Bigclam
Function
This API is used to run the Bigclam algorithm based on input parameters.
The BigClam algorithm is an overlapping community detection algorithm that models the relationship between vertices and communities as a bipartite graph. Assuming that the edges between vertices in the graph are generated based on community relationships, it can detect overlapping communities in the graph.
URI
POST /ges/v1.0/{project_id}/hyg/{graph_name}/algorithm
Parameter |
Mandatory |
Type |
Description |
---|---|---|---|
project_id |
Yes |
String |
Project ID. For how to obtain the project ID, see Obtaining a Project ID. |
graph_name |
Yes |
String |
Graph name |
Request Parameters
Parameter |
Mandatory |
Type |
Description |
algorithmName |
Yes |
String |
Algorithm name |
parameters |
Yes |
parameters Object |
Algorithm parameters |
Parameter |
Mandatory |
Type |
Description |
convergence |
No |
Double |
Convergence accuracy. The value range is (0, 1). The default value is 0.00001. |
max_iterations |
No |
Integer |
Maximum iterations. An integer ranging from 1 to 2147483647. For frontend calls, the range is [1, 2000]. The default value is 100. |
community_num_space |
Yes |
String |
Search space for the number of communities, with multiple integer values separated by commas (,). There can be a maximum of 100 values, and each integer value must be within the range of [1, 10000]. |
learning_rate |
No |
Double |
Model learning rate. The value is greater than 0. The default value is 0.01. |
Response Parameters
Parameter |
Type |
Description |
errorMessage |
String |
System prompt. If execution succeeds, this parameter may be left blank. If the execution fails, this parameter is used to display the error message. |
errorCode |
String |
System prompt. If execution succeeds, this parameter may be left blank. If the execution fails, this parameter is used to display the error code. |
jobId |
String |
ID of the algorithm execution job. This parameter is left blank when the request fails.
NOTE:
You can use the returned job ID to view the task execution status and obtain the algorithm return result. For details, see Querying Job Status on the Service Plane (1.0.0). |
jobType |
Integer |
Job type. This parameter is left blank when the request fails. |
Example Request
POST http://{SERVER_URL}/ges/v1.0/{project_id}/hyg/{graph_name}/algorithm { "algorithmName":"bigclam", "parameters":{ "community_num_space":"3,2,10" } }
SERVER_URL: Address for accessing a graph. For details about its value, see Using Service Plane APIs.
Example Response
Status code: 200
Example response for a successful request
Http Status Code: 200 { "jobId": "4448c9fb-0b16-4a78-8d89-2a137c53454a001679122", "jobType": 1 }
Status code: 400
Example response for a failed request
Http Status Code: 400 { "errorMessage":"graph [demo] is not found", "errorCode":"GES.8402" }
Query Result
- Example Request
GET http://{SERVER_URL}/ges/v1.0/{project_id}/graphs/{graph_name}/jobs/{jobId}/status?limit=limit
- Query result response
Table 4 response_data parameter description Parameter
Type
Description
community_num
Integer
Number of communities
communities
List
Community corresponding to each vertex. The format is as follows:
[{vertexId1:communityIds},{vertexId2:communityIds}],
where
vertexId: string type
communityIds: list type. Each element is of the int type.
log_likelihood
Double
Logarithmic value of the joint probability of edges between model vertices calculated by the current graph generation model. The larger the value, the more accurate the estimated number of communities.
Status Codes
Return Value |
Description |
---|---|
400 Bad Request |
Request error. |
401 Unauthorized |
Authorization failed. |
403 Forbidden |
No operation permissions. |
404 Not Found |
No resources found. |
500 Internal Server Error |
Internal server error. |
503 Service Unavailable |
Service unavailable. |
Error Codes
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.
For any further questions, feel free to contact us through the chatbot.
Chatbot