DNA Insight: A Machine Learning Framework for Hair-Based Forensic DNA Matching in Crime Investigation

Shalini R; Supreetha Gowda H D

doi:10.70528/IJLRP.v7.i7.2298

DNA Insight: A Machine Learning Framework for Hair-Based Forensic DNA Matching in Crime Investigation

Author(s)	Ms. Shalini R, Dr. Supreetha Gowda H D
Country	India
Abstract	Forensic DNA analysis is among the most reliable techniques available to criminal investigators, and hair evidence recovered from crime scenes is a particularly common and informative biological source because a hair strand containing its root can yield DNA sufficient for individual identification. Conventional DNA profiling, however, depends on extensive laboratory procedures—extraction, amplification, sequencing, and expert comparison—that are accurate but slow, costly, and difficult to scale when large numbers of samples must be processed. This paper presents DNA Insight, a machine-learning-based decision-support framework that automates the comparison of hair-derived short tandem repeat (STR) DNA profiles against suspect records to predict whether a sample constitutes a match or a non-match. DNA profile values extracted from forensic hair evidence are converted into structured tabular features—including per-locus allele overlap indicators, exact and partial locus match counts, average and maximum allele differences, and an aggregate similarity score—and are processed through a pipeline of cleaning, Min-Max normalization, and feature scaling before being supplied to a Random Forest classifier. The system was developed and evaluated on a synthetic forensic hair STR dataset of 2,000 records with 28 attributes, partitioned into an 80:20 train-test split, and was deployed behind a web-based interface that allows an investigator to submit DNA profile values and receive a prediction together with a confidence score. Experimental results show that the trained Random Forest model achieves an accuracy of 96.8%, precision of 96.5%, recall of 97.2%, and an F1-score of 96.8% on the held-out test set, correctly classifying 387 of 400 test samples. These results indicate that ensemble machine learning classifiers can provide a fast, consistent, and reasonably accurate preliminary screening tool to support—rather than replace—forensic experts during hair-based DNA investigation workflows
Keywords	-Forensic Science; DNA Profiling; Hair Evidence; Machine Learning; Random Forest; Crime Investigation; Short Tandem Repeat (STR); Classification.
Field	Computer Applications
Published In	Volume 7, Issue 7, July 2026
Published On	2026-07-13
DOI	https://doi.org/10.70528/IJLRP.v7.i7.2298

View / Download PDF File

About IJLRP Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics	Join as a Reviewer Editors & Reviewers Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijlrp.com

International Journal of Leading Research Publication

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Monthly Scholarly International Journal

DNA Insight: A Machine Learning Framework for Hair-Based Forensic DNA Matching in Crime Investigation

Share this