gradient descent

This article discusses the practical aspects of applying logistic regression for binary data classification. Logistic regression determines the probability of an object belonging to one of two classes. This probability is calculated with the help of a sigmoid function, the argument of which is a linear convolution of the feature vector of the object with the weighting coefficients obtained during the minimization of the logarithmic loss function. Predicted class labels are determined by comparing the calculated probability with a given threshold value.

This article reviews the theoretical aspects of logistic regression for binary data classification, including data preparation processes, training, testing, and model evaluation metrics.

Requirements for input data sets are formulated, methods of coding categorical data are described, methods of scaling input features are defined and substantiated.

In this article is considered the algorithm of logistic regression and construction of the neural network for the recognition of handwritten symbols in the image. Examples of implementation of two approaches for solving the problem of numerical recognition are given. The efficiency of using a neural network, as the provision of the most reliable recognition results, is explored.