The Analysis Of Mass Travel Data Based On Cloud Computing

With the rapid progrss of tourism and transport, the explosive growth of travel data has formed a huge amount of information space. How to do the customer relation analysis and business conditions analysis effectivly on this mass travel data have became an impotant research problem. This research work is extremely significant for clarifying tourism market, forecast the impact of tourism related industries, adjust tourism macroeconomic policy, and etc.Focused on the mass travel data, this paper presents a constraint-based, parallel association rule mining algorithm. The algorithm aimed to analyse the mass travel data with longer features frequent itemsets, which are hardly processed by current existing data mining methods. Through this association rules based algorithm, a variety of travel data can be reorganized to a very good form. That makes the further data mining process more efficient, especialy for analysing particular region, specific group of people or particular needs, and can provide more valuable results.Oriented to dynamic characteristics of large-scale customer relationships network formed by the mass travel data, we developed a visual analysis framework based on dynamic link. The framework realized efficient network analysis algorithms, including statistical analysis of large-scale networks, network nodes rate the importance and the role of the network community discovery and network community display, the visualization of large-scale network, thereby achieved more effective analysis of the characters relationship from mass travel data.In summary, this paper provides a massive travel data analysis system based on cloud computing technic. The system based on Hadoop platform that integrates a parallel association rule mining algorithm orentied to travel data, thes system also provides a visual analysis framework based on dynamic link analysis. For practical application, the system can support the whole data proceing procedures, from data preprocessing, data management, data mining to decision support. At last, we complete the experiment used the real massive travel data. The results verified the efficiency and effectiveness of the systems and methods we proposed.