Systems and methods for recommendation generation
One or more computing devices, systems, and/or methods for generating and providing recommendations of products are provided. For example, content is extracted from a message sent to a user. The content is evaluated to identify a product identifier corresponding to a product title of a product. If the product identifier is a truncated version of the product title, then a database of product titles and frequencies of occurrence of the product titles is used to complete the product title. A model is used to infer a product category for the product title. Matching scores are assigned to products within a product category based upon weighted attributes. A recommendation is provided to the user for a product having a matching score greater than a matching threshold.
1 . A computing device comprising:
a processor; and
memory comprising processor-executable instructions that when executed by the processor cause performance of operations, the operations comprising:
extracting content from an email of an email account of a user;
evaluating the content to identify a product identifier as a product title of a product;
training a model using a hierarchical taxonomy of product categories and product titles;
inferring a product category for the product title identified for the product using the model trained using the hierarchical taxonomy of product categories and product titles;
assigning matching scores to products within a product catalog using weighted attributes corresponding to at least one of categories, brands, product titles, and product descriptions, wherein the matching scores correspond to probabilities that products within the product catalog match the product title; and
transmitting a recommendation of a product having a matching score greater a matching threshold over a computing network to a device associated with the user.
2 . The computing device of claim 1 , wherein the operations comprise:
training the model using at least one of a database, crawled website data, one or more product catalogs or labeled product data.
3 . The computing device of claim 1 , wherein the operations comprise:
aggregating email data of the email account over a time period, wherein the email data corresponds to emails from a plurality of product providers.
4 . The computing device of claim 3 , wherein the operations comprise:
matching product categories of product titles occurring within the email data to products within the catalog for creating recommendations.
5 . The computing device of claim 4 , wherein the operations comprise:
aggregating the email data to track frequency and recency of product titles occurring within the email data.
6 . The computing device of claim 5 , wherein the operations comprise:
utilizing the frequency and recency for assigning matching scores to products within the product catalog.
7 . The computing device of claim 4 , wherein the operations comprise:
aggregating the email data to track at least one of frequency or recency of product titles occurring within the email data.
8 . The computing device of claim 7 , wherein the operations comprise:
utilizing at least one of the frequency or the recency for assigning matching scores to products within the product catalog.
9 . A non-transitory machine readable medium having stored thereon processor-executable instructions that when executed cause performance of operations, the operations comprising:
extracting content from messages of an account of a user;
evaluating the content to identify a product identifier as a product title of a product;
identifying a product category for the product title identified for the product;
assigning matching scores to products within a product catalog using weighted attributes corresponding to at least one of categories, brands, product titles, and product descriptions, wherein the matching scores correspond to probabilities that products within the product catalog match the product title; and
generating a recommendation of a product having a matching score above a matching threshold.
10 . The non-transitory machine readable medium of claim 9 , wherein the operations comprise:
assigning the matching scores based upon a recency of the product occurring within at least one of email data, catalog data, or crawled data.
11 . The non-transitory machine readable medium of claim 10 , wherein the recency corresponds to the product occurring within email data.
12 . The non-transitory machine readable medium of claim 10 , wherein the recency corresponds to the product occurring within catalog data.
13 . The non-transitory machine readable medium of claim 10 , wherein the recency corresponds to the product occurring within crawled data.
14 . The non-transitory machine readable medium of claim 9 , wherein the operations comprise:
assigning the matching scores based upon a frequency of the product.
15 . The non-transitory machine readable medium of claim 14 , wherein the frequency corresponds to the product occurring within email data.
16 . The non-transitory machine readable medium of claim 14 , wherein the frequency corresponds to the product occurring within catalog data.
17 . The non-transitory machine readable medium of claim 14 , wherein the frequency corresponds to the product occurring within crawled data.
18 . A method, comprising:
executing, on a processor of a computing device, instructions that cause the computing device to perform operations, the operations comprising:
extracting content from an email of an email account of a user;
evaluating the content to identify a product identifier as a product title of a product;
training a model using a hierarchical taxonomy of product categories and product titles;
inferring a product category for the product title identified for the product using the model trained using the hierarchical taxonomy of product categories and product titles;
assigning matching scores to products within a product catalog using weighted attributes corresponding to at least one of categories, brands, product titles, and product descriptions, wherein the matching scores correspond to probabilities that products within the product catalog match the product title; and
transmitting a recommendation of a product having a matching score greater a matching threshold over a computing network to a device associated with the user.
19 . The method of claim 18 , comprising:
modifying weights for the weighted attributes based upon user interaction data of content associated with products within the product catalog.
20 . The method of claim 18 , comprising:
aggregating email data of the email account over a time period, wherein the email data corresponds to emails from a plurality of product providers.