如果想構(gòu)建自己的數(shù)據(jù)集，應(yīng)該繼承dgl.data.DGLDataset類，并且實(shí)現(xiàn)下面的方法：

• __getitem__(self,i)：得到數(shù)據(jù)集的第i個(gè)數(shù)據(jù)，
• __len__(self)：數(shù)據(jù)集的大小
• process(self)：從硬盤加載和處理原始數(shù)據(jù)

這里使用一個(gè)小數(shù)據(jù)集Zachary’s Karate Club network，包含：

• menbers.csv文件包含每個(gè)成員的屬性
• interactions.csv文件包含兩個(gè)成員的關(guān)系

import urllib.request import pandas as pd urllib.request.urlretrieve('https://data.dgl.ai/tutorial/dataset/members.csv', './members.csv') urllib.request.urlretrieve('https://data.dgl.ai/tutorial/dataset/interactions.csv', './interactions.csv')members = pd.read_csv('./members.csv') members.head()interactions = pd.read_csv('./interactions.csv') interactions.head()

我們將成員視作節(jié)點(diǎn)，關(guān)系視作邊，年齡視作節(jié)點(diǎn)的屬性，加入的club作為節(jié)點(diǎn)的標(biāo)簽，邊的權(quán)重作為變的屬性：

import pandas as pd import dgl from dgl.data import DGLDataset import torch import osclass KarateClubDataset(DGLDataset):def __init__(self):super().__init__(name='karate_club')def process(self):nodes_data = pd.read_csv('./karate/members.csv')edges_data = pd.read_csv('./karate/interactions.csv')node_features = torch.from_numpy(nodes_data['Age'].to_numpy())node_labels = torch.from_numpy(nodes_data['Club'].astype('category').cat.codes.to_numpy()) # 將Club屬性變?yōu)閏ategory類型，往往作為label 并且轉(zhuǎn)為0,1edge_features = torch.from_numpy(edges_data['Weight'].to_numpy())edges_src = torch.from_numpy(edges_data['Src'].to_numpy())edges_dst = torch.from_numpy(edges_data['Dst'].to_numpy())self.graph = dgl.graph((edges_src, edges_dst), num_nodes=nodes_data.shape[0])self.graph.ndata['feat'] = node_featuresself.graph.ndata['label'] = node_labelsself.graph.edata['weight'] = edge_features# If your dataset is a node classification dataset, you will need to assign# masks indicating whether a node belongs to training, validation, and test set.n_nodes = nodes_data.shape[0]n_train = int(n_nodes * 0.6)n_val = int(n_nodes * 0.2)train_mask = torch.zeros(n_nodes, dtype=torch.bool)val_mask = torch.zeros(n_nodes, dtype=torch.bool)test_mask = torch.zeros(n_nodes, dtype=torch.bool)train_mask[:n_train] = Trueval_mask[n_train:n_train + n_val] = Truetest_mask[n_train + n_val:] = Trueself.graph.ndata['train_mask'] = train_maskself.graph.ndata['val_mask'] = val_maskself.graph.ndata['test_mask'] = test_maskdef __getitem__(self, i):return self.graphdef __len__(self):return 1dataset = KarateClubDataset() graph = dataset[0]print(graph)

總結(jié)

以上是生活随笔為你收集整理的DGL教程【五】使用自己的数据集的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯(cuò)，歡迎將生活随笔推薦給好友。