Automatic multilabel classification using machine learning

A system comprising:

• a memory;
• and
• a processor coupled with the memory, the processor is configured to perform a method for automatic multilabel classification, the method comprising:
• receiving a label vector y comprising a plurality of labels, each label representing a respective category;
• computing an intermediate matrix yyT, where yT is a transpose of the label vector y;
• computing a basis matrix H by a non-negative matrix factorization of the intermediate matrix yyT;
• generating a group testing matrix A by sampling the basis matrix H;
• generating a reduced label vector z by computing a product of the group testing matrix A and the label vector y;
• and
• predicting multiple labels associated with an input based on the reduced label vector z.

A computer program product comprising a computer-readable storage medium having program instructions embodied therewith, the program instructions when executed by a processor cause the processor to perform a method for automatic multilabel classification, the method comprising:

• receiving a label vector y comprising a plurality of labels, each label representing a respective category;
• computing an intermediate matrix yyT, where yT is a transpose of the label vector y;
• computing a basis matrix H by a non-negative matrix factorization of the intermediate matrix yyT;
• generating a group testing matrix A by sampling the basis matrix H;
• generating a reduced label vector z by computing a product of the group testing matrix A and the label vector y;
• and
• predicting multiple labels associated with an input based on the reduced label vector z.

A computer-implemented method for automatic multilabel classification, the computer-implemented method comprising:

• receiving, by a processor, a label vector y, which comprises d labels, each label representing a respective category;
• computing, by the processor, a reduced label vector z from the label vector y, the reduced label vector z comprises m labels, m=k·log(d), where k is an average sparsity of the label vector y;
• training, by the processor, a plurality of m binary classifiers respectively corresponding to the m labels from the reduced label vector, each binary classifier is trained to identify features associated with a corresponding label;
• and
• in response to receiving an input, predicting, by the processor, multiple labels that are applicable to the input based on the m binary classifiers.

A system comprising:

• a memory;
• and
• a processor coupled with the memory, the processor configured to perform a method comprising:
• receiving a label vector y, which comprises d labels, each label representing a respective category;
• computing a reduced label vector z from the label vector y, the reduced label vector z comprises m labels, m=k·log(d), where k is an average sparsity of the label vector y;
• training a plurality of m binary classifiers respectively corresponding to the m labels from the reduced label vector, each binary classifier is trained to identify features associated with a corresponding label;
• and
• in response to receiving an input, predicting multiple labels that are applicable to the input based on the m binary classifiers.