📄 中文摘要
RFX-Fuse（随机森林 X [X=压缩] -- 森林统一学习与相似性引擎）实现了 Breiman 和 Cutler 的完整愿景，提供了原生的 GPU/CPU 支持。Breiman 和 Cutler 的原始随机森林不仅仅是一个集成预测器，而是一个统一的机器学习引擎，具备分类、回归、无监督学习、基于相似性的邻近度、异常检测、缺失值插补和可视化等功能。这些功能在现代库如 scikit-learn 中并未得到实现。现代机器学习管道通常需要五个以上的独立工具，例如 XGBoost 用于预测，FAISS 用于相似性，SHAP 用于解释，Isolation Forest 用于异常检测，以及自定义代码用于重要性评估。RFX-Fuse 提供了一个或两个模型对象的替代方案，简化了这一过程。
📄 English Summary
RFX-Fuse: Breiman and Cutler's Unified ML Engine + Native Explainable Similarity
RFX-Fuse (Random Forests X [X=compression] -- Forest Unified Learning and Similarity Engine) delivers the complete vision of Breiman and Cutler with native GPU/CPU support. The original Random Forest by Breiman and Cutler was designed as a unified ML engine, not merely an ensemble predictor, encompassing capabilities such as classification, regression, unsupervised learning, proximity-based similarity, outlier detection, missing value imputation, and visualization—features that modern libraries like scikit-learn have not implemented. Modern ML pipelines typically require five or more separate tools, including XGBoost for prediction, FAISS for similarity, SHAP for explanations, Isolation Forest for outliers, and custom code for importance. RFX-Fuse provides a streamlined alternative with one or two model objects, simplifying this complex process.