Construction Of Mobile Application User Behavior Analysis System Based On Log Mining

In recent years, Mobile Internet has already become the most important direction of the Internet development.The users of mobile application have generated a lot of data in the process of human-computer interaction,which contains a huge business value. Any mobile applicaiton need to work out a data system,when it designed and operated. In addition, we can also use the analysis results of user behavior data to make a better decision for application design.Traditional web log mining technology has come up for many years, but the research about mobile application user behavior analysis method is still in its infancy. Whether the interface structure design and the original link relationship in mobile applications to meet the requirement and expectations or not, it could come up with the answer through analysis of user behavior data. The operations on the mobile terminal, such as slide between interface and click between different widget,don’t send a request to the server. Therefore, the traditional analysis method based on access to the web site cann’t apply for mobile application. In order to take full advantage of original operating data,in this paper, the data collection work based on the mobile terminal, and promptly uploaded to the server for data processing, analysis and the results displaying.The work of this paper can be divided into the following three aspects:I. In the traditional FP-Growth algorithm, the data items are out of order, which does not conform to the order of user access behavior, so we need to improve the algorithm. Before that, the maximum forward reference algorithm is used to preprocess the sequence in the user session log, so as to filter the meaningless behavior path.II. Come up with the weight of interest. The reaults of association rules algorithm,with the support value and confidence value filtration, still exists a lot of redundant and spurious sequences. Therefore, it is necessary for different items to add interest weight to get the association rules that customers are really interested in.Then, effectiveness of the algorithm is verified by experiments.Ⅲ. The concept of the interface distance. The interface distance combines the characteristics of the mobile application interface, which is determined by the selected probability between different interface type.In the frequent access sequence, there is a strong correlation between the interface,which should consider whether to add links.In this paper.author use traditional log mining method as the basis, introduces the basic way for mobile application user behavior analysis and contrastive analysis with the core algorithm, detailing described the system architecture design, such as logic architecture and technical architecture design. Finally,author described the design and development of the key function module of the system.Finally, author uses the real case as the experimental object to verify the feasibility and effectiveness of the system, so as to prove the application value of the research on mobile applications user behavior analysis.