Goal: Using Learning Time-series Shapelets to learn shapelets in a human sensor dataset and classify it as either 'state_0' or 'state_1'.
Shapelets are subsequences of time series which could tell us immediately what group is the right one for an input sequence. Learning Time-series Shapelets (LTS) is one of many algorithms for shapelets. LTS inventors make the problem of finding shapelets as a learning problem. LTS is implemented by Python library tslearn. Watch my presentation for further explanation: https://youtu.be/u69v5gm_zBk