Night club recommendation system based on decision trees

The following document presents the use of a recommendation system designed for nightclubs in Aguascalientes City, Mexico, which offers customized services according to different profiles based on user’s information such as age, gender, and preferred type of music. The advantage of this system is that it improves users’ experience. A preference questionnaire was applied to 75 young people in order to collect their preferences. One decision tree per nightclub was modeled to predict the user’s evaluation based on their profile. Based on the data collected, the most important features for choosing a nightclub are age and type of music that the customer likes, being reggeaton, banda, and electronic the most preferred ones. The absolute error average of predictions was 0.96 using a scale of 4 points. Besides, data from INEGI and the Kernel Density Estimation technique were used to locate the nightclubs. It was found that nightclubs are located mainly in two points in the city: north and downtown.