Association Rule Mining Method Of Resource Service Feature Sequence For Collaborative Tasks

With the acceleration of economic integration,collaborative task model is has been more and more widely applied.Supported by emerging technologies such as cloud computing and the Internet of Things,collaborative task systems manage business processes and resource services more efficiently than ever,and coordinate different organizations to complete a task together.Resource services are provided to business processes in the form of "service flows" and follow certain association rules.However,the distribution characteristics of a large number of resource services and the autonomy of various organizations in the selection of resource services make it difficult to reflect the relationship between the resource services.In the environment of collaborative tasks,in order to better support the interaction between different organizations,this paper proposes a method for mining association rules between resource services by analyzing the sequence of resource services with the support of workflow and related technologies.From the perspective of resource service feature,the problem is converted to mine the trend patterns and association rules of resource service feature sequence.The main research works are as follows.(1)Mining method of the trend of resource service feature value sequence based on clustering algorithm.By analyzing business data sets,the trend of resource service feature value sequence is studied from the perspective of time series.Specifically,firstly,a method based on sliding window linear regression is used to segment the feature values sequence.Then,the trend similarity between the feature value subsequences is analyzed,and an improved method for calculating the trend distance of the binary change is presented.Finally,through hierarchical clustering method,the trend of feature values sequence from resource service is discovered.The verified of the proposed method is verified by taking the design and manufacturing process of the collaborative products as an example of simulation experiments.(2)Mining method of feature sequence association rules based on FP-Tree.In order to optimize the combination of resource services,a method for mining feature sequence association rules based on FP-Tree is proposed.Firstly,through alignment time scale of time series,the trend in the same time range is regarded as a transaction and the trend transaction set is obtained.Then,FP-Tree is constructed in an orderly manner according to thetrend time span and time-sequence dependency of the workflow.Finally,the index considering the characteristics of time span distribution is used to output association rules,and validity of the proposed method is verified by comparing with the association rule mining method based on support degree.Finally,the application of the theoretical method proposed in this paper is discussed in a cloud manufacturing service platform project for small-and medium-sized enterprises.