Table 2 Summary of parameter estimation premise settings

Data Segmentation

Divided into 30% discount

the training dataset 2/3, the test dataset 1/3

Data set Description

Logistic regression

The entire training set, with 858 data

Particle filtering initial parameters

The first 250 data in the training set

Filtering process

The remaining 608 data in the training set

Other parameters description

A variance of the initial particle set A

A ~ N(0,1)

Initial particle set P

P ~ N(initial parameters, A)

State equation noise covariance Q

A matrix of differentiation and standardized

Measurement equation noise covariance R

0.1

Number of particles

100

Cycle Time K

608