Abstract:

The article concerns data mining in the process of extracting data to find anomalies in the security of maritime transport. The important issue in this context is also the demand for information on maritime transport security. The aim of the investigation is the extraction of data and the identification of sources of variability (anomalies) in the security of maritime transport. The research task was to decompose the data into the following components: trend, seasonality, and irregular component, also taking into account the influence of working days, moving holidays, and various types of outlier observations. The study was based on secondary data from Google Trends. The research is conducted on data from January 2004 to May 2021 and is related to the case of Poland. Data science - specific techniques were used, combining statistical, econometric, artificial intelligence, and data mining methods. It was validated that the anomalies were caused by characteristic outliers that were connected with changes in law and reaction of society. In addition, seasonality, the Easter effect, and the trend affect the demand for information about maritime transport security.