I am trying to use the personality insights to predict the personality traits. While reading the document of the API, the detail of the science behand the API is not clearly stated.
The service first tokenizes the input text to develop a representation in an n-dimensional space. The service uses an open-source word-embedding technique, GloVe, to obtain a vector representation for the words in the input text (Pennington and others, 2014). It then feeds this representation to a machine-learning algorithm that infers a personality profile with Big Five, Needs, and Values characteristics. To train the algorithm, the service uses scores from surveys that were conducted among thousands of users along with data from their Twitter feeds.
from the document of the personality insights API.
As above, a machine-learning algorithm is used to infer a personality profile with Big Five. But the algorithm is like a black box with no other information.
Does someone know the detail with this part?
Thanks all!
I guess you are referring to the 'science behind the service' documentation - https://cloud.ibm.com/docs/personality-insights?topic=personality-insights-science
That provides an overview, but you need to click on all the reference links, that the doc is littered with, for the theory.