UCI Machine Learning Repository

该数据集包含了多种源自UCI存储库的真实世界数据集，并加入了合成生成的异常数据，以供实验评估之用。这些数据集展示了异常样本与正常样本在不同方面的特性，如异常比例和维度等，它们被用于评估在线和离线异常检测算法的性能。所涉及的任务是异常检测。

This dataset suite contains multiple real-world datasets sourced from the UCI Repository, supplemented with synthetically generated anomalous data for experimental evaluation. These datasets exhibit the characteristics of anomalous and normal samples across various aspects such as anomaly ratio and data dimensionality, and are employed to evaluate the performance of both online and offline anomaly detection algorithms. The core task involved is anomaly detection.

搜集汇总

背景与挑战

背景概述

UCI Machine Learning Repository是一个大型的机器学习数据集存储库，目前维护689个数据集，涵盖分类、回归、聚类等多种任务类型。它提供广泛的数据集资源，包括经典数据集（如Iris）和新兴数据集（如医疗音频数据），并持续更新以支持全球机器学习研究和应用。