A Naïve Hopfield Neural Network based Approach for Multiclass Classification of Customer Loyalty

Customer classification is an area of utmost interest for all businesses. For any organization retaining customer is more important than making new customers. In this paper, a simple idea based on Hopfield Neural Network (HNN) is proposed for multiclass classification of customer loyalty. Initially, transformation and k-medoid clustering algorithm preprocesses the training example dataset. Then, classifier model (HNN) learns patterns from this training set. After training is done, patterns are stored and classifier is ready to classify the unclassified examples using weighted matrices and Euclidean norm. It learns from its environment and does not need to be reprogrammed. The proposed classifier is tested over a real dataset collected through an online survey and it is 87.5% accurate, which is an encouraging result. General Terms Neural Networks, Clustering, Multiclass Classification, Training Dataset, Learning Algorithm