GNSS/Low-Cost MEMS-INS Integration Using Variational Bayesian Adaptive Cubature Kalman Smoother and Ensemble Regularized ELM

Algorithm 2

Ensemble RELM (ERELM) based AdaBoost-R.T.

Input: Sequence of samples

        where output

Output: Final hypothesis .

(1)    Initialize:

            Machine number or iteration

            Distribution for all

            Weak learning algorithm

            Maximum number of iterations

             (machines) .

       Threshold for demarcating correct and

       Incorrect predictionst

(2)   while    do

(3)        Call weak learner (i.e., Algorithm 1),

        providing it with distribution

(4)        Build the regression model:

(5)        Compute absolute relative error for each

       training sample as

                         () 

(6)        Compute the error rate of as follows:

                         () 

(7)        Set

(8)        Update distribution as follows:

              ( )

(9)        Compute contribution of to the

       final result:

(10)      Normalize such that .

(11)       Set .

(12) end while