Preprints
Preprints
iCost: A Novel Instance Complexity-Based Cost-Sensitive Learning Framework
Summary: In this work, a modification to the original cost-sensitive (weighted) classifiers has been proposed that uses instance-level complexity to penalize the misclassifications made by the model, in contrast to the traditional way of naively weighting all instances equally.
Research Question: How can machine learning models be made more generalizable when training data is scarce, noisy, or imbalanced?
Status: Under review in IEEE Transactions on Knowledge and Data Engineering.
2. An End-to-End Deep Learning Framework for Arsenicosis Diagnosis Using Mobile-Captured Skin Images
Summary: A non-invasive DL framework that diagnoses arsenicosis from mobile-captured skin images, enabling early detection and screening in low-resource settings.
Status: Under review in Computer Methods and Programs in Biomedicine.
3. Toward Accessible Dermatology: Skin Lesion Classification Using Deep Learning Models on Mobile-Acquired Images
Status: Accepted for publication in the International Conference on Signals and Systems (ICSigSys-2025).
An ML-based decision support system for reliable diagnosis of ovarian cancer by leveraging explainable AI [ DOI ] [Code - GitHub]
Authors: Asif Newaz, Abdullah Taharat, Md Sakib Ul Islam, Khairum Islam, A.G.M. Fuad Hasan Akanda
Journal: Informatics in Medicine Unlocked, Q2 (cite score = 9.5)
Summary: In this work, new distinctive biomarkers have been identified for OC diagnosis in pre- and post-menopausal women. XAI tools (SHAP) have been incorporated to provide a more transparent prediction. SHAP is further utilized to discern the stochastic nature and associated bias from the ML and feature selection algorithms.
Machine Learning Enabled Multimode Fiber Specklegram Sensors: A Review [ DOI ]
Authors: Asif Newaz, Md Omar Faruque, Rabiul Al Mahmud, Rakibul Hasan Sagor, M. Zahed Mustafa Khan.
Journal: IEEE Sensors Journal, Q1 (cite score = 8.2)
An intelligent decision support system for the accurate diagnosis of cervical cancer [ DOI ]
Authors: Asif Newaz, Sabiq Muhtadi, Farhan Shahriyar Haq
Journal: Knowledge-Based Systems; Q1 (cite score = 14.8)
Survival prediction of heart failure patients using machine learning techniques [ DOI ]
Authors: Asif Newaz, Nadim Ahmed, Farhan Shahriyar Haq
Journal: Informatics in Medicine Unlocked
Multi-year dataset on daily electricity demand, generation, load shedding, and external conditions in Bangladesh curated via web scraping [ DOI ]
Authors: Md. Ikrama Hossain, Tasnia Nafs, Sakif Yeaser, Asif Newaz
Journal: Data in Brief
iBRF: Improved Balanced Random Forest Classifier [ DOI ] [ Code - GitHub ]
Authors: Asif Newaz, Md Salman Mohosheu, Abdullah al Noman, Dr. Taskeed Jabid
Conference: 35th Conference of Open Innovations Association FRUCT, Finland.
Summary: The work presents a modified version of the original BRF classifier, designed to improve its generalizability and performance in handling imbalanced data.
A Comprehensive Evaluation of Sampling Techniques in Addressing Class Imbalance Across Diverse Datasets [ DOI ]
Authors: Md Salman Mohosheu, Abdullah al Noman, Asif Newaz, Al-Amin, Dr. Taskeed Jabid
Conference: 6th International Conference on Electrical Engineering and Information & Communication Technology 2024, MIST, Bangladesh.
Summary: A detailed experimental analysis has been conducted in this study to understand the efficacy of different sampling techniques and to identify their weaknesses.
Survival Analysis of Breast Cancer Patients: A Population-based Study from SEER [ DOI ]
Authors: Md. Saifur Rahman, Aseer Imad Keats, Mohammad Abrar Kabir, Asif Newaz, Md Mahfuzul Islam
Conference: 3rd International Conference on Electrical, Computer, and Energy Technologies, South Africa
Performance Improvement of Heart Disease Prediction by Identifying Optimal Feature Sets Using Feature Selection Technique [ DOI ]
Authors: Asif Newaz, Sabiq Muhtadi
Conference: 2021 International Conference on Information Technology (ICIT), Egypt