3 ms·

It would be nice to mine Pubmed to search for tuples (plant, disease, administration mode, contraindications, side effects)

• hosh 4y ago
  Being able to do stuff like this was the point of the semantic web. That was supposed to be Web 3.0. To be fair, according to a write up by a guy deep in the semantic web, there were several choices made that lead to the collapse of the first attempt.
  • JPLeRouzic 4y ago
    >> semantic web In principle this is a great topic. I got some interest in the semantic web technologies in early 2000'. My conclusions were that: - The approach was extremely rigid, basically all info (data and metadata) had to be classified mostly in a hierarchical way (first order logic) with only a few exceptions when most real life cases were unable to fit in this scheme. - Companies in Semantic Web just sold search. Classification was often done by hand or by "connectors" that had to be chosen or even tailored by hand. A few years later search technologies were much more powerful and AI enabled impressive results in classification. That said some projects like Cyc were impressive. https://en.wikipedia.org/wiki/Cyc https://en.wikipedia.org/wiki/Cyc
    • hosh 4y ago
      I think the query language was capable of remapping data into a different schema, but it suffered from its own design.