Piecewise Stationary Linear Bandit Data

该数据集来源于分段的静止线性老虎机模型，涉及在变化点从未知分布中抽取的上下文。该数据集特别关注PSεBAI+算法的实证性能。数据集包含了宽松度参数ε从0.04到0.6的变化，用于评估所提出算法的表现。规模上，每个算法和实例都进行了20次独立的试验。任务是对最佳臂进行识别。

This dataset is derived from piecewise stationary linear multi-armed bandit models, involving contexts sampled from unknown distributions at change points. It specifically focuses on the empirical performance of the PSεBAI+ algorithm. The dataset varies the relaxation parameter ε from 0.04 to 0.6 to evaluate the performance of the proposed algorithm. For experimental scale, each algorithm and instance is tested with 20 independent trials. The primary task of this dataset is best arm identification.