用于荧光光谱识别的自适应训练及删剪算法

Adaptive training and pruning algorithm for recognizing the fluorescence spectrum

提出利用基于自适应训练及删剪算法的网络模型对荧光光谱进行识别。首先采用基于递归最小方差的自适应学习算法对网络模型进行学习训练,由于该算法的学习步长能够自行调整,初始参数少,所以收敛速度很快;再利用删剪算法对学习后的网络结构进行删剪,优化网络的拓扑结构,降低网络的计算复杂度,提高网络的泛化能力;然后对优化后的网络进行再学习,使优化后的网络具有最佳参数;最后利用优化后的网络对测试样本进行识别。仿真实验表明,与删剪前的网络结构相比,在降低了网络的计算复杂度的同时,删剪优化后的正确识别率依然是100%。

A neural network based on adaptive learning and pruning algorithm is proposed to recognizing the fluorescence spectrum.Firstly employed adaptive learning algorithm based on recursive least square to train the tapped delay neural network, because this algorithm＇s learning step can be auto-conditioning and the number of it＇s tunable parameters is few, the convergence rate is fast.Secondly the architecture of neural network which has been trained is optimized by utilizing pruning algorithm to reduce the computational complexity and enhance network＇s generalization.And then the optimized network is retrained so that it has the optimum parameters.At last the test samples are predicted by the ultimate network.The simulation and comparison show that this optimized neuron network can not only reduce the calculating complexity greatly, but also the correct recognition rate is up to 100%.