Abstract: Techniques for identifying similar queries based on their overall similarity and partial similarity of time series of frequencies of the queries are provided. To identify queries that are similar to a target query, the query analysis system generates, for each query, an overall similarity score for that query and the target query based on the time series of the query and the target query. The query analysis system also generates, for each query, partial similarity scores for the query and the target query based on various time sub-series of the overall time series of the queries. The query analysis system then identifies queries as being similar to the target query based on the overall similarity scores and the partial similarity scores of the queries.

Abstract: Techniques for analyzing and modeling the frequency of queries are provided by a query analysis system. A query analysis system analyzes frequencies of a query over time to determine whether the query is time-dependent or time-independent. The query analysis system forecasts the frequency of time-dependent queries based on their periodicities. The query analysis system forecasts the frequency of time-independent queries based on causal relationships with other queries. To forecast the frequency of time-independent queries, the query analysis system analyzes the frequency of a query over time to identify significant increases in the frequency, which are referred to as “query events” or “events.” The query analysis system forecasts frequencies of time-independent queries based on queries with events that tend to causally precede events of the query to be forecasted.

Abstract: Techniques for analyzing and modeling the frequency of queries are provided by a query analysis system. A query analysis system analyzes frequencies of a query over time to determine whether the query is time-dependent or time-independent. The query analysis system forecasts the frequency of time-dependent queries based on their periodicities. The query analysis system forecasts the frequency of time-independent queries based on causal relationships with other queries. To forecast the frequency of time-independent queries, the query analysis system analyzes the frequency of a query over time to identify significant increases in the frequency, which are referred to as “query events” or “events.” The query analysis system forecasts frequencies of time-independent queries based on queries with events that tend to causally precede events of the query to be forecasted.

Abstract: Described is a behavioral targeting technology for online advertising, by which an original attribute is uniformly expanded. Users that meet an original attribute are aggregated into a mid-result used to determine similarity relative to candidate attribute types. The most similar candidate attributes are selected for the expanded attribute. A URL/URL pattern suggestion technology is provided, with similarity computed from users/URLs visited by the users. URLs are separated into URL tree nodes, for calculating the number of users who have visited each URL and the number of users who have visited the URL on a sub-tree whose root is the node. URL/URL patterns are output based on similarity. Domains are also suggested based on user-visits. Similarities between pairs of domains may be computed (e.g., offline), with an output for a given domain provided in based on its similarity with each other domain.

Abstract: Described is a technology by which online advertisements for returning with a query response are ranked according to reputation. The reputation may correspond to a product or service and/or seller reputation. In one example, a set of relevant advertisement items are located and ranked using reputation data as a factor. For example, for each item, a ranking value is based on a mathematical combination of a product reputation score, a seller reputation score and a relevance score, with the items ranked by their computed values. The scores may be weighted differently. The reputation data may be mined from a review source, such as customer reviews available on the web. In one example implementation, a 3-gram model that considers terms in the review along with the two terms proceeding each term is used to analyze the reviews to determine whether each review is positive or negative with respect to the reputation.

Abstract: Described is an efficient retrieval mechanism that quickly locates documents (e.g., corresponding to online advertisements) based on query term discrimination. A topmost subset (e.g., two) of search terms is selected according to their ranked importance, e.g., as ranked by inverted document frequency. The topmost terms are then used to narrow the number of rows of an inverted query index that are searched to find document identifiers and associated scores, such as computed offline by a BM25 algorithm. For example, for each document identifier of each important term, a fast search within each of the narrowed subset of rows (that also contain that document identifier) may be performed by comparing document identifiers to jump a pointer within each other row, followed by a binary search to locate a particular document. The scores of the set of particular documents may then be used to rank their relative importance for returning as results.

Abstract: A method for ranking online advertisements using retailer reputation and product reputation. In one implementation, a query may be received. Advertisements may be selected by determining a level of relevance between the query and each advertisement and selecting the advertisements with a level of relevance above a pre-determined level of relevance. A predicted reputation for a retailer and a predicted reputation for a product may be retrieved for each of the selected advertisements. The selected advertisements may then be ranked based on the predicted reputation for the retailer and the predicted reputation of the product. The ranking of the selected advertisements may be accomplished by calculating a ranking score for each selected advertisement based on the retailer predicted reputation and the product predicted reputation. The selected advertisements may then be displayed according to the ranking.

Abstract: Described is a technology by which blocks of web pages may be selected, such as for building a user-personalized web page containing selected blocks. A selection mechanism, such as a browser toolbar add-on, provides a user interface for selecting blocks, and records information about selected blocks. A block tracking mechanism (e.g., a daemon program) uses the information to locate selected blocks of the web pages, including when the web page containing the block is updated with respect to content and/or layout. The block tracking mechanism may update a local gadget that when invoked, such as by browsing to a particular web page, which shows updated versions of the block on a personalized web page. Blocks may be efficiently located by processing trees representing web pages into reduced trees, and then by performing a minimum distance mapping algorithm on the reduced trees.

Abstract: Systems and methods are described for an Internet visualization system and related user interfaces. In one implementation, the system analyzes Internet search logs to determine most popular search queries across the world at a current time. A user interface displays a keyword of each of the most popular queries in a single visual display that relates each query to a geographical location of greatest popularity. The system can also filter queries according to demographics. In one implementation the user interface provides a 3-dimensional Internet visualization that adopts an ocean or seascape theme. The ocean floor displays a map of the world, and query bubbles rise from geographical locations on the map. The size and duration of each query bubble denotes the relative popularity of a given query.

Abstract: A method and system for ranking messages of discussion threads based on relationships between messages and authors is provided. The ranking system defines an equation for attributes of a message and an author. The equations define the attribute values and are based on relationships between the attribute and the attributes associated with the same type of object, and different types of objects. The ranking system iteratively calculates the attribute values for the objects using the equations until the attribute values converge on a solution. The ranking system then ranks the messages based on attribute values.

Abstract: Systems and methods for related term suggestion are described. In one aspect, term clusters are generated as a function of calculated similarity of term vectors. Each term vector having been generated from search results associated with a set of high frequency of occurrence (FOO) historical queries previously submitted to a search engine. Responsive to receiving a term/phrase from an entity, the term/phrase is evaluated in view of terms/phrases in the term clusters to identify one or more related term suggestions.

Abstract: An exemplary method for use in information retrieval includes, for each of a plurality of terms, selecting a predetermined number of top scoring documents for the term to form a corresponding document set for the term; receiving a plurality of terms, optionally as a query; ranking the plurality of terms for importance based at least in part on the document sets for the plurality of terms where the ranking comprises using an inverse document frequency algorithm; selecting a number of ranked terms based on importance where each selected, ranked term comprises its corresponding document set wherein each document in a respective document set comprises a document identification number; forming a union set based on the document sets associated with the selected number of ranked terms; and, for a document identification number in the union set, scanning a document set corresponding to an unselected term for a matching document identification number. Various other exemplary systems, methods, devices, etc.

Abstract: An exemplary web browser system includes a selection module for selecting a webpage block and recording information about a selected webpage block; a tracking module for tracking changes to a selected webpage block based at least in part on the recorded information for that webpage block; and a display module for displaying a selected webpage block wherein the tracking module updates the display module as to changes to the selected webpage block. Various other exemplary systems, methods, devices are also disclosed.

Abstract: Systems and methods for implementing diverse topic phrase extraction are disclosed. According to one implementation, multiple word candidate phrases are extracted from a corpus and weighed. One or more documents are re-weighed to identify less obvious candidate topics using latent semantic analysis (LSA). Phrase diversification is then used to remove redundancy and select informative and distinct topic phrases.

Abstract: A method and system for classifying display pages based on automatically generated summaries of display pages. A web page classification system uses a web page summarization system to generate summaries of web pages. The summary of a web page may include the sentences of the web page that are most closely related to the primary topic of the web page. The summarization system may combine the benefits of multiple summarization techniques to identify the sentences of a web page that represent the primary topic of the web page. Once the summary is generated, the classification system may apply conventional classification techniques to the summary to classify the web page. The classification system may use conventional classification techniques such as a Naïve Bayesian classifier or a support vector machine to identify the classifications of a web page based on the summary generated by the summarization system.

Abstract: A method and system are provided for identifying groups in large-scale networks. The large-scale networks include a collection of nodes and edges that may represent relationships between entities or individuals. The large-scale network is split into a number of fractions satisfying an edge threshold. In turn, the nodes in each fraction are merged to generate one or more clusters based on a specified similarity metric. The large-scale network is recursively split and clustered until distinct groups are identified.

Abstract: A system, method, computer-readable media, and related techniques are disclosed for predicting demographic information of a user. A social link network is created and a search request for demographic information related to a first user within the social link network is received. The requested demographic information based on the demographic information of other users connected to the first user within the social link network is provided.

Abstract: A social network visualization and mining system that includes a visualization application for mining social networks of users in an online social network. This visualization can be used to mine the social network for additional information and intelligence. The social network is displaying in graphical form, such as a node-link graph, with a center node representing the social network of a user being examined, and secondary nodes represent the primary user's friends. Lines represent links between the primary user and his friends, while various visualization features such as line thickness, line color, and text size are used to easily identify the type of relationship between users. The system also includes a topics visualization module, which builds and displays a social network based on a certain topic or keyword that is entered by the application user. A demographic prediction module examines a user's social network to predict demographics of users.

Abstract: A method and system for generating and using a combined model to identify whether a bid term is relevant to an advertisement is provided. A relevance system trains a combined model that includes an initial model and a decision tree model that are trained using features that represent relationships between bid terms and advertisements. The relevance system trains the initial model to map initial model features to a modeled relevance. The relevance system trains the decision tree model to map the decision tree features and the modeled relevance to a final relevance. The trained initial model and decision tree model represent the combined model. The relevance system then uses the combined model to determine the relevance of bid terms to advertisements.

Abstract: Systems and methods for clustering-based text classification are described. In one aspect text is clustered as a function of labeled data to generate cluster(s). The text includes the labeled data and unlabeled data. Expanded labeled data is then generated as a function of the cluster(s). The expanded label data includes the labeled data and at least a portion of unlabeled data. Discriminative classifier(s) are then trained based on the expanded labeled data and remaining ones of the unlabeled data.