Classification based on a permanental process with cyclic approximation
提出一种双随机标记点过程模型用于监督分类，仅需2-3个协方差参数，通过多项式时间循环近似计算永久比率，有效处理高维和交错特征区域，在DNA微阵列分析中显著降低预测误差。
We introduce a doubly stochastic marked point process model for supervised classification problems.\nRegardless of the number of classes or the dimension of the feature space, the model requires\nonly 2–3 parameters for the covariance function. The classification criterion involves a permanental\nratio for which an approximation using a polynomial-time cyclic expansion is proposed. The approximation\nis effective even if the feature region occupied by one class is a patchwork interlaced\nwith regions occupied by other classes. An application to DNA microarray analysis indicates that\nthe cyclic approximation is effective even for high-dimensional data. It can employ feature variables\nin an efficient way to reduce the prediction error significantly. This is critical when the true\nclassification relies on non-reducible high-dimensional features.