The Mutual Information Approach for Determining the Strength of Associations Between Features of Clinical Depression

Abstract

This study addresses a crucial challenge relating to key predictors and
their association with human depression using information theory,
focusing on mutual information. Mutual information is a well-known
technique for determining the strength of statistical relationships
between variables in healthcare and many other research fields.
Finding mutual information using unbalanced and limited dataset
data set is a demanding task. The results from the mutual information
and information gain indicate high mutual relationship between
“depression” and “alcohol or other drug consumption”;
“depression” and family support and availability of
accommodation”. But a low mutual relationship between
“depression” and “cigarette smoking”. The results also indicate
significant mutual relationship between “depression” and a synergy
of “impaired function and alcohol and other drug consumption”.
Given the challenges posed by depression, it is hoped that the findings
from the study will be among the current universal study for the
inclusion of ICT model in the identification of the predictors of
depression.