基于开源生态知识图谱的软件舆情分析研究

The high-efficient utilization of open source resources has become the key factor for software enterprises to keep being competitive. While the open source software has developed and formed a kind of widely-connected and dynamics evolving ecosystem which covers the software, the participants, and the software knowledge and so on. In this ecosystem, different types of entities are highly dispersed, continuously changing and extensively connected. These characteristics introduce great challenge for high-efficient utilization of open source resources. This project proposes a new problem of software situation analysis based on the knowledge graph of open source ecosystem. We view the open source ecosystem as an organic whole, and try to construct a high-quality knowledge graph based on the huge amounts of widely dispersed open source data in various communities. We aim to make breakthroughs on a set of software situation analysis problems including open source software evaluating and monitoring, domain experts discovering and tracking, hot technique topics analyzing and predicting and so on. This project tries to establish a feasible analysis framework and corresponding theory for analyzing open source software situation. From the correlation instead of isolation point of view we target at acquiring more precise and comprehensive results by multi-dimensions and multi-levels analyzing; from the dynamic instead of static point of view we target at revealing the evolution patterns and predicting the development trends by multi-granularities timing sequence analysis. Based on the developed theory and framework, we aim to provide guidelines and support for high-efficient open source resource utilization.

高效开源资源利用已成为软件企业保持持续竞争力的关键因素。开源软件的不断发展形成涵盖软件、参与者以及软件知识的广泛关联且动态演化的开源生态，在开源生态中各类实体高度分散、动态变化和相互关联的特点使得开源资源高效利用面临巨大挑战。本项目首次提出基于开源生态知识图谱的软件舆情分析问题，将开源生态看做有机整体，力图利用高度分散的开源社区数据构建面向开源生态的高质量知识图谱，并基于开源生态知识图谱突破开源软件的度量与监测、领域专家的发现与跟踪以及技术热点的分析与预测等等一系列软件舆情分析技术。本项目旨在建立起切实可行的开源生态舆情分析理论和框架，从联系而不是孤立的角度出发，通过多维度、多层次的关联分析，获得更为准确、全面的研究结果；从发展而不是静态的角度出发，通过多粒度的时序分析，发现演化规律，预测发展趋势，从而为开源资源高效利用提供指导和支撑。

高效开源资源利用已成为软件企业保持持续竞争力的关键因素。开源软件的不断发展形成涵盖软件、参与者以及软件知识的广泛关联且动态演化的开源生态，在开源生态中各类实体高度分散、动态变化和相互关联的特点使得开源资源高效利用面临巨大挑战。本项目首次提出基于开源生态知识图谱的软件舆情分析问题，将开源生态看做有机整体，构建面向开源生态的高质量知识图谱，并基于开源生态知识图谱突破开源软件的度量与监测、领域专家的发现与跟踪以及技术热点的分析与预测等等一系列软件舆情分析技术。.本项目通过对全球开源社区数据的实时采集分析，构建形成了全球开源软件的监测分析平台OSSEAN，实现了对全球开源软件的检索、分析与评估排行，并基于开源知识图谱实现对开源软件与开发者的精准推荐，OSSEAN平台已对外提供公共服务。本项目研究从两类社区及用户等不同维度的关联建立开源知识图谱，为开源软件和开源生态的准确分析评估提供了全新视角。