Scholarlink

<p>The biomolecules DNA and RNA are essential in the study of genetics. Genes play a crucial role in regulating cellular activities and their associated processes. Cancer is one of the numerous disorders that can affect cells, which are the body’s basic building blocks. Determining the biomolecule features that are initially impacted is also essential for cancer diagnosis and treatment. This article investigates the nucleotide distribution inside DNA biomolecules for cancer detection. Using the Fuzzy Markov Chain of Nucleotides (FMCN) for feature extraction and selection, we present a novel approach to analyzing nucleotide distribution characteristics as fuzzy Markovian features across genomic data. We thoroughly investigate these nucleotide features in both cancerous and non-cancerous DNA sequences. We compared classification models for cancer identification, such as SVM (Support Vector Machine) with its different kernels, DT (Decision Trees), KNN (K-Nearest Neighbors), RF (Random Forest), GBDT (Gradient Boosting Decision Tree), XGBoost (Extreme Gradient Boosting), and Neural Network (NN) models. FMCN utilizes fuzzy Markov models to predict nucleotide probabilities, thereby minimizing both computing time and the number of required features. Various case studies have demonstrated the effectiveness of this method through 10-fold cross-validation and accuracy evaluation. The explicit relationship between the classifier and the feature reduction strategy is essential to improving the accuracy of cancer predictions. We employ genomic areas to recognize cancerous and non-cancerous DNA sequences. SVM with a linear kernel and MLP achieved 100% accuracy for cancer and 97% for non-cancer cases, but KNN, DT, RF, GBDT, and XGBoost beat SVM with sigmoid and polynomial kernels.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Hybrid Fuzzy-Markov chain model for feature reduction in cancer DNA sequence classification

  • 期刊论文

  • 发表日期：

    2025年9月8日

查看全文
  • Vijay Kalal,
  • Brajesh Kumar Jha

摘要

The biomolecules DNA and RNA are essential in the study of genetics. Genes play a crucial role in regulating cellular activities and their associated processes. Cancer is one of the numerous disorders that can affect cells, which are the body’s basic building blocks. Determining the biomolecule features that are initially impacted is also essential for cancer diagnosis and treatment. This article investigates the nucleotide distribution inside DNA biomolecules for cancer detection. Using the Fuzzy Markov Chain of Nucleotides (FMCN) for feature extraction and selection, we present a novel approach to analyzing nucleotide distribution characteristics as fuzzy Markovian features across genomic data. We thoroughly investigate these nucleotide features in both cancerous and non-cancerous DNA sequences. We compared classification models for cancer identification, such as SVM (Support Vector Machine) with its different kernels, DT (Decision Trees), KNN (K-Nearest Neighbors), RF (Random Forest), GBDT (Gradient Boosting Decision Tree), XGBoost (Extreme Gradient Boosting), and Neural Network (NN) models. FMCN utilizes fuzzy Markov models to predict nucleotide probabilities, thereby minimizing both computing time and the number of required features. Various case studies have demonstrated the effectiveness of this method through 10-fold cross-validation and accuracy evaluation. The explicit relationship between the classifier and the feature reduction strategy is essential to improving the accuracy of cancer predictions. We employ genomic areas to recognize cancerous and non-cancerous DNA sequences. SVM with a linear kernel and MLP achieved 100% accuracy for cancer and 97% for non-cancer cases, but KNN, DT, RF, GBDT, and XGBoost beat SVM with sigmoid and polynomial kernels.