11
$\begingroup$

Is the Semantic Web dead? Are ontologies dead?

I am developing a work plan for my thesis about "A knowledge base through a set ontology for interest groups around wetlands". I have been researching and developing ontologies for it but I am still unclear about many things. What is the modeling language for ontologies?

Which methodology for ontologies is better? OTK or METHONTOLOGY?

Is there any program that does as does

Cratilo is a software for analyzing of textual corpora and for extraction of specific terms of the domain of study (it is developed by professors Jorge Antonio Mejia, Francisco Javier Alvarez and John Albeiro Sánchez, Institute of Philosophy the University of Antioquia). It enables lexical analysis of texts, identifying the words that appear their frequency and location in the text. Through a process of recognition, Cratylus identifies all the words in the text and builds a database becomes the draft analysis of the work. Are there other similar tools?

Can the terms found by Cratilo be used to create a knowledge base?

What are the existing open semantic frameworks that can be used for such things?

Is there software that automatically creates RDF, OWL, and XML? How does Tails work? Jena? Sesame?

$\endgroup$
16
$\begingroup$

(While I think that your question is usually not regarded as a good question for this site, I think that it should be preserved because almost everyone new to the field is similarly confused at first and it's hard to find straight, sober, and balanced information about it.)

Is the Semantic Web dead? Are ontologies dead?

Many people moved away from the Semantic Web. On the other hand, there are still many using it. There's always been a lot of confusion about where its value lies. There are use cases that genuinely benefit from the semantic web technologies but they are quite few and far between.

Linked Open Data

That's partially why the subfield of Linked Open Data was created. You can see Linked Data as a more pragmatic stripped down version of the Semantic Web (or as a necessary enabler of the grander semantic web vision).

Linked Data, while more pragmatic, still uses ontologies. It is just not so uptight about using OWL and designing your ontologies first and in a very formal way.

A knowledge base such as Freebase (now being replaced by Wikidata) doesn't even build on Semantic Web technologies (it's always been available also in the form of RDF dumps though). On the other hand, even Freebase builds on concepts similar to those in the Semantic Web and it does have a form of ontologies.

Software to generate ontologies

There is no software that would create high-quality ontologies automatically for you. At least not in the sense of OWL/first-order logic ontologies. On the other hand, many kinds of software, including Cratilo mentioned by you, can help you to build a lower step in the semantic spectrum such as a glossary or maybe even a folksonomy which can later be transformed into more of a taxonomy or ontology by other algorithms. There might be even software to create fullblown OWL ontologies but I think that's still rather an active area of research.

Ontologies and methodologies

As always, it depends on what exactly your needs and goals are. In fields such as biomedicine and life sciences in general, they create complex formal ontologies because they have data with highly varied structure and ontologies help them keep things organized and usable. In contrast, an eshop might be better off sticking with normal relational database modelling and adding only global identifiers (URIs) in the spirit of Linked Data if they want to build a proper knowledge graph later.

Even if you need ontologies you might want to skip the methodologies first. Especially if you are really new to the field and have little idea what your needs are.

Jena, Sesame, ...

Jena, Sesame, Virtuoso, etc. are triplestores - they are used to store and query RDF. Most ontologies can be represented in RDF. Even those written in OWL. RDF has the semantic part, RDF/S, which allows you to formulate some basic ontologies. That might well be enough for start. Even RDF/S can get confusing and convoluted when you start to think about blank nodes, named graphs (are they fixed? are they dynamic?), etc.

While I think that your question is usually not regarded as a good question for this site, I think that it should be preserved because almost everyone new to the field is similarly confused at first and it's hard to find straight, sober, and balanced information about it.

| improve this answer | |
$\endgroup$
  • $\begingroup$ thank you very much, sorry for so many questions, is the first time I use stack exchange, I am confused by my graduation project, I have no degree advisors or directors :( and I'm lost so about information ontologies, I'm trying to structure my graduation project on wetlands and ontologies and knowledge management, but not It is if I'm doing well. Thanks again for your help. sorry for my English, I'm trying to improve it $\endgroup$ – Antonio Edgar Martinez May 8 '15 at 14:59
  • $\begingroup$ @AntonioEdgarMartinez please, accept one of the answers. I suggest accepting jkbkot's answer as it's more comprehensive than mine. $\endgroup$ – Wojciech Walczak May 9 '15 at 18:19
2
$\begingroup$

There are too many too general problems in your post.

We're definitely in an AI summer era right now (as opposed to AI winter), and the research on Semantic Web receives less attention.

Still, there are many projects related to building ontologies. Google has Knowledge Graph and Knowledge Vault. Both of these are using Freebase (among other sources).

There are dozens of links I can give you to answer some of your questions, but the best thing you can do is to browse W3C Semantic Web pages.

Take a look at RDF, SPARQL, OWL, Virtuoso, Protege - these are de facto standards.

In terms of extracting ontologies from textual corpora - there are various tools out there. Neither of them is perfect, so you really have to do some research and find something that suits your needs. For example, there's the OntoLearn Reloaded (this paper is relatively new, so you can check out the bibliography to seek out for other approaches).

| improve this answer | |
$\endgroup$
  • $\begingroup$ thank you very much, sorry for so many questions, is the first time I use stack exchange, I am confused by my graduation project, I have no degree advisors or directors :( and I'm lost so about information ontologies, I'm trying to structure my graduation project on wetlands and ontologies and knowledge management, but not It is if I'm doing well. Thanks again for your help. sorry for my English, I'm trying to improve it. $\endgroup$ – Antonio Edgar Martinez May 8 '15 at 15:05
2
$\begingroup$

There is also the academic DBPedia project, an RDF version of a subset of Wikipedia.

According to the Alexa search engine, which ranks web sites according to their popularity on the internet, it is among top 100.000 of websites - not bad, not completely irrelevant.

General interest for the DBpedia.org website in the period 2015-2016 seems to have been fairly constant. See screen grab.

dbpedia timeseries 2015

Other similar sites (according to Alexa) are much less popular.

So I'd say Semantic Web Acceptance is not dead, it just smells funny.

| improve this answer | |
$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.