Business Model Discovery And Optimization Techniques With The Support Of Data Mining

Nowadays,application of the internet has penetrated into all walks of life,especially in the field of finance,e-commerce and etc.Online business has become more and more frequent and is gradually replacing the traditional business.However,it is difficult to form a unified online business pattern due to the openness and flexibility of online transactions.Some enterprises lost trade opportunities because of the complexity of the trading process.These transaction processes produced many event logs and every event log recorded someone did something at somewhere at some time,such as shopping online,applying for a driver's license.Online applications and audit for exams and booking tickets and etc.Event logs contain much high valuable information.The information helps to identify bottlenecks and insight abnormal,optimize and make decision,etc.However,we can't gain enough information only by collecting log data,processing and analyzing them is necessary.At the same time,privacy of users possibly Leak and loss as the flexibility of trade,therefore security should be considered when mining process pattern so that users could trust trade online.Currently methods of mining process model are not so much and also to some extent some have some shortcomings.For example,they can't find repeated and invisible activities and etc.Some could find existed loop structure.So process model discovery also needs to explore and research deeply to find process model which match with true model in reality.In this way,we can analyze and improve process model.So the issue will start from the following several aspects: 1.Mining frequent process models 2.Check the consistence between model and logs 3.Manage and optimize process models.1)Mining frequent process models.Simulate logs when handle transactions in enterprises.According to the collected logs to filter infrequent processes and infrequent business units with Apriori algorithm.Mining process models by using process mining algorithms.2)Check the consistence between model and logs.In order to the accuracy of process models,we compared process models and event logs to verify their consistency.As graph corresponds to tree,in addition tree is more unique and better recognition ability.We use divide and conquer strategy after graph is transformed into tree.Then decompose process tree to more subtrees and compare the consistency between subtrees and logs.In this way,to a large extent it reduced the complexity of the problem,reduced calculated amount and improved computing efficiency.3)Manage and optimize process models.To simplify process of trading and improve reliability of trade online and trade efficiency,we considered process model from the perspective of management process model.On the basis of original models and mined frequent models to optimize process models.In the end,extracted process model which has high transaction efficiency,strong security,high evaluation.4)Experiment validation.Through the simulation experiment of log data to validate the effectiveness of the proposed FPM mining algorithm.FPM mining algorithm can effectively mine the simple,frequent business process model,and compared with the event log to verify its correctness.To a certain extent,it improved the efficiency of the transaction and improved the user experience.