Deep Learning-Based Sentiment Analysis of Customer Reviews on Hotels Through Tweets

EasyChair Preprint no. 9629

Abstract

Tweets are continually produced and short. Sentiments play a vital role in comprehending personal feelings. Multitudes show their opinions on any subject or object on social media. The public’s sentiment is split into different categories namely, positive, negative and neutral. In this paper, data is acquired from hotel reviews on Twitter using Python’s Tweepy library and pre-processed using Python scripts. Inconsistent data or noise, retweets, tags, URLs along with hashtag symbols and duplicate entries are removed. The tweets are upsampled and split by using scikit-learn in Python. The textual data is converted into vectors using Keras Tokenizer in Python. Bi-Sense Emoji Embedding (BSEE) is used for performing SA. The sentiments are categorized using Support Vector Machine (SVM) and Random forest (RF), and performance is compared with BSEE based on Accuracy, Recall, F-measure, Precision and Time period. It is seen that the proposed classifier offers better results.

Keyphrases: 1 SUPPORTVECTORMACHINE, 2 RANDOMFOREST, 3 BISENSEEMOJIEMBEDDING

@Booklet{EasyChair:9629,
  author = {M Srisankar and K P Lochanambal},
  title = {Deep Learning-Based Sentiment Analysis of Customer Reviews on Hotels Through Tweets},
  howpublished = {EasyChair Preprint no. 9629},

  year = {EasyChair, 2023}}