NAFM: Neural and Attentional Factorization Machine for Web API Recommendation

Abstract

With the wide adoption of SOA (Service Oriented Architecture), a massive amount of innovative applications emerge on the Internet. One of the popular representations is Mashup composed of multiple Web APIs. Recommending desirable Web APIs to develop Mashup applications has attracted much attention. A dozen of service recommendation approaches are proposed by incorporating multi-dimensional features extracted from service repository into recommendation models. Among the existing works, factorization machine based models show better performance than traditional collaborative filtering techniques in accuracy. However, they either model factorized interactions with the same weight or neglect the non-linear and complex inherent structure of real-world data. In real-world applications, different predictor variables usually have different predictive power, and not all features contain useful signal for estimating the target. Moreover, higher-order feature interactions are usually underlain in real-world data. To address these drawbacks, this paper proposes a hybrid factorization machine model with a novel neural network architecture named NAFM by integrating deep neural network to capture the non-linear feature interactions and attention mechanism to capture the different importance of feature interactions. Comprehensive experiments on a real-world dataset show that the proposed approach outperforms the other state-of-the-art models for service recommendation.