Seguí Mesquida, SantiBernabé Constans, Pau2022-06-272022-06-272022-01-22https://hdl.handle.net/2445/187023Treballs Finals de Grau d'Enginyeria Informàtica, Facultat de Matemàtiques, Universitat de Barcelona, Any: 2022, Director: Santi Seguí Mesquida[en] Predicting certain aspects of real estate through machine learning is a topic that has already been seen in several articles. However, each problem is a challenge that needs to be analyzed and put into context. This project aims to predict the price of overnight stays and occupancy for August 2021 of the apartments announced on the Airbnb platform in the city of Barcelona with their respective data set. The task is to use regression methods, such as k-NN, random forest, linear regression, and neural network, along with other techniques and methodologies, such as feature selection and text mining, among others, to get a prediction with a reasonable error. Another important aspect of the project is sentiment analysis. A sentiment analysis model is developed to determine the opinion of the different reviews of the apartments, establishing whether a review has a positive or negative comment. This information is also used when performing the main task. The results obtained show the ability of the various methods and techniques to solve problems in different areas in a remarkably satisfactory way.75 p.application/pdfcatmemòria: cc-nc-nd (c) Pau Bernabé Constans, 2022codi: GPL (c) Pau Bernabé Constans, 2022http://creativecommons.org/licenses/by-nc-nd/3.0/es/http://www.gnu.org/licenses/gpl-3.0.ca.htmlAprenentatge automàticAllotjaments turísticsProgramariTreballs de fi de grauXarxes neuronals (Informàtica)Sistemes classificadors (Intel·ligència artificial)Machine learningTourist accommodationComputer softwareNeural networks (Computer science)Learning classifier systemsBachelor's thesesinfo:eu-repo/semantics/bachelorThesisinfo:eu-repo/semantics/openAccess