真实数据集扩产与被动语料接入

- scripts/expand_real_dataset.py:AutoWeakS 全量 2215 条 + 500 条分层评测集 + 简历扩量(seed=2026 可复现)
- pipeline 被动语料机制(passive_sources):真实无日期 JD 参与接入/清洗/抽取验证(2915 条命中率 99.8%),不进图谱/演化/发现
- 新岗位发现第四指标'差异度'与双年窗口增长率口径
- 评测新增 real_jd_500 维度(99.91%);测试会话隔离 jd_extra_dirs 防止本机配置拖慢套件
This commit is contained in:
RoyirinRosmontis
2026-08-03 15:07:54 +08:00
parent 441d555caf
commit 1722c00249
9 changed files with 412 additions and 18 deletions
+22 -3
View File
@@ -318,12 +318,31 @@
"samples": 200,
"bad_lines": 0
},
"real_jd_500": {
"accuracy": 0.9991,
"title_accuracy": 1.0,
"fields": {
"required_skills": {
"precision": 1.0,
"recall": 0.998,
"f1": 0.999
},
"preferred_skills": null,
"scenarios": {
"precision": 1.0,
"recall": 1.0,
"f1": 1.0
}
},
"samples": 500,
"bad_lines": 0
},
"real_resume": {
"accuracy": 1.0,
"name_accuracy": 1.0,
"phone_accuracy": 1.0,
"samples": 50,
"phone_samples": 50,
"empty_skill_resumes": 28
"samples": 200,
"phone_samples": 200,
"empty_skill_resumes": 113
}
}