| CPC G06F 16/2455 (2019.01) [G06F 16/22 (2019.01); G06F 16/24542 (2019.01); G06F 16/248 (2019.01); G16H 10/60 (2018.01)] | 15 Claims |
|
1. A method for identifying time series data using a time series retrieval system, comprising:
receiving, by the time series retrieval system, a plurality of time series, each time series comprising a plurality of datapoints obtained over a time period, wherein a least one of the plurality of times series comprises datapoints obtained at irregular time intervals within the time period;
storing the received plurality of time series in a time series database of the time series retrieval system;
for each of the received plurality of time series:
normalizing each of the plurality of datapoints of the time series;
concatenating, with each of the normalized plurality of datapoints, a time interval between the respective datapoint and an immediately preceding datapoint;
generating from the normalized and concatenated time series, using a trained time series encoder/decoder of the time series retrieval system, a context vector for the time series; and
storing the generated context vector in a context vector database of the time series retrieval system;
receiving a request for identification of one or more of the plurality of time series based on similarity to a time series query, comprising: (i) receiving, via a user interface of the time series retrieval system, a query time series; and (ii) generating, using the trained time series encoder of the time series retrieval system, a query time series context vector;
identifying, based on similarity to the query time series context vector, one or more of the stored generated context vectors;
retrieving, from the time series database, each of the time series represented by the identified one or more stored generated context vectors; and
providing, via the user interface of the time series retrieval system, the retrieved time series.
|