sentiment mining
Sentiment lexicon; scale.
sentiment classifiers
sentiment annotation

Training data: Polarity score - Negative, positive, neutral

"the educated guess"

Styles/tools for annotation

Every application needs different approach, classes ("there is no free lunch")what do you mean?
Also depending on available data.

Some Data sourcesEnglish
·       Sentiment analysis:

·       Personality, age, gender:
·       Personality: and

·       Authorship recognition, attribution: Project Gutenberg:
·       Deception Detection:
·       Personality, age, gender:
Software - Pattern
·       Example script + data

machine learning classifiers
feed them with data, (make sure the corpus is 50%-50% with/without illness, and that all other parameters the same)
<author id="{author-id}"
          extraverted="-0.5 to +0.5"
          stable="-0.5 to +0.5"
          agreeable="-0.5 to +0.5"
          conscientious="-0.5 to +0.5"
          open="-0.5 to +0.5"
Looking for a conscientious stable female over 45 :D
Chat with conscientious stable females in your area now!
(an agreeable one, please)

About support factor machines (?) "as soon as the hyperplane is found, classification is pretty easy"

The Gold Standard: An annotated data-set, "that is correct".
Use it over an unknown data-set

Danger of over-fitting: fitting your data-science to a stable data-set

10-fold cross validation

[[Leo, Gijs, Catherine, Femke]]


[[wikifoo]] => Michael, Silvio, Valerio, Johnny, Dave