Hybrid neural network

About: Hybrid neural network is a research topic. Over the lifetime, 1305 publications have been published within this topic receiving 18223 citations.

Neuro-genetic approach to optimize parameter design of dynamic multiresponse experiments

Abstract: Engineers have widely applied the Taguchi method, a traditional approach for robust experimental design, to a variety of quality engineering problems for enhancing system robustness. However, the Taguchi method is unable to deal with dynamic multiresponse owing to increasing complexity of the product or design process. Although several alternative approaches have been presented to resolve this problem, they cannot effectively treat situations in which the control factors have continuous values. This study incorporates desirability functions into a hybrid neural network/genetic algorithm approach to optimize the parameter design of dynamic multiresponse with continuous values of parameters. The objective is to find the optimal combination of control factors to simultaneously maximize robustness of each response. The proposed approach is based on three stages which (1) use neural networks for constructing a response function model of a dynamic multiresponse system, (2) use exponential desirability functions for evaluating overall performance of a specific factor combination, and (3) use a genetic algorithm to optimize parameter design. Effectiveness of the proposed approach is illustrated with a simulated example. Analysis results reveal that the approach has higher performance than the traditional experimental design.

A hybrid neural network/genetic algorithm applied to breast cancer detection and recurrence

Abstract: Genetic algorithms (GAs) and neural networks (NNs) are both inspired by computation in biological systems and many attempts have been made to combine the two methodologies to boost the NNs performance. This paper deals with the evolutionary training of a feedforward NN for both breast cancer detection and recurrence. A multi-layer perceptron (MLP) has been designed for this purpose, using a GA routine to set weights, and a Java implementation of this hybrid model has been made. Four databases concerning cancer detection and recurrence have been used, two databases containing numerical attributes only, one database containing ordinal (categorical) attributes solely and one database with mixed attributes. In comparison to some standard NNs, the performance of this approach using the same databases is shown to be superior. Moreover, this hybrid MLP/GA model is very flexible in terms of providing accurate classification, even with different types of attributes, which is usually found in medical studies.

Multi-Model Fusion Short-Term Load Forecasting Based on Random Forest Feature Selection and Hybrid Neural Network

Abstract: In an increasingly open electricity market environment, short-term load forecasting (STLF) can ensure the power grid to operate safely and stably, reduce resource waste, power dispatching, and provide technical support for demand-side response Recently, with the rapid development of demand side response, accurate load forecasting can better provide demand side incentive for regional load of prosumer groups Traditional machine learning prediction and time series prediction based on statistics failed to consider the non-linear relationship between various input features, resulting in the inability to accurately predict load changes Recently, with the rapid development of deep learning, extensive research has been carried out in the field of load forecasting On this basis, a feature selection algorithm based on random forest is first used in this paper to provide a basis for the selection of the input features of the load forecasting model After the input features are selected, a hybrid neural network STLF algorithm based on multi-model fusion is proposed, of which the main structure of the hybrid neural network is composed of convolutional neural network and bidirectional gated recurrent unit (CNN-BiGRU) The input data is obtained by using long sliding time windows of different steps, then multiple CNN-BiGRU models are trained respectively The forecasting results of multiple models are averaged to get the final forecasting load value The load datasets come from a region in New Zealand and a region in Zhejiang, China, are used as load forecast examples Finally, a variety of load forecasting algorithms are introduced for comparison The experimental results show that our method has a higher accuracy than comparison models