Welcome to AutoCoder!
AutoCoder is a project designed for detecting the type of node failures in distributed systems, which aims to satisfy both minimalist use and deep customization as much as possible. It can perform indicator analysis based on samples with fault labels, and train fault classification models, which can diagnose faults on input data and provide various data analysis functions.
您好!这是一个用于检测分布式系统的节点故障类型的项目,
名为AutoCoder,它可以基于带有故障标签的样本进行指标分析,训练故障分类模型,
该模型能够对输入的数据进行故障诊断,并提供多样的数据分析功能。
本项目旨在尽可能地同时满足极简使用与深度自定义两种需求,在降低使用门槛的同时允许您按照自己的需要进行定制。
Function
- Data Preprocessing: Clean and transform raw data to be suitable for model training and prediction, including steps such as deduplication, normalization, and balancing of datasets.
- 数据预处理:清理和转换原始数据以适合模型训练和预测,包括去重复、标准化、平衡数据集等步骤。
- Model Training: Highly automated model training, supports nine algorithms, including
Baseline, Linear, Random Forest, Extra Trees, LightGBM, Xgboost, CatBoost, Neural Networks, Nearest Neighbors
, and can be highly customized in the advanced mode to meet your needs in different scenarios. - 训练模型:高度自动化的模型训练,支持多达九种算法,包括
Baseline, Linear, Random Forest, Extra Trees, LightGBM, Xgboost, CatBoost, Neural Networks, Nearest Neighbors
,并且在高级模式中可以进行高度的自定义,以满足您在不同场景下的需求。
- Model Evaluation: After the model training is completed, AutoCoder provides a variety of indicators and visualization methods to evaluate the model, including but not limited to a variety of algorithm model indicator comparison charts, correlation heat maps, feature heat maps, etc.
- 模型评价:模型训练完成后,AutoCoder提供多种指标和可视化方法来评估模型,包括但算法指标对比图、相关热图、特征热图等。
建议
- Self-training: We provide a self-training method for use when your data set has fewer labels, or when there is a difference in the distribution of training data and prediction data.
- 自训练方法:AutoCoder提供一种自训练方法,以供您的数据集标签较少,或训练数据与预测数据存在分布差异时使用
请注意!
- 请谨慎使用自训练功能,使用不当将会使模型严重过拟合,请您仔细阅读自训练介绍,了解自训练方法的适用场景和实现逻辑后再考虑使用
- Data Analysing: We provide an easy-to-use data analysis method. You only need to select the data set to be analyzed to generate a complete data analysis report. At the same time, it supports the generation of comparison reports for two to three data sets.
- 数据分析:AutoCoder提供一种易于使用的数据分析方法,您只需要选择需要分析的数据集,即可生成完善的数据分析报告,同时支持生成两到三个数据集的对比报告。
Getting Started
- To get started with AutoCoder, you can visit our website.
- 请访问我们的网站体验AutoCoder的便利!