An ML-based Cancer Genome Profile Drug Prediction Framework

Main Article Content

Rahul Chauhan

Abstract

This article discusses predictive modelling for individualised cancer therapy. We employ machine learning to predict pharmaceutical reactions, drug synergies, drug target-interactions, and cancer classification. This work aims to construct machine learning prediction models for drug sensitivity prediction, medicine combination therapy, drug target interaction prediction, and cancer classification. C-HMOSHSSA, a cancer classification framework using multi-objective meta-heuristic and machine learning, predicts both recognised and new cancer biomarkers. A hybrid feature selection algorithm (HMOSHSSA) for gene selection improves on the multi-objective spotted hyena optimizer (MOSHO) and salp swarm algorithm (SSA). Four classifiers are trained using the HMOSHSSA dataset. The approach uncovers informative gene groupings. KSRMF also predicted missing drug response values. The BE-DTI framework uses dimensionality reduction and active learning to predict drug-target interactions. Active learning helps under-sampling bagging ensembles. High-dimensional data demands unique dimension reduction strategies. Five existing (RF, SVM) feature-based approaches are compared to the proposed framework's performance..

Downloads

Download data is not yet available.

Metrics

Metrics Loading ...

Article Details

How to Cite
Chauhan, R. . (2018). An ML-based Cancer Genome Profile Drug Prediction Framework. Turkish Journal of Computer and Mathematics Education (TURCOMAT), 9(2), 488–495. https://doi.org/10.17762/turcomat.v9i2.13853
Section
Research Articles