Hi,
I am trying to build my database with a metric index. I am using a Fixed Query Array.
I have got a table with many sentences of natural languages. I'm using the Dice Coefficient for calculating the metric
space.
I am going to take a certain number of sentences and use them as pivots by calculating their distance to all of my
sentences.
The table of distances looks like this:
id INTEGER, pivot TEXT, distance1 REAL, distance2 REAL, distance3 REAL, ... Eine Query sieht so aus:
A query looks like this:
SELECT id FROM distance
WHERE dice('sentence', 'pivot1') BETWEEN (distance1 - radius) AND (distance1 + radius)
AND dice('sentence', 'pivot2') BETWEEN (distance2 - radius) AND (distance2 + radius)
AND dice('sentence', 'pivot3') BETWEEN (distance3 - radius) AND (distance3 + radius)
...
My issue is to find appropriate pivots now.
I cannot take all existing words, that would be too much.
Does any of you have an idea?
TX very much for any support!
Janek
