A novel multi-attribute decision-making for ranking mobile payment services using online consumer reviews.

The onset of the COVID-19 pandemic has changed consumer usage behavior towards mobile payment (m-payment) services. Consumer usage behavior towards m-payment services continues to increase due to access to usage experiences shared through online consumer reviews (OCRs). The proliferation of massive OCRs, coupled with quick and effective decisions concerning the evaluation and selection of m-payment services, is a practical issue for research. This paper develops a novel decision evaluation model that integrates OCRs and multi-attribute decision-making (MADM) with probabilistic linguistic information to identify m-payment usage attributes and utilize these attributes to evaluate and rank m-payment services. First and foremost, the attributes of m-payment usage discussed by consumers in OCRs are extracted using the Latent Dirichlet Allocation (LDA) topic modeling approach. These key attributes are used as the evaluation scales in the MADM. Based on an unsupervised sentiment algorithm, the sentiment scores of the text reviews regarding the attributes are calculated. We convert the sentiment scores into probabilistic linguistic elements based on the probabilistic linguistic term set (PLTS) theory and statistical analysis. Furthermore, we construct a novel technique known as probabilistic linguistic indifference threshold-based attribute ratio analysis (PL-ITARA) to discover the weight importance of the usage attributes. Subsequently, the positive and negative ideal-based PL-ELECTRE I methodology is proposed to evaluate and rank m-payment services. Finally, a case study on selecting appropriate m-payment services in Ghana is examined to authenticate the validity and applicability of our proposed decision evaluation methodology.