三支决策理论是传统二支决策上的拓展,具有三种决策规则,即接受、拒绝和不承诺。三支决策广泛适用于不确定或不完整信息的处理。基于覆盖算法的三支决策模型能够自动确定三个域,但是,传统覆盖算法的覆盖中心选取是个不可控的随机过程,单次实验的精度无法保证。因此,本文提出了一种优化覆盖算法中心的三支决策模型(optimal center in constructive covering algorithm,简称OCCCA)。该模型结合最近均值思想,在获取覆盖中心时,先求取数据集同类样本的均值,然后选取与均值最近的样本作为覆盖中心,从而实现优化覆盖算法中心的三支决策模型。实验表明,OCCCA比传统覆盖算法在三支决策模型分类准确率上有平均5%的提高。
The concept of deep learning has been applied to many domains, but the definition of a suitable problem depth has not been sufficiently explored. In this study, we propose a new Hierarchical Covering Algorithm(HCA) method to determine the levels of a hierarchical structure based on the Covering Algorithm(CA). The CA constructs neural networks based on samples' own characteristics, and can effectively handle multi-category classification and large-scale data. Further, we abstract characters based on the CA to automatically embody the feature of a deep structure. We apply CA to construct hidden nodes at the lower level, and define a fuzzy equivalence relation R on upper spaces to form a hierarchical architecture based on fuzzy quotient space theory. The covering tree naturally becomes from R. HCA experiments performed on MNIST dataset show that the covering tree embodies the deep architecture of the problem, and the effects of a deep structure are shown to be better than having a single level.
Multiple-Instance Learning(MIL) is used to predict the unlabeled bags' label by learning the labeled positive training bags and negative training bags. Each bag is made up of several unlabeled instances. A bag is labeled positive if at least one of its instances is positive, otherwise negative. Existing multiple-instance learning methods with instance selection ignore the representative degree of the selected instances. For example, if an instance has many similar instances with the same label around it, the instance should be more representative than others. Based on this idea, in this paper, a multiple-instance learning with instance selection via constructive covering algorithm(MilCa) is proposed. In MilCa, we firstly use maximal Hausdorff to select some initial positive instances from positive bags, then use a Constructive Covering Algorithm(CCA) to restructure the structure of the original instances of negative bags. Then an inverse testing process is employed to exclude the false positive instances from positive bags and to select the high representative degree instances ordered by the number of covered instances from training bags. Finally, a similarity measure function is used to convert the training bag into a single sample and CCA is again used to classification for the converted samples. Experimental results on synthetic data and standard benchmark datasets demonstrate that MilCa can decrease the number of the selected instances and it is competitive with the state-of-the-art MIL algorithms.
