Journey Optimization Platform Release Notes
|Release Date||Release Type||Restrictions||AIX||Linux||Solaris||Windows|
This is a restricted release for this component. For availability of this release, contact your Genesys representative. This release includes only resolved issues.
This release contains the following resolved issues:
Printing into Workers logs has been reduced, in order to reduce the size of the log file. Previously, excessive logging caused the log file to grow rapidly during long job executions, sometimes causing the container to run out of space. (PRR-2004)
The time required for the Lift Estimation Report to complete has been significantly reduced. Previously, because agent data was vectorized for each interaction, the number of vectorization calls equalled (num_of_samples) x (num_of_agents) x (num_of_group_by). As a result, the time for the report to complete grew exponentially with growing numbers of features and samples. Starting with this release, agent vectorization is performed only once, and the agent vector is then reused for all subsequent sample runs. Since the number of agents grows at the same rate as the number of samples, reusing the agent vector reduces report completion time by an order of magnitude.
In addition, this release improves the core Lift Estimation algorithm as follows:
- Processing speed has been increased by reusing the random selection of agents from higher availability to lower availability pools.
- The algorithm now supports exclusion of agents for an interaction based on NULL scores.
- If no interaction was processed for a given agent availability fraction, the value of the estimate in the output is set to NULL.
The following improvements have been made to the Feature Analysis reports displayed in the application interface:
- All the features you selected during report configuration and which have a relative weight greater than or equal to one percent (1%) now display. Previously, reports displayed no more than 15 features.
- All feature columns from feature scores are now considered, so the feature scores are more accurate.
- By default, tabs for feature sub-reports are visible only for features with a weight greater than 0.5. To access sub-reports for features weighted less than 0.5, click the corresponding bar in the bar chart.
The following optimizations reduce the likelihood of encountering out-of-RAM database errors when viewing the status of jobs in the UI:
- An index has been added to the MongoDB database, to avoid in-memory sorting.
- The number of parameters passed to jobs has been reduced, to improve the memory footprint of the application.
Previously, database errors might have occurred when the default MongoDB limit of 32 MB of RAM was exceeded. (PRR-1850)
No special procedure is required to upgrade to release 9.0.007.05.