BEYOND TRADITIONAL TIME SERIES ANALYTICS

Discover patterns and anomalies on your data and predict the future

DATA MINING TIME SERIES AT SUPERSPEED

Shapelets incorporates the latest and best performance time series algorithms created by academics so you can distribute and scale them in a parallel fashion

Dot
1

Find patterns and unveil the important data

Traditional analytics platforms focus on metrics instead of changes in data


Shapelets offers key accelerated analytics such as pattern recognition (motifs and chains) and anomaly detection (discords) in a horizontal and scalable offer

Analysis 1 Chart Analysis 2
2

Forecast in a precise, fast and easy way

Predict future trends in your data series in just milliseconds


Shapelets algorithms harness the GPU’s parallel processing power and returns super fast results without any prior training

3

Discover the cause of changes

Changes can be caused by multiple variables that can be interconected


Shapelets' algorithms allow you to establish a cause and effect relationship between a collection of time series and an output in such way that you can explain linear relationships between the inputs that cause an effect on the output

Forecast in a precise, fast and easy way
4

Segmentation, clustering and many other algorithms

Explore sequences and find regularities in your data in a domain agnostic manner


Classify your time series and identify variations in sequences according to their shape

Run all these analytics in a parallel fashion to exploit the resources available

Segmentation, clustering and many other algorithms

TRY KHIVA!

   Github

Join our community and contribute with the largest library of algorithms for time series

Khiva
This website collects personal data and uses cookies to improve services. By clicking "Yes I agree", I hereby consent to the use of cookies as described in the Privacy Policy.