USING THE SCIKIT-LEARN LIBRARY IN MACHINE LEARNING CLASSIFICATION METHODS

Abstract

The article analyzes the selection of an appropriate machine learning algorithm, which depends on several factors, such as the amount of data, its quality and diversity, as well as the awareness of the business goals that need to be achieved from this data. Machine learning in scikit-learn is all about importing the right modules and running the model fitting method. The object of the study is the application of various classification algorithms for grouping the results of machine learning models in cases of binary and multi-class classification. Therefore, it is necessary to test different algorithms, evaluating their effectiveness on test data sets, and choose the best one. In this regard, it is important to select algorithms that are most suitable for the given task. The authors of the paper focused on accuracy, training time, and data characteristics. Therefore, the choice of the optimal algorithm is a combination of business requirements, technical specifications, experimental activities and taking into account the available time. The implementation of machine learning methods in various fields was also investigated. The machine learning process is described, which includes the following stages: data preparation, training set creation, classifier development, classifier training, forecasting, classifier performance evaluation, and parameter setting. An analysis of the application of various classification algorithms was carried out using the Scikit-learn library with Python. An analysis of the method of model selection, calculation, formatting and data preparation was also carried out, as well as the selection of optimal input values and models. Several approaches to classifier evaluation are evaluated. The main goal of the work is to study the library to assess its practical effectiveness. Classification methods in machine learning using Scikit-Learn are described. A comparison of different classification methods was made using the Scikit-learn library for machine learning models.