clefourrier HF staff commited on
Commit
f73543f
1 Parent(s): e12f39d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +110 -0
README.md CHANGED
@@ -1,3 +1,113 @@
1
  ---
2
  license: unknown
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: unknown
3
  ---
4
+
5
+ # Dataset Card for IMDB-BINARY (IMDb-B)
6
+
7
+ ## Table of Contents
8
+ - [Table of Contents](#table-of-contents)
9
+ - [Dataset Description](#dataset-description)
10
+ - [Dataset Summary](#dataset-summary)
11
+ - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
12
+ - [External Use](#external-use)
13
+ - [PyGeometric](#pygeometric)
14
+ - [Dataset Structure](#dataset-structure)
15
+ - [Data Properties](#data-properties)
16
+ - [Data Fields](#data-fields)
17
+ - [Data Splits](#data-splits)
18
+ - [Additional Information](#additional-information)
19
+ - [Licensing Information](#licensing-information)
20
+ - [Citation Information](#citation-information)
21
+ - [Contributions](#contributions)
22
+
23
+ ## Dataset Description
24
+
25
+ - **[Homepage](https://dl.acm.org/doi/10.1145/2783258.2783417)**
26
+ - **[Repository](https://www.chrsmrrs.com/graphkerneldatasets/IMDB-BINARY.zip):**:
27
+ - **Paper:**: Deep Graph Kernels (see citation)
28
+ - **Leaderboard:**: [Papers with code leaderboard](https://paperswithcode.com/sota/graph-classification-on-imdb-b)
29
+
30
+ ### Dataset Summary
31
+
32
+ The `IMDb-B` dataset is "a movie collaboration dataset that consists of the ego-networks of 1,000 actors/actresses who played roles in movies in IMDB. In each graph, nodes represent actors/actress, and there is an edge between them if they appear in the same movie. These graphs are derived from the Action and Romance genres".
33
+
34
+ ### Supported Tasks and Leaderboards
35
+
36
+ `IMDb-B` should be used for graph classification (aiming to predict whether a movie graph is an action or romance movie), a binary classification task. The score used is accuracy, using a 10-fold cross-validation.
37
+
38
+
39
+ ## External Use
40
+ ### PyGeometric
41
+ To load in PyGeometric, do the following:
42
+
43
+ ```python
44
+ from datasets import load_dataset
45
+
46
+ from torch_geometric.data import Data
47
+ from torch_geometric.loader import DataLoader
48
+
49
+ dataset_hf = load_dataset("graphs-datasets/<mydataset>")
50
+ # For the train set (replace by valid or test as needed)
51
+ dataset_pg_list = [Data(graph) for graph in dataset_hf["train"]]
52
+ dataset_pg = DataLoader(dataset_pg_list)
53
+
54
+ ```
55
+
56
+
57
+ ## Dataset Structure
58
+
59
+ ### Data Properties
60
+
61
+ | property | value |
62
+ |---|---|
63
+ | scale | medium |
64
+ | #graphs | 1000 |
65
+ | average #nodes | 19.79 |
66
+ | average #edges | 193.25 |
67
+
68
+ ### Data Fields
69
+
70
+ Each row of a given file is a graph, with:
71
+ - `edge_index` (list: 2 x #edges): pairs of nodes constituting edges
72
+ - `y` (list: 1 x #labels): contains the number of labels available to predict (here 1, equal to zero or one)
73
+ - `num_nodes` (int): number of nodes of the graph
74
+
75
+ ### Data Splits
76
+
77
+ This data comes from the PyGeometric version of the dataset.
78
+ This information can be found back using
79
+ ```python
80
+ from torch_geometric.datasets import TUDataset
81
+ cur_dataset = TUDataset(root="../dataset/loaded/",
82
+ name="IMDB-BINARY")
83
+ ```
84
+
85
+ ## Additional Information
86
+
87
+ ### Licensing Information
88
+ The dataset has been released under unknown license, please open an issue if you have this information.
89
+
90
+ ### Citation Information
91
+ ```
92
+ @inproceedings{10.1145/2783258.2783417,
93
+ author = {Yanardag, Pinar and Vishwanathan, S.V.N.},
94
+ title = {Deep Graph Kernels},
95
+ year = {2015},
96
+ isbn = {9781450336642},
97
+ publisher = {Association for Computing Machinery},
98
+ address = {New York, NY, USA},
99
+ url = {https://doi.org/10.1145/2783258.2783417},
100
+ doi = {10.1145/2783258.2783417},
101
+ abstract = {In this paper, we present Deep Graph Kernels, a unified framework to learn latent representations of sub-structures for graphs, inspired by latest advancements in language modeling and deep learning. Our framework leverages the dependency information between sub-structures by learning their latent representations. We demonstrate instances of our framework on three popular graph kernels, namely Graphlet kernels, Weisfeiler-Lehman subtree kernels, and Shortest-Path graph kernels. Our experiments on several benchmark datasets show that Deep Graph Kernels achieve significant improvements in classification accuracy over state-of-the-art graph kernels.},
102
+ booktitle = {Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining},
103
+ pages = {1365–1374},
104
+ numpages = {10},
105
+ keywords = {collaboration networks, bioinformatics, r-convolution kernels, graph kernels, structured data, deep learning, social networks, string kernels},
106
+ location = {Sydney, NSW, Australia},
107
+ series = {KDD '15}
108
+ }
109
+ ```
110
+
111
+ ### Contributions
112
+
113
+ Thanks to [@clefourrier](https://github.com/clefourrier) for adding this dataset.