What Is It?

Explainable AI (XAI) analyzes relationships that define each community within a network and provides succinct descriptions of these insights to users.

Why Is This Important?

This tool empowers understanding of network graph communities and promotes further exploration. This is critical to drive insights from network analysis by providing necessary additional context. Consider this example involving the eCommerce sample dataset, where NetXAI informs us that the user’s hobby and source of purchase are strong factors distinguishing communities of users:

How?

XAI can be run on any type of network visualization within the platform, regardless of whether the Network Extractor was used to construct the network or not.

Steps for running Explainable AI on a Network Graph:

1. Open the XAI panel by selecting Explainable AI from the Data Analytics menu.
2. Select the XAI method you would like to use:
   • Relative Edge Density (default) - This method will attempt to determine which characteristics are most descriptive for each community. The results in each card may be interpreted as saying that if a node falls into community N, then it likely has certain values for the other features.
   • Identification Tree - This method will attempt to determine which characteristics are most specific to each community. The results in each card may be interpreted as saying that if a node has certain values for the input features, then it likely falls into community N.
3. Drag and drop Louvain Community into the target input in the XAI panel.
4. Add other features by dragging and dropping them into the Add Features area or using the Input All button.
   • Tip: If you used the Network Extractor to generate your network, try using your Associative Features as inputs here.
5. Click the “Run” Button.
6. The panel will populate with cards that describe the communities. Click the cards to explore which communities contain certain relationships.