Syntactic Openings in Online Reviews: A Computational Pipeline for Cross-Domain Register Taxonomy and Helpfulness Analysis
Zoom - please contact: tpasumarthi@umassd.edu or josedomingo.mora@umassd.edu for Zoom information
Jose Domingo Mora
josedomingo.mora@umassd.edu
Advisor:
Dr. José Domingo Mora - Associate Professor & Chairperson, Management & Marketing
Committee Members:
Dr. Donghui Yan - Data Science Co-Director & Department of Mathematics
Dr. Yuchou Chang - Department of Computer & Information Science
Date: Wednesday, August 19th, 2026
Time: 11:00AM – 12:00 PM (Eastern Time)
Location: Zoom (please contact: tpasumarthi@umassd.edu or josedomingo.mora@umassd.edu for Zoom information)
Committee Members:
Dr. Donghui Yan - Data Science Co-Director & Department of Mathematics
Dr. Yuchou Chang - Department of Computer & Information Science
Abstract:
Online reviews are consumed in volume and evaluated rapidly, making the linguistic properties of their earliest words consequential. Mora and Izadi (2024) demonstrated that the grammatical and syntactic composition of a review's opening carries diagnostic information about the register of the full text, and that this register co-occurs with perceived helpfulness. This thesis operationalizes and extends that account through a reproducible seven-stage computational pipeline applied to 9,999 Amazon reviews drawn equally from the Books and Electronics domains. Review openings were parsed for dependency and constituency structure, abstracted into canonical syntactic templates, embedded as sentence vectors, and clustered using k-means. An eight-class taxonomy of opening strategies was selected on the basis of clustering evaluation metrics and stability across random initializations (mean adjusted Rand index = 0.99). The taxonomy was validated against manual annotation and tested for association with helpfulness using nested negative binomial regression and for cross-domain generalizability using chi-square and Kruskal–Wallis tests. Opening class was significantly associated with helpfulness after controlling for review length, star rating, domain, and reviewer activity, and this association varied by domain. Taxonomy composition was broadly stable across domains, confirming cross-domain generalizability. Validation further revealed that the pipeline's embedding space captures semantic-functional organization rather than strictly syntactic structure, an empirical finding about how computational methods represent register. The thesis contributes an automated, evaluated, and reproducible alternative to semi-manual register analysis.
For further questions, please contact Professor José Domingo Mora at josedomingo.mora@umassd.edu.