Google Inc.

Baris Yuksel

Douglas Richard Grundman

Craig Nevill-Manning

Anurag Adarsh

Piyush Janawadkar

Lev Ratinov

Ashutosh Kulshreshtha

Andrea Lynn Frome

Ethan G Russell

Paul Eastlund

Diego Ariel Gertzenstein

Howard Wellington Trickey

Daniel J Yehuda

Melanie Clements

Ana Krulec

Samantha Merritt

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for document ranking. One of the methods includes receiving a request for one or more documents, obtaining a set of documents responsive to the request, and obtaining, from a user profile associated with a source of the request, representations of one or more topics of interest to a user. The method also includes selecting, from the set of documents, at least one document associated with a particular topic that matches at least one of the one or more topics of interest to the user, for the at least one selected document, obtaining a value corresponding to an inverse document frequency of documents associated with the particular topic in a corpus of documents, and generating a score for the at least one document based at least in part on the value corresponding to the inverse document frequency.

Document ranking based on entity frequency

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for providing recommendations to users. One of the methods includes receiving data indicating a selection, by a user, of a notification option relating to a first entity, adding the first entity to an interest list for the user, determining, based at least on adding the first entity to the interest list for the user, to provide a notification relating to an entity, and providing, based on determining to provide the notification relating to the entity, the notification.

Notification engine

Systems and methods are provided for determining whether given business listings are spam or are closed businesses. In one example, a system receives a business listing from a user. The business listing contains a geolocation and the name of a business. The system also receives one or more street level images related to a geographic object that is at or near the same geolocation as the business listing. The system determines if the business listing is spam or is closed based on the received street level images by determining if the geographic object is the business or by analyzing whether the business associated with the listing is shown in any of the street level images.

Systems and methods of correlating business information to determine spam, closed businesses, and ranking signals

The present invention relates generally to identifying fraudulent businesses and business listings. More specifically, the invention relates to determining a 'surprisingness' value for a particular combination of words in a business title based on the likelihood that the combination has appeared in legitimate business titles. The value may be used to determine whether the business or business listing is legitimate or fraudulent. For example, third party hijackers may 'keyword-stuff' business titles or attempt to include words associated with prominent businesses in a title of a less prominent business associated with the third party in order to have the less prominent business displayed more often in search results for the prominent business. For example, if a business title has too many surprising word combinations or a particular combination is highly unlikely, the business listing is likely to be fraudulent or &#x201c;keyword-stuffed&#x201d; and may be withheld, excluded, removed from search results.

Systems and methods of detecting keyword-stuffed business titles

Methods, systems, and apparatus, including computer programs encoded on computer storage media, are described for identifying target terms, e.g., spam terms within a collection of documents. In one aspect, methods can include identifying spam terms by calculating a blacklist term frequency-inverse document frequency (BTF-IDF) score for multiple terms, and by selecting, as the spam terms, the terms that have scores above or below a threshold score. The multiple terms may be derived from documents that are associated with accounts that have been designated as spam accounts.

Identifying terms

A method and system as described identifies business records. A business listing stores one or more business records. In order to filter spam records, the method and system operate to identify whether a target business record is a spam business record. To identify the target record as a spam record, the method and system compare extracted characteristics of the target record with characteristics of known spam or good business records. Depending upon the similarity with the known business records, the target business record is identified as a spam record or a good record.

Method and system for identifying business records

False or spam business listings may be identified by searching reviews associated therewith. The associated business reviews may be searched for one or more keywords that suggest the business listing is spam. Moreover, business listings may be categorized based on a type of business in the listing, and the associated reviews for business listings in each category may be searched for characteristics particular to that category. Further business reviews may be searched for hits matching portions of business reviews associated with known spam business listings. Once a business listing is identified as potential spam using these techniques, the listing may be flagged for further review.

Aspects of the disclosure provide for detection of spam business listings. Aspects operate to identify business listing characteristics in trusted sources and untrusted sources. As untrusted sources are likely to contain more spam, characteristics that are present in untrusted sources but not present in trusted sources are typically indicative of spam listings, and vice versa. Thus, statistical analysis of the frequency of characteristics within each source may be used to identify common characteristics of spam listings. These characteristics may further be analyzed in specific listing contexts, as different listing contexts (e.g., different types of businesses) typically use different terms and vocabularies, such that terms that are indicative of spam in one context may not be indicative of spam in another. Various methods for leveraging this context-specific statistical information to improve spam detection operations are disclosed.

Detection of spam using contextual analysis of data sources

Aspects of the disclosure provide for detection of spam attacks. In order to filter spam listings among business listings associated with a particular geographical region, a method and system operate to analyze the frequency of particular characteristics of the business listings, such as words within a listing title, phone numbers, or websites, to identify a normal frequency of each characteristic. Business listings may be periodically analyzed to identify anomalous increases in the frequency of particular characteristics. Characteristics that exhibit these anomalies may be identified as suspicious characteristics, and listings that contain suspicious characteristics may be identified as possible spam listings.

Method and system for identifying business listing characteristics

A determination of whether a mapped business listing that is produced as a search result corresponds to an actual location of operation is based on different factors. One factor identifies whether the business listing is associated with a business category that appears as search results for a particular geographic area in numbers that exceed average proportions for the same business category density in similarly situated geographic areas. Another factor determines whether different business listings in the same geographic area include the same identifying data. Specific characteristics of a neighborhood where the business listing is mapped provide an additional factor for identifying whether a search result for a business listing is map spam. The different factors may be considered together to determine the likelihood that a mapped search result is spam.

Map spam detection

A system and method are provided that detect spam business listing data. In one aspect, a plurality of clusters containing previously captured spam business data is formed. The attributes of new incoming business data may be compared to the attributes of the clustered spam business data. If the incoming business data is associated with any one of the plurality of clusters, the incoming business data may be deemed spam.

Growing community of inventors

New York, NY, United States of America

Baris Yuksel