Methodology for evaluating the effectiveness of a commercial organization

Cozac Eugeniu

Modern conditions of the digital economy are characterized by a large flow of information data coming from numerous sources that human intelligence is not able to process. The reason for this is not only a sufficiently large amount of data, but also their uncertainty, as well as not infrequently their insufficiency. The resulting problem of effective pattern recognition, necessary for analysis and forecasting in the economy, began to be solved by using neural networks in mathematical programming. Over the past few years, there has been a significant increase in the use of artificial intelligence capabilities in the economy as a result of the development of neural network technology. The creation of new algorithms based on the fundamental principles of the work of the ordinary human brain made it possible to process much more economic information. Due to the fact that neural networks are not programmed, but are independently trained using the analysis of previous experience, their use allows to reduce the percentage of errors, increasing the efficiency of economic entities. The article discusses the types of neural networks that are currently used in the economy in order to solve various problems. Six classification features of such neural networks are identified. In recent years, convolutional neural networks, which have great capabilities compared to other types of neural network models, have become the most widespread in the economy.

Keywords:neural network technologies, neural networks, pattern recognition, digital technologies.