Machine learning algorithm to analyse tipping behaviour based on decision tree model

Nowadays, with the improvement in people’s quality of life, more and more people choose to travel abroad in leisure time. Therefore, the great difference in lifestyles could cause culture shock, which make tourists’ trip a relative awful experience, especially in the popular resorts of some countries such as Europe and America, where tips are given whenever there is a service being provided, and this behaviour confused lots of tourists from Asia, Africa or even European countries. Based on decision tree, this study aimed to share troubles of those tourists with needs by taking the New York City Taxi and Limousine Commission’s database as research’s target, thereby analysing factors that may cause variation in tipping rates. Finally, it is found that distance and duration contributed most to trip fare. Also, besides accurately predicting the tip amount, how each factor contributes to the trip needs to be known; thus, linear regression is used to check the validity of the result, and a 14.1% rate of all distance as a fee is achieved. This study provides a model – which can also be used in many different places – that tremendously improve people’s quality of life, and the main idea is relative to the fuzzy system in solving social problems.