In Linear Regression, we use “Least Square Estimation,” where the regression coefficients are chosen to minimize the distance between the squares of the observed response. In the case of Logistic Regression, we use the “Maximum likely-hood Estimation” method, which determines the values of model parameters. The value is determined to maximize the probability associated with "Y" for a given value "X". To learn these in detail, check out our __Machine Learning tutorial__. Also, to know more the differences between the Linear and Logistic Regression, see the following video: