Kuala Terengganu homestay recommender system using Content Based filtering algorithm / Nur Atierah Nor Azman

Kuala Terengganu's tourism industry is expanding quickly, which has increased demand for acceptable hospitality and lodging alternatives. However, finding the ideal homestay that aligns with user preference can be time taking task. Hence, ContentBased filtering is capable to find the matching homestay based on the user preferrable criteria of the homestay such as price, locations and availability. Then in order to identify the distinguishing features or term of the homestay, TF-IDF will be calculating the importance term in the homestay review or description. By analyzing this textual content, TF-IDF will creates a representation of each homestay in the form of a corpus vector. It is required to find and compares the similarity between the user’s preferences and the characteristics of the homestay. The similarity measure needs to be done in order to generate the recommendations. Cosine Similarity and TF-IDF will allow the prototype to identify the homestays that align closely with the user’s desired features. The system's performance is meticulously evaluated through precision at 79.3%, recall at 93%, F1-score at 86%, RMSE at 9.95 and MAE at 9.42. These results demonstrate the system capability to offer accurate and relevant homestay suggestions. The project has successfully achieved all of the objectives and able to demonstrated the practical application of content-based filtering algorithm.