I need a point in the right direction for the problem I'm trying to solve:
I have a lot of already classified short articles. The articles themselves or a reference to them should be stored in some sort of database and a Ai or algorithm should allow smart and recommended navigation through said articles.
The navigation should allow four directions: Random next,more similar, less similar, back.
My first guess would be a vector database, because the distance of the articles in the coordinate system should allow all needed assumptions.
My questions:
Is a vector database the best approach?
In what way should I add these data to a database? (Preprocessing / Training)
Do I need to do NLP or word embedding over the complete article and store the whole text in the database or is there a faster approach?
Example: A user is interested in random sea battles. Than there is one table for this class and he gets a random battle between two western ships around 1910. The user is interested in the time but is more interested in Eastern battles. Now the algorithm suggest another one from 1912 between Western Partys. Now he goes back and wants another similar one. How to use this information to train a model.
There is so much information out there and I'm only searching for the techniques to use.
Thank you all in advance!
