基于数据挖掘算法的通信领域故障信号识别性能分析

Performance analysis of fault signal recognition in communication field based on data mining algorithm

针对通信领域故障信号识别的技术问题,提出了基于数据挖掘算法的通信故障信号识别方法。通过Kmeans聚类算法,使得用户从诸如温度、振动、电网故障、负荷、湿度、谐波、磁场、电网纹波等影响通信质量的样本数据中,根据通信类别样本的某些属性或某类特征,确定聚类簇数K,把通信样本类型归为已确定的某一类别中,使得簇内的通信样本数据能够紧密分布在一起,并通过欧几里得距离公式计算出某个类别范围内的数据,通过对数据进行聚类,使用户快速对影响因子进行分析、计算,大大提高了分类效果及稳定性。然后在聚类的数据中,采用BP神经网络模型再次对获取的聚类数据进行训练、计算,能够映射、处理不同聚类类别故障信息数据之间的复杂非线性关系,更加精确、及时处理数据,使用户对评估故障信号的精确度大大提高,减少了计算误差。

Aiming at the technical problem of fault signal identification in communication field,this paper proposes a communication fault signal recognition method based on data mining algorithm.Through the K-means clustering algorithm,the user can make sample data from the communication quality such as temperature,vibration,grid fault,load,humidity,harmonics,magnetic field,grid ripple,etc.the cluster number K of clusters is determined according to certain attributes or certain types of characteristics of the communication category sample.the communication sample type is classified into a certain category,so that the communication sample data in the cluster can be closely distributed together.The Euclidean distance formula is used to calculate the data within a certain category.By clustering the data,the user can quickly analyze and calculate the impact factors,which greatly improves the classification effect and stability.Then,in the clustered data,the BP neural network model is used to train and calculate the acquired cluster data again,which can map and process complex nonlinear relationships between different cluster categories of fault information data,and process data more accurately and timely.The accuracy of the evaluation of the fault signal is greatly improved,and the calculation error is reduced.