Variable Length Motif Discovery

Computes the Matrix Profile and Profile Index for a range of query window sizes

valmod(
  ...,
  window_min,
  window_max,
  heap_size = 50,
  exclusion_zone = getOption("tsmp.exclusion_zone", 1/2),
  lb = TRUE,
  verbose = getOption("tsmp.verbose", 2)
)

Arguments

...: a matrix or a vector. If a second time series is supplied it will be a join matrix profile.
window_min: an int. Minimum size of the sliding window.
window_max: an int. Maximum size of the sliding window.
heap_size: an int. (Default is 50). Size of the distance profile heap buffer
exclusion_zone: a numeric. Size of the exclusion zone, based on window size (default is 1/2). See details.
lb: a logical. (Default is TRUE). If FALSE all window sizes will be calculated using STOMP instead of pruning. This is just for academic purposes.
verbose: an int. See details. (Default is 2).

Value

Returns a Valmod object, a list with the matrix profile mp, profile index pi

left and right matrix profile lmp, rmp and profile index lpi, rpi, best window size w

for each index and exclusion zone ez. Additionally: evolution_motif the best motif distance per window size, and non-length normalized versions of mp, pi and w: mpnn, pinn and wnn.

Details

This algorithm uses an exact algorithm based on a novel lower bounding technique, which is specifically designed for the motif discovery problem. verbose changes how much information is printed by this function; 0 means nothing, 1 means text, 2 adds the progress bar, 3 adds the finish sound. exclusion_zone is used to avoid trivial matches; if a query data is provided (join similarity), this parameter is ignored.

Paper that implements skimp() suggests that window_max / window_min > than 1.24 begins to weakening pruning in valmod().

References

Linardi M, Zhu Y, Palpanas T, Keogh E. VALMOD: A Suite for Easy and Exact Detection of Variable Length Motifs in Data Series. In: Proceedings of the 2018 International Conference on Management of Data - SIGMOD '18. New York, New York, USA: ACM Press; 2018. p. 1757-60.

Website: http://www.cs.ucr.edu/~eamonn/MatrixProfile.html

Examples

mp <- valmod(mp_toy_data$data[1:200, 1], window_min = 30, window_max = 40, verbose = 0)
# \donttest{
ref_data <- mp_toy_data$data[, 1]
query_data <- mp_toy_data$data[, 2]
# self similarity
mp <- valmod(ref_data, window_min = 30, window_max = 40)
#> 

#> Pruning  [====================>--------]  73% at 39 it/s, elapsed:  0s, eta:  0s
#> 

#> Pruning  [====================>--------]  73% at 38 it/s, elapsed:  0s, eta:  0s
#> 

#> DPs  1   [====================>--------]  73% at 36 it/s, elapsed:  0s, eta:  0s
#> 

#> MASS     [====================>--------]  73% at 35 it/s, elapsed:  0s, eta:  0s
#> 

#> Re Check  [===================>--------]  73% at 35 it/s, elapsed:  0s, eta:  0s
#> 

#> Pruning  [=======================>-----]  82% at 39 it/s, elapsed:  0s, eta:  0s
#> 

#> DPs  5   [=======================>-----]  82% at 36 it/s, elapsed:  0s, eta:  0s
#> 

#> MASS     [=======================>-----]  82% at 36 it/s, elapsed:  0s, eta:  0s
#> 

#> STOMP 2  [=======================>-----]  82% at 36 it/s, elapsed:  0s, eta:  0s
#> 

#> STOMP 3  [=======================>-----]  82% at 35 it/s, elapsed:  0s, eta:  0s
#> 

#> MASS     [=======================>-----]  82% at 35 it/s, elapsed:  0s, eta:  0s
#> 

#> STOMP 2  [=======================>-----]  82% at 35 it/s, elapsed:  0s, eta:  0s
#> 

#> Re Check  [======================>-----]  82% at 35 it/s, elapsed:  0s, eta:  0s
#> 

#> Pruning  [=========================>---]  91% at 39 it/s, elapsed:  0s, eta:  0s
#> 

#> DPs  4   [=========================>---]  91% at 37 it/s, elapsed:  0s, eta:  0s
#> 

#> MASS     [=========================>---]  91% at 37 it/s, elapsed:  0s, eta:  0s
#> 

#> STOMP 2  [=========================>---]  91% at 37 it/s, elapsed:  0s, eta:  0s
#> 

#> MASS     [=========================>---]  91% at 37 it/s, elapsed:  0s, eta:  0s
#> 

#> MASS     [=========================>---]  91% at 36 it/s, elapsed:  0s, eta:  0s
#> 

#> Re Check  [========================>---]  91% at 36 it/s, elapsed:  0s, eta:  0s
#> 

#> Pruning  [=============================] 100% at 40 it/s, elapsed:  0s, eta:  0s
#> 
#> max_number_motifs_found 480
#> min_number_motifs_found 4
#> Finished in 0.28 secs
# join similarity
mp <- valmod(ref_data, query_data, window_min = 30, window_max = 40)
#> 

#> Pruning  [====================>--------]  73% at 38 it/s, elapsed:  0s, eta:  0s
#> 

#> Pruning  [=======================>-----]  82% at 41 it/s, elapsed:  0s, eta:  0s
#> 

#> Pruning  [=======================>-----]  82% at 40 it/s, elapsed:  0s, eta:  0s
#> 

#> DPs  2   [=======================>-----]  82% at 39 it/s, elapsed:  0s, eta:  0s
#> 

#> MASS     [=======================>-----]  82% at 38 it/s, elapsed:  0s, eta:  0s
#> 

#> STOMP 2  [=======================>-----]  82% at 38 it/s, elapsed:  0s, eta:  0s
#> 

#> Re Check  [======================>-----]  82% at 38 it/s, elapsed:  0s, eta:  0s
#> 

#> Pruning  [=========================>---]  91% at 42 it/s, elapsed:  0s, eta:  0s
#> 

#> DPs  1   [=========================>---]  91% at 39 it/s, elapsed:  0s, eta:  0s
#> 

#> MASS     [=========================>---]  91% at 38 it/s, elapsed:  0s, eta:  0s
#> 

#> Re Check  [========================>---]  91% at 38 it/s, elapsed:  0s, eta:  0s
#> 

#> Pruning  [=============================] 100% at 42 it/s, elapsed:  0s, eta:  0s
#> 
#> max_number_motifs_found 438
#> min_number_motifs_found 1
#> Finished in 0.27 secs
# }