物理嵌入递归卷积神经网络（Physics-encoded Recurrent Convolutional Neural Network，PeRCNN）

1. Open source ecosystem

• Productivity: To evaluate the ability of open-source projects to output software artifacts and open-source value.
• Innovation: Used to evaluate the degree of diversity of open source software and its ecosystem.
• Robustness: Used to evaluate the ability of open-source projects to resist internal and external interference and self recover in the face of changing development environments.

2. Collaboration, People, Software

• Collaboration: represents the degree and depth of collaboration in open source development behavior.
• Observe the influence of core personnel in open source projects, and examine the evaluations of users and developers on open source projects from a third-party perspective.
• Software: Evaluate the value of products exported from open-source projects and their ultimate destination. It is also a concrete manifestation of "open source software", one of the oldest mainstream directions in open source evaluation.

3. Evaluation model

Based on the dimensions of "open source ecosystem" and "collaboration, people, and software", identify quantifiable indicators directly or indirectly related to this goal, quantitatively evaluate the health and ecology of open source projects, and ultimately form an open source evaluation index.