LSTM Abnormal Cable Temperature Prediction Algorithm Based on SVM with Particle Swarm Optimization
To address the issue of abnormal cable temperature prediction,this paper proposes an abnormal cable tem-perature prediction.Firstly,the algorithm employs support vector machine to construct a classifier for abnormal cable tem-perature judgment and uses particle swarm optimization for adaptive parameter settings.Secondly,the cable data is classi-fied by month to obtain twelve datasets,and all cable data is used as one dataset.Subsequently,long short-term memory neural networks are used to train these thirteen datasets respectively,and the thirteen trained networks are integrated by training prediction accuracy weighting.Finally,the cable temperature data is used as input,and the integrated network is used to determine whether the cable is in an abnormal state in the future time.Experimental results show that the proposed prediction algorithm has good performance and high application value.
万方数据
维普
