Abstract:
"Sentiment analysis, a specialized field within NLP, plays a central role in this research. It
involves the extraction and interpretation of sentiments from textual customer reviews, ranging
from enthusiastic praise to constructive criticism and covering aspects such as food quality,
service, ambiance, and overall customer satisfaction. By assigning sentiment scores to various
food establishments based on the analysis of these reviews, the system quantifies the collective
sentiments of customers.
Within the theoretical landscape of food search applications, a notable gap exists in the
integration of sentiment analysis and machine learning techniques to develop a comprehensive
search engine that leverages customer sentiments for enhanced decision-making. While current
platforms predominantly offer proximity-based outcomes, the theoretical gap resides in the
unexplored application of sentiment analysis models to aggregate and evaluate customer
reviews systematically (Moschitti, 2016). This untapped theoretical territory underscores the
potential for a paradigm shift in culinary decision-making that transcends geographical
constraints (Xing Fang, 2015). Subsequently, the data were trained using different machine
learning (ML) models such as Multinomial Naïve Bayes (MNB), Random Forest (RF), Support
Vector Machine (SVM), Decision Tree (DT).
During the training phase, word embedding techniques such as Term Frequency-Inverse
Document Frequency (TF-IDF) were employed, and pre-processing techniques like stemming
were applied. Among various techniques, SVM achieved an accuracy score of 88%. This
implies that 88% of the entire search results aligned with the expected output based on
sentiment reviews."
