This is our first "Upcoming Feature" blog post to give you a sneak peek into what's on the horizon for the Nexosis Machine Learning API. Today's upcoming feature is an enhancement to the ability to forecast time series data at different granularity.
Forecast time series at different granularity
We are continually looking to improve our API time series capabilities which is why we want developers to be able to forecast their time series data at different granularity. Today, Nexosis API users can forecast at the hourly, daily, weekly, monthly, or yearly levels. However, the source data is often not at the same granularity as the forecast, nor should every datatype be treated the same.
In this update we will be making smarter decisions about how we roll up your data before we forecast by allowing you to specify a broader range of data types that better direct our data aggregation strategies. No two time series data sets look alike so we are excited to add this feature as a precursor to others to continue our mission of empowering developers to build smart applications with our machine learning API.