A cognitive buying decision-making process in B2B e-commerce using Analytic-MLP

Complexities in the decision-making process are a major problem in business-to-business (B2B) e-commerce due to the fact that an organisation may involve several individuals with different backgrounds and incentives in the buying process, resulting in inconsistency when making decisions. In this paper, a novel buying decision-making process, using an analytic hierarchy process (AHP) and a Multilayer Perceptron Neural Network (MLP-NN) with a library-publishers scenario as a case study is proposed to address the challenge of inconsistency in buyers' opinions. The AHP uses library books criteria and assigns a weight to each, using a pairwise comparison matrix with an arrangement based on priority. Consistency in judgement levels was measured using congruence and dissonance. The MLP-NN was used to develop specific non-linear mapping by adjusting network weights using a learning algorithm. The weights were used to adjust coefficients using the least-squared error and gradient descent method to increase the consistency of buyers' opinions and consequently improve the choice of indexes that conformed to the common agreement of the buyers. The results show that the intensity of the process, in terms of number of alternatives and criteria, is dependent on both decision-making and knowledge acquisition.