Two Statistical Approaches to Justify the Use of the Logistic Function in Binary Logistic Regression

Logistic regression is a commonly used classification algorithm in machine learning. It allows categorizing data into discrete classes by learning the relationship from given set of labeled data. learns linear dataset and then introduces nonlinearity through an activation function to determine hyperplane that separates points two subclasses. In case logistic regression, most perform binary classification. The choice for classifications justified its ability transform any real number probability between 0 1. This study provides, different approaches, rigorous statistical answer crucial question torments us, namely where does this on which neural network algorithms are based come from? Moreover, it determines computational cost using theoretical experimental approaches.