摘要:
Technologies are described herein for extracting quotes from customer reviews for a collection of items in order to provide salient comments for a particular item and/or upsell the collection. Individual sentences contained the customer reviews regarding the collection of items are parsed into a collection of sentences. A list of topics is generated from the collection of sentences, and each sentence in is assigned to one or more of the topics. A number of the topics from the list of topics are identified as related to the particular item, and quotes are extracted from the sentences assigned to the identified topics to be displayed as customer reviews of the particular item. Similarly, a number of topics in the list of topics are identified as related to a customer satisfaction with the collection of items, and quotes are extracted from the sentences assigned to the identified topics to be displayed in conjunction with an offer to upsell the collection of items.
摘要:
Technologies are described herein for classifying sentences or phrases as expressing positive or negative sentiment based on machine learning from training data comprising sentences manually labeled as to sentiment. A list of terms is generated from the manually labeled sentences and sentiment scores are determined for the terms in the list of terms based on the manually labeled sentences. A collection of sentences or phrases may then be classified as to sentiment utilizing one or more logistic regression classifiers trained on the sentiment scores determined for the terms in the list of terms. The classified collection of sentences may be further analyzed to determine an overall majority sentiment regarding a topic discussed in the sentences and/or to extract specific sentences or phrases expressing a particular sentiment for display to a customer.
摘要:
Technologies are described herein for extracting quotes for an item from customer reviews. Individual sentences or phrases contained in customer reviews regarding an item or category of items are parsed into a collection of sentences. A list of topics is generated from the collection of sentences, and each sentence or phrases is assigned a topic from the list of topics. The most relevant topics from the list of topics are identified for a particular item, and one or more sentences or phrases from customer reviews for the particular item are selected for each of the most relevant topics that best represent the sentences associated with the item and assigned to that topic, according to content and sentiment.
摘要:
Technologies are described herein for extracting quotes from customer reviews regarding collections of items. An identifier of an individual item is received and customer reviews regarding a collection of items containing the individual item are retrieved. A collection of sentences is parsed from the retrieved customer reviews, and sentences that discuss the individual item are identified. A quote is extracted from the identified sentences and displayed to a customer in conjunction with information regarding the individual item. Similarly, individual items belonging to a collection of items are identified, and customer reviews regarding the individual items are retrieved. A collection of sentences are parsed from the retrieved customer reviews, and those sentences that discuss topics relevant to the collection of items are identified. A number of quotes are extracted from the identified sentences and displayed to a customer in conjunction with information regarding the collection of items.
摘要:
The top rated products in a product category are identified by computing a score for each product in the category. The score is computed by calculating a time decay factor for each product category. The time decay factor is utilized to decay the ratings of each product as a function of time. A relevance value is also computed for each product. The relevance value is utilized to modify the ratings for each product in view of the recency of the ratings. The score may be computed for each product using the ratings for the product, the computed time decay factor, and/or the relevance value. The computed scores are utilized to rank the products in the product category.
摘要:
Disclosed are various embodiments for increasing diversity of search results. A computing device performs a search of items in an electronic catalog to produce a plurality of search results. The electronic catalog is stored in a data store accessible to the computing device. A latent group is identified which includes a most relevant one of a plurality of search results. Each of the search results has a plurality of attributes. Membership in the latent group is determined by similarity of a non-indexed one of the attributes. A relevance score of those search results in the identified latent group other than the most relevant one of the search results is reduced.
摘要:
Disclosed are various embodiments for selecting a subset of categories of product items to be used in merchandising. The subset of categories may be selected on a basis of a measured level of interest in the product items. Based on the subset of subcategories that have been selected, merchandising presentations may be automatically formulated and presented to a customer.
摘要:
A system and method for selectively displaying selection indication fields is described. A user may select multiple items using various selection indication fields. The fields may be displayed all the time or may only appear if a pointing device or cursor is near the field or an associated item.
摘要:
Disclosed are various embodiments for tracking user behavior relative to a network page and identifying user interest in various content items of the network page according to the user behavior. A network page that includes multiple content items is rendered for display in a client. A user action is obtained relative to the network page. A user behavior report is sent to one or more servers. The user behavior report indicates the user action, a timestamp associated with the user action, and one or more of the content items that are associated with the user action.
摘要:
A computer user enters free-form electronic ink at run-time of an application program for manipulating the trajectories of mobile objects within the application program. Collisions between mobile objects and ink are processed by determining a collision angle that bisects a first and last intersecting point along a periphery of the first object. After the collision, the first object moves away from the ink as if the first object collided with a flat surface oriented perpendicular to the collision angle.