<![CDATA[Customer reviews are views and comments given by customers after they experience a product or service. Customer reviews can be a valuable source of information for owners, but there are certain problems related to reviews piling up and not having time to read the meaning of the contents of the reviews, as well as paying attention to what aspects are highlighted by customers. The research was conducted using deep learning, data collection using web scrapping techniques with the Python-based Selenium tool, data obtained for 2065 reviews consisting of 1955 Google review data and 110 TripAdvisor data. The aspects discussed are service quality, food quality, environment and price using the Artificial Neural Network (ANN) algorithm with word weighting using TF-IDF. Implementation of dataset imbalance, random undersampling technique applied. Hyperparameter tuning was done via the GridsearchCV function from the scikit-learn library. The model testing results were evaluated using a confusion matrix, producing an accuracy value of 89%. Next, a negative review ranking process was carried out to identify the negative reviews most frequently given by customers and the aspects that accompany them.]]>
Copyrights © 2024
