Sales and print publications prediction using fuzzy regression neural network

  • Ramyar keyvaninejad, Morteza Romoozi, Hamideh Babaei


One of the main problems in predicting artificial neural networks is to provide the necessary data for prediction because neural networks need a lot of data to obtain accurate results. However, it should be noted that collecting the data required by the network is firstly very expensive and secondly requires a long time. Therefore, due to rapid changes in real environments, especially economic and financial systems, forecasting in these environments with a small amount of data must be effective and efficient. Fuzzy forecasting methods require less data than other methods due to the use of fuzzy numbers instead of definite numbers, but their performance is not always satisfactory. In this paper, artificial neural networks have been proposed by fuzzy regression to limit the amount of data required for networking and to obtain more accurate results for predicting publication sales. Experimental results show 95% of the effect in the proposed method and in predicting sales.