ABSTRACT
Air pollution is one of the most significant concerns of the present era, which has severe and alarming effects on human health and the environment, thereby escalating the climate change issue. Hence, in-depth analysis of air pollution data and accurate air quality forecasting is crucial in controlling the growing pollution levels. It also aids in designing appropriate policies to prevent exposure to toxic pollutants and taking necessary precautionary measures. Air quality in Delhi, the capital of India, is inferior compared to other major cities in the world. In this study, daily and hourly concentrations of air pollutants in the Delhi region were collected and analyzed using various methods. A comparative analysis is performed based on months, seasons, and the topography of different stations. The effect of the Covid-19 lockdown on the reduction of pollutant levels is also studied. A correlation analysis is performed on the available data to show the relationships and dependencies among different pollutants, their relationship with weather parameters, and the correlations between the stations. Various machine learning models were used for air quality forecasting, like Linear Regression, Vector Auto Regression, Gradient Boosting Machine, Random Forest, and Decision Tree Regression. The performance of these models was compared using RMSE, MAE, and MAPE metrics. This study is focused on the dire state of air pollution in Delhi, the primary reasons behind it, and the efficacy of calculated lockdowns in bringing down pollution levels. It also highlights the potential of Linear Regression and Decision Tree Regression models in predicting the air quality for different time intervals. © 2022 IEEE.
ABSTRACT
Predicting the corona virus can be divided into several phases, including a state-wide analysis that includes active, confirmed, cured, deaths as well as an increase in cases on a daily basis that includes each and every state of India as well as Union Territories. This also includes a thread of new corona virus cases from throughout India and forecasts the outbreak's conclusion in the next days. Machine learning algorithms like SVM, Linear Regression and Decision Tree Regression are used to analyze this data and improve this model's outcome. In this study, Jupyter notebook is used which provides an environment that is suited for machine learning principles. This technique provides for a comprehensive analysis of the virus's spread, including total and active cases, as well as forecasting future outbreaks and a weekly study epidemic. © 2021 IEEE
ABSTRACT
Machine Learning is a seeing gradual acceptance in every other field and hence healthcare is not an exception. Machine Learning in healthcare is useful for various works like suggesting judgments, presenting timely risk scores and specific resource allocation. Thus Machine Learning has become an important factor during the pandemic period. A rapid growth in the amount of covid cases has led to the corresponding upsurge in demand for ICU care which has also put an immense pressure on the healthcare workers and hospital resources. The risk here is that the medical care workers will find it difficult to give frequent monitoring for the covid patients at high threat of clinical disintegration. This paper focuses on the study and experiments of Machine Learning Algorithms towards predicting whether a confirmed COVID patient will need to use the ICU or not. It also addresses some drawbacks while using the Machine Learning models in real-world problems. Several ML models like XGBoost, Random forest, Classification and Regression decision tree (CART) are used to predict patient deterioration. The Machine Learning models gave good competence in prognosticating critical COVID-19 and therefore it can be used to accurately predict risks in patients, monitor the hospital resource allocation, arrangement of medical resources, and enhance the administration of the unprecedented pandemic. ©2021 IEEE