UDC 330.342
DOI: 10.36871/ek.up.p.r.2024.01.06.017
Authors
Murad G. Isaev,
Federal State Budgetary Educational Institution of Higher Education
“Dagestan State University”, Makhachkala
Abstract
The article presents a mathematical apparatus for predicting the effectiveness of managing the digital transformation of economic business systems. The methodological basis of the study is an adapted version of the KNN (k-Nearest Neighbors) method or the random forest method when working with a large amount of data. The proposed approach allows us to determine the type of digitalization of the business ecosystem (strong, medium-strong, etc.). Particular attention is paid to the applicability of the KNN method for predicting digital transformation management parameters. The stages of implementation of the method are presented, starting from collecting initial data and ending with assessing the quality of the model using the RMSE (Root Mean Square Error) or R-square metrics. The possibilities of using the algorithm are described using specific examples, including determining the optimal number of nearest neighbors (k) and choosing a distance metric. The analysis is performed on the basis of a multidimensional feature space in which the distance between objects is measured using the Euclidean metric. The algorithm for dividing data into training and test samples, as well as the computational aspects of applying the method, including program code in Python using the scikit-learn, NumPy and SciPy libraries, are discussed in detail. The study has practical significance for specialists in the field of digital transformation management, providing tools for qualitative and quantitative analysis of management parameters and making informed decisions.
Keywords
digital transformation, innovation management, k-nearest neighbors (KNN), parameter forecasting, machine learning.
