Logo image
Decision making using social network analysis and topic modeling in service firms
Dissertation   Open access

Decision making using social network analysis and topic modeling in service firms

Jin Fang
Doctor of Philosophy (Ph.D.), Drexel University
Jun 2021
DOI:
https://doi.org/10.17918/00000817
pdf
Fang_Jin_20214.13 MBDownloadView

Abstract

Professional corporations Operations Research
A service firm is a business that makes its facilities available to others for a fee. Nowadays, many of the management tools and techniques that people use cannot meet the challenges of complex decision-making problems in service firms. This dissertation, therefore, explores two research streams, service firm selection, and service technology planning. The first research stream, the service firm selection problem, is to identify the optimal service firm, which satisfies several predefined selection factors. This selection process typically includes two major stages. During the first stage, the selection criteria identify a limited number of candidate service firms. The second stage involves a detailed examination of candidate service firms to find the most appropriate one. Understanding how to optimize decisions in each of these two selection stages motivates this dissertation's first and second parts. The first part of the dissertation focuses on the first stage of service firm selection, "how to determine the selection criteria in the decision-making problems." The model is developed based on latent Dirichlet allocation (LDA) and the analytic hierarchy process (AHP). We validate our model by comparing the ranking results with reviews in TripAdvisor.com. The second part of the dissertation is relevant to the second stage, "how to select the optimal service firm." We build a network model based on data envelopment analysis (DEA) and the weighted hyperlink-induced topic search (HITS) algorithm and demonstrate its applicability by case analysis. The second research stream, the service technology planning problem, strategically selects technologies to invest in the service firm, which is discussed in the third part of this dissertation. We construct a network model based on HITS and PageRank algorithms and then present its application with a numerical example.

Metrics

38 File views/ downloads
46 Record Views

Details

Logo image