Missing values imputation of sequential data

A non-transitory computer readable storage medium tangibly embodying a computer readable program code having computer readable instructions that, when executed, causes a computer device to carry out a method of imputing corrupted sequential data, the method comprising:

• receiving a plurality of input data vectors of a sequential data;
• for each input data vector of the sequential data:
• corrupting the input data vector;
• mapping the corrupted input data vector to a staging hidden layer to create a staging vector;
• reconstructing the input data vector based on the staging vector to provide an output data vector;
• iteratively training a parameter of the staging hidden layer until the output data vector is within a predetermined tolerance of a loss function;
• predicting a next input data vector of the sequential data based on the staging vector;
• and
• storing the predicted next input data vector in the storage device.

A method of imputing corrupted values in a sequential data, the method comprising:

• receiving a plurality of input data vectors of a sequential data;
• for each input data vector of the sequential data:
• corrupting the input data vector;
• mapping the corrupted input data vector to a staging hidden layer to create a staging vector;
• reconstructing the input data vector based on the staging vector to provide an output data vector;
• iteratively training a parameter of the staging hidden layer until the output data vector is within a predetermined tolerance of a loss function;
• predicting a next input data vector of the sequential data based on the staging vector;
• and
• storing the predicted next input data vector in the storage device.