<![CDATA[Travel planning popularized by influencers sparks audiences to review destinations based on image and attractiveness before designing a trip alone, with family, with friends, or as a couple. This research offers ideas to analyze the sentiment of travelers who use hotel accommodation services in Makassar based on Tripadvisor website review data. The method used is the Cross-Industry Standard Procedure for Data Mining (CRISP-DM) with the following stages: business understanding stage, data understanding stage, data preparation stage, modeling stage, evaluation stage, and deployment stage. The results of this study show that the business understanding stage shows the importance of identifying and analyzing tourist behavior related to the assessment of location, cleanliness, service, and hotel value as a strategic step to develop a hotel imaging program to attract tourists with guest characteristics (solo, couple, business, family). The data understanding stage refers to the features processed according to the CRISP-DM method and supporting data for analysis and discussion of the context of Tripadvisor's Top 10 recommended hotels in Makassar. At the data preparation stage, the amount of text data that has been processed using the Decision Tree (DT) algorithm and Support Vector Machine (SVM) is 1,138 through data pre-processing (tokenize, transform cases, filter tokens by length, stopwords, stemming). At the modeling stage, the algorithm that shows the best performance is SVM, with an accuracy value of 98.98%, a precision value of 100%, a recall value of 97.96%, an f-measure discount of 98.97%, and an AUC value of 100%. At the evaluation stage, it can be seen that the classification of reviews is dominant on positive sentiment compared to negative sentiment. Thus, the recommendation for the deployment stage is optimizing products and services related to room amenities, room features, room type, cleanliness, service, value, and location. ]]>
Copyrights © 2023
