Time Prediction Of Electronic Equipment Failure Based On Hawkes Point Process

Today in the 21st century,with the continuous innovation of Internet technology,various applications play an indispensable role in our food,clothing,housing and transportation,now.They have become indispensable tools in our lives,making the entire society more development in the direction of intelligence.How to make use of the information generated by equipment with embedded application software has become a key issue that needs to be discussed by large hardware companies.The prediction of equipment failure time is one of the important directions.Efficient and accurate forecasting can help the company take effective countermeasures in time to eliminate the risk points of equipment and achieve the purpose of maximizing profits.At present,in academia,the prediction of equipment failure time has become one of the important research topics of many experts and scholars.The most common methods are mainly divided into two categories,namely methods based on point processes and methods based on traditional machine learning.The method based on traditional machine learning does not take into account the impact of historical data on events at the present moment,but considers this problem as a general regression or classification problem considering feature factors,precisely because of the lack of chronological order of events in the modeling process Considering the current impact,the prediction accuracy of this type of method is generally not very high.The second type of method is based on the Point process.The Point process is a kind of random process.It is a commonly used and effective method for time series modeling.Modeling the conditional intensity function is the core of this method,such as the common Poisson point in the random process Process and Hawkes point process,this article focuses on the method based on Hawkes point process.Modeling based on the Poisson point process is a relatively simple method.The conditional intensity function of the Poisson point process is a fixed constant and has nothing to do with time,although the conditional intensity function in the Poisson point process becomes a continuous function related to time is used,but the time sequence of events is not considered.The modeling method based on the Hawkes point process considers the influence of historical events on the decision at the present moment on the basis of the Poisson point process.The accuracy has been improved,but the influence of other characteristic factors on the time of equipment failure is ignored.Based on the above facts,this paper has carried out the following research work:1.This paper uses the advantages of Hawkes point process in time series modeling to improve the accuracy of time prediction of electronic equipment failure is short.2.Aiming at the problem that the prediction accuracy is not high when the electronic equipment fails for a long time,this paper adds equipment aging factor and voltage factor to the model,so that the accuracy is further improved.3.After verifying that equipment aging factor and voltage influencing factor have an impact on the time prediction of equipment failure,the influence of these two factors on the final result is discussed immediately.The main method is through the linear and Non-linear combination,use the constructed model to predict,take the comparison result of the prediction effect between the modules and the expressive power of the combined model.In the final experimental data display,it is found that for this article,the linear model is more expressive,and the two influencing factors tend to be independent of each other.But this article does not rule out the possibility of a nonlinear combination model with better expressiveness.4.This paper conducts experiments on the above model on the data set of eight real modules to verify the prediction accuracy of the model.