<img src="https://d5nxst8fruw4z.cloudfront.net/atrk.gif?account=J5kSo1IWhd105T" style="display:none" height="1" width="1" alt="">

Nexosis @ Work & Play


Data Imputation Aggregation and Transformation Strategies, bug fixes

August 8, 2017 @ 8:00 AM | Release Notes


New Feature

  • Data Imputation/Transformation (Time Series Regression) - text fields and categorical data is transformed automatically and can also be controlled using column metadata.
  • Imputation/aggregation strategies - (impute min, max, average and aggregate sum, mean, mode).
  • Change Resolution of Time Series.  Users can now submit data at one time granularity and request forecasts at another (e.g. submit hourly data but forecast at the daily, monthly, etc).

Bug Fixes

  • PUT data endpoint now properly updates metadata in datasets.
  • Number of historical observations used in model-building are not improperly trunkated.
  • Forecasts are no longer floored at 0 when they should go negative.
  • Imputation fails on yearly dataset with numeric measure feature

Ready to start building machine learning applications?

Get your free API key  Talk to an expert