基于知识图谱的个性化API推荐及组合技术研究

With the continually expanding scale of the Internet applications, the number of Web APIs is increasing dramatically. It becomes more and more difficult to select appropriate API or construct Mashups from the vast amounts of APIs manually. As a result, API recommendation and mashup technology have attracted more and more attention from both academia and industry. But the large scale of the data and the complex network environment bring new challenges to API recommendation and mashup. That is, the vast amounts of data, the crossover application environment, and the multiple relations during API interactions have limited the accuracy and recall of the API recommendation and mashup algorithms. To this end, the project is to explore a new solution based on the knowledge graph. We will study the construction of knowledge graph for API recommendation and mashup, the feature models for APIs based on the fusion knowledge, and the API recommendation and mashup algorithm based on link prediction. The project intends to provide theoretical foundation and technical support for the API recommendation and mashup in the “Big data era” and the complex environment.

随着互联网应用系统的规模持续扩张，以Web API形态存在的服务数量急剧增长，在海量的API资源中人工选择服务并构造API组合（Mashup）的难度越来越大。因此，API推荐及组合技术越来越受到学术界和产业界的关注。但是不断扩大的数据规模和日趋复杂的网络环境对API推荐及组合技术研究带来了新的挑战，包括海量的数据规模，跨界的应用环境，多元的交互关系等影响因素都在一定程度上限制了API推荐与组合算法的准确率和召回率。为此，本项目探索了一条基于知识图谱的全新技术路线，通过研究面向API推荐与组合的知识图谱构建技术、基于知识融合的API特征建模技术以及基于关系预测的API推荐与组合技术，为大数据时代复杂网络环境下的个性化API推荐与组合提供理论支撑和技术支持。

随着互联网应用系统的规模持续扩张，以Web API形态存在的服务数量急剧增长，在海量的API资源中人工选择服务并构造API组合（Mashup）的难度越来越大。因此，API推荐及组合技术越来越受到学术界和产业界的关注。但是不断扩大的数据规模和日趋复杂的网络环境对API推荐及组合技术研究带来了新的挑战，包括海量的数据规模，跨界的应用环境，多元的交互关系等影响因素都在一定程度上限制了API推荐与组合算法的准确率和召回率。为此，本项目探索了一条基于知识图谱的全新技术路线，通过研究面向API推荐与组合的知识图谱构建技术、基于知识融合的API特征建模技术以及基于关系预测的API推荐与组合技术，为大数据时代复杂网络环境下的个性化API推荐与组合提供理论支撑和技术支持。