0

Shouldn't the system be able to scan these book names automatically from the content of the question?

For this, all you need is a naive and simple program that scans the question for bible book name keywords and selects the corresponding tags. A first-year CS student can do this. This simple method does not require any training corpus and it will work for this site.

Being a specialist in AI and Machine Learning training procedures, I know what I'm talking about here. I have published in The international journal of Artificial Intelligence. See https://sites.google.com/site/tonychanytcv/publications.

Here, I am talking about a specific problem of automatically inputting Bible book names as tags. They are talking about a sophisticated problem for the machine to learn to tag generally, not just book names. They are two distinct problems.

One begins with a naive program. After testing, you improve it and so on, version 1, version 2, etc. This is called the software development cycle.

2 Answers 2

Reset to default
6

This feature exists; it's called Suggested Tags but the algorithm isn't deemed effective enough for a smaller Stack Exchange site like Biblical Hermeneutics:

The training corpus should be at least around 20k questions.

(source; @balpha is a former Stack Overflow developer)

There are currently 8.5k questions on the main site, so that's way below the threshold.

2

This is not a good idea for this case. It's quite common to mention a Biblical book that is not actually the focus or core of the question. A dumb scan for book names would clutter the tag in a way that makes them nearly useless as a taxonomy. If you want every question that mentions a book, you can search for it instead.

You must log in to answer this question.