Representation and visualization of multivariate sensory time series data

A system for transforming a set of data sequences, the system comprising:

• at least one processor configured to:
• receive a plurality of data sequences from respective signal sources;
• identify a shortest data sequence of the plurality of data sequences that corresponds to a first interval;
• calculate, for pairs of data sequences of the plurality of data sequences, a similarity value indicative of a degree of similarity between the pairs of data sequences over the first interval;
• form a first segment of a representative data sequence that is a weighted combination of the plurality of data sequences over the first interval wherein each of the plurality of data sequences is weighted based on similarity values associated with that data sequence;
• truncate at least a subset of data sequences to exclude elements corresponding to the first interval;
• identify a shortest data sequence of the subset of data sequences that corresponds to a second interval;
• calculate, for pairs of data sequences of the subset of data sequences, a similarity value indicative of a degree of similarity between the pairs of data sequences over the second interval;
• form a second segment of the representative data sequence that is a weighted combination of the subset of data sequences over the second interval wherein each of the subset of data sequences is weighted based on similarity values associated with that data sequence;
• and
• concatenate the first segment of the representative data sequence and the second segment of the representative data sequence.

A non-transitory computer readable medium containing computer executable instructions that, when executed by a processor, cause the processor to perform a method of transforming a set of data sequences, the method comprising:

• receiving a plurality of data sequences from respective signal sources;
• identifying a shortest data sequence of the plurality of data sequences that corresponds to a first interval;
• calculating, for pairs of data sequences of the plurality of data sequences, a similarity value indicative of a degree of similarity between the pairs of data sequences over the first interval;
• forming a first segment of a representative data sequence that is a weighted combination of the plurality of data sequences over the first interval wherein each of the plurality of data sequences is weighted based on similarity values associated with that data sequence;
• truncating at least a subset of data sequences to exclude elements corresponding to the first interval;
• identifying a shortest data sequence of the subset of data sequences that corresponds to a second interval;
• calculating, for pairs of data sequences of the subset of data sequences, a similarity value indicative of a degree of similarity between the pairs of data sequences over the second interval;
• forming a second segment of the representative data sequence that is a weighted combination of the subset of data sequences over the second interval wherein each of the subset of data sequences is weighted based on similarity values associated with that data sequence;
• and
• concatenating the first segment of the representative data sequence and the second segment of the representative data sequence.