SVM Prediction Model Interface for Plant Contaminates Shilpi Aggarwal 1* , Madhulika Bhatia 2 , Rosy Madaan 1 , Hari Mohan Pandey 3 1 Faculty of Engineering and Technology, Manav Rachna International Institute of Research and Studies, Faridabad 121004, Haryana, India 2 Department of Computer Science Engineering, Amity University, Noida 201313, Uttar Pradesh, India 3 Department of Computer Science, Edge Hill University, Lancashire L394QP, United Kingdom Corresponding Author Email: shilpiaggarwal.phd2017@mriu.edu.in https://doi.org/10.18280/ts.380412 ABSTRACT Received: 14 June 2021 Accepted: 19 July 2021 In today's time, our nature is fighting against many life-threatening problems which can even threaten the existence of life on the Earth. Pollution is one of the deadliest problems among them. It is caused primarily by means of air, water and land but air pollution is the most severe and dreadful among them. It is caused by introduction of toxic substances like oxides of Sulphur, nitrogen and carbon into the atmosphere making it unfit for living beings. Along with humans, plants have also become a victim to it, and this fact is mostly ignored. A model has been designed to predict the effect of pollution on plants. Image samples of 5 Indian oxygen rich plants namely Ocimum Tenuiflorum, Sansevieria Trifasciata, Chlorophytum Comosum, and Azadirachta Indica have been taken for analysis and various properties like shape, color, corners and texture of the plants were considered from these input RGB images. As a consequence of these properties and the pollution index value, certain calculations have been performed and the results are compared with the threshold values. Based on the range in which the calculated results lie, the plants will be categorized into a category which depicts the severity level of pollution in the environment. After applying the model on the images, a dataset was prepared and SVM classification model has been trained on it which predict with an accuracy of 85%. It has been presented in the form of an interactive user interface to predict the effect of pollution on plants. Plants are an integral part of nature and should not be ignored. Keywords: pollution, plants, prediction, classification, air quality index, GUI 1. INTRODUCTION The presence of hazardous elements into the environment is known as pollution. Air, water and land pollution are the three prominent categories of pollution, air pollution being the most severe and dangerous among them. Burning of fossil fuels is the primary cause of air pollution. It includes the smoke from industries, automobiles and other machines. Burning crackers and fire accidents are also among the contributing factors. The pollutants comprise of particulate matter and a number of gases. Sulphur dioxide, nitrogen dioxide, carbon dioxide, carbon monoxide etc. are regarded as highly toxic gases. These polluting substances have pathetic consequences on human health and plant life. They lead to global warming which in turn triggers the melting of glaciers leading to floods, soil erosion, droughts, fires, loss of wildlife and as a result, the complete ecological balance gets disturbed. The quality of air can be checked by the Air quality index (AQI) value. It indicates the degree of pollution in the air. The higher the AQI value, the greater the level of air pollution and the greater the health concern. It depends upon the concentration of various pollutants like Sulphur dioxide (SO2), nitrogen dioxide (NO2), particulate matter etc. The algorithm to calculate AQI is defined by the Indian National Air Quality Standards (INAQS) [1]. There are around 390,800 different plant species surviving on earth. According to a botanist, about 21 percent of the total plant species on earth are on the path of extinction [2]. Amidst so much deforestation, sustaining the plant species is very difficult but the need of the hour. In order to achieve this goal, recognition and classification of endangered species is indispensable. Different researchers have proposed several algorithms for the recognition of plants. These algorithms include the techniques like deep learning [3, 4] and machine learning [5, 6] and they are widely used for the identification process. During the 18th century, some classification schemes for plants were developed by Carolus Linnaeus [7]. There are many other classification techniques which are used to classify objects. The algorithms like Naïve Bayes, Support Vector Machine, K Nearest Neighbor along with the hybrid approaches were also applied by many researchers [8-11]. This paper starts the sequence with an introduction, continuing with the related work. Next section speaks about the material and methods required to implement the research. Third section explains the proposed model in detail. In continuation to this section, results obtained after implementation of the model and the graphical user interface have been discussed which show the effect of pollution on plants. Then the final section gives the conclusion of the paper. 2. RELATED WORK As per a report published by the World Health Organization, the average life span of humans has got diminished by one year Traitement du Signal Vol. 38, No. 4, August, 2021, pp. 1023-1032 Journal homepage: http://iieta.org/journals/ts 1023