Topic Modeling and Sentiment Analysis of Online Review for Airlines

Abstract

The purpose of this study is to conduct topic modeling and sentiment analysis on the posts of Skytrax (airlinequality.com), where there are many interests and participation of the people who have used or are willing to use it for airlines. The purpose of people gathering at Skytrax is to make better choices using the actual experiences of other customers who have experienced airlines. Online reviews written by customers with experience using airlines in Asia were collected. The data collected were online reviews from 27 airlines, with more than 14,000 reviews. Topic modeling and sentiment analysis were used with the collected data to figure out what kinds of important words are in the online reviews. As a result of the topic modeling, ‘seat’, ‘service’, and ‘meal’ were significant issues in the flight through frequency analysis. Additionally, the result revealed that delay was the main issue, which can affect customer dissatisfaction while ‘staff service’ can make customers satisfied through sentiment analysis as the result shows the ‘staff service’ with meal and food in the topic modeling.

Topic Modeling for Amharic User Generated Texts

“…The authors reported that their proposed approach with the seeded model improved topic detection compared to models that use seed naïve information. Kwon et al (2019) [33] developed topic modeling and sentiment analysis of an airline using with over 14,000 online reviews collected from 27 airlines. First, significant topics are judged using frequency analysis, word cloud and topic modeling and, then, for each of these topics, sentiment analysis was carried out to measure the level of customer satisfaction of an airline.…”

“…The topic modeling related works are organized by approaches such as topic modeling (unsupervised) [13,14,28,[33][34][35][36], seeded topic modeling (semi-supervised) [29][30][31][32], and supervised topic modeling [17,[37][38][39].…”

Determination of the LNG Bunkering Optimization Method for Ports Based on Geometric Aggregation Score Calculation

“…For the LNG bunkering infrastructure, we studied 115 papers (e.g., journal publications and related reports) published since 2010, via web crawling for 10 days from 23 March 2021. The main keywords were analyzed using text mining with the R Studio program (version: R-4.0.5) [29]. Table 1 shows the classification of 115 data (articles, reports, papers, etc.)…”

“…Ha (hazard analysis) represents the result of evaluating the risks of LNG leaks in surrounding facilities and vessels during bunkering in the target port facilities for each LNG bunkering optimization model. The Korea Marine Equipment Research Institute analyzed the distance of the risk sections based on the possibility of explosion when LNG fuel is leaked during bunkering, using ANSYS Inc. (Pittsburgh, PA, USA) for Busan Port in 2020 [29]. The risky sections according to the LNG leak were distinguished into red, yellow, and green zones based on the probability of explosion; the details are summarized in Table 6.…”

A Survey on Sentiment Analysis and Opinion Mining in Greek Social Media

“…Feature learning techniques have been proven to be better than vector-based approaches, since they are able in capturing semantic term-document information, as well as rich sentiment content [14]. Topic modeling is a form of unsupervised learning that has been extensively employed for the extraction and categorization of aspect from online product reviews [15] and for sentiment analysis [16]. Topic modeling techniques (e.g., Latent Semantic Analysis (LSA), Probabilistic Latent Semantic Analysis (PLSA), Linear Discriminant Analysis (LDA)) jointly model sentiment and semantic information of terms, but they do not generalize to new, unseen documents or words and they are prone to overfitting.…”