Andre Salvaro Furtado; Despina Kopanaki; Luis Otavio Alvares; Vania Bogorny.

Transactions in GIS, v. 20, p. 280-298, 2016.

• publication: https://onlinelibrary.wiley.com/doi/abs/10.1111/tgis.12156
• code: https://github.com/bigdata-ufsc/furtado-2016-msm
• data:
  Artificial Dataset

MSM is the first similarity measure that can deal with the dimensions space, time and semantics, using weights for each dimension. It ignores the order and allows partial matching of dimensions, overcoming the limitations of EDR and LCSS. Finds objects that do the same things or visit the same places, but in different order.

Andre Salvaro Furtado, Luis Otávio Alvares, Nikos Pelekis, Yannis Theodoridis, Vania Bogorny

International Journal of Geographical Information Science 32(1): 140-168 (2018)

• publication: https://www.tandfonline.com/doi/full/10.1080/13658816.2017.1372763
• code: https://github.com/bigdata-ufsc/furtado-2018-ums
• data:
  Crawdad Project
  Geolife

UMS is a parameter free spatial similarity measure that solves the threshold problem, using ell
ipses between trajectory points to measure the spatial distance, solving the problem irregular point distribution. It does not consider time but the order of the points (direction), and outperformed LCSS and EDR. Finds objects that move close in space even if the points have different sampling rates or different spatial distribution.

Lucas May Petry, Carlos Andres Ferrero, Luis Otavio Alvares, Chiara Renso, Vania Bogorny. 

Transactions in GIS. 2019; 23: 960– 975.

• publication: https://onlinelibrary.wiley.com/doi/abs/10.1111/tgis.12542
• code: https://github.com/bigdata-ufsc/petry-2019-muitas
• data:
  ASL
  TagMyDay

MUITAS bridges the gap between MSM and EDR. It considers the dimensions space, time, and semantics, allowing the definition of dimensions/attributes that are independent and should be analyzed individually (e.g. weather condition) or dependent dimensions that must be considered together because they are related to each other (e.g. hotel and review). This measure is interesting for data enriched with a high number of semantic dimensions.

Andre L. Lehmann, Luis Otávio Alvares, Vania Bogorny. 

International Journal of Geographical Information Science 33(9): 1847-1872 (2019).

• publication: https://www.tandfonline.com/doi/full/10.1080/13658816.2019.1605074
• code: https://github.com/bigdata-ufsc/lehmann-2019-SMSM
• data:
  Geolife
  Crawdad
  Synthetic Dataset extracted with Hermopoulis

SMSM is the first similarity measure that considers heterogeneous elements as stops and moves, and partial sequence.It considers all dimensions of space, time, and semantics, and evaluates the move similarity only the start stop and end stop of a subsequence have a spatial match.

Andre Salvaro Furtado; Laercio Pilla Lima; Vania Bogorny.

Data & Knowledge Engineering, v. 5, p. 1, 2018.

• publication: https://www.sciencedirect.com/science/article/abs/pii/S0169023X17302021
• code: https://github.com/bigdata-ufsc/furtado-2018-ftsm-msm
• data:
  UBER-SF 
  Trucks
  The UCI Trajectory 
  Park
  Geolife
  AIS Brest

FTSM is an approach to improve the performance of sequence comparison, that instead of using a dynamic programming approach, adopts a solution inspired by the branch and bound paradigm to optimize the exact computation of threshold-based and dynamic threshold-based similarity measures.