- We now provide more details for each field in the input data and the certainty fortheautomatic mapping to help you explore your input data and validate itsrecognizedstructure.
We’re constantly working to improve your Datomize experience and are happy to share our latest version,1.8,released08 Aug 2022
- To enable a more straightforward and advanced analysis of the generated data using ourrule-based engine (Reinvent task), we’ve added an overview option emphasizing themain differences between the generated data and the origin data.
For each rule, you can explore the main difference between the generated andthe inputpopulation; The left graph shows the fields where the most significant difference is foundbetween the two populations, and the right chart shows the magnitude and direction ofchange for numeric fields (e.g., in the graph above, the experienceyears in the generatedpopulation is much lower compared to the input population, and the wage is slightlyhigher).
- The rule-based engine (Reinvent) now supports defining both Normal and Uniformdistributions for conditions based on numeric fields.
- To improve user experience, we made various adjustments to Datomize’s mappingscreen. The maintain sequence data and partial synthetization options can now beactivated under the Advanced Options.
The Partial synthetization can now be set by checking on the relevant toggleand thenselecting the condition fields that will be used with their original values and used as acondition to generate the rest of the fields:
- We have various improvementsto our core generator and an improved Recharge taskwith wiser and faster automatic balancing for a given prediction target.