tsmp (version 0.4.15)

mass_v3: Calculates the distance profile using MASS_V3 algorithm

Description

Mueen's Algorithm for Similarity Search is The Fastest Similarity Search Algorithm for Time Series Subsequences under Euclidean Distance and Correlation Coefficient.

Usage

mass_v3(
  query_window,
  data,
  window_size,
  data_size,
  data_mean,
  data_sd,
  query_mean,
  query_sd,
  k = NULL,
  ...
)

Value

Returns the distance_profile for the given query and the last_product for STOMP algorithm.

Arguments

query_window

a vector of numeric. Query window.

data

a matrix or a vector.

window_size

an int. Sliding window size.

data_size

an int. The length of the reference data.

data_mean

precomputed data moving average.

data_sd

precomputed data moving standard deviation.

query_mean

precomputed query average.

query_sd

precomputed query standard deviation.

k

an int or NULL. Default is NULL. Defines the size of batch. Prefer to use a power of 2.

...

just a placeholder to catch unused parameters.

Details

This is a piecewise version of MASS that performs better when the size of the pieces are well aligned with the hardware.

References

• Abdullah Mueen, Yan Zhu, Michael Yeh, Kaveh Kamgar, Krishnamurthy Viswanathan, Chetan Kumar Gupta and Eamonn Keogh (2015), The Fastest Similarity Search Algorithm for Time Series Subsequences under Euclidean Distance

Website: https://www.cs.unm.edu/~mueen/FastestSimilaritySearch.html

See Also

mass_pre() to precomputation of input values.

Examples

w <- mp_toy_data$sub_len
ref_data <- mp_toy_data$data[, 1]
query_data <- mp_toy_data$data[, 1]
d_size <- length(ref_data)
q_size <- length(query_data)

pre <- tsmp:::mass_pre(ref_data, query_data, w)

dp <- list()
for (i in 1:(d_size - w + 1)) {
  dp[[i]] <- tsmp:::mass_v3(
    query_data[i:(i - 1 + w)], ref_data,
    pre$window_size, pre$data_size, pre$data_mean, pre$data_sd,
    pre$query_mean[i], pre$query_sd[i]
  )
}