Bulgarian Academy of Sciences - National Committee for Mathematics
Citation:
Mathematica Balkanica New Series, Vol. 24, Fasc 3-4 (2010), 199p-205p
Abstract:
This study examines different approaches of binary classification applied to the prob-
lem of making distinction between former and current smokers. Prediction is based on data
collected in national survey performed by the National center for health statistics of America
in 2000. The process consists of two essential parts. The first one determines which attributes
are relevant to smokers status, by using methods like basic genetic algorithm and different
evaluation functions [1]. The second part is a classification itself, performed by using methods
like logistic regression, neural networks and others [2]. Solving these types of problems has its
real contributions in decision support systems used by some health institutions.