InfraNodus uses the [[back link]] or [[wikilink]] syntax to identify the phrases and entities that should be processed as separate nodes in the network graph and linked together.
It is also used in a similar way in personal knowledge management tools like Obsidian, LogSeq or Roam Research to link ideas together.
However, there is a difference.
Let's say you have a document called "Network Science" in Obsidian / LogSeq which contains the following sentence:
[[Betweenness centrality]] is a measure of [[influence]]. It is used to understand which [[nodes]] appear most often on the [[shortest path]] between any two randomly chosen nodes in a network.
Both of these tools will build the following graph as a result:
The "host" page [[network science]] is connected to all the other [[backlinks]] mentioned on that page.
However, those pages are not connected to each other.
This is a problem, because, in fact, we're using [[betweenness centrality]] in the same context as the [[shortest path]]. Roam Research is the only software that takes account of that by providing the information about the backlinks, but it still doesn't show this in a graph.
That's why in InfraNodus we treat these connections slightly differently. We do not only connect the concepts that appear inside a page to that page, but also all the [[backlinks]] that are mentioned in the same context, following the 4-gram logic (a window scan that contains 4 words at a time). So, as a result, you get this graph:
As you can see, the concepts that are used in the same context, are connected to one another, as well as to the page where they are mentioned.
This provides a much more precise representation of how the different ideas are connected.
You can also switch the 4-gram connection to 2-gram connections (so only the concepts that directly follow each other are connected) in the Global Settings of InfraNodus.
If you need to convert any text into the format where the main ideas are the [[wiki links]] to visualize the relations between them, you can use our custom GPT [[Wiki Links]] Converter module or use the entity detection module inside InfraNodus directly.
Comments
0 comments
Please sign in to leave a comment.