A Machine-Learning Based Nano-Biosensing Study on Cancer Diagnosis and IoT Applications

Abstract

Cancer is a very big cause of death & an extremely costly illness to treat. The likelihood of a cure & survival rates increase with early cancer discovery, but sadly, most tumours are only discovered after they have spread to other parts of the body. Biosensors are tools created to identify a particular biological analyte by translating the intricate biological interactions into an electrical signal whose strength is related to the analyte's concentration. Nanotechnology with nanoparticles enhances & modifies the bio-recognition element portion to improve the bio-sensing phenomenon & makes it one of the hottest issues attracting the scientific fraternity. The primary goal of this study is to examine several nanostructures that have been applied to bio-sensing, & certain implementations in the areas of cancer diagnosis & IoT, & also a brief introduction to machine-learning-based bio-sensing. To categorise microarray data in this study, IOT & machine learning (ML) methods were applied. Two sets of data were used to generate them: one having 1,919 protein types & the other with 24,481 protein types for 97 individuals, 46 of whom had a reoccurring illness & 51 of whom did not. According to the study's findings, before feature reduction, logistic regression (LR) yielded the highest outcomes (90.23%) & also Random Forest yielded good outcomes (67.22%). Support Vector Machine had the best accuracy rates-99.23% in both techniques in the first data & 87.87% in Random Logistic Regression (RLR) & 88.82% in LTE in the second data. To conclude, nanotechnology development surely helped biosensors to advance to new heights. © 2023, Ismail Saritas. All rights reserved.