# 第三届阿里云磐久智维算法大赛（亚军） ## [官网地址]() ## [官方数据地址](https://tianchi.aliyun.com/competition/entrance/531947/information?lang=zh-cn) ## 项目目录结构 ``` ├── Dockerfile ├── README.md ├── code │   ├── catboost_fs.py +++++++++++++++++++++++++++++++ 模型训练代码 │   ├── generate_feature.py ++++++++++++++++++++++++++ 特征生成代码 │   ├── generate_pseudo_label.py ++++++++++++++++++++ 伪标签代码 │   ├── get_crashdump_venus_fea.py +++++++++++++++++++ 新数据特征生成代码 │   ├── requirements.txt +++++++++++++++++++++++++++++ python包版本 │   ├── stacking.py ++++++++++++++++++++++++++++++++++ 模型融合代码 │   └── utils.py +++++++++++++++++++++++++++++++++++++ 小工具脚本 ├── data │   ├── preliminary_a_test +++++++++++++++++++++++++++ 初赛A榜测试数据集 │   ├── preliminary_b_test +++++++++++++++++++++++++++ 初赛B榜测试数据集 │   └── preliminary_train ++++++++++++++++++++++++++++ 训练集数据 ├── docker_push.sh +++++++++++++++++++++++++++++++++++++++++ Docker镜像构建、push脚本 ├── feature │   └── generation +++++++++++++++++++++++++++++++++++ 特征生成文件夹 ├── log ++++++++++++++++++++++++++++++++++++++++++++++++++++ 日志文件夹 │   ├── catboost.log +++++++++++++++++++++++++++++++++ 模型运行日志 ├── model ++++++++++++++++++++++++++++++++++++++++++++++++++ 模型文件 ├── prediction_result ++++++++++++++++++++++++++++++++++++++ 模型预测结果文件夹 │   ├── cat_prob_result.csv ++++++++++++++++++++++++++ CATBOOST模型预测概率 │   ├── catboost_result.csv ++++++++++++++++++++++++++ CATBOOST模型预测结果 │   └── stacking_result.csv ++++++++++++++++++++++++++ 模型融合结果 ├── run.log ++++++++++++++++++++++++++++++++++++++++++++++++ 代码运行日志 ├── run.sh ++++++++++++++++++++++++++++++++++++++++++++++++ 代码运行脚本 ├── tcdata ++++++++++++++++++++++++++++++++++++++++++++++++ 复赛测试集数据文件夹(具体文件请使用初赛相关文件更改文件名替换) │   ├── final_crashdump_dataset_b.csv ++++++++++++++++ 复赛B榜新数据文件 │   ├── final_sel_log_dataset_b.csv ++++++++++++++++++ 复赛测试集日志文件 │   ├── final_submit_dataset_b.csv +++++++++++++++++++ 复赛测试集ID │   └── final_venus_dataset_b.csv ++++++++++++++++++++ 复赛B榜新数据文件 ├── user_data │   └── tmp_data +++++++++++++++++++++++++++++++++++++ 临时文件 └── 答辩PPT └── 悦智AI实验室_20220525.pdf ``` ## 运行环境 Python版本为3.8，各个Python包版本见requirements.txt，使用如下命令即可安装： ``` pip install -r code/requirements.txt ``` ## 构建镜像运行代码 ### 构建镜像 ``` docker build -t [你的镜像仓库]:[TAG] . ``` ### 运行镜像 ``` docker run [你的镜像ID] sh run.sh ``` ### push 镜像 ``` docker push [你的仓库地址]:[TAG] ``` ### 运行&push 镜像 ``` bash docker_push.sh ``` ## 运行代码 ``` bash run.sh ``` ## 终榜亚军 亚军盖章