Parallel modular neural network-based byproduct gas real-time prediction method

Abstract

The invention relates to a parallel modular neural network-based byproduct gas real-time prediction method. According to the method, according to the principle of state space segmentation of a neural network, Fuzzy c-means (FCM) clustering is adopted to divide sample data into a plurality of categories; each category is corresponding to the subspace (namely, module) of one state space; the data are reconstructed, so that a prediction model can be established; in a modeling process, an improved echo state network is provided, a modular method is adopted to segment the state space of the neural network into a plurality of independent sub spaces, wherein each subspace is a sub network; a reserve pool sharing method is used in combination, so that the training of all networks is completed in the same reserve pool, each sub space is corresponding to an output weight matrix, and therefore, the operation rules of a system can be better simulated; a network training problem is simplified into a parallel training problem of a plurality of small networks, so that the calculation process of the model can be accelerated; and a big data sample containing more useful information is introduced, so that the prediction precision of the model can be improved; and a Map Reduce computing framework is adopted to parallelize solution problems, so that a high speed-up ratio can be obtained, and real-time prediction of the metallurgical gas system can be realized.

Description

technical field [0001] The invention belongs to the field of information technology, relates to an echo state neural network, fuzzy clustering and parallel computing, and is a real-time prediction method for a metallurgical gas system driven by a large data set. The present invention utilizes a large amount of existing data at the metallurgical enterprise site, and first uses Fuzzy c-means (FCM) clustering to divide the sample data into several categories according to the principle of neural network state space segmentation, and then reconstructs the data to establish a prediction model . In the modeling process, a parallel modular neural network is constructed, and the state space of the neural network is divided into multiple independent subspaces through a modular method, and each subspace is a subnetwork. The network training problem is simplified to the parallel training problem of multiple small networks to speed up the model calculation process, the introduction of lar...

AI Technical Summary

Problems solved by technology

If the traditional time series-based forecasting algorithm has to deal with such a large number of samples, there will be problems such as slow training speed, complex algorithm, and low efficiency.

How to solve the contradiction between accuracy and solution efficiency is the main problem to be dealt with based on the data prediction method, that is, if the gas system cannot be predicted in a timely and effective manner, it will affect the subsequent balance adjustment of the gas system and affect the production efficiency of the enterprise

Images