Skip to content

Commit

Permalink
docs: update global openrank sigmod
Browse files Browse the repository at this point in the history
Signed-off-by: frank-zsy <[email protected]>
  • Loading branch information
frank-zsy committed Oct 14, 2024
1 parent 77b9623 commit 6acaffe
Show file tree
Hide file tree
Showing 3 changed files with 4 additions and 4 deletions.
2 changes: 1 addition & 1 deletion docs/user_docs/metrics/global_openrank.md
Original file line number Diff line number Diff line change
Expand Up @@ -40,7 +40,7 @@
| :------------- | :---- | :---------- | :--- |
| 仓库 OpenRank 默认值 | $0$ | 协作网络中新增仓库节点的默认值 | 该算法认为协作网络中的价值均应来自开发者的活跃,因此仓库节点默认不提供价值 |
| 开发者 OpenRank 默认值 | $0$ | 协作网络中新增开发者节点的默认值 | 该算法认为协作网络中的价值来自开发者活跃,因此开发者默认不提供价值 |
| 开发者 OpenRank 增值 | $min(1,a/88.17)$ | 每次计算时开发者节点的 OpenRank 增值 | 网络中所有价值均来自开发者当月活跃带来的 OpenRank 增值。其中 88.17 为全域开发者每月活跃度分布的 75% 分位数,大于该值则均为 1,小于该值则线性衰减 |
| 开发者 OpenRank 增值 | $\frac{1}{1+e^{0.1*(-x+44.08)}}$ | 每次计算时开发者节点的 OpenRank 增值 | 网络中所有价值均来自开发者当月活跃带来的 OpenRank 增值。其中 44.08 为全域开发者每月活跃度分布上四分位数的一半,这里使用 Sigmod 归一化减少低活跃开发者的影响 |
| 开发者继承比例 | $0.5$ | 开发者节点对于上个月 OpenRank 的依赖比例 | 该算法认为相较于仓库,开发者的价值更应体现出开源中的长期价值,因此开发者对于历史价值的依赖度较高 |
| 仓库继承比例 | $0.3$ | 仓库节点对于上个月 OpenRank 的依赖比例 | |
| OpenRank 衰减系数 | $0.85$ | 对于当月不活跃的开发者和仓库节点的 OpenRank 衰减比例 | 所有节点无论是否活跃,均会在当月有一定衰减,从而需要额外的开发者活跃来维持网络价值 |
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -40,7 +40,7 @@ The parameters included in the calculation of Global OpenRank are as follows:
| :------------------------- | :------------------------- | :----------------------------------------------------------------- | :------------------------------------------------------------------- |
| Repository OpenRank Default Value | $0$ | Default value for new repository nodes in the collaboration network | The algorithm assumes that the value in the collaboration network should come from developer activity, so repository nodes do not provide initial value by default. |
| Developer OpenRank Default Value | $0$ | Default value for new developer nodes in the collaboration network | The algorithm assumes that the value in the collaboration network comes from developer activity, so developers do not provide initial value by default. |
| Developer OpenRank Increment | $min(1,a/88.17)$ | Developer node OpenRank increment in each calculation | All value in the network comes from the OpenRank increment generated by developer activity during the month. 88.17 is the 75th percentile of global developer monthly activity distribution; values greater than this are capped at 1, while lower values decay linearly. |
| Developer OpenRank Increment | $\frac{1}{1+e^{0.1*(-x+44.08)}}$ | Developer node OpenRank increment in each calculation | All value in the network comes from the OpenRank increment generated by developer activity during the month. 44.08 is the 75th percentile of global developer monthly activity distribution; Using Sigmod here to reduce the influence of low activity developers. |
| Developer Inheritance Ratio | $0.5$ | Developer node dependency ratio on the previous month's OpenRank | The algorithm assumes that, compared to repositories, the value of developers should better reflect long-term value in open source, so developers have a higher dependency on historical value. |
| Repository Inheritance Ratio | $0.3$ | Repository node dependency ratio on the previous month's OpenRank | |
| OpenRank Decay Coefficient | $0.85$ | OpenRank decay ratio for inactive developers and repository nodes | All nodes, whether active or not, will experience some decay each month, requiring additional developer activity to maintain the network's value. |
Expand Down
4 changes: 2 additions & 2 deletions static/img/global_openrank.svg
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.

0 comments on commit 6acaffe

Please sign in to comment.