Methods of machine learning in choosing a strategy of an enterprise

 Download PDF | Downloads: 29 | Citations: 5

Journal paper

Russian Journal of Innovation Economics ^{(РИНЦ, ВАК)}
опубликовать статью | оформить подписку

Volume 9, Number 1 (January-March 2019)

Citation:

Indexed in Russian Science Citation Index: https://elibrary.ru/item.asp?id=37649386
Cited: 5 by 24.01.2023

Abstract:
The choice of enterprise strategies is particularly difficult in the context of a changing economic environment, inaccurate definitions of variables affecting a company's actions, and incomplete information about competitors' behavior. In such a situation, it is considered preferable to use the methods that are part of machine learning when choosing a strategy. The purpose of the study is to develop a method for choosing a strategy using machine learning tools, which should include ways to analyze and select the most important indicators of an enterprise, and test the efficiency of the method created using simulated or real data. The principal components method is used as data preprocessing. Among the tools of machine learning, the most suitable for solving the task are neural systems. With the help of the neural network system, the Statistica software package implements a mechanism for selecting an appropriate enterprise strategy. A trained neural network in the form of a perceptron allows for a set of selected variables that influence a strategy to choose the scheme of the organization’s actions that best suits the situation at the enterprise.

Keywords: enterprise strategy, machine learning, principal components, neural network system, feature selection, strategy selection

JEL-classification: D81, С45, С65

References:

Страница обновлена: 04.04.2025 в 20:42:10