Interest-Aware Service Association Rule Creation for Service Recommendation and Linking Mode Recommendation in User-Generated Service

In the area of user-generated service, service recommendation and service-linking mode recommendation can help end users select suitable service components for the creation of a new service application. However, existing service recommendation and service-linking mode recommendation methods typically focus on analyzing user-service dualistic relationship and predicting the quality of service values for service composition; the information about the latent interests of the user and the service is largely underexplored, resulting in large computational costs and weak recommendation explanations. To address these issues, we propose an interest-aware service association rule creation (SARC-IA) approach for recommending suitable service components and service linking modes to end users. The proposed solution includes two phases: 1) extracting the latent interests of the user and the service by utilizing the latent Dirichlet allocation model to train the invocation records of users; 2) forming interest subsets by collecting the users with the same interest; and 3) SARC algorithm is employed to generate the service-linking modes in the interest subsets. The performance of the proposed SARC-IA is evaluated via extensive experiments with the ProgrammableWeb data set. The experimental results show that the SARC-IA method outperforms the reference schemes in terms of its support, confidence, recall, precision, and F1-measure results.