Xiaoyuan Wu

A first variant of a listing and a second variant of the listing are received. The listing describes an item for sale in an electronic marketplace. The first variant describes a different iteration of the item relative to the second variant. First user activity data of the first variant is generated and second user activity data of the second variant is generated. The first user activity data corresponds to user input metrics associated with the first variant. The second user activity data corresponds to user input metrics associated with the second variant. A search engine receives a first query. Based at least in part on the first user activity data relative to the second user activity data, a first search result associated with the first variant is ranked higher than a second search result associated with the second variant.

A categorization analysis system is provided. The categorization analysis system includes one or more hardware processors, a memory including a first plurality of listings categorized in a first target category, and a categorization analysis engine executing on the one or more hardware processors. The categorization analysis engine is configured to determine a label for each listing including performing a search on title, select a set of training listings based on the determined labels, train a first model using the set of training listings and the determined labels, the first model being a classification model configured to classify categorization of listings, identify a suspect listing categorized in the first target category, apply the suspect listing to the first model, thereby generating a categorization result for the suspect listing, the categorization result indicating miscategorization of the suspect listing, and identify the suspect listing in the memory as miscategorized.

A categorization analysis system is provided. The categorization analysis system includes one or more hardware processors, a memory including a first plurality of listings categorized in a first target category, and a categorization analysis engine executing on the one or more hardware processors. The categorization analysis engine is configured to determine a label for each listing including performing a search on title, select a set of training listings based on the determined labels, train a first model using the set of training listings and the determined labels, the first model being a classification model configured to classify categorization of listings, identify a suspect listing categorized in the first target category, apply the suspect listing to the first model, thereby generating a categorization result for the suspect listing, the categorization result indicating miscategorization of the suspect listing, and identify the suspect listing in the memory as miscategorized.

A first variant of a listing and a second variant of the listing are received. The listing describes an item for sale in an electronic marketplace. The first variant describes a different iteration of the item relative to the second variant. First user activity data of the first variant is generated and second user activity data of the second variant is generated. The first user activity data corresponds to user input metrics associated with the first variant. The second user activity data corresponds to user input metrics associated with the second variant. A search engine receives a first query. Based at least in part on the first user activity data relative to the second user activity data, a first search result associated with the first variant is ranked higher than a second search result associated with the second variant.

Techniques for sharing item images based on a similarity score are described. For example, a machine receives an input item listing for an item from a user device. The machine accesses an existing image that may be associated with one or more existing item listings. The machine identifies a degree of similarity between the input item listing and the one or more existing item listings. The machine generates a similarity score for the existing image based on the degree of similarity. The machine selects one or more existing images corresponding to the one or more existing item listings based on the one or more existing images having corresponding similarity scores that exceed a threshold value. The machine transmits a communication to the user device in response to the receiving of the input item listing. The communication may include the one or more existing images.

A categorization analysis system is provided. The categorization analysis system includes one or more hardware processors, a memory including a first plurality of listings categorized in a first target category, and a categorization analysis engine executing on the one or more hardware processors. The categorization analysis engine is configured to determine a label for each listing including performing a search on title, select a set of training listings based on the determined labels, train a first model using the set of training listings and the determined labels, the first model being a classification model configured to classify categorization of listings, identify a suspect listing categorized in the first target category, apply the suspect listing to the first model, thereby generating a categorization result for the suspect listing, the categorization result indicating miscategorization of the suspect listing, and identify the suspect listing in the memory as miscategorized.

A method and a system for making contextual recommendations to users on a network-based system. For example, activity associated with a user interacting with a network-based system is tracked. Based, at least in part, on the tracked user activity on the network-based system, a recommendation relationship type is selected. The recommendation relationship type can be either a substitute relationship type or a complement relationship type. A recommended object can be selected based at least in part on the recommendation relationship type and a first object accessed by the user interacting with the network-based system. A recommendation can be generated for the recommended object for presentation to the user interacting with the network-based system.

A categorization analysis system is provided. The categorization analysis system includes one or more hardware processors, a memory including a first plurality of listings categorized in a first target category, and a categorization analysis engine executing on the one or more hardware processors. The categorization analysis engine is configured to determine a label for each listing including performing a search on title, select a set of training listings based on the determined labels, train a first model using the set of training listings and the determined labels, the first model being a classification model configured to classify categorization of listings, identify a suspect listing categorized in the first target category, apply the suspect listing to the first model, thereby generating a categorization result for the suspect listing, the categorization result indicating miscategorization of the suspect listing, and identify the suspect listing in the memory as miscategorized.

Techniques for determining search results based on session based refinements are presented herein. A method is disclosed that includes receiving a query in a user session, the query comprising one or more search parameters, detecting, in the user session and after receiving the query, a user event associated with a property of an item, updating a record in a table that associates the query with the property, the table comprising a plurality of records that associate the query with respective item properties, the record comprising the query, the property, and a score, and ranking search results for a subsequent query based on the associated properties indicated in the plurality of records, the subsequent query including the one or more search parameters.

Some embodiments provide personalized search services. The personalized search services operate to contextualize what is being searched for and to provide at least some search results in that context that are specific to the user to whom the results are to be provided.

Methods and systems to suggest information to be used with an item being listed for sale may include analyzing a plurality of stored entries of words and identifying certain entries based on desirability scores associated with those entries. The entries and the desirability scores may be identified based on their frequency of use in the plurality of stored entries and/or in the title information or other information associated with items offered for sale.

Techniques for sharing item images based on a similarity score are described. For example, a machine receives an input item listing for an item from a user device. The machine accesses an existing image that may be associated with one or more existing item listings. The machine identifies a degree of similarity between the input item listing and the one or more existing item listings. The machine generates a similarity score for the existing image based on the degree of similarity. The machine selects one or more existing images corresponding to the one or more existing item listings based on the one or more existing images having corresponding similarity scores that exceed a threshold value. The machine transmits a communication to the user device in response to the receiving of the input item listing. The communication may include the one or more existing images.

Methods and systems to suggest information to be used with an item being listed for sale may include analyzing stored search queries and identifying co-occurring terms or terms based on their desirability scores (or desirable terms). The co-occurring terms and the desirable terms may be identified based on their frequency of use in the search queries and/or in the title information or other information associated with items offered for sale.

Techniques for sharing item images based on a similarity score are described. For example, a machine receives an item listing for an item from a user device. The machine generates a similarity score for an existing image corresponding to one or more existing item listings by comparing the item listing received from the user device with the one or more existing item listings. The similarity score may indicate a degree of similarity between the item listing received from the user device and the one or more existing item listings corresponding to the existing image. The machine selects one or more existing images corresponding to the one or more existing item listings based on the one or more existing images having corresponding similarity scores that exceed a threshold value. The machine transmits a communication to the user device in response to the receiving of the item listing from the user device. The communication may include the one or more existing images.

A method and a system for making contextual recommendations to users on a network-based system. For example, activity associated with a user interacting with a network-based system is tracked. Based, at least in part, on the tracked user activity on the network-based system, a recommendation relationship type is selected. The recommendation relationship type can be either a substitute relationship type or a complement relationship type. A recommended object can be selected based at least in part on the recommendation relationship type and a first object accessed by the user interacting with the network-based system. A recommendation can be generated for the recommended object for presentation to the user interacting with the network-based system.

In some example embodiments, a system and method are illustrated to associate an item listing with one or more corresponding images. The system and method include receiving an item listing for an item from a user device. The system and method include generating a similarity score for a respective existing image associated with one or more existing item listings. The generating may be done by comparing the item listing with the existing item listings received from the user device. The similarity score may indicate a degree of similarity between the item listing and the existing item listings associated with the respective existing image. The system and method include proposing a specified number of existing images with the highest similarity scores to the user device. The system and method further include associating the item listing with one or more of the specified number of existing images accepted by the user device. In some example embodiments, a system and method are illustrated to associate an item listing with one or more corresponding images. The system and method include receiving an item listing for an item from a user device. The system and method include generating a similarity score for a respective existing image associated with one or more existing item listings. The generating may be done by comparing the item listing with the existing item listings received from the user device. The similarity score may indicate a degree of similarity between the item listing and the existing item listings associated with the respective existing image. The system and method include proposing a specified number of existing images with the highest similarity scores to the user device. The system and method further include associating the item listing with one or more of the specified number of existing images accepted by the user device.

A method and a system for making contextual recommendations to users on a network-based system. For example, a first and second related category are selected and a relationship score is calculated representing an aspect of the relationship between the first and second related categories. A relationship type is selected to represent the relationship between the first and second categories using the relationship score. The relationship type is one of either a complement or a substitute. A complement relationship indicates a complementary relationship between the first and second related categories. While a substitute relationship indicates a substitutionary relationship between the first and second categories. The relationship type is stored associated with the first and second related categories. Using the selected relationship type, a recommendation is provided to a user on the network-based system.

A method and a system for making contextual recommendations to users on a network-based system. For example, a first and second related category are selected and a relationship score is calculated representing an aspect of the relationship between the first and second related categories. A relationship type is selected to represent the relationship between the first and second categories using the relationship score. The relationship type is one of either a complement or a substitute. A complement relationship indicates a complementary relationship between the first and second related categories. While a substitute relationship indicates a substitutionary relationship between the first and second categories. The relationship type is stored associated with the first and second related categories. Using the selected relationship type, a recommendation is provided to a user on the network-based system.

A computerized method and a system for providing geographic demand distribution and forecast are provided. Example embodiments may include receiving a criterion to identify an item and searching a database for data associated with a geographic distribution of demand for the item identified by the criterion. The computerized method may also include presenting the geographic distribution of the demand for the item on a map. The map may display multiple geographic areas; each one of the multiple geographic areas may be displayed with a first and a second attribute. The first and the second attributes may be associated with the demand.