java - Lucene: Overwrite Term Frequency at Index Time -


I am very new to Lusen's index structure, please let me know if it makes sense or if I use it Trying to drill holes to drill a hole

main point / observation

I believe that I need to overwrite the frequency of Lusen with a number of my own (I.e. a value [0,100] that represents a probability (i.e. [0,1]) or any other number which acts as an evidence proof that can replace the period of time. It is possible to overwrite the word frequency value so that the number In reality:

I have such a kind of lucene index (using the word Lucene instead of the normal word frequency)?

Most of the digital artifacts (or supposedly), with meta information instead, are files that contain text or very little text. This meta information learns perceptual possibilities, which are derived from classifier and other machine learning methods (e.g. Based ON object recognition, color histogram, or a combination of evidence). Here is a very simple example where an image was classified (with high probability) with a tree and depiction of a house.

  file path: /pics/1.jpg Meta: tree = 0.9 meta: home = 0.8 meta: dog = 0.0 ... (up to 10000 meta fields)  

One more dog, one house, but no tree shows.

  filepath: /pics/2.jpg Meta: tree = 0.0 meta: home = 0.3 meta: dog = 1.0 ... (max 10000 meta fields)  

Each meta tag is stored in a separate document field called 'meta' so that the search can be searched to give instructions. Each field has a concept as a word or phrase and it is considered to be a token

So, I have primarily the external sources of evidence about which images are about 1 and 2, I know that this is mostly out of the realm of classic TF-IDF paradigm. I would like to incorporate these probabilities (for the 'Meta' field) to find these meta information tokens in Lucene's skocking scheme and to bring throbbing probability to score like TF-IDF. If I search for Meta: Trees and Meta: Dogs I want to find another document and it can be achieved when scoring uses these new possibilities. Therefore, if I can modify the TF of each of these meta concepts (tree, house and dog) with these probabilities, then I can include it in Lausanne without changing everything else.

Does this make sense? Does Lucene provide such a low level amendment on the index? Am I going in the right direction?

How about sub-cloning and override of the method?

Have you read?


Comments

Popular posts from this blog

java - Can't add JTree to JPanel of a JInternalFrame -

javascript - data.match(var) not working it seems -

javascript - How can I pause a jQuery .each() loop, while waiting for user input? -