Article Preview
TopIntroduction
According to a survey, more than 55 billion smart loT devices will used by the year 2025 (Hejazi et al., 2018). Most industrial IoT giants, (such as Google Cloud IoT Edge, Microsoft Azure IoT, and Amazon AWS IoT) now support machine learning for predictive capabilities. When a sensor detects disproportionate heat or vibration inputs, e.g., it sends out an alarm. If same sensor is linked to the internet, the data it collects can be used to gain additional insights and perform analytics for the later use.
As loT becomes more common and widely used where various numbers of devices and sensors generate massive amount of data, and numerous IoT applications are built to provide users with fine-grained and more precise services (Routh & Pal, 2018). These big data from the IoT can additionally be processed and analyzed to deliver insights to IoT service providers and consumers. Many data-driven analytic procedures are used in emerging loT applications to efficiently use large IoT sensing data (Sharma & Wang, 2017). AI algorithms have recently been implemented into loT data analytics procedures. The following main components make up a smart IoT system:
- •
Electrical and mechanical components.
- •
Ports, antennas, and protocols.
- •
Sensors, processors, storage, and applications.
- •
Using analytics to train and execute ML models.
The hundreds of millions of devices that are installed at the edge, in homes and workplaces, warehouses, oil fields and agricultural fields, planes and cars, and vehicles are critical to the success of an loT solution (Chau et al., 2011; Klaine et al., 2017; Suthaharan, 2014). Machine learning (ML) is an impactful method for uncovering insights in data from, for example, Internet of Things (IoT) devices. It consists of several procedures that work smartly to enhance different processes such decision-making in a variety of fields, including education, defense, business, healthcare industry and many more. ML enables IoT to decipher secret patterns in large amounts of data for optimum prediction and recommendation systems. Healthcare has adopted IoT and ML, allowing virtual computers to create medical records, forecast illness diagnosis, and, most notably, track patients in real time.
In healthcare field, machine learning approaches use the growing amount of health data generated by the Internet of Things to optimize patient outcomes. These methods offer both exciting applications and major challenges (Aggarwal, Bhamrah, and Ryait 2016; Aggarwal Karan and Bhamrah 2020; Durga, Nag, and Daniel 2019). Medical imaging, natural language interpretation of medical records, and genetic knowledge are the three primary fields where machine learning is used. Many of these fields are concerned with diagnosis, monitoring, and estimation. A vast infrastructure of medical devices currently produces data, but the enabling infrastructure to properly use those data is often lacking. The many formats in which medical records can be found poses several difficulties in data formatting and can increase noise. Descriptive, exploratory, inferential, statistical, and causal data processing methods are the most common. An exploratory analysis establishes correlations between variables in a dataset while a descriptive analysis provides summaries of results without explanation. Finally, a causal study identifies how changes in one variable cause changes in another. In this article, we look at a short overview of machine learning and IoT, some fundamental methods, and the present state of this technology in healthcare.
TopMain Contribution Of The Paper
A predictive research aims to quantify the likelihood of an event at the level of a person, whereas an inferential study attempts to quantify the degree to which an observed correlation in a population would hold beyond the sample from which it was derived: