Key takeaways
- Neo4j has launched the updated Graph Data Science Library, GDS 1.7, enhancing link prediction capabilities.
- The new library simplifies the graph data science process, making it more accessible and reducing the risk of data leakage.
- Users can now easily construct predictive models by specifying node and link characteristics without the worry of data distortion.
- Progress logging features allow users to monitor long-running jobs and track their status effectively.
- The new enterprise system monitoring feature enhances observability of shared resources and active procedures.
Graph database platform Neo4j has released an updated version of its Graph Data Science Library -- GDS 1.7. It adds machine learning pipelines for predicting graph native links.
GDS 1.7 placed a premium on making graph data science accessible, easy, and foolproof. No more worrying about ensuring that you have the correct features in your input data when creating new predictions or about whether or not you followed the correct methods for splitting and generating your features to avoid data leakage.
In practice, this enables the construction of a pipeline by specifying the node and link characteristics and how the data must be separated. Hence, it's easier than ever to get from raw data to a highly predictive model with certainty that you haven't mistakenly miss-sampled your edges, distorted your data, or spilt information using the new link prediction pipelines.
Similarly, the new progress logging processes allow you to rapidly check on long-running jobs and see how far they have progressed and how long they have been running.
Finally, system monitoring is a new enterprise feature that takes observability to the next level: a simple command that shows how your team uses shared resources. You can rapidly examine the amount of available memory and threads and see active procedures and their progress.





