Unveiling the Ambivalence from the Airline Reviews: the Airline Recommendation System Using CNN

EasyChair Preprint no. 8596

Abstract

There are several advantages to traveling when it comes to crossing international boundaries. People have the right to write reviews on websites or online platforms to express their opinions. Reviews have a direct impact on consumer relationships. These opinions could be expressed within a single review (positive or negative) or across reviews (conflicting). Conflicting online reviews, a little-studied topic, have exploded in prominence in recent years. These reviews not only offer assistance to customers in selecting a suitable airline, but also assist airline firms in identifying and correcting flaws in their service. We address this gap by proposing a research paradigm that conceptualizes the characteristics of conflicting airline reviews, which identify the traveler perceptions that cause their attitudinal ambivalence or uncertainty in developing their actions. We examined how conflicting attributes of airline reviews trigger travelers' attitudinal ambivalence, which leads to indecisiveness. In this work, we suggested different ways to remove this ambivalence and provide recommendations to customers and airlines. The contribution of this work is twofold: first, we used NLP (Natural Language Processing) techniques to preprocess traveler reviews in the recommender system. Second, the Convolutional Neural Network model was implemented, which proposes data collection on different social networks. This machine learning approach recommends an appropriate airline to travelers. This unique strategy to utilize online social networks, promote low-cost airlines to travelers, and machine learning model increases the CNN model's recommendation accuracy.

Keyphrases: CNN model, EDA, Natural Language Processing, Scraping

@Booklet{EasyChair:8596,
  author = {Muhammad Usman Javeed and Waheed Ramay and Shafqat Maria Aslam and Aqib Rehman Pir Zada},
  title = {Unveiling the Ambivalence from the Airline Reviews: the Airline Recommendation System Using CNN},
  howpublished = {EasyChair Preprint no. 8596},

  year = {EasyChair, 2022}}