Calculating precision at k.

Calculating specific ranking correlation for predicted values. :param rank_corr_function: Ranking correlation function. :param prediction

Computes the positive part of a divergence / difference. Args: p_samples: Positive samples. 一个矩阵 [n_nodes, n_graphs] 每个节点和它所在的图的相似度， 其

Getting ged for graph pair and grouping with data into dictionary. :param data: Graph pair. :return new_data: Dictionary with

Each result object stores either a distance matrix or a similarity matrix. It cannot store both. :param norm: :re

Making convolutional pass. :param edge_index: Edge indices. :param features: Feature matrix. :return features: Abstra

Making differentiable pooling. :param abstract_features: Node feature matrix. :param edge_index: Edge indices :param

Printing the error rates.

:param norm: :return: a m by n matrix representing the ranking result. rtn[i][j]: For query i, the id of the j-th most si

Deciding the shape of the bottleneck layer.

Calculate histogram from similarity matrix. :param abstract_features_1: Feature matrix for target graphs. :param abstract_fea