why not include "title" field in the OccupationClassifier

[cmangani]

Hey guys,

incredible stuff!!! I am overwhelmed by how awesome this is..
Quick questions:

1. I see that you included the following fields as the input data to the classifier.. Why not the "title" field? That seems like the most obvious/strong predictor.? I think I am missing something obvious..
   document_schema_fields = [
   'description',
   'experienceRequirements',
   'qualifications',
   'skills']

2. The evaluation results of the Classifiers (e.g. the Major group around .61 ).. seems suprisingly low.. Do you think that is a function of the little amount of data you were using (e.g. the 4000 documents)? Do you think we will get much larger numbers when using whole corpus? Have you done that yet? of so, what are the stats on that..

3. Do you have a published API for this classifier? if so, I think my company I am consulting with might pay for it to be user.

Thanks