Diagonal Recurrent Neural Network Based Prediction Model for Sales Forecasting

Autor:	Rajat Gera, Smriti Srivastava, Rajesh Kumar, R.S. Tiwari
Rok vydání:	2020
Předmět:	Sales forecasting Radial basis function network Recurrent neural network Artificial neural network Diagonal recurrent neural network Computer science Feature (computer vision) business.industry Pattern recognition Artificial intelligence Perceptron business Gradient descent
Zdroj:	Metaheuristic and Evolutionary Computation: Algorithms and Applications ISBN: 9789811575709
Popis:	This article describes the novel application of Diagonal Recurrent Neural Network (DRNN) for solving the sales forecasting problem. The proposed model is having temporal features since it consists of feedback adjustable connections that provide it with the memory feature. Gradient descent based Dynamic Back-Propagation (DBP) method is applied for generating the tuning equations. The performance evaluation of the results obtained from the proposed model is compared with that of the well-known models such as Multi-Layered Perceptron (MLP) containing single hidden layer, Radial Basis Function Network (RBFN) and Legendre-Functional Link Neural Network (LeFLNN). The simulation results obtained shows the superior performance of the DRNN model over the other considered neural network models.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::bc1012247c5eb9259315da3453c3714e https://doi.org/10.1007/978-981-15-7571-6_29 Zobrazit plný text záznamu