Differences between MTStudio and metatagger

p_kipling

Hi.

I have built a classifier index in metatagger and deployed it successfully on the server (running metatagger).

However when I test a file in MTStudio and compare it to the same file in metatagger the suggested terms are different.

The rack script (tokenize ;classify -K 6 -cutoff 0) is the same in both MTStudio and in metatagger. This is probably not the best rack script to use but at this stage I am just intrigued as to why I am getting 2 different results and how I can synchronise the two.

Regards
Paul

Migrateduser

There are two possible reasons for this that I can think of:

1) A preprocessor is being used on the test document in the live site that is selecting only a portion of the text to be categorized.

2) If this is an updated/revised, classifier, then it is possible that the classifier indexes on the production site were not updated. If that's the case, then after committing the project from Studio, you must login to the Admin GUI and deploy the indexes again. The project should be marked as "update" in the list of projects that are available to be deployed.

ABC.ssd