Computational concept modeling for student centric lifestyle analysis: A technical report on socialite case study
Authors
Rahul Sharma
Bernardete Ribeiro
Alexandre Miguel Pinto
F. Amilcar Cardoso
Duarte Raposo
Marcelo
André Rodrigues
Jorge Sá Silva
Fernando Boavida
Bernardete Ribeiro
Alexandre Miguel Pinto
F. Amilcar Cardoso
Duarte Raposo
Marcelo
André Rodrigues
Jorge Sá Silva
Fernando Boavida
Abstract
Students are an important part of our society, andtheir way of life has a great effect in their performance. There
has been a remarkable contribution by psychological research
in studying students’ behavior and its impact over aspects like
studies, and health. Nowadays Internet of Everything (IoE)
sources (such as smart-phones, social networks, sensor networks
etc.) are capturing person specific data which can contribute to
study humans over various dimensions (for example behavior
analysis, activity monitoring etc.), but it’s challenging to handle
such colossal data. Though, Computational modeling techniques
can play an essential role comprehending such data. In this
work we used three datasets obtained from IoE source, out
of which two datasets (Sleep-Detection, Student-Activity) are
from Smart-phone, and one data (Room) is from Sensor-boxes.
Further, we developed computational models for each dataset
using Regulated Activation Networks (RANs) modeling approach.
Model generated with dataset Sleep-Detection is used to make
inferences about active-subjects through evaluation metrics (pre-
cision, recall, f1-score, and accuracy), and ROC curve analysis,
whereas for models generated with the other two datasets
(Student-Activity, and Room) observations are from statistical
methods (such as mean, standard deviation etc.). Results of
Sleep-Detection dataset indicate that life scenarios (or classes) of
students depicted by data are very well represented with RANs
model. The outcome of models from other two datasets drives us
in identifying numbers of categories exhibited by each dataset.