Modeling the Customer's Contextual Expectations Based on Latent Semantic Analysis Algorithms

This paper presents the approach of modeling the system of customer’s contextual expectations. The novelty consists in systematic usage of the different technic. The first technic – the theory of Benefits Language as an instrument of forming Concept of Benefits for studied product. The second one – the combination of advantages the probabilistic Latent Dirichlet Allocation (LDA) and Linear Algebra based Latent Semantic Analysis (LSA) methods as an instrument of textual data retrieval. The verification of the proposed approach for specifies type of product – films – was conducted. The main research plan was realized: Contextual Summary using the LDA-based algorithm was formed; the Contextual Frameworks using LSA-based approach were performed; the Manually Created Contextual Expectations Dictionary was built. The results of case study, based on the Polish-language film reviews corpora analysis, allowed to make the conclusions about the possibility to use proposed approach for building the system of Customer’s Contextual Expectations.