Abstract:
The most important criteria for selecting a candidate is interview. Here I proposed
a predictive system that helps to predict the hire ability and rate different traits of the
interviews. The paper which I referred for the developed system is camera. The system
can also be used for a fully automated screening of candidates in online interviews. It
helps for initial shortlisting of suitable candidates for further domain specific evaluation.
To build such an automated predictive system, the first step is data gathering. Here the
system analyses the facial expression, hand gestures and body postures with vocal
content. Then extracting the features like verbal, visual and audio cues. The next step is
to test different classification algorithm to create a predictive model for the system that
would classify and rate different traits of individuals.
The system produce a verity of result from single personality trait to the overall
hireability recommendation of candidate. The CNN architecture is used for feature
extraction in the system. The SVR algorithm is used as the predictive framework for
create the model. Also statistical results are produced by the system. The proposed system
produces the user interface results which helps the interviewer to observe in real time.
Finally, at the end of interview the system make predictions and produces a series of
ratings to interviewer based on various parameters like visible, calmness, eye contact,
friendliness etc. The system is an automated expertise system which predicts the overall
and recommended ratings of the interviewee.
