Volume 9, Number 2

A Sentiment Lexicon-based Analysis for Food and Beverage Industry Reviews.
The Greek Language Paradigm


Anastasios Liapakis, Theodore Tsiligiridis and Constantine Yialouris, University of Athens, Greece


The purpose of this research is to implement a methodology to detect and quantify customers’ opinions which referred to the Food and Beverage (F&B) sector using the Greek language. Due to the large and continuously opinionative data produced by the evaluations of the customers’ reviews, the F&B companies, and/or other stakeholders face difficulties to extract all the necessary data and to proceed to further analysis. As far as the Greek market is concerned, the F&B sector is one of the most dynamic sectors. Delivery or take away food or coffee is very common, with the vast majority of consumers to order from aggregators’ platforms (online digital markets). In this study, 8,950 customers’ reviews are extracted from 690 companies selected randomly from a total of 6,795 companies covering the most popular capitals of Greece and presented in the most used common e-platform. The mining of customers’ reviews covers a month period during the year of 2018 and the evaluated functions are the quality of food, the customer service, the image of the company, the pricing, and the quantity of food. As it appears, the sentiment analysis in an aspect-level using the lexicon-based technique should approach methodologically the problem by identifying not only the relevant information but also the particular expressions and phrases the evaluators use over the Internet. The extracted keywords and phrases from the customers’ reviews are used to form the corresponding dictionaries of the functions and to proceed in the sentiment classification. The method is tested in an annotated dataset of 2,000 customers’ reviews and, overall, the findings are expected to contribute towards the design and implementation issues of a sentiment lexicon particularly devoted to the Greek F&B industry.


Sentiment analysis, modern Greek, Food & Beverage Industry, Aspect-level, lexicon-based, corpus-based