Using Recurrent Neural Networks for Data-Centric Business

Abstract

Over the last decade, print and online publications have published a huge number of articles with proposals for the introduction and use of neural networks in existing business and research. And just two years ago, many leading IT companies showed the world already created smart applications in the field of neural networks, which indicates the uniqueness and relevance of this technology. This is easily explained by the fact that systems based on neural networks are able to perform complex business tasks more efficiently and cheaper than the people. While working with big data, the probability of error remains relatively low. Unlike humans, neural networks are more stable. With long-term high loads, the efficiency of solving problems by the neural network does not sag. Finally, neural networks make free the people from monotonous computing operations and enable creative implementation. However, the issue of choosing the most appropriate topology and type of neural networks remains extremely relevant. The best results are demonstrated by recurrent neural networks. This paper is devoted to the review and proposing on the use of recurrent neural networks in data-centric business. The tests and analysis of their results demonstrated the relevance of the use of modern architectures of artificial neural networks.