A contemporary review on soft computing techniques for thyroid identification and detection

This paper is aimed to review various soft computing techniques for identification and detection of thyroid disorder. The papers are extracted from reputable databases like Science Direct, ACM etc. More than 400 papers are downloaded since last one decade. The inclusion criteria include papers from SCI/Scopus databases. Soft computing techniques are divided into three categories namely simple, improved and hybrid techniques. Out of 113 papers, 51%, 37% and 12% papers are based on simple, improved and hybrid soft computing techniques. This review will answer the following research questions for thyroid disorder like what are the (1) various soft computing techniques? (2) reasons to adopt soft computing techniques? (3) existing research gaps? (4) feature selection techniques? and (5) different performance metrics? It sheds light on less attention of researchers on hybrid soft computing techniques for thyroid nodule disorder. It also provides future research directions to develop novel techniques for thyroid disorder.