Weighted Domain Transfer Extreme Learning Machine and Its Online Version for Gas Sensor Drift Compensation in E-Nose Systems

Algorithm 3

Weighted Domain Transfer Extreme Learning Machine.

Input:

  the training data set;

  the unlabeled data set;

  the number of clusters;

  the number of hidden layer nodes;

Output:

  the output weight matrix for target domain classifier;

(1) Initialize a base classifier ELM with nodes using ;

(2) Set the output weight to ;

(3) Clustering into clusters;

(4) Select the label and unlabeled samples as and using Algorithm 1;

(5) for each cluster do

(6) Update the weight matrix using Algorithm 2.

(7) end for

(8) Initialize a new ELM with nodes;

(9) if the number of samples in then

(10) Calculate the output weight matrix using (11);

(11) else

(12) Calculate the output weight matrix using (16);

(13) end if

(14) return ;