Chapter 12 - Machine learning in surface plasmon resonance for environmental monitoring

Abstract

Biosensors are being used for diverse purposes, including food safety, biological studies, drug screening, medical diagnosis, and environmental sensing. They have numerous advantages such as high speed, high sensitivity, low costs, and so forth. Among a various range of biosensors, optical biosensors, in particular, surface plasmon resonance (SPR), have brought new advantages like real-time and label-free detection with higher levels of sensitivity and cost-effectiveness. Considering environmental hazards endangering human health and applications of SPR in environmental monitoring, SPR has indicated great promise, especially in detecting environmental hazards with low molecular weights in complex matrices. Apart from the all as-mentioned positives, there exist some relevant concerns such as data processing, sensor reliability and accuracy, and poor signal-to-noise ratios, all of which can be improved using machine learning (ML) approaches. ML can evaluate large data, produce appropriate results, even from noisy and low-resolution sensing data, and find interrelationships between signals and bioevents. In this chapter, SPR principles and its use for environmental monitoring will be discussed. Finally, different ML algorithms and their applications in different SPR sensors are aimed to be reviewed.