An efficient query processing with approval of data reliability using RBF neural networks with web enabled data warehouse

Abstract

To rise above the limitation of the Traditional load forecasting method using data warehousing system, a new load forecasting system basing on Radial Basis Gaussian kernel Function (RBF) neural network is proposed in this project. Genetic algorithm adopting the actual coding, crossover and mutation probability was applied to optimize the parameters of the neural network, and a faster growing rate was reached. Theoretical analysis and models prove that this model has more accuracy than the traditional one. There are several methods available to integrate information source, but only few methods focus on evaluating the reliability of the source and its information. The emergence of the web and dedicated data warehouses offer different kinds of ways to collect additional data to make better decisions. The reliable and trust of these data depends on many different aspects and metainformation: data source, experimental protocol. Developing generic tools to evaluate this reliability represents a true challenge for the proper use of distributed data. In this project, RBF neural network based approach to evaluate data reliability from a set of criteria has been proposed. Customized criteria and intuitive decisions are provided, information reliability and reassurance are most important components of a data warehousing system, as their power in a while retrieval and examination.