As a result of global urbanization and the drive for more sustainable and livable cities, smart cities are becoming increasingly significant. It is widely known that air pollution exists all around us as a result of numerous automobiles on the road, stubble burning, and industrial air pollution, all of which impair our health, particularly our lungs. Chronic diseases are growing more common as our society ages. Chronic respiratory problems such as asthma and chronic obstructive pulmonary disease (COPD) impact millions of people globally, and the number is growing every day. COPD, for example, afflicted roughly 251 million individuals worldwide in 2016, and claimed the lives of 3.17 million people (i.e., 5% of the population). There are many distinct sorts of the pulmonary diseases: COPD, pulmonary fibrosis, asthma, and lung cancer, to name a few. This chapter explains the three machine learning algorithms that helped to diagnose pulmonary diseases based on the dataset.
TopSupport Vector Machine
These introduced in the 1960s and improved in the 1990s are supervised ML techniques that can be used for classification as well as regression (ElearningProviders, 2006),. They are, nevertheless, most typically used in categorization problems. SVMs have a distinct implementation strategy when compared to other ML algorithms. Because of their ability to handle both continuous and categorical variables, they've recently garnered a lot of traction. This model is essentially a representation of separate classes in a hyperplane in a multidimensional space. To minimize the error, SVM will iteratively construct the hyperplane. The goal of SVM is to divide datasets into classes to find the “Maximum Marginal Hyperplane (MMH)”.
The following concepts are important in SVM:
- •
Support Vectors: These are the data points that are closest to the hyperplane, and between them, a dividing line will be drawn.
- •
A “hyperplane” is a decision plane or space that divides a collection of things into discrete classes.
- •
Margin is defined as the distance between two lines on a set of data points from different classifications. It is possible to compute the perpendicular distance between the line and the support vectors. A wide margin is considered a good margin, whereas a tiny margin is considered a bad margin.
The main goal of SVM is to divide the datasets into classes to find a “Maximum Marginal Hyperplane (MMH)”.
TopK-Nearest Neighbor (K-Nn) Algorithm For Machine Learning
The Supervised Learning Technique is used in K-Nearest Neighbor, a simple ML methodology. It assumes that the new case/data and previous cases are comparable, and it assigns the new case to the category that is closest to the existing categories It saves all available data and groups fresh data points into categories based on their similarity. As new data is generated, the K-NN algorithm can quickly classify it into the appropriate category. This method can be applied to both Regression and Classification problems, but it is more commonly employed for Classification. It's a non-parametric method, which means no assumptions about the underlying data are made. It's also known as a lazy learner algorithm because it doesn't learn from the training set right away, instead of storing the data and acting on it later during classification. During the training phase, this method simply saves the dataset, and when new data is received, it is categorized into a group that is quite similar to the new data.
Consider the following scenario: there is a snapshot of an animal that could be a cat or a dog, and we want to know if it is a cat or a dog. The K-NN technique can be used for this identification because it is based on a similarity metric. It will compare the new data set to photos of cats and dogs and divide it into two groups based on the most similar qualities.
Need for K-NN Algorithm
Assume there are two categories, A and B, and we have a new data point X1 to which we wish to assign it. This type of challenge necessitates the use of a K-NN algorithm. K-NN is used to determine the category or class of a dataset. Consider Figures 1 and 2: