2024/05/01 - Amazon Personalize Runtime - 2 updated api methods
Changes This release adds support for a Reason attribute for predicted items generated by User-Personalization-v2.
{'personalizedRanking': {'reason': ['string']}}
Re-ranks a list of recommended items for the given user. The first item in the list is deemed the most likely item to be of interest to the user.
See also: AWS API Documentation
Request Syntax
client.get_personalized_ranking(
campaignArn='string',
inputList=[
'string',
],
userId='string',
context={
'string': 'string'
},
filterArn='string',
filterValues={
'string': 'string'
},
metadataColumns={
'string': [
'string',
]
}
)
string
[REQUIRED]
The Amazon Resource Name (ARN) of the campaign to use for generating the personalized ranking.
list
[REQUIRED]
A list of items (by itemId) to rank. If an item was not included in the training dataset, the item is appended to the end of the reranked list. If you are including metadata in recommendations, the maximum is 50. Otherwise, the maximum is 500.
(string) --
string
[REQUIRED]
The user for which you want the campaign to provide a personalized ranking.
dict
The contextual metadata to use when getting recommendations. Contextual metadata includes any interaction information that might be relevant when getting a user's recommendations, such as the user's current location or device type.
(string) --
(string) --
string
The Amazon Resource Name (ARN) of a filter you created to include items or exclude items from recommendations for a given user. For more information, see Filtering Recommendations.
dict
The values to use when filtering recommendations. For each placeholder parameter in your filter expression, provide the parameter name (in matching case) as a key and the filter value(s) as the corresponding value. Separate multiple values for one parameter with a comma.
For filter expressions that use an INCLUDE element to include items, you must provide values for all parameters that are defined in the expression. For filters with expressions that use an EXCLUDE element to exclude items, you can omit the filter-values.In this case, Amazon Personalize doesn't use that portion of the expression to filter recommendations.
For more information, see Filtering Recommendations.
(string) --
(string) --
dict
If you enabled metadata in recommendations when you created or updated the campaign, specify metadata columns from your Items dataset to include in the personalized ranking. The map key is ITEMS and the value is a list of column names from your Items dataset. The maximum number of columns you can provide is 10.
For information about enabling metadata for a campaign, see Enabling metadata in recommendations for a campaign.
(string) --
(list) --
(string) --
dict
Response Syntax
{
'personalizedRanking': [
{
'itemId': 'string',
'score': 123.0,
'promotionName': 'string',
'metadata': {
'string': 'string'
},
'reason': [
'string',
]
},
],
'recommendationId': 'string'
}
Response Structure
(dict) --
personalizedRanking (list) --
A list of items in order of most likely interest to the user. The maximum is 500.
(dict) --
An object that identifies an item.
The and APIs return a list of ``PredictedItem``s.
itemId (string) --
The recommended item ID.
score (float) --
A numeric representation of the model's certainty that the item will be the next user selection. For more information on scoring logic, see how-scores-work.
promotionName (string) --
The name of the promotion that included the predicted item.
metadata (dict) --
Metadata about the item from your Items dataset.
(string) --
(string) --
reason (list) --
If you use User-Personalization-v2, a list of reasons for why the item was included in recommendations. Possible reasons include the following:
Promoted item - Indicates the item was included as part of a promotion that you applied in your recommendation request.
Exploration - Indicates the item was included with exploration. With exploration, recommendations include items with less interactions data or relevance for the user. For more information about exploration, see Exploration.
Popular item - Indicates the item was included as a placeholder popular item. If you use a filter, depending on how many recommendations the filter removes, Amazon Personalize might add placeholder items to meet the numResults for your recommendation request. These items are popular items, based on interactions data, that satisfy your filter criteria. They don't have a relevance score for the user.
(string) --
recommendationId (string) --
The ID of the recommendation.
{'itemList': {'reason': ['string']}}
Returns a list of recommended items. For campaigns, the campaign's Amazon Resource Name (ARN) is required and the required user and item input depends on the recipe type used to create the solution backing the campaign as follows:
USER_PERSONALIZATION - userId required, itemId not used
RELATED_ITEMS - itemId required, userId not used
For recommenders, the recommender's ARN is required and the required item and user input depends on the use case (domain-based recipe) backing the recommender. For information on use case requirements see Choosing recommender use cases.
See also: AWS API Documentation
Request Syntax
client.get_recommendations(
campaignArn='string',
itemId='string',
userId='string',
numResults=123,
context={
'string': 'string'
},
filterArn='string',
filterValues={
'string': 'string'
},
recommenderArn='string',
promotions=[
{
'name': 'string',
'percentPromotedItems': 123,
'filterArn': 'string',
'filterValues': {
'string': 'string'
}
},
],
metadataColumns={
'string': [
'string',
]
}
)
string
The Amazon Resource Name (ARN) of the campaign to use for getting recommendations.
string
The item ID to provide recommendations for.
Required for RELATED_ITEMS recipe type.
string
The user ID to provide recommendations for.
Required for USER_PERSONALIZATION recipe type.
integer
The number of results to return. The default is 25. If you are including metadata in recommendations, the maximum is 50. Otherwise, the maximum is 500.
dict
The contextual metadata to use when getting recommendations. Contextual metadata includes any interaction information that might be relevant when getting a user's recommendations, such as the user's current location or device type.
(string) --
(string) --
string
The ARN of the filter to apply to the returned recommendations. For more information, see Filtering Recommendations.
When using this parameter, be sure the filter resource is ACTIVE.
dict
The values to use when filtering recommendations. For each placeholder parameter in your filter expression, provide the parameter name (in matching case) as a key and the filter value(s) as the corresponding value. Separate multiple values for one parameter with a comma.
For filter expressions that use an INCLUDE element to include items, you must provide values for all parameters that are defined in the expression. For filters with expressions that use an EXCLUDE element to exclude items, you can omit the filter-values.In this case, Amazon Personalize doesn't use that portion of the expression to filter recommendations.
For more information, see Filtering recommendations and user segments.
(string) --
(string) --
string
The Amazon Resource Name (ARN) of the recommender to use to get recommendations. Provide a recommender ARN if you created a Domain dataset group with a recommender for a domain use case.
list
The promotions to apply to the recommendation request. A promotion defines additional business rules that apply to a configurable subset of recommended items.
(dict) --
Contains information on a promotion. A promotion defines additional business rules that apply to a configurable subset of recommended items.
name (string) --
The name of the promotion.
percentPromotedItems (integer) --
The percentage of recommended items to apply the promotion to.
filterArn (string) --
The Amazon Resource Name (ARN) of the filter used by the promotion. This filter defines the criteria for promoted items. For more information, see Promotion filters.
filterValues (dict) --
The values to use when promoting items. For each placeholder parameter in your promotion's filter expression, provide the parameter name (in matching case) as a key and the filter value(s) as the corresponding value. Separate multiple values for one parameter with a comma.
For filter expressions that use an INCLUDE element to include items, you must provide values for all parameters that are defined in the expression. For filters with expressions that use an EXCLUDE element to exclude items, you can omit the filter-values. In this case, Amazon Personalize doesn't use that portion of the expression to filter recommendations.
For more information on creating filters, see Filtering recommendations and user segments.
(string) --
(string) --
dict
If you enabled metadata in recommendations when you created or updated the campaign or recommender, specify the metadata columns from your Items dataset to include in item recommendations. The map key is ITEMS and the value is a list of column names from your Items dataset. The maximum number of columns you can provide is 10.
For information about enabling metadata for a campaign, see Enabling metadata in recommendations for a campaign. For information about enabling metadata for a recommender, see Enabling metadata in recommendations for a recommender.
(string) --
(list) --
(string) --
dict
Response Syntax
{
'itemList': [
{
'itemId': 'string',
'score': 123.0,
'promotionName': 'string',
'metadata': {
'string': 'string'
},
'reason': [
'string',
]
},
],
'recommendationId': 'string'
}
Response Structure
(dict) --
itemList (list) --
A list of recommendations sorted in descending order by prediction score. There can be a maximum of 500 items in the list.
(dict) --
An object that identifies an item.
The and APIs return a list of ``PredictedItem``s.
itemId (string) --
The recommended item ID.
score (float) --
A numeric representation of the model's certainty that the item will be the next user selection. For more information on scoring logic, see how-scores-work.
promotionName (string) --
The name of the promotion that included the predicted item.
metadata (dict) --
Metadata about the item from your Items dataset.
(string) --
(string) --
reason (list) --
If you use User-Personalization-v2, a list of reasons for why the item was included in recommendations. Possible reasons include the following:
Promoted item - Indicates the item was included as part of a promotion that you applied in your recommendation request.
Exploration - Indicates the item was included with exploration. With exploration, recommendations include items with less interactions data or relevance for the user. For more information about exploration, see Exploration.
Popular item - Indicates the item was included as a placeholder popular item. If you use a filter, depending on how many recommendations the filter removes, Amazon Personalize might add placeholder items to meet the numResults for your recommendation request. These items are popular items, based on interactions data, that satisfy your filter criteria. They don't have a relevance score for the user.
(string) --
recommendationId (string) --
The ID of the recommendation.