Pattern-aware transformation of time series data to multi-dimensional data for deep learning analysis
Methods and apparatus are provided for pattern-aware transformation of time series data to multi-dimensional data for Deep Learning analysis. An exemplary method comprises: obtaining time series data and an indication of seasonal components in the time series data; obtaining the time series data separated into data chunks of a predefined length based on at least one seasonal component; aligning the data chunks based on the at least one seasonal component; generating an image and/or a multi-dimensional vector using the aligned data chunks; and applying the image and/or the multi-dimensional vector to a Deep Learning module to obtain a prediction, a classification and/or a profiling of parameters associated with the time series data. The classification of the parameters comprises, for example, an anomaly detection. The profiling of the parameters comprises, for example, a clustering of the parameters and/or a behavior identification.
1. A method, comprising:
obtaining time series data and an indication of one or more seasonal components in said time series data;
separating the time series data into data chunks of a predefined length based at least in part on said indication of at least one of said seasonal components;
aligning the data chunks based on said at least one seasonal component;
generating one or more of an image and a multi-dimensional vector using the aligned data chunks; and
applying one or more of said image and said multi-dimensional vector to a Deep Learning module to obtain one or more of at least one prediction, at least one classification and at least one profiling of one or more parameters associated with said time series data,
wherein the method is performed by at least one processing device comprising a processor coupled to a memory.
2. The method of claim 1 , wherein said one or more seasonal components in said time series data are obtained from one or more of a subject matter expert and a pre-processing algorithm.
3. The method of claim 1 , wherein said one or more seasonal components comprise one or more of an hourly, a daily, a weekly, a monthly, and an annual seasonality in said time series data.
4. The method of claim 1 , wherein said aligning step positions time series data in a vicinity of related seasonal data.
5. The method of claim 1 , further comprising the step of stacking said data chunks based on said at least one seasonal component in a plurality of dimensions, prior to said generating step.
6. The method of claim 1 , further comprising the step of converting an output said Deep Learning module back to a time series data format.
7. The method of claim 1 , wherein said at least one classification of said one or more parameters associated with said time series data comprises an anomaly detection.
8. The method of claim 1 , wherein said at least one profiling of said one or more parameters associated with said time series data comprises one or more of a clustering of said one or more parameters and a behavior identification.
9. The method of claim 1 , wherein one or more parameters of said Deep Learning module are predefined by an external training routine prior to the generating step.
10. A computer program product, comprising a non-transitory machine-readable storage medium having encoded therein executable code of one or more software programs, wherein the one or more software programs when executed by at least one processing device perform the following steps:
obtaining time series data and an indication of one or more seasonal components in said time series data;
separating the time series data into data chunks of a predefined length based at least in part on said indication of at least one of said seasonal components;
aligning the data chunks based on said at least one seasonal component;
generating one or more of an image and a multi-dimensional vector using the aligned data chunks; and
applying one or more of said image and said multi-dimensional vector to a Deep Learning module to obtain one or more of at least one prediction, at least one classification and at least one profiling of one or more parameters associated with said time series data.
11. The computer program product of claim 10 , wherein said aligning step positions time series data in a vicinity of related seasonal data.
12. The computer program product of claim 10 , wherein said at least one classification of said one or more parameters associated with said time series data comprises an anomaly detection.
13. The computer program product of claim 10 , wherein said at least one profiling of said one or more parameters associated with said time series data comprises one or more of a clustering of said one or more parameters and a behavior identification.
14. An apparatus, comprising:
a memory; and
at least one processing device, coupled to the memory, operative to implement the following steps:
obtaining time series data and an indication of one or more seasonal components in said time series data;
separating the time series data into data chunks of a predefined length based at least in part on said indication of at least one of said seasonal components;
aligning the data chunks based on said at least one seasonal component;
generating one or more of an image and a multi-dimensional vector using the aligned data chunks; and
applying one or more of said image and said multi-dimensional vector to a Deep Learning module to obtain one or more of at least one prediction, at least one classification and at least one profiling of one or more parameters associated with said time series data.
15. The apparatus of claim 14 , wherein said aligning step positions time series data in a vicinity of related seasonal data.
16. The apparatus of claim 14 , further comprising the step of stacking said data chunks based on said at least one seasonal component in a plurality of dimensions, prior to said generating step.
17. The apparatus of claim 14 , further comprising the step of converting an output said Deep Learning module back to a time series data format.
18. The apparatus of claim 14 , wherein said at least one classification of said one or more parameters associated with said time series data comprises an anomaly detection.
19. The apparatus of claim 14 , wherein said at least one profiling of said one or more parameters associated with said time series data comprises one or more of a clustering of said one or more parameters and a behavior identification.
20. The apparatus of claim 14 , wherein one or more parameters of said Deep Learning module are predefined by an external training routine prior to the generating step.