Scalable discovery of multidimensional motifs in time series.
The EvaluationSuite
Jupiter Notebook offer an interactive small evaluation against the SoTA for exact discovery other than many usage examples without the need to download anything. For further evaluation refer to source/test.py
The Thesis_Plots
notebook ensures repeatability for all the figures included in the thesis.
Discretized Random Projections
has been implemented with the use of tensoring to minimize the computations.
MinHash
is currently offered by the datasketch package.