Introduction
中文版
AMiner Knowledge Graph is a structured entity network extracted from AMiner. It is comprised of over 500,00 entities and about 290,000,000 links among them. The knowledge graph can be used as a benchmark to study knowledge graph construction and also used as an external resource for search/recommendation.
This knowledge graph is used for research purpose. It contains four types of nodes (Tag, Person, Organization and Venue) and four types of relationships (Tag-Tag, Person-Tag, Organization-Tag and Venue-Tag). To construct this knowledge graph, we extracted top 100,000 frequent keywords, e.g., data mining, computer vision, social network analysis, from the publications in AMiner, then calculated the weights among those tags based on the co-occurrence statistics. Furthermore, we assigned those tags to scholars, research organizations, and publication venues based on their publications. The statistics of the dataset are listed as follows:
| #tags | 100,000 |
|---|---|
| #person | 318,406 |
| #organizations | 63,068 |
| #venues | 23,709 |
Demo
Here is a simple demo where you can semantically query the relationships of the entities in the knowledge graph.
Slides: [PPT][PDF]
Description
The data is in JSON Lines text format, where each line is a valid JSON object. There are four types of entities in the knowledge graph, namely Tag, Person, Organization, and Venue. The properties of each are listed as follows:
Tag
| property | type | description | example |
|---|---|---|---|
| id | string | tag identifier | "tag:1" |
| type | string | entity type | "tag" |
| name | string | tag name | "data mining" |
| nbrs | object | related tags and the corresponding weights | {"feature extraction": 18433.9, "machine learning": 16314.7, "support vector machine": 12936.0, ...} |
Person
| property | type | description | example |
|---|---|---|---|
| id | string | person identifier | "person:1256" |
| type | string | entity type | "person" |
| name | string | person name | "Jiawei Han" |
| name_dab | string | person name used for disambiguation | "Jiawei Han:1" |
| tags | list | person tags | ["data mining", "relational database", "association rule", ...] |
Organization/Venue
| property | type | description | example |
|---|---|---|---|
| id | string | entity identifier | "venue:1071" |
| type | string | entity type | "venue" |
| name | string | entity name | "kdd" |
| tags | list | entity tags | ["data mining", "social netowork", "machine learning", ...] |
Note:
- The nbrs property of Tag entity is formatted as: <tag_name>: <weight>, where the <weight> is calculated based on the co-occurrence statistics.
- The tags property of Person entity may include tags that are not included in the Tag entities due to their low frequency.