A collaborative fuzzy-neural approach for forecasting the price of a DRAM product

Abstract

This paper presents a collaborative fuzzy-neural approach for accurately and precisely forecasting the price of a Dynamic Random Access Memory (DRAM) product, which is considered as one of the most important semiconductors widely used in various applications. In the collaborative fuzzy-neural approach, some experts from the application domain are invited to form a council. For some aspects, these experts put forward different points of view. These views are incorporated into the Back Propagation Network and Nonlinear Programming (BPN-NP) approach, and result in different fuzzy-valued price forecasts. To derive a single representative value from these fuzzy price forecasts, the Fuzzy Intersection and Radial Basis Function network (FI-RBF) approach is employed. The effectiveness of the collaborative fuzzy-neural approach is validated with a real example containing the 256-day price data of a 1G DRAM product.

An Agent-Based Fuzzy Collaborative Intelligence Approach for Predicting the Price of a Dynamic Random Access Memory (DRAM) Product

“…Tarui and Tarui [1] found that for DRAM long-term trends do exist and could be approximately modeled. The same phenomenon can also be observed in other industries [2][3][4][5][6][7][8][9][10]. Cupertino [11] was not so optimistic.…”