MIL(multi-instance learning)

MIL算法在71 mil基准数据集上进行了测试。这是最大的实验MIL存储库，用于算法比较。数据集的应用领域是分子活性预测，图像注释，文本分类，网页分类和音频记录分类 (n miproblems.org提供数据集的mat文件)。
每个数据集文件都是一个逗号分隔值 (CSV) 格式的文件，它具有实例数许多行和特征数许多列以及两个附加列。第一个附加列对应于传播到实例的bag类标签。第二列是bag ID列，其中每个实例接收其所有者bag的bag id号。其余列分别存储实例的特征值。

MIL algorithms were tested on the 71 MIL benchmark datasets. This is the largest experimental MIL repository for algorithm comparison. The application fields of these datasets include molecular activity prediction, image annotation, text classification, web page classification, and audio recording classification. The MAT format files of the datasets are provided via miproblems.org. Each dataset file is in Comma-Separated Values (CSV) format, containing multiple rows corresponding to the number of instances, multiple columns corresponding to the number of features, and two additional columns. The first additional column corresponds to the bag-level class label of the bag that the instance belongs to. The second column is the bag ID column, where each instance is assigned the bag ID number of its parent bag. The remaining columns respectively store the feature values of the instance.