Modeling And Classification Method Of The Sequence Data Of Point Process

In clinical research, the reserch of the censord data is more common.Due to theuncertain death time, it can only be seen as a large range. Because the point process theoryassumes some events happened in a small scope, the censord data can not be considered asa point process.However, the censord data contains some information. The point processof the censord data in survival data modeling and analysis is an important topic. In thispaper, it is proposed the method based on point process which used to estimate thesurvival function.Then use this method to estimate survival curves of the patients withbreast cancer, and compare the result with the traditional method of SC.Entropy is widely applied the detection and classification of EEG signals andelectrical signals.Because entropy ignores the big difference of amplitude between the twosignals which has the same chaotic degree, the classification of the entropy effect isgreatly reduced.Adding the large amplitude information to the sample entropy method,this paper proposes a signal classification method based on multiple point processentropy.To apply this method to the Bonn epilepsy EEG data.And then compared theclassification results with traditional multivariate multiscale entropy, the results ofcomparison show that the proposed method accuracy is higher than traditional multivariatemultiscale entropy.Finally Bonn epileptic EEG data added with different levels of whitenoise was classified by the method and the traditional multivariate multiscale entropy.Theresult shows that: the multiple point process entropy has better anti-interference abilitythan the traditional methods.