Scribd
Upload
13 views5 pages

Air Quality Prediction Using Artificial Neural Networks

Uploaded by

Meghana Vemuganti
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
13 views5 pages

Air Quality Prediction Using Artificial Neural Networks

Uploaded by

Meghana Vemuganti
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 5

Abstract:

Air quality has been an enormous health concern in recent decades as the
place has become further industrialized and more and more of its citizens
have begun driving automobiles. The occurrence of air pollution takes place
in the following ways.

1. Release and generation of pollutants from their source.

2. Carry of pollutants in the atmosphere.

3. Penetrating and negatively impacting human health and ecosystems.

We tend to minimise the effects of these emissions as there is no practical,


economical or technical method for zero emissions. PM 2.5 is especially
dangerous because it can pass through the human body's natural filters and
enter the lungs. Health concerns related to PM 2.5 include heart and lung
disease, asthma, bronchitis, and other respiratory problems. Machine
learning, as one of the most accepted techniques, is capable to efficiently
train a model using regression models to predict the hourly air pollution
concentration. Following six regressors chosen for this problem were Linear
Regression, K-Nearnest Neighbor, Stochastic Gradient Descent, Decision
Tree, Random Forest and Multi-layer Perception. Although performance of all
models was comparable, Multi-layer Perception Algorithm model successfully
bring about better accuracy and true positive rate with 95.4 accuracy.
Objective:

The main objectives of this study were

(i) (PM 2.5) and meteorological conditions (air temperature, wind speed,
relative humidity ) as the evaluator, and (ii) to compare the predictive and
generalization abilities of this mode using a limited data set Subsequently,
linear and nonlinear (MLPN) developed and performance criteria parameters

Existing System:

In first approach monitoring of real-time Air Quality Monitoring and another is


developing statistical models using historical data. This paper summarizes
Air Quality Prediction studies in two categories. The first study is on
prediction of PM2.5 / PM10 concentration and on prediction of air pollutants
like CO2, O3, NO2 and then inferring Air Quality Index (AQI) using machine
learning techniques whereas second study is on monitoring real-time AQI
using sensor devices

Proposed System:
Supervised learning repressors Time series forecasting can also be cast as a
supervised learning regression problem where the target variable ’y’ is the
feature pm25 and the predictive variables ’x and y’ are ’y’ at various lags. A
significant benefit to using this approach is the ability to incorporate
multivariate time series into the predictive variable. For this project we use
pm25 at hourly lags up to one week, and deep at the same lags as our
predictive variables. We have chosen the following six repressors for this
problem:

Linear Regression

The benefit of linear regression is that it is the simplest repressor and is


readily interpretable. Linear Regression 20is used for finding linear
relationship between target and one or more predictors. It is generally
applied to one Independent variable and one dependent variable. The
regression has five important assumptions:

1. Linear relationship: Linear regression assumes the relationship between


the independent and dependent variables to be linear.

2. Multivariate normality: The data is normally distributed.

3.Little or No multicollinearity: Independent variables are not correlated to


one another.

4. No Auto correlation: Auto correlation occurs when the residuals are not
independent from each other.

5. Homoscedasticity: If the residuals are equal along the regression line then
the scatter plot is a exemplary way to check homoscedasticity. .

Linear Regression is all about fitting a line into data point.

The mathematical equation of a line is y = mx + c where m is the slope of


gradient and c is y intercept where the line crosses y axis.
To train a model we first check how well the model fits the training data. Its
parameters are then set to fir the training set. Generally RMSE value is used
to evaluate the performance of a regression model .Although practically it is
easier to minimise the MSE value than RMSE value and they both lead to the
same result.

SOFTWARE & HARDWARE REQUIREMENTS

Hardware Requirements:
 Processor : Intel i3 and above

 RAM : 4GB and Higher

 Hard Disk : 500GB: Minimum

Software Requirements:

 Programming Language/Platform : Python

 IDE : pycharm / jupyter

You might also like