Citation

Nagarajaiah, Kavyashree and Pun, Ooi Chee and Tan, Wooi Haw ENHANCED CHROMOSOMAL ABNORMALITY PREDICTION VIA ADVANCED FEATURE SELECTION AND MACHINE LEARNING: A GEO DATASET APPROACH. Journal of Environmental Protection and Ecology, 26 (2). pp. 506-516.

Full text not available from this repository.

Abstract

Fetal chromosomal abnormalities are significant genetic disorders that can result in severe developmental and health challenges, with early detection being critical for effective prenatal care and intervention. This study presents a novel framework for predicting fetal chromosomal abnormalities using gene expression data obtained from the Gene Expression Omnibus (GEO) dataset. The framework integrates several advanced computational techniques to preprocess the data, extract relevant features, select optimal attributes, and classify chromosomal abnormalities with high accuracy. The process begins with pre-processing, where a shift mean draft filter is applied to standardise the data, reducing noise and outlier influence, and ensuring that the analysis focuses on the most relevant gene expression signals. Subsequently, feature extraction is performed using Independent Quasi Attribute Highlight Component Analysis (IQAHCA), which identifies and emphasises the key features that are most indicative of chromosomal abnormalities in fetal development. For feature selection, the Geo Attribute Wave Optimisation Algorithm (GAWOA) is employed to select the most informative attributes while minimising redundancy, improving the efficiency and accuracy of the classification process. To predict fetal chromosomal abnormalities, a Probability Rooted Neural Tree Classifier (PRNTC) is used. This classifier, based on a probabilistic decision-tree architecture, effectively handles the complexities of gene expression data and accurately predicts the presence of chromosomal abnormalities in fetal samples. The proposed framework is evaluated on the GEO dataset, demonstrating promising results in terms of prediction accuracy and robustness. This methodology offers a powerful tool for early detection of fetal chromosomal abnormalities, which can significantly enhance the clinical decision-making process in prenatal care and genetic counseling. © 2025, Scibulcom Ltd.. All rights reserved.

Item Type:	Article
Uncontrolled Keywords:	chromosomal abnormality; Geo Attribute Wave Optimisation Algorithm; Independent Quasi Attribute Highlight Component Analysis; machine learning; Probability Rooted Neural Tree Classifier
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines
Divisions:	Faculty of Engineering (FOE)
Depositing User:	Ms Suzilawati Abu Samah
Date Deposited:	29 May 2025 06:24
Last Modified:	30 May 2025 02:21
URII:	http://shdl.mmu.edu.my/id/eprint/13859

Downloads

Edit (login required)