Title: Mathematical modelling of doctors' perceptions in the diagnosis of depression: a novel approach

Authors: Subhagata Chattopadhyay

Addresses: Department of Computer Science and Engineering, Camellia Institute of Engineering, Kolkata 700129, West Bengal, India

Abstract: This paper is a novel attempt to model a perception process in the diagnosis of depression. In order to do so, two Fuzzy Clustering Techniques (FCT), such as Fuzzy C-means (FCM) and Fuzzy k-Nearest Neighbour (FkNN) are applied. Both the techniques have a special parameter called as 'cluster fuzzifier' (m), which determines the degree of fuzziness between any two clusters. Hence, by varying 'm', one can manipulate the partition between the clusters. Thus, appropriate tuning of 'm' is critical to obtain the desired number of good quality clusters. The paper proposes that 'm' mathematically mimics doctors' diagnostic perceptions, which needs to be tuned appropriately for making a correct diagnosis, i.e. assigning appropriate class labels of depression. Having proposed this, the paper examines how 'm' influences the clustering task, on a sample of real-world depression cases.

Keywords: mathematical modelling; depression diagnosis; fuzzy clustering; cluster fuzziness; distance measures; doctor perceptions; fuzzy C-means clustering; FCM; fuzzy kNN; k-nearest neighbour clustering; FkNN; diagnostic perceptions.

DOI: 10.1504/IJBET.2013.053702

International Journal of Biomedical Engineering and Technology, 2013 Vol.11 No.1, pp.1 - 17

Published online: 27 Sep 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article