The API concludes the process by producing an analysis that provides summary and detail of the various steps.
The process report provides a summary and detailed alignment step report.
The summary describes the result of the process in terms of the following metrics:
- input nodes: the set of nodes that assume to belong to the same domain
that most likely contain duplicated, connected and overlapping nodes. - merged nodes: the set of nodes that are mapped and merged onto other
nodes. - connected nodes: the set of nodes that are not merged onto other nodes,
but are connected to the hierarchy - dangling nodes: the set of nodes that are neither merged nor connected.
Figure target to alignment step report
shows the outcome of each alignment step: how many unmapped nodes where at the start, how many mappings can be found for the given unmapped node set, how many mappings were dropped due to multiplicity, and finally how many nodes were merged as the result of the step.
Pandas profiling provides a convenient way to examine table data. OntoMerger generates profiling reports for each tables. Analysing these reports can provide insights (e.g. what are the missing mappings, and /or hierarchies) for fine tuning the inputs in order to further reduce duplication, as well as to increase connectivity.
For more detail please read the Data testing
section.