Towards a hybrid approach of Primitive Cognitive Network Process and Self-Organizing Map for computer product recommendation

Products have similarities which can be analyzed to recommend products to consumers with different preferences. This paper combines Primitive Cognitive Network Process (PCNP) and Self-Organizing Map (SOM) to cluster products into appropriate categories on the basis of consumer preferences and product similarities. PCNP is an ideal alternative of Analytic Hierarchy Process (AHP) to quantify the weights for the attributes used in SOM. To demonstrate the applicability of PCNP-SOM, an example of computer product recommendation is illustrated.