Monitoring and Prediction of Air Quality using Neural Network

Abstract

Air Quality monitoring is very important in today s world. There are many harmful pollutants present in the air which are very harmful for human health. Prolonged consumption of such air may lead to severe death and harmful diseases. It is also harmful for crops as well as animals which may damage natural environment. There are several pollutants which are present in the air that decreases the quality of air such as sulfur oxide, nitrogen dioxide, carbon monoxide and dioxide, and particulate matter. Neural Network can be used for prediction of population for short term as well as long term using a deep learning technologies. Neural network specify two types of predictive models. the first one is the a temporal which is for short-term forecast of the pollutants in the air for the short coming or nearest days and the second one is a spatial forecast of atmospheric pollution index in any point of city. The artificial neural networks takes initial information and consider the hidden dependencies are used to improve the efficiency and accuracy of the ecology management decisions. In this paper the forecasting of atmospheric air pollution index in industrial cities based on the neural network model has proposed.

Prediction and Optimization of Air Pollution-A Review Paper

“…Feed Forward Neural Network [1] has been generally used for prediction in short term the neuron network consists of a single layer of output nodes; the inputs are feed directly to the outputs via a series of weights. The neuron network model has specified in the figure which shows the topology of feed forward neural network.…”

“…If we consider the Computer Science Branch then without the Machine Learning it is incomplete. There are various algorithms which have been discussed is going to be discussed is a part of Machine Learning such as ANN [1], Genetic Algorithm [5,6] etc. According to the definition of Wikipedia 3 without depending on instructions, patterns and inference to produce effective output is carried out by Machine Learning.…”

“…Another method is wavelet Neural Network (WNN). The Wavelet Neural Network is a combination of artificial neural network [1][2] and wavelet analysis. The network avoids falling into local optimal, and also has the characteristics of time-frequency local analysis.…”

“…To predict air pollution many researchers are doing many research works to get proper and accurate result with the help of which we can be able to take right steps at the right time. According to the Shwetal Raipure [1] in Monitoring and Prediction of Air Quality using Neural Network stated two concepts of predicting the values of air pollutants they are first using the neural network and the other one is using the spatial Prediction of air pollutant particle index in the atmosphere. The main aspect to improve the efficiency and accuracy is to take into account the hidden values and take initial information as input to artificial neural network Surajit Chattopadhya and Goutami Bandyopadhya stated the prediction of mean monthly total ozone concentrations over the calculated period [3], they used the concept of Sigmoid activation function over single hidden layer and two hidden layer perceptron that are they used two neural network model to predict the ozone mean at 0.9 rates of learning.…”

“…Forward Neural Network1 For activation functions of the neurons, there are used hyperbolic tangent in the neurons of the first layer and linear function in the neurons of the second layer. It has various parameters where, 1) C -current value of the pollutant concentration; 2) T -air temperature; 3) W -wind direction; 4) U -wind speed; 5) P -atmospheric pressure; 6) M -unfavorable meteorological conditions mode; 7) Cp -forecasted value of the pollutant concentration There is another method called nonlinear Auto Regressive model (NARX)[2] based Artificial Neural Network (ANN)[2] used to predict air pollution particles for major two particles such as CO and NOx.…”