Abstract:
Currently, in many industrial and urban area of Bangladesh, analyzing and predicting air quality has
become one of the most essential activities. The nature of air is favorably influenced due to
Different types of pollution caused by transportation, power, fuel utilizes and so on. The testimony
of hurtful gases is making a genuine danger for the quality of life in smart cities. With expanding air
pollution, we need to implement proficient air quality analyzing and predicting models which gather
data about the convergence of air poisons and give evaluation of air pollution in every zone. In this
paper, we utilize a famous machine learning technique, to figure pollutant and particulate levels and
to predict and analysis the air quality. The quality of air is influenced by multi-dimensional
components including area, time, and uncertain factors. The aim of this paper is to investigate various
machine learning based techniques for air quality analyzing and prediction.
Air quality is surveyed dependent on a banding framework which gauges the degrees of poisons,
specifically Ozone (O3), Nitrogen dioxide (NO2) and Particulate issue - PM10 and PM2.5. The
general air quality list at a specific time is given as the greatest band for any poison. PM2.5 is fine
particulate matter of size under 2.5 micrometers and is considered to impact sly affect wellbeing
going from cellular breakdown in the lungs to cardiovascular illnesses.
This project intends to predict the air quality band for PM2.5 utilizing present and authentic
contamination information in blend with predicted air information which is promptly accessible. To
tackle this issue, initially, exploratory information investigation will be directed on accessible air and
pollution datasets to find the relationship between various highlights. Subsequent to utilizing
appropriate information cleaning and highlight designing strategies dependent on the perceptions
made, the practicality of utilizing different Machine Learning procedures, for example, LSTM will
be investigated.