Generating touring path suggestions using time-interval sequential pattern mining

Guiding service plays an important role for visitors to comprehend exhibits in exhibitions or museums. Without appropriate guiding service, visitors might miss some popular exhibits or cannot complete the visiting trip in time. Therefore, how to provide visitors customized touring service becomes an important task for exhibitions. To bridge the gap, this research takes previous popular visiting behaviors as the suggestion foundation and develops a sequential pattern mining based touring path suggestion system to generate personalized tours. In the exhibition, visiting sequences are persistently collected through a Radio-Frequency Identification (RFID) system. Next, the time-interval sequential pattern mining (I-PrefixSpan) algorithm is applied to obtain popular touring paths containing not only the visiting sequences but also the visiting time between exhibits. Based on the visitor’s personal profile, the system retrieves a set of candidate touring paths. Then, the suggestion priority of each candidate touring path is evaluated based on the number of sections the path passed through, the number of exhibits the path passed through, and time closeness between the path spend and the visitor’s intended visiting time. Finally, the Nation Museum of History in Taiwan is used as an example to show the feasibility of the proposed method.