Linear regression and k-means methods for forecasting and clustering of production indicators in Orange Data Mining

Abstract

The paper presents the specifics of the Orange Data Mining software system in the field of data analytics, namely, its practical application for forecasting and clustering of production indicators of enterprises. A linear regression model has been developed and tested, which has universal properties and can be used by enterprises for forecasting and adjusting data. Additionally, the model is supplemented with the K-Means clustering algorithm, which allows obtaining accurate clusters and analysing the results. The obtained results are visualised using internal software tools. Intermediate and general recommendations for applying the model with different types of data are proposed. The experimental results show that the Orange Data Mining software system can be successfully used for forecasting and clustering production indicators