Open AccessApplication of neural network based on self-organizing incremental learning in anti-interference
Author(s) -
Shaodong Hu,
Xuechen Wu,
Bo Xue
Publication year - 2020
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/914/1/012028
Subject(s) - computer science , artificial intelligence , artificial neural network , competitive learning , robustness (evolution) , dimensionality reduction , machine learning , content addressable memory , nonlinear dimensionality reduction , process (computing) , biochemistry , chemistry , gene , operating system
In view of the shortcomings of traditional neural network in the actual information fusion process, a data anti-interference method based on self-organizing incremental learning neural network is proposed, which is used to cluster and represent the dynamic input data online without prior knowledge. In this paper, the neuron distribution, dynamic node adjustment, topology representation and denoising process of the network are described in detail. Finally, the multi-source heterogeneous data collected by the sensor can be self-adaptive dimensionality reduction and self-organization learning. At the same time, it has strong robustness to noise data, so that it can continuously learn new patterns in data flow. SOINN model is suitable for supervised learning, associative memory, pattern-based reasoning, manifold learning and other learning scenarios. It can also be extended to the application fields of unbalanced, highly complex and nonlinear big data prediction.