Assessing pattern recognition or labeling in streams of temporal data. Marteau, P. In 2nd ECML/PKDD Workshop on Advanced Analytics and Learning on Temporal Data, Riva del Garda, Italy, 2016.
Assessing pattern recognition or labeling in streams of temporal data [link]Paper  bibtex   
@InProceedings{Marteau2016b,
  author    = {Marteau, Pierre-Fran{\c c}ois},
  title     = {{Assessing pattern recognition or labeling in streams of temporal data}},
  booktitle = {{2nd ECML/PKDD Workshop on Advanced Analytics and Learning on Temporal Data}},
  year      = {2016},
  address   = {Riva del Garda, Italy},
  month     = Sep,
  file      = {streamLabellingAssessment-hal.pdf:https\://hal.archives-ouvertes.fr/hal-01403948/file/streamLabellingAssessment-hal.pdf:PDF},
  keywords  = { sequential data ; automatic labelling assessment ; assessment measure ; dynamic programming ; temporal data },
  review    = {This algorithm determines if two time-series data streams are similar by using adopting a word edit approach via dynamic programming. Assuming the data has already been segmented and labelled, the distance between the two data streams are calculated by assigning a penalty value if primitives match (ie segment i from stream 1 and segment j from stream 2 overlap), have been inserted, or deleted. A distance matrix can be constructed between the first i^th segment of stream 1, and the first j^th segment of stream 2, via dynamic programming and the shortest path is taken.},
  timestamp = {2017-06-03},
  url       = {https://hal.archives-ouvertes.fr/hal-01403948},
}
Downloads: 0