Help Center> RES> API Reference> APIs (Old Version)> Online Service-related APIs> Prediction APIs
View PDF

Prediction APIs

Function

This API is used to perform online prediction.

URI

POST Prediction address returned after the service is successfully deployed

Request

Table 1 describes the request parameters.

Table 1 Request parameters

Parameter

Mandatory

Type

Description

rec_num

No

Integer

Number of returned requests. The default value is 50.

user_id

Yes

String

User ID

items

No

List

List of item IDs. This parameter needs to be used together with the Item-Based Collaborative Filtering algorithm during service deployment and provides the function of querying the similarity between items.

others_users

No

List

List of user IDs. This parameter provides the function of querying the candidate sets of multiple users.

context

No

String

Context information, which can be used to configure the online filtering blacklist. Items in the blacklist will be excluded from the final candidate set.

filter_items

No

List

List of item IDs. Items in the online blacklist will not be in the returned recommendation candidate set.

online_tags

No

List

List of data formats that consist of attributes, attribute values, and attribute weights. The attribute values or attribute weights are optional. If the weight information is not provided, the default value 0.01 is used. This function must be used together with online retrieval during service deployment. The format is as follows: [{"name": "age","value": " middle-aged ","weight": 0.6},{"name": "sex","value": " male "}].

online_override

No

Boolean

This parameter must be used together with online_tags. Indicates whether the online prediction tags overwrite the tags and weights configured when a service is released. The default value is false. If the value of online_override is false, the existing tags and weights are updated by the new ones.

priority_tags

No

List

List of data formats that consist of attributes, attribute values, attribute weights, and attribute data types. The attribute values or attribute weights are optional. If the weight information is not provided, the default value 0.01 is used. This parameter must be used together with attribute matching and ranking during service deployment. The format is as follows: [{"name":"age","value": "middle-aged ","item_data_type":"strArray"},{"name":"sex","value":[" male "," female "] ,"item_data_type":"strArray"}].

priority_override

No

Boolean

This parameter must be used together with priority_tags. Indicates whether the online prediction tags overwrite the tags and weights related to attribute matching and ranking configured when a service is released. The default value is false. If the value of priority_override is false, the existing tags and weights are updated by the new ones.

topping_sources

No

List

This parameter is used to move the candidate sets of some strategies to the top as required. For example, if topping_sources is set to ["A", "B"], all selected candidate sets of A, all selected candidate sets of B, and candidate sets that are not in A or B are output in sequence.

Response

Table 2 describes the response parameters.
Table 2 Response parameters

Parameter

Mandatory

Type

Description

is_success

Yes

Boolean

Whether the request is successful. The value can be true or false. true indicates the request is successful, and false indicates the request has failed.

data

Yes

JSON

Data parameter. For details, see Table 3.
Table 3 data parameters

Parameter

Mandatory

Type

Description

trace_id

Yes

String

Message tracing ID, which identifies the recommendation result of the current request

rec_num

Yes

Integer

Number of returned results

user_id

Yes

String

User ID

content

Yes

List

Content of recommended items. For details, see Table 4
Table 4 content parameters

Parameter

Mandatory

Type

Description

score

Yes

Float

Score of an item. A larger value indicates a higher probability of being recommended.

item

Yes

JSON

Recommended item. For details, see Table 5.
Table 5 item parameters

Parameter

Mandatory

Type

Description

id

Yes

String

ID of a recommended item

Example

  • Example request 
    {
	"user_id": "u1",
	"rank_items": ["i1", "i2", "i3"],
	"rec_num": 20,
	"items": ["i4", "i5"],
	"other_users": ["u2", "u3"],
	"context": "{\"blackItemList\":[\"i6\", \"i7\", \"i8\"]}",
	"topping_sources": ["A", "B"]
}
  • Example of a successful response 
    {
	"is_success": true,
	"data": {
		"trace_id": "a94495c062c6443b87c097427d99232d",
		"user_id": "260086000000100759",
		"rec_num": 2,
		"content": [{
				"score": 0.51,
				"item": {
					"id": "3101010003401"
				},
				"source": "A"
			},
			{
				"score": 0.52,
				"item": {
					"id": "3101010003402"
				},
				"source": "B"
			},
			{
				"score": 0.53,
				"item": {
					"id": "3101010003403"
				},
				"source": "C"
			}
		]
	}
}
  • Example of a failed response 
    {
    "is_success": false,
    "error_msg": "Backend service found error. Failed to respond due to backend service not found or failed to respond, please check the service you request. "
}

Status Code

For details about status codes, see Status Codes.