This thesis applies an enhanced progressive clustering approach, involving fuzzy clustering algorithms and fuzzy neural networks, to solve some practical problems of pattern
recognition. A new fuzzy clustering framework, referred to as Cluster Prototype Centring by Membership (CPCM), has been developed. A Possibilistic Fuzzy c-Means algorithm(PFCM), which is also new, has been formulated to investigate properties of fuzzy clustering. PFCM extends the useability of the Fuzzy c-Means (FCM) algorithm by
generalisation of the membership function.