Estimation of Ground Water Levels by Using Linear Regression and Logistic Regression Algorithms

Abstract

An indispensable resource that is given naturally to every living being on this planet is water. The Blue Planet contains 71 percentage of water. Water is available in many forms which are not suitable for consumption purpose. The 98% of groundwater that is available on earth is fresh and relevant for drinking. To estimate the amount of water present inside the earth at any particular zones which will help us to manage usage of water for the future generations. In this paper, we are implementing a machine learning model which has the potential to estimate the water levels inside the ground. This machine learning model is trained by using Logistic Regression Algorithm, and it is tested by using a linear regression algorithm which gives more accurate results than any other machine learning algorithms. The model is trained by taking various conditions like the recharge from rainfall during monsoon season and non-monsoon season, other sources to recharge during monsoon season and non-monsoon season, annual replenishable resource, net annual ground water availability, draft due to irrigation needs, domestic and industrial water supply needs, total annual draft, projected demand for domestic and industrial uses up to 2025, ground water availability for future irrigation use, stage of ground water development in real time to predicts the water availability in particular area..

 Keywords:Machine Learning, Decision Tree, Linear Regression, Ground Water levels