Termite Fly Optimization Algorithm towards Measuring and Predicting Air Pollution in California
-
1
University of California, Davis, John Muir Institute of the Environment, United States
-
2
University of California, Davis, Department of Land, Air, and Water Resources, United States
Background and Objective:
Despite systematic improvements since the 1970s, large areas of California’s rural interior continue to experience prolonged episodes of poor air quality, leading to upper-respiratory disease and other human health risks. One factor that contributes to unhealthy air is geo-topographic; the valley which forms along the central spine of California, known as the Central Valley, collects and concentrates diffuse sources of air pollution from both domestic and overseas sources. Such geomorphology, coupled with climatic inversion during hot periods, traps particulate and gaseous pollution for prolonged periods of time and thereby increases people’s exposure to unhealthy air. Sources of pollution include vehicles, stationary fossil fuel combustion, wildfires, natural soils and agriculture; however distinguishing among them, and the varieties of air pollutants, remains challenging. Connecting sources to conditions is a key component of driving solutions and improving public health in California as a whole. Here we use an AI approach based on distributed sensors of air pollution to develop new predictive algorithms and spatial air-quality index for California.
Statement of Methods:
We present an efficient and automated approach (TFO-Termite Fly Optimization Algorithm) towards predicting the amount of air pollution in various locations in California. In order to develop this model, several volumes of data spanning 5 years from various sources including meteorology data on climatic changes and concentrations of various pollutants were collected from OpenAQ for California. The Air quality index (AQI) is computed using this automatic approach that links the amount of air pollution ascribed to different pollutants. The TFO algorithm is a newly designed nature-inspired algorithm, which is based on observations of termite flight patterns and their affinity for bright lights. In computer science, TFO is a computational method that is used to optimize a problem iteratively by improving individual solutions vis-à-vis an offset distance that measures the quality of performance against observations. We use MATLAB to construct the TFO algorithm and compare the accuracy of results to alternative AI approaches, such as Random Forest, Support Vector Machine, and Auto-Regressive Moving Average.
Result and Discussion:
Input data includes the concentration of PM2.5, SO2, NO2, and CO. The entire raw data has additional properties including time, location, longitude, latitude, and meteorological properties. The size of the dataset used in the experiment is 15580 rows, collected for two months from 15-09-2018 to 15-11-2018. During the experimental analysis various results are computed and the performance of the TFO algorithm is compared with other machine learning algorithms such as Multi-Layer Perceptron (MLP), Random Forest (RF) and Auto Regressive Moving Average (ARIMA). Initially, TFO algorithm is used for extracting the parameter and value features from the 15,580 data, which is having 8 features such as PM2.5, PM10, O3 (1hr avg), O3 (8hrs avg), SO2 (1hr avg), SO2 (24hrs avg), CO and NO2. With the extracted feature TFO clustered and calculated the amount of pollutants mixed in the dataset.
The proposed TFO algorithm produces the high accuracy of 0.99 in predicting the air quality while other existing algorithms such as MLP, RF and ARMIA produced the accuracy of 0.908, 0.782 and 0.608 respectively.
Conclusion:
The main conclusions of this analysis are:
• TFO approach was used to classify levels of pollution such as Good, Mild, Moderate, Unhealthy, Highly Unhealthy and Hazardous
• From the obtained results, TFO algorithm provides high accuracy in predicting air pollution for California
• Hence, this algorithm is concluded as a best algorithm for prediction and it can also be used for various other prediction applications
Keywords:
Termite Fly Optimization Algorithm,
Air pollution (SO2, NO2, NO),
California,
OpenAQ,
Machine learning algorithm
Conference:
GeoVet 2019. Novel spatio-temporal approaches in the era of Big Data, Davis, United States, 8 Oct - 10 Oct, 2019.
Presentation Type:
Poster-session
Topic:
Emerging GIS, data science and sensor technologies adapted to animal, plant and human health, including precision medicine and precision farming
Citation:
Manogaran
G and
Houlton
BZ
(2019). Termite Fly Optimization Algorithm towards Measuring and Predicting Air Pollution in California.
Front. Vet. Sci.
Conference Abstract:
GeoVet 2019. Novel spatio-temporal approaches in the era of Big Data.
doi: 10.3389/conf.fvets.2019.05.00089
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
13 Jun 2019;
Published Online:
27 Sep 2019.
*
Correspondence:
Dr. Gunasekaran Manogaran, University of California, Davis, John Muir Institute of the Environment, Davis, United States, gmanogaran@ucdavis.edu