<![CDATA[Batam, as the third largest city in Sumatra, has many attractive beaches such as Melur Beach that attract both local and foreign tourists. However, the abundance of choices and the lack of complete information often make novice travelers confused in choosing the right destination. The variety of attractions and activities at tourist sites also adds to the hesitation. As a solution to this problem, an automated chatbot was developed that is able to provide services as if visitors were interacting directly with staff or officers without any time constraints. This research aims to design and implement a chatbot system capable of providing beach tourism recommendations in Batam using the Feedforward Neural Network (FFNN) method. The dataset used includes descriptive information about beach tourist attractions in Batam as well as reviews from visitors. The model achieved the best accuracy with 80% dataset division for training and 20% for testing, with 600 epochs, batch size 10, and learning rate 0.002, which resulted in 97.1% accuracy. Evaluation of this model shows a macro precision value of 79.71%, macro recall of 80%, and macro F1 score of 79.76%. Overall, the model showed high effectiveness in classification tasks, with high accuracy and a good balance between detecting correct instances and minimizing prediction errors. This design shows that a chatbot with feedforward neural network model can be used effectively which is able to provide beach tourism recommendations in Batam with high accuracy and appropriate response to the user.]]>
Copyrights © 2024
