International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 09 Issue: 08 | Aug 2022
p-ISSN: 2395-0072
www.irjet.net
A Smart air pollution detector using SVM Classification M.Meghana1, Dr.R.Maruthamuthu2 1student,
Department of Computer Applications, Madanapalle Institute of Technology and science, India Department of Computer Applications, Madanapalle Institute of Technology and science, India ---------------------------------------------------------------------***--------------------------------------------------------------------the Bayes theorem to solve classification issues. It is mostly Abstract - One of the top priorities for the governments of 2Asst.Professor,
employed in text classification tasks with high-dimensional training data.
developing nations, especially India, is the control of the fast rising levels of air pollution. People can take action to reduce pollution by becoming more aware of the degree of pollution in their immediate surroundings. Fossil fuel combustion, travel habits, and industrial elements like power plant emissions all have a big impact on air pollution. The total amount of particulate matter (PM) that affects air quality. When it is concentrated heavily in the aerial medium, it poses serious health risks to people. It must therefore be controlled by regularly checking its atmospheric concentration.
LITERATURE SURVEY [1] A Machine Learning Approach for Air Quality Prediction: Model Regularization and Optimization. Dixian Zhu, Changjie Cai, Tianbao Yang, and Xun Zhou In this study, we address the problem of air quality forecasting by predicting the hourly concentration of air pollutants, such as ozone, particle matter (PM 2.5), and sulfur dioxide. One of the most used techniques, machine learning, can effectively train a model on massive amounts of data by employing powerful optimization algorithms. Although some studies have used machine learning to predict air quality, most of the earlier research has only used data from a few years and has only trained basic regression models (either linear or nonlinear) to predict the hourly air pollution concentrationBy defining the prediction across 24 hours as a multi-task learning (MTL) issue, we offer improved models in this study to forecast the hourly air pollution concentration based on meteorological data from previous days. This makes it possible for us to choose a suitable model using various regularization methods. We suggest a practical regularization by mandating that the prediction models for consecutive hours be near one another and contrast it with other common regularizations for MTL, such as ordinary Frobenius norm regularization, nuclear norm regularization, and l 2, 1 -norm regularization. Our tests demonstrated that the suggested parameter-reducing formulations and consecutive-hour-related regularizations outperform existing standard regression models and existing regularizations in terms of performance
Key Words: Particulate matter, SVM classifier, Regression, and Quality
1.INTRODUCTION There can be both naturally occurring and artificial particles. Examples include dust, ash, and sea spray. Burning of solid and liquid fuels, such as when creating energy, heating a home, or driving a car, releases particulate matter (including soot). The size of the particles varies (i.e. the diameter or width of the particle). The term "PM2.5" refers to the quantity of airborne particles per cubic meter of air that have an average diameter of less than 2.5 micrometers Another name for it is fine particulate matter, or PM2.5. When airborne levels of tiny particulate matter (PM2.5) are quite high, it poses a substantial risk to people's health and is a significant portion of the pollutant index. PM2.5, or particulate matter 2.5, lowers visibility and causes the air to appear hazy when concentrations are high. The identification of air pollution and forecasting of PM2.5 levels have been accomplished using a variety of machine learning models based on a data set made up of daily atmospheric conditions. Dan Wei forecasted Beijing's air quality using the Naive Bayes classification and support vector machine algorithms to get the lowest possible error. José Juan Carbajal developed the fuzzy inference technique, which he then applied to categorize parameters using logic and include them in an air quality score.
[2]. Sachit Mahajan, Ling-Jyh Chen, and Tzu-Chieh Tsai are the authors of "An Empirical Study of PM2.5 Forecasting Using Neural Network”. In most industrialized and developing nations, significant efforts have been undertaken in recent years to restrict air pollution levels. Many efforts are being undertaken to control the levels of fine particulate matter (PM2.5), which is thought to be one of the main causes of declining public health. Forecasting PM2.5 levels accurately is a difficult undertaking that has relied heavily on model-based approaches. In this study, we investigate fresh approaches to PM2.5 hourly forecasting. In order to increase prediction
1.1 Naïve Bayes Classification A group of classification methods built on the Bayes Theorem is known as naive Bayes classification. Every pair of features being categorised independently from one another is not a common principle shared by all of the algorithms. It is a supervised learning algorithm that uses
© 2022, IRJET
|
Impact Factor value: 7.529
|
ISO 9001:2008 Certified Journal
|
Page 1529