Machine Learning based model to Detect Anomaly in the Water
Gaurav Kumar, Saroj Kumar Pandey, Dhirendra Prasad Yadav, Kamred Udham Singh, Teekam Singh, Ankit Kumar
Abstract
The environmental impact of technology is only one area where the field is expanding rapidly without necessarily good results. [1] Water contamination is one such critical problem. Because water is essential to human survival, polluting it has serious consequences for civilization. [1] Knowing what harmful substances are in the water and being able to identify them is a crucial step toward reducing the damage they cause. While Image classification and sparse coding have been widely employed for water pollution identification [1,] they have unintended implications when applied to bacterial information detection. The current technologies used to identify contaminants in photos and water pipes, such as Convolutional neural networks (CNN), natural language processing (NLP), and the Internet of Things (IoT), are expensive and inefficient when compared to 3D microscopy data. It is a model that uses categorization and data visualization to estimate the total number of bacteria in a given environment. Accuracy scores may be seen in the model's studied graphs, decision tree, confusion matrix, and classification report. By the use of these data sets, the crucial factors for clean water were evaluated, and a model was built that can categorize the massive amounts of information, find meaningful patterns, and make reliable predictions. It has been determined which method yields the highest accuracy prediction (87 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup> ) by thoroughly analyzing the available models' accuracy scores. This study thus provides a valid evaluation of bacterial water vs unadulterated water.