (bias, hallucination, goal misalignment, data poisoning, lack of explainability) and surveys emerging detection/containment strategies.

Retrieval-Augmented Generation and Hallucination in Large Language Models: A Scholarly Overview

SJET, Vol 13, Issue 5, pp. 328–330 — published May 19, 2025

Explains why LLMs hallucinate and how retrieval-augmented generation (grounding responses in external documents) helps reduce the problem.

Crowdsource Activity as Applications of Neural Networks

IJSRET, Vol 11, Issue 4, Jul–Aug 2025

Explores how neural networks improve crowdsourcing platforms — validating annotations, predicting worker reliability, and personalizing task assignment.

Evolution of a Neural Network in ERP Implementations

IJSRET, Vol 11, Issue 4, Jul–Aug 2025

Traces how neural networks have been progressively integrated into ERP systems, from early forecasting bolt-ons to today's AI-native platforms, with case studies on SAP, Oracle, and Microsoft Dynamics.

Recurrent Neural Networks in Complex Finance Applications

IJSRET, Vol 11, Issue 4, Jul–Aug 2025

Surveys how RNNs/LSTMs/GRUs are used in trading, credit risk, fraud detection, and portfolio optimization, with case studies on their strengths and limitations.

Comparative Analysis of Generic and Specialized Natural Language Processing Models Using Prompt Engineering

IJSRET, Vol 11, Issue 4, Jul–Aug 2025

Tests whether prompting alone can let general-purpose models (GPT, BERT) match domain-specific models (BioBERT, FinBERT) in healthcare, finance, and legal tasks — finds specialized models still win on precision-critical work.

Essential Strategic Factors for Ensuring a Successful ERP Implementation

Scholars Journal of Engineering and Technology (SJET), Vol 13, Issue 3, pp. 163–167 — published March 19, 2025

A closely related but distinct paper, going deeper into technical execution factors (system architecture, data migration, QA) alongside the organizational ones.

Critical Success Factors for ERP Implementation

International Journal of Engineering & Scientific Research (IJESR), Vol 13, Issue 4, April 2025

A literature review identifying the factors (change management, executive sponsorship, data governance, risk planning) most associated with successful ERP rollouts.

ERP Post-implementation Challenges and Solutions

IJSRET, Vol 11, Issue 1, Jan–Feb 2025

Reviews recurring problems companies face after go-live on ERP systems and summarizes best practices for change management, data quality, and vendor partnerships to resolve them.

The Impact of Artificial Intelligence on Accounting Practices

IJNREFM, Vol 2, Issue 6, Nov–Dec 2024

Surveys how AI/ML and generative AI are changing accounting workflows (AP/AR, close, tax, audit) and proposes a governance framework for using AI safely in financial controls.

A Multi-Agent System for Extracting and Querying Financial Documents

IJNREFM, Vol 2, Issue 5, Sep–Oct 2024

Describes a pipeline of cooperating AI agents that read financial filings (10-Ks, invoices, statements), pull out structured facts, and answer questions about them with traceable citations.

Cross-Modal Extrinsic Calibration of LiDAR and RGB Cameras Using Compact Planar Targets and Joint Geometric Optimization

Journal of Information Systems Engineering and Management (JISEM), 6(3) — Received Aug 1, 2021 · Accepted Sept 28, 2021

Proposes a low-cost way to align LiDAR and camera sensors using small checkerboard-style targets instead of bulky calibration rigs, aimed at robotics and autonomous-vehicle applications.