TWI501096B - Ranking user generated web content - Google Patents

Description

Ranking user generated web content

This specification generally describes techniques for analyzing web content, including user-generated web content.

A website or electronic community (eg, a collection of websites hosted by the same entity) may host one or more types of electronic content created and/or uploaded by users of the website. For example, forums, e-books, and video-sharing sites provide users with the ability to post or upload user-generated content to share with other users. Some hosting sites require each user to log in using identifying information before contributing content. In this way, content can be positively associated with a login user. In some cases, users can interact with each other. For example, in a forum, a first user can post a question or comment, and other users can respond to postings by the first user.

This specification describes techniques for weighting interactions between users in an electronic community and generating user credential scores based on interactions between users. In general, user-generated content items (eg, comments on a blog or an answer posted on a question and answer website) can be analyzed to assign a quality factor. User-generated content items may be additionally analyzed to assess the quality of the input and identify individual interactions between the user (eg, uploading one or more users of the response to the question posted by the first user). The interactions may be represented in a social or user activity map having weights assigned to directed links between pairs of users based on the assigned quality factors. These weights and corresponding interactions can be used to generate user credential scores.

In general, one aspect of the subject matter described in this specification can be embodied in a method that includes the following actions: by a computer operation to identify multiple interactions between users via an electronic network, each interaction Between a pair of users; and assign each interaction a weighting factor that represents the quality of the interaction. User credential scores for each of a plurality of users are generated using one or more processors. The user credential scores are based on the weighting factors for each of the plurality of interactions. The user credential scores are stored in association with a user identifier on a computer readable storage device. Other embodiments of this aspect include corresponding systems, devices, and computer programs that are configured to perform the actions of the methods, encoded on a computer readable storage device.

These and other embodiments may each include one or more of the following features, individually or in combination, as appropriate. Receiving a search query and using a processor to identify content items or items generated by a user responsive to the search query, and ranking the identified users based at least in part on user credential scores associated with the items Content item. Identifying a user-generated content item responsive to the search query may involve assigning a measure of content relevance for each item, and ranking the user-generated content items based on the user credential scores may involve combining The user credentials associated with the projects are scored and the measurements relating to the content of the content of the projects. The measurements of the user credential scores and content relevance may be normalized, wherein the combinations of the user credential scores associated with the items and the content relevance for the items may be combined These normalized measurements involving the combination of such normalized user voucher scores and content relevance are involved. A processor may be used to generate a user activity map that identifies links between users based on interactions between users, determine an authority score for each user, and determine a contribution score for each user. The authority score of a particular user may be based on the contribution score of the user to which the particular user is linked in the user activity map, and the contribution score may be based on the particular user in the user activity map The authority rating of the user linked to it. The authority scores and the contribution scores are generated using a repeated update procedure until the repeated update procedure reaches a predetermined convergence threshold. The user interactions may correspond to user generated content on at least one of a question and answer website, a bulletin board website, a blog, or a social networking website. The weighting factor can include a combination of a plurality of quality factors, and the quality factor can include one of a previous content item associated with one of the user's content items and one of the other users, and one content item relative to the other content item The originality, a content item corresponds to a coverage of one of the non-generic terms in the content item, the richness of the content item, or the timeliness of the content item. The rewarding user can be scored based on the user credentials.

In general, another aspect of the subject matter described in this specification can be embodied in a method comprising: receiving and distributing content generated by a user for accessing a network, storing user generated content, and Identifying interactions between user pairs associated with the content generated by the stored user. An objective measure for each interaction based on one of the quality of each interaction produces a weighting factor for the interaction, based on the identified interactions and the weighting factors for the interactions to generate a user voucher score for each user And ranking the user or user generated content based on the user credentials.

These and other embodiments may each include one or more of the following features, as appropriate. The user-generated content can be received and published using one or more servers, and the user-generated content can be stored in one or more storage devices, and the interaction can be identified using one or more processors to generate a weighting factor. The user credentials are scored and the content generated by the user or user is ranked. A user credential score for each user is generated by repeatedly updating an authority score and a related contribution score based on the identified interactions and the weighting factors. A search query can be received and content items generated by a plurality of users responsive to the search query can be identified. Ranking the identified user-generated content items based at least in part on the user's user credential score associated with each user-generated content item, and generating a set of search results based on the ranking of the content items. Ranking the identified users based on a weighted combination of user's user voucher scores associated with each user-generated content item based on a measure of relevance associated with each user-generated content item Content item. The objective measure of the quality of the interaction is derived from a combination of one of the following factors: a relevance of a user's content item to one of the other users' previous content items, and a content item relative to other content items. Originality, and coverage of one of the non-generic terms in the content project. One of the interactions may include the first user electronically responding to one of the electronic messages posted by a second user, and the weighting factor for the particular interaction may be related to: the first user The electronic response is related to the electronic information posted by the second user, one of the relatively non-universal information in the electronic response, or one of the electronic responses is relatively original.

Particular embodiments of the subject matter described in this specification can be implemented to realize one or more of the following advantages. The quality of posted content or other user-generated content can be assessed and used for the purpose of generating search results, rewarding users for high quality input, or restricting access based on low quality contributions. Interactions or relationships between users can be assessed to identify relatively authoritative or reliable contributors. The search engine can use the user credential score to generate search results. For example, it can be adjusted by user B or related to user B by using the strength of the relationship between user A and user B along with the score of user B (for example, by user B, or by user B, etc.) The content appears in a relative position in the results of the search performed by User A, thereby personalizing the search results. Using this personalization capability, for a different user C, the adjustment may vary depending on the strength of the relationship between User B and User C.

The details of one or more embodiments are set forth in the accompanying drawings and claims. Other features, objectives, and advantages of the subject matter will be apparent from the description and drawings.

Many of the media content that is available electronically today is user generated, that is, generated by people who are not employed by the website hosting the content. The term "media content" (or simply "content") may refer to a single electronic document, a collection of related electronic files that may include audio, visual, and/or text components, or a portion of an electronic document. In some instances, text, photos, or video may be added to a web portal by a logged-in user for sharing with other (logged in or not logged in) users. The quality of the media content supplied by the users and the relevance of the subject matter may vary.

Typically, websites hosting user-generated content provide a framework for interacting with each other with respect to user-generated content to generate additional user-generated content. For example, the first user can post a question to the forum and the second user can post a response to the question. This exchange of information between two users can be described as interaction. Other exemplary user activities that can contribute to one or more interactions include submitting questions or answers on the Q&A website, posting topics or discussions on the blog or online bulletin board, and submitting content available on the website. Rating, or view new content provided by the first user operating the first user device (eg, the second user views the new content on the second computing device). Systems and methods for ranking user-generated content can provide quality weighting of interactions between users in an electronic community, and generate user credential scores based on interactions between users. Each user and each user generated content item can be associated with a relative quality value based on the interaction between the users. For example, an authority score can be assigned to a user based on an analysis of the quality of the response to the user-generated content and a contribution score of the user posting the content, and can be based on the quality of the content generated by the user posted or uploaded. The authoritative score of the user who analyzes and responds to the content assigns a contribution score to the user. In addition, user interaction can be viewed as a trust vote, such that many interactions (particularly involving interactions with authoritative and/or contributing users) tend to increase user credential scoring. The interactions may be represented in a social or user activity map having a directed link (eg, a link between the user posting the question and the user answering the question) assigned to the user pair. Weights. These weights and corresponding interactions can be used to generate user credential scores.

The described techniques may provide one or more benefits, such as ranking user credibility to improve the entire community within the electronic environment. For example, users who abuse the special rights of the electronic environment (eg, spamming spam or harassing comments to others) can be automatically discovered, and users who frequently contribute to high quality interactions can be upgraded or rewarded. The techniques can also be used to improve search results including user-generated content by providing quality measures (eg, topic relevance, contribution, descriptive, or reputation of the content source) associated with user-generated content. Ranking. Pre-existing user relationships (eg, based on the number and/or quality of previous interactions) may also be used, for example, to score by correlating search results associated with users who have previously interacted with a particular user conducting a search. (For example, based on a combination of the user credentials of the other users and the strength of the relationship between the user being searched and the other users) to personalize the search results for that particular user. In some implementations, user credential scoring (and any ranking based on such scoring) can be different for different categories or tags associated with user generated content. For example, the user credential score can be calculated independently for content generated by the user associated with the gardening and content generated by the user associated with the meal. Thus, a particular user participating in a forum related to gardening and dining may have different user credential scores associated with each category.

1 is a block diagram of a system 100 for ranking search results for searching for content generated by a user. Within system 100, one or more users operating user device 104 can connect to network server 106 and search server 108 via network 102 to search for, upload, and retrieve electronic content. Network server 106 and search server 108 are each coupled to interactive processing server 110 either directly or via network 102. The interactive processing server 110 can cooperate with the web server 106 to analyze the electronic content generated by the user and derive a quality metric associated with the electronic content generated by the user. The search server 108 can additionally cooperate with the interactive processing server 110 to enforce ranking of search results for user-generated content based on the derived quality measurements.

Computer network 102 (e.g., a local area network (LAN), a wide area network (WAN), the Internet, or a combination thereof) connects network server 106, search server 108, and user devices 104a, 104b, and 104c. The example user device 104 includes a personal computer, a mobile communication device, a smart phone, a personal data assistant (PDA), and a television set box. Although only one search server 108 and one network server 106 are shown, system 100 can include any number of network servers, search servers, and user devices.

The web server 106 can be a universal content server that receives the request for content and retrieves the requested content in response to the request. In some examples, web server 106 can be associated with a new content provider, retailer, standalone blog, social networking website, or any other entity that provides and/or receives content via network 102. The web server 106 includes a repository 112 of user generated content and a user profile repository 126 (eg, each repository includes one or more electronic storage devices, the one or more electronic storage devices being included in the web server Within 106 or coupled to the network server 106). In some implementations, the user-generated content can include the content item itself (eg, uploading one or more files by a particular user), as well as interpretive material associated with the content generated by the individual user (eg, user identification, relevance) A reference to the content generated by the user, a category of content generated by the user, a date uploaded by the user to the web server, a count indicating the number of times the user content item has been requested/viewed, and the like.

When the user provides the user-generated content 112 to the web server 106, a user identifier (user ID) can be established, or the user can log in to the web server 106 using the user ID and password. Each user can provide additional profile information (eg, gender, age, location, email address, etc.). The data associated with each user can be stored in the user profile repository 126. As an example, each user record may include a reference to a user content item uploaded by the user to the web server 106 and stored in the repository 112 of the user generated content. User generated content stored in the repository 112 of user generated content may be provided by the web server 106 in response to a request from the user device 104 via the network 102.

The content available from the web server 106 can be searched using the search server 108. The search server 108 includes a search engine 114 and a search index repository 116. To facilitate the generation of search results in response to user queries, search engine 114 may generate or access an index (eg, an index of a web page) of content provided by an issuer such as web server 106 for later searching and identifying responses. The content item for the query. A user operating a user device (e.g., user device 104a) can submit a search query to the search server 108. The search engine 114 can search for web content in response to a search query within the search index repository 116. The search server 108 can then return the search results to the user in response to the search query. The search results, for example, may include web content references that link to web pages available via web server 106 (eg, a list of web page titles, segments of text extracted from such web pages, or hypertext links to their web pages). ).

The search results can be presented to the user in a manner that is grouped by a predetermined number (eg, ten) of search results. The search results are based in part on the scores associated with the content items identified by the search results (e.g., information capture ("IR") scores) and are optionally provided by the relevance score module 120. The file rank is ranked by the independent ranking of each file.

The search server 108 can communicate with the interactive processing server 110 to obtain data that can be applied to relevance scores for content items generated by the user. The interactive processing server 110 includes a content analysis module 122. The content analyzer 122 can analyze the content and interactions represented within the user-generated content (eg, user-generated content stored in the repository 112 of user-generated content) and generate correlations with individual content items (interactive materials). Sexuality and quality related scoring, and the authority or contribution of each user who provides user-generated content (user data). The interactive processing server 110 can store data associated with each user (e.g., stored in the voucher data repository 138) and associated with content and interactions represented by the content generated by the user (e.g., stored in interactive material) Within repository 124).

In operation, the interactive processing server 110 retrieves user generated content collected by the web server 106. Although web server 106 is shown as having a connection to interactive processing server 110, in some implementations, interactive processing server 110 may be able to directly access repository 112 of user generated content. In other implementations, the interactive processing server 110 stores the user generated content in a temporary storage location or interactive data repository 124. The interactive processing server 110 additionally receives or has access to a portion of the user profile record (eg, user identification and related upload content) that may be included in the credential repository 138.

The article analyzer 128 can analyze each user generated content item or article to generate a quality value. In some instances, within the venue of the online forum, a question is analyzed to determine its relevance to the forum topic, language appropriateness (eg, no swearing), and/or originality relative to previously posted questions. In some instances, the analysis responds to a question within the forum to determine its relevance to the question, the specificity of the response, the originality of the response relative to the previous posting, or the timestamp of the original posting relative to the question. Timeliness. The article analyzer 128 can assign a quality score to each content item.

The social mapping module 132 within the content analyzer generates an indication of different users based on an association between two users regarding a user-generated content item (eg, a posting by a user to a blog posted by another user). Interactive social or user activity map. The user activity map can be constructed from nodes representing individual users and directed and/or undirected links representing interactions between the two users. For example, the first user may upload questions about the best pizza available in Atlanta Georgia, and then the second user may upload a response to the pizza restaurant in Atlanta. Each user will be represented by one of the nodes in the user activity map, and the interaction will be represented by a directed link between the two nodes representing the posting of the second user's response to the first user's question. In some implementations, the user activity map also identifies relationships between users based on other types of connections between users. For example, user A may be a friend of user B, or user C may be interested in any content posted by user D. These types of relationships may be explicit (user A declares User B as their friend) or implicit (based on the history of many previous interactions, content analysis, or the like, it may be inferred that User A is a friend of User B) ). Interactions reflected in social or user activity diagrams may also be based on such relationships.

The quality weighting module 130 can assign weights to each link within the user activity map generated by the social mapping module 132. Such weights, for example, may represent the quality of one or both of the user content items involved in the interaction (eg, the quality of the question relative to the forum, or the quality of the answer relative to the question).

The voucher scoring module 134 can analyze the user activity map generated by the social mapping module 132 to assign a voucher score to a user associated with the user-generated content item. User credential scores can be used to indicate the credibility or credibility of a particular user. The user credential score may be based in part on the quality of the content item provided by the user (e.g., as determined by article parser 128) and the quality of the interaction in which the user has been involved. For example, when a user responds to a high quality question with a high quality answer, the interaction may positively affect the user's voucher score. Alternatively, if a user responds to a question with a low quality answer, the interaction may negatively impact the user's voucher score. If the user responds to a question posted by a user with a high voucher score, the interaction may result in a more positive score for the user's voucher than if the user responded to a question posted by a user with a low voucher score. The impact. In another example, if a high quality response from a user with a high voucher score is received by a user posting question, the interaction can have a positive impact on the voucher score of the user posting the question. In some implementations, the user credential score can be calculated, in part, by using a Hypertext Induced Top Selection (HITS) algorithm that considers weighted links.

Based on the user credential score, the user ranking module 136 can rank users who have contributed user generated content to the web server 106. In some implementations, user rankings can be used to upgrade or reward users who contribute a large number of high quality content items. For example, the highest ranked user gift certificate, prize, or other incentive item may be awarded. In another example, the previous user or users may be reported to the web server 106, wherein the indication of the user status (eg, "top 10", "top 100", "star contributor", etc.) may be embedded Within a web page that includes content generated by users created by their users. In other implementations, user rankings can be used to help identify poor quality users. For example, the lowest ranking user can be further evaluated as a potential spammer.

When the user submits the user-generated content to the web server 106 (eg, per project or batch submission), the user-generated content can be provided to the interactive processing server 110, at the interactive processing server 110, The content can be analyzed and scored by the article analyzer 128 based on the quality of the contribution. The user credential score may be continuously or periodically updated by the credential scoring module 134 based on the addition of the user activity map generated by the social mapping module 132 and the associated interactive weights calculated by the quality weighting module 130. The interactive data generated by the article analyzer 128, the quality weighting module 130, and the social mapping module 132 can be stored in the interactive data repository 124. The user credential score generated by the voucher scoring module 134 can be stored in the voucher data repository 138.

When the search server 108 receives a search query having a responsive web page result including user generated content, the results ranking module 118 can provide the interactive processing server 110 with each user generated content item included in the query result list. Identification. The interactive processing server 110 may retrieve user credential scores associated with the identified user generated content and provide the user credential scores to the search server 108. The results ranking module 118 can normalize the user credential scores and normalize the relevance scores calculated by the relevance scoring module 120 as appropriate. The results ranking module 118 can then combine the user credential scores and relevance scores to generate a ranked search results list. In some implementations, the information retrieval (IR) score can be calculated from the inner product of the feature vectors corresponding to the query and the content item, and the ranking of the search results can be based on the correlation of the combination of the IR score and the user credential score. Score. The search server 108 can then respond to the user who initiated the query for the ranked search results list. The user can select one of the results in the results list to retrieve user generated content or other content from the web server 106.

Although the web server 106, the interactive processing server 110, and the search server 108 are each represented as an individual machine, in some implementations, one or more of the servers 106, 108, or 110 are combined into a single unit. Inside the server. Similarly, in some implementations, user profile repository 126 and voucher repository 138 are combined in the same record that can reside within the same storage device. In some implementations, the article analyzer 128, the quality weighting module 130, the social mapping module 132, the voucher scoring module 134, or the user ranking module 136 are functionally combined within a single software application. Other implementations are possible.

2 is an example screen shot of a web page 200 of a question and answer electronic discussion board related to scuba travel. Users can access or log into the eDiscovery Edition to post questions related to dive trips or respond to questions posted by another user. Web page 200 includes a series of questions 202, each of which is owned by a particular user identifier 204. The last posted line 206 lists which user 204 has recently responded to the associated question 202. The total view line 208 lists the number of users who have reviewed the Q&A exchange associated with each question 202. The total response line 210 lists the number of responses provided to each question 202. Time line 212 provides a timestamp of the most recent response to each question 202. The search block 214 provides an opportunity for the user interacting with the web page 200 to search for an electronic discussion board. In some implementations, web page 200 depicts a particular forum for one of the larger electronic communities. For example, an electronic discussion website may be related to many diving topics (eg, underwater photography, equipment advice, safety reminders, etc.), with an independent forum dedicated to each topic.

For example, by selecting a question 202, the user may be presented with all of the text including the question (if it is truncated within the web page 200 due to length) along with another web page of a series of answers related to the question. In some implementations, the user may also select a user identifier (eg, user identifier techdiver 204a) to receive additional information about the user. In some examples, selecting the user identifier techdiver 204a may launch an additional web page displaying profile information about the user techdiver 204a (eg, location, gender, age, diving interest, date of login to the electronic discussion board, most recent A list of timestamps for accessing the eDiscovery, user's voucher score/rank, etc.) and/or each article (question and answer) contributed by the user techdiver 204a. In some implementations, viewing information about another user 204 by one user 204 can constitute an interaction that affects user credential scoring.

Each question and answer pair can be referred to as an interaction between users who create content generated by the user. For example, according to the total view line 208, the question 202b contributed by the user headed4mex 204b has been reviewed 46 times, with a total of 11 responses according to the total response line 210. Each of the 11 responses (and in some cases, each of the 46 views) establishes an interaction between the author of the response and the user headed4mex 204b. Users can be involved in more than one interaction. For example, user maddog 204e submits a fifth question 202e and an eighth question 202h. In another example, user cheapdives 204c contributes a third question 202c and contributes to the most recent response to the tenth question 202j.

3A and 3B are diagrams of an example user activity diagram 300 depicting an exemplary interaction involving logging into a user 204 of an electronic discussion board as described with respect to FIG. 2. For example, user activity map 300 may be generated by social mapping module 132 as described with reference to FIG. User activity diagrams (also known as social graphs) represent interactions (connections) between entities and entities. In this example, the user is represented as a node in the graph and the interaction is represented as the line connecting the nodes. Each of the nodes and connections may be stored as an object or otherwise defined in a data structure stored on a computer readable storage device. For example, the interaction may involve communication between two individual users, for example, a response by the user lobstahdive 204n to the question submitted by the user yellowbcd 204m. A pair of users can be involved in multiple interactions.

As shown in FIG. 3A, user 204 is connected by a line representing a question-and-answer interaction that models the interactions between users involved in question 202 and provided by users listed in last posted line 206 (shown in FIG. 2). Response to the question. For example, question 202a posted by user techdiver 204a is answered by user dive4fun 204j. The edge 302 within the user activity map 300 represents this interaction. In another example, the user lobstahdive 204n and the user yellowbcd 204m are represented by three individual interactions (eg, the question posted by the user lobstahdive 204n is answered by the user yellowbcd 204m, or the question posted by the user yellowbcd 204m is lobstahdive 204n by the user Respond to) the three sides of the 304 connection. In some implementations, the equilateral edges are directional. For example, edge 302 can be directed from user dive4fun 204j to user techdiver 204a, which is related to a response to question 202a.

Not all users are interconnected by interactions within the user activity diagram. For example, users beachdweller 204g, underh2o 204p, headed4mex 204b, and diverdan 204k are separated from the rest of the group of users 204. Within a user activity map, any number of interactions can be represented between users. In some implementations, the edges within a user activity map can be weighted. For example, a quality score can be associated with each interaction between a pair of users 204. In some implementations, the quality weighting module 130 as described with respect to FIG. 1 can be weighted in part based on quality scores associated with user-generated content contributed by one or both of the users involved in the interaction. (eg, 0.3 or 0.5) is applied to the side of the user activity map 300.

When a user posts a new question to the electronic discussion board, there are no other users involved in the interaction. In some implementations, as shown in FIG. 3B, a default user forum_travel 352 is introduced to establish an interaction between the user who posted the new question and the forum. For example, the quality weighting associated with the interaction between the preset user forum_travel 352 and the user beachdweller 204g may be based in part on the relevance of the issue posted by the user beachdweller 204g relative to the subject of the electronic discussion board (eg, diving tour). Sex. For example, a low quality rating may be awarded to a question posted by the user beachdweller 204g regarding a low-priced mail-order prescription drug, and to the underground dark river posted near the Playa del Carmen (by Playa de Carmen) posted by the user headed4mex ( The issue of cenote) is awarded a high quality rating.

4 is a flow diagram of a procedure 400 for generating user credential scores based in part on interactions involving each user. In some implementations, program 400 is executed by content analyzer 122 as described in FIG.

At 402, a quality measure associated with the content of the user interaction is calculated. In some instances, user interaction includes uploading new user-generated content by submitting a question or answer on a Q&A website, or by posting a topic or discussion on a blog or online bulletin board. Each user generated content item can be evaluated to determine a quality score. This quality score can be based on a number of factors, such as the relevance of the content item, the originality of the content item, or the degree of focus of the content item. In some implementations, the quality score of a user-generated content item includes a collection of individual factor scores.

The quality of the user-generated content item can be evaluated based in part on the context in which the content item is introduced. In some instances, a correlation factor is used to assess the relevance of a user-generated content item relative to a website, bulletin board, or topic forum, or to a question responsive to a content item. In some implementations, a correlation factor is calculated using Latent Dirichlet Allocation (LDA).

Another exemplary evaluation factor is coverage, which refers to the versatility or specificity of the vocabulary used by the posted content item within the content item. Coverage can be used as an indication of the degree of focus of user contributions. In some implementations, the coverage factor is measured using an inverse document frequency (IDF) of words within a user generated content item.

User content items can additionally be evaluated based on their originality. For example, a question posted to a Q&A website can be compared to other questions posted to the website, or an answer posted to one of the Q&A websites can be compared to other answers to the same or similar questions. In some implementations, the originality of a content item can be based in part on comparisons with other content items contributed by the same user. For example, a user who posts the same message (eg, spam) across multiple topics of the Q&A website can be discovered by examining the originality of all content provided by the user. In some implementations, the original score is measured via, for example, a Bilingual Evaluation Understudy (BLEU) scoring method. In some implementations, the article analyzer 128 as depicted in FIG. 1 applies a quality score to a user-generated content item.

At 404, links between users are identified based on interactions between users. A link between these individual users can be established using an interaction between two users, such as a "reply" activity. A common type of interaction may involve the first user uploading new user generated content and the second user responding in a certain manner (eg, viewing, ranking, rating, or uploading a new user generated content response) to the user. content. In another example, when a user uploads a new question to a Q&A website or posts a discussion on the online bulletin board, the interaction is considered to be between the user and the website or bulletin board. In some examples, the link connects two users based on the following: a response posted by the second user in response to the question posted by the first user, and a rating posted by the second user referenced by the content provided by the first user, Or a comment posted by the second user in response to the blog entry posted by the first user. In some implementations, a virtual user (eg, a node) represents one end of an interaction. For example, if the first user posts a new question to the bulletin board, the bulletin board can be represented by a virtual user. For example, in this way, if a user posts an irrelevant topic to a discussion board, and no user responds to an irrelevant article, the program 400 has a content for applying a quality score to the user and adjusts the content accordingly. The method of scoring a user's voucher.

At 406, the interaction is assigned a weighting factor that represents the quality of the interaction. Quality scores relating to the aforementioned correlations, coverage and originality factors and/or other factors (eg, timeliness of contributions, including multimedia, including rich media, etc.) may be individually weighted and combined to produce content items for use in content items. Single total quality score. In some examples, the quality score is divided into a numerical rating (eg, a value between 1 and 10 or 1 and 100, etc.), or a category rating (eg, positive, neutral, or negative). In some implementations, quality scores associated with individual quality factors can be weighted based on learned coefficients and combined to produce a quality weight. For example, the learned coefficients can be generated based on the relative importance of each quality factor relative to the electronic community and applied to individual quality factors. These factors may vary depending on the type of e-community. For example, in some implementations, the timeliness of responses to user articles is more important than the coverage of user articles. In some implementations, the quality weighting module 130 as depicted in FIG. 1 calculates a weighting factor associated with the interactive link.

At 408, a user credential score is calculated. For example, user credential scoring attempts to quantify the value of a contribution made by a particular user to an e-community. In summary, a user can contribute new content or provide feedback about content contributed by another user. Each character employed by the user can be independently considered when generating a user's voucher score. First, the user can be assigned a contribution score based on the quality of the contribution that the user has made to the e-community (eg, website, discussion board, forum, etc.). For example, a contribution score indicates how many questions a particular user has asked and how much other authoritative users in the e-community have generated interest in the questions. Second, the user can be assigned an authority score based on the quality of the response or rating of a user's content that has been generated by other users contributing to the electronic community. For example, the authority score can represent a measure of the number of useful responses a user has provided to the e-community. Depending on how the user interacts with the e-community, the user may have a high contribution score and a low authority score, or have a low contribution score and a high authority score. In some implementations, the scores can be calculated based on a time decay such that the older interaction has less impact on weighting and/or scoring than the most recent interaction.

In some implementations, a modified version of the Hyperlink-Induced Topic Search (HITS) algorithm (as described in conjunction with Figure 5 below) is used to generate a contribution score and an authority for each user. Minute. For example, a particular user's contribution score can be related to the interaction with the user by scaling and weighting (eg, by quality scores of content items generated by users involved in the interaction) User authority scores are generated. Similarly, a particular user's authority score can be related to other users in the interaction with the user by scaling and weighting (eg, by quality scores of user-generated content items involved in the interaction). The contribution score is generated. In this manner, a positive (high quality) response to a content item generated by the first user posted by an authoritative second user may increase the contribution score of the first user. Alternatively, when calculating the contribution score of the first user, the negative (low quality) response posted by the second user by a non-trust (spammer) may be reduced. Once the contribution score and authority score have been assigned, a total user reputation score can be calculated based on the combination of the two user credential scores (eg, the average or some other linear or non-linear combination).

FIG. 5 is a flow diagram of a routine 500 for generating user credential scores (eg, contribution scores and authority scores). For example, the routine 500 can be executed using a modified version of a ranking algorithm based on a HITS link, where the user's authority score is used instead of the content as a conventional HITS estimate of the value of an authoritative source, and using the user contribution Sub-replacement estimates the user's known HITS center score as a value of authoritative reference. In addition, the modified HITS algorithm may also use weighted links when generating user credential scores. In some implementations, content analyzer 122 (as described with respect to FIG. 1) can execute program 500.

At 502, a weighted user activity map is constructed. Based on the interactive links derived within the program 400 (as described with respect to FIG. 4), an affinity matrix defining the weighted interactions between the users can be constructed, wherein each directed link between the two users is related to The quality factor associated with the content item generated by the user in the interaction is weighted. For example, the matrix elements A(u _i , u _j ) may be populated by a combination of interaction quality weighting factors (eg, sum, weighted sum, etc.) between the first user u _i and the second user u _j . In some implementations, a user activity map is generated by a social mapping module 132 (as shown in FIG. 1).

At 504, the contribution score and authority score associated with each user are initialized. A base value initialization system for contribution scoring and authoritative scoring so that relative contribution scoring and authority scoring can be obtained based on the neutral starting point. A random initial value can be assigned. In some implementations, the contribution score and the authority score are each represented by a vector equation. For example, the contribution score can be viewed as a representation of a user activity map having a link to the first orientation (eg, from the responder to the questioner), and the authority score can be considered to have A representation of the user activity map of the link to the reverse direction (eg, from the sender to the responder). For example, a random initial value can be obtained via random sampling of content item quality scores. In some implementations, the initial value is equal to, approximated, or otherwise based on a value generated in a previous execution of a HITS-based algorithm (eg, updating a user credential value after an additional interaction occurs).

At 506, a convergence threshold is defined. The convergence threshold can be used to define the stopping point for the authority score and contribution score for each user.

At 508, the convergence threshold is used as a criterion to update the contribution score and the authority score based on the weighted user activity map. For example, vector scoring can be calculated using an algorithm involving first normalizing each column of the user activity graph matrix and then normalizing each row of the user activity graph matrix.

The algorithm is repeated if it is found at 510 that the score change (between the randomly initialized value and the calculated value) is greater than or equal to the convergence threshold. At the end of each iteration, the change in the score between the previously calculated value and the current calculated value is compared to the convergence threshold. In some implementations, the initialization and calculation of user scores is performed by the voucher scoring module 134. In an example of a system in which a large number of user-generated content items exist, user credential scores are repeatedly calculated in parallel across multiple computer processors.

Once the score change is less than the convergence threshold, the user credential score is assigned at 512. For example, the contribution scores and authority scores associated with each user may be stored in a voucher repository 138 (as shown in FIG. 1) associated with the corresponding user identifier. Contribution scores and authority scores can be combined as appropriate to generate a total user reputation score. Based on user credential scores (contribution, authorization, and/or reputation), such users may be ranked as appropriate (eg, by user ranking module 136 of FIG. 1).

The program 500 can be repeated whenever needed to update the user credential score based on new user-generated content items added to the e-community. For example, the process 500 can be repeated daily or weekly to generate updated user credential scores and, as appropriate, update the rankings of individual users within the electronic community. In some implementations, program 500 can be executed incrementally, for example, by incrementally adjusting user credential scores after each new interaction. For example, the user's voucher score (authority and contribution) for users in the vicinity of the interaction in the user activity map can be adjusted using the quality of the new interaction. In some implementations, a propagation distance parameter can be defined to limit the propagation of the scoring adjustments in the overall user activity map. For example, a propagation distance parameter value of two may limit the incremental recalculation of user credential scores for nodes within the two links corresponding to the link of the new interaction. Therefore, in this example, if the user does not have an interaction history with the user involved in the new interaction or with the user involved in the new interaction, the node representing the user will not be influences. Thus, for each new interaction, the social activity map can be updated by the new link and can also update the user's user voucher score within the travel distance on the social activity map. In some implementations, instead of limiting the propagation distance, or in addition to limiting the propagation distance, the number of affected nodes may be limited. For example, the node to be updated may be selected based on the proximity to the new link and the number of interactions that link the node directly or indirectly to the node connected by the new link. Incremental calculations of user credential scores can be used to provide substantially instant user ranking or access to user credential scores. Even in the case of incremental scoring updates, the calculation of the scoring for the total user activity map can be periodically calculated for all nodes using the above-described repeated methods to ensure that incremental updates do not cause global accuracy. The score is divergent.

6 is a flow diagram of a procedure 600 for generating a ranking of search results using user credential scores. For example, the program 600 can combine the activity of the search server 108 (shown in FIG. 1) with the activity of the interactive processing server 110 to enhance the ranking of traditional search results by product quality measurements of the content generated by the user. At 602, a quality measure associated with a user-generated content item involved in the interaction is calculated. In some instances, these quality measures are based on a number of factors including, but not limited to, relevance, coverage, originality, timeliness of contributions (eg, by a quick response that is much greater than a slow response value, or And vice versa), multimedia content, or rich media content.

At 604, links between users are identified based on interactions between users. A directed link may correspond to a user's reaction to a content item (eg, a question, a blog, a multimedia content item, etc.) generated by a user provided by another user (eg, posting a response, ranking, viewing, etc.). In some implementations, such interactions occur within an electronic community such as a Q&A website, a forum, or an online bulletin board.

At 606, a user credential score is calculated. User voucher scoring can include more than one individual scoring, such as contribution scoring and authority scoring. User credential scores may include composite scores such as credit scores. For example, the user credential score can be calculated using the procedure 500 as described in connection with FIG. In some implementations, the user credential score is calculated by the voucher scoring module 134 as shown in FIG.

At 608, the user credentials are normalized. In order to rank the search results based in part on the voucher scores of the users who supply the content items, the user voucher scores are first normalized to provide a basis for comparison between the search result scores and the user credential scores. In some implementations, the user credential score can be transformed into a standard Gaussian distribution. In some implementations, the user credential score is normalized by the ranking module 118 as shown in FIG.

At 612, a search query is received at the same time or at different times when steps 602 through 608 are performed. The search query may be in response to one or more user generated content items. In some instances, the search query is submitted to an electronic community containing user generated content items or submitted to a general search engine. In some implementations, the search query is received by a search engine 114 as shown in FIG.

At 614, a content entry generated by a user responsive to the query is identified. For example, the responsive content items can be located based on a keyword match between the search query and the text of the content item.

At 616, the relevance score is applied to the identified content item as appropriate based on the first ranking method. For example, search engine 114 can apply an information capture (IR) score to a user-generated content item to determine a first ranking of items that are responsive to the search query. In one example, the Okapi BM25 ranking function can be used to determine the relevance score of each user-generated content item relative to the search query.

If the relevance score has been applied to the identified content item, the correlation score is normalized at 618. For example, the normalization technique can be selected such that the normalized user credential score from step 608 can be combined with the normalized correlation score. In some implementations, the correlation score can be transformed into a standard Gaussian distribution. In some implementations, the user credential score is normalized by the ranking module 118 as shown in FIG.

At 610, the normalized user credential score from step 608 and the normalized correlation score from step 618 are used, based on the user credential score associated with the identified content item plus The normalized correlation score is chosen to rank the entries. In some implementations, the type of user credential score used is based on the type of content item generated by the user referenced by the query search results. For example, if the content item is a response to a question, the user authority score can be normalized and combined with the relevance score for ranking. On the other hand, if the content item is a problem, the user contribution score can be normalized and combined with the relevance score for ranking. The weighting can be applied to one of the scores when combining the normalized user credential score with the normalized correlation score. For example, a weighting factor between zero and one may be applied to the normalized user credential score to indicate the relative importance of the user's reputation as a user-generated content item within the ranking query search results. These query search results can now be sorted by ranking and returned to the requester.

Example: Integration of Quality Factor Scoring

Given a set of scores associated with individual quality factors (eg, coverage, originality, relevance, or timeliness), the following illustrative equations can be applied to generate a single quality indicator. This quality indicator can be used as a weighting within the user activity graph that illustrates the interaction between users.

Applicable coefficient vector The linear combination produces a combined score com ( q _i , r _ij ) involving the response r to the problem q :

Com ( q _i , r _ij )= α ₀ + α ₁ ‧ rel ( q _i , r _ij )+ α ₂ ‧cov( r _ij )+ α ₃ ‧ ori ( r _ij )

Where rel is the correlation factor, cov is the coverage factor, and ori is the originality factor.

Using combined scoring, a quality score qua ( q _i , r _ij ) can be generated by introducing a combined score into the following equation (eg, involving the interaction weighting of question q and response r ):

In some implementations, any standard acquisition coefficient algorithm can be used to learn the alpha coefficients.

Example: Calculating User Credential Scores

Based on the user activity graph generated as the affinity matrix A (where each element A ( u _i , u _j ) contains the sum of the weights from the user u _i to the user u _j ), the contribution scores And authority score Can be used and The random initial values of the two start and are calculated repeatedly.

Where i is an all-one vector, A _row is the same as A, where A _row is normalized to a total of 1, A _col is normalized to a total of 1 A , and ε is used to ensure the convergence of the algorithm. Probability. The algorithm repeats until the following convergence conditions are met:

Where l is a predefined threshold.

The embodiments of the subject matter and operation described in this specification can be implemented in a digital electronic circuit, or in a computer software, firmware or hardware (including the structures disclosed in the specification and their structural equivalents), or one of them. Or a combination of many. The embodiments of the subject matter described in this specification can be implemented as one or more computer programs (ie, computer program instructions) for use on a computer storage medium for execution by a data processing device or for controlling the operation of a data processing device. One or more modules). Alternatively or additionally, the program instructions may be encoded on a manually generated propagating signal (eg, a machine-generated electrical, optical, or electromagnetic signal) that is encoded for transmission to a suitable receiver device for use. Information that is executed by the data processing device. The computer storage medium can be, or be included in, a computer readable storage device, a computer readable storage substrate, a random or serial access memory array or device, or a combination of one or more thereof. In addition, although the computer storage medium is not intended to propagate signals, the computer storage medium may be the source or destination of computer program instructions encoded in the artificially generated propagation signals. The computer storage medium can also be included in or included in one or more separate physical components or media (eg, multiple CDs, discs, or other storage devices).

The operations described in this specification can be implemented as operations performed by a data storage device on data stored on one or more computer readable storage devices or received from other sources.

The term "data processing apparatus" encompasses all types of devices, devices and machines for processing data, including (as an example) a programmable processor, a computer, a system on a wafer, or a plurality of the foregoing, or each of the foregoing a combination of people. The apparatus can include dedicated logic circuitry, such as a field programmable gate array (FPGA) or a special application integrated circuit (ASIC). The device may include, in addition to the hard body, code that generates an execution environment for the computer program in question, for example, a processor firmware, a protocol stack, a database management system, an operating system, a cross-platform execution environment, a virtual machine, The code of a combination of one or more of them. The device and execution environment enable a variety of different computing model infrastructures, such as web services, decentralized computing, and raster computing infrastructure.

Computer programs (also known as programs, software, software applications, scripts or code) can be written in any form of programming language (including compiled or interpreted languages, announcements or procedural languages) and can be in any form. Deployment, including deployment as a stand-alone program or a module, component, subroutine, object, or other unit suitable for use in a computing environment. A computer program can (but need not) correspond to a file in a file system. The program can be stored in a portion of a file that holds other programs or materials (for example, one or more scripts stored in a markup language file), a single file dedicated to the program under discussion, or multiple collaborative files. (for example, a file that stores one or more modules, subroutines, or parts of a code). A computer program can be deployed to execute on one computer or on multiple computers located at a single point or on multiple computers that are spread across multiple sites and interconnected by a communication network.

The procedures and logic flows described in this specification can be performed by executing one or more computer programs or a plurality of programmable processors to perform actions by operating on input data and generating output. The programs and logic flows may also be performed by dedicated logic circuits such as field programmable gate arrays (FPGAs) or special application integrated circuits (ASICs), and the devices may also be implemented as, for example, field programmable gate arrays (FPGAs). ) or a dedicated logic circuit for a special application integrated circuit (ASIC).

Processors suitable for the execution of a computer program include, by way of example, both a general purpose microprocessor and a special purpose microprocessor, and any one or more processors of any kind of digital computer. Generally, the processor will receive instructions and data from a read-only memory or a random access memory or both. The main components of the computer are used to execute the processor according to the actions of the instructions and one or more memory devices for storing the instructions and data. In general, the computer will also include one or more mass storage devices (eg, magnetic disks, magneto-optical disks, or optical disks) for storing data, or operatively coupled to receive from one or more mass storage devices. Data or transmission of data to one or more mass storage devices, or both reception and transmission. However, computers do not need to have such devices. In addition, the computer can be embedded in another device, such as a mobile phone, a personal digital assistant (PDA), a mobile audio or video player, a gaming machine, a global positioning system (GPS) receiver, or a portable storage device (eg, Universal Serial Bus (USB) flash drive (to name a few). Devices suitable for storing computer program instructions and data include all forms of non-volatile memory, media and memory devices, including (as an example): semiconductor memory devices such as EPROM, EEPROM and flash memory devices; Discs, for example, internal hard drives or removable discs; magneto-optical discs; and CD-ROM and DVD-ROM discs. The processor and memory can be supplemented by dedicated logic circuitry or incorporated in dedicated logic circuitry.

In order to provide interaction with a user, embodiments of the subject matter described in this specification can be implemented on a computer having a display device (such as a cathode ray tube (CRT) or liquid crystal display) for displaying information to a user. (LCD) monitor), and a keyboard and indicator device (such as a mouse or trackball) that the user can provide input to the computer. Other types of devices may be used to also provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback, such as visual feedback, audible feedback, or tactile feedback; and may receive the user from any form. Inputs, including acoustic, voice, or tactile inputs. In addition, the computer can interact with the user by sending a file to the device used by the user and the device used by the free user; for example, by sending a web page to the user's client in response to a request received from a web browser. A web browser on the device.

Embodiments of the subject matter described in this specification can be implemented in a computing system that includes a backend component (eg, a data server), an intermediate software component (eg, an application server), or a front end component (eg, A client computer with a graphical user interface or web browser that interacts with the implementation of the subject matter described in this specification via a graphical user interface or web browser, or such backend, intermediate software or front end components Any combination of one or more of them. The components of the system can be interconnected by any form or medium of digital data communication (eg, a communication network). Examples of communication networks include regional networks ("LAN") and wide area networks ("WAN"), inter-networks (eg, the Internet), and peer networks (eg, special-purpose peer networks) .

The computing system can include a client and a server. The client and server are typically remote from each other and typically interact via a communication network. The relationship between the client and the server is generated based on a plurality of computer programs executing on respective computers and having a client-server relationship with each other. In some embodiments, the server transmits the material (eg, an HTML page) to the client device (eg, to display data to and receive user input from the user interacting with the client device). The data generated at the client device (eg, the result of user interaction) can be received at the server from the client device.

While the specification has been described with a particular embodiment of the invention, the specific details of the invention are not to be construed as limiting the scope of the invention. Particular features described in the context of a plurality of separate embodiments may also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can be implemented in various embodiments or in any suitable sub-combination. Moreover, although features may be described above as acting in a particular combination and even initially claimed, in some cases one or more features from the claimed combination may be deleted from the claimed combination, and the claimed combination Changes may be made to sub-combinations or sub-combinations.

Similarly, although the operations are depicted in a particular order in the drawings, this should not be construed as being required . In certain situations, multitasking and parallel processing may be advantageous. In addition, the separation of the various system components in the above-described embodiments should not be construed as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated or integrated in a single software product. Packaged into multiple software products.

Thus, specific embodiments of the subject matter have been described. Other embodiments are within the scope of the following patent claims. In some cases, the actions recited in the scope of the claims can be performed in a different order and still achieve the desired results. In addition, the procedures depicted in the drawings are not necessarily in a particular order or order of order to achieve the desired results. In certain implementations, multitasking and parallel processing may be advantageous.

100. . . System for ranking search results for searching for content generated by users

102. . . network

104a. . . User device

106. . . Web server

108. . . Search server

110. . . Interactive processing server

112. . . Repository/user-generated content of user-generated content

114. . . Search engine

116. . . Search index repository

118. . . Result ranking module

120. . . Correlation score module

122. . . Content Analysis Module / Content Analyzer

124. . . Interactive data repository

126. . . User profile repository

128. . . Article parser

130. . . Quality weighting module

132. . . Social drawing module

134. . . Voucher scoring module

136. . . User ranking module

138. . . Voucher data repository

200. . . Web page of the Q&A e-discussion related to diving tourism

202a. . . problem

202b. . . problem

202c. . . problem

202d. . . problem

202e. . . problem

202f. . . problem

202g. . . problem

202h. . . problem

202i. . . problem

202j. . . problem

204a. . . User identifier

204b. . . user

204c. . . user

204d. . . user

204e. . . user

204f. . . user

204g. . . user

204h. . . user

204i. . . user

204j. . . user

204k. . . user

204l. . . user

204m. . . user

204n. . . user

204o. . . user

204p. . . user

206. . . Last posted

208. . . Total inspection line

210. . . Total response line

212. . . Time line

214. . . Search block

300. . . User activity diagram

302. . . side

304. . . side

352. . . Default user

1 is a block diagram of a system for ranking search results for searching for content generated by a user.

Figure 2 is a screenshot of an example of a web page for questions in the Q&A e-Discussion.

3A and 3B are diagrams of an example user activity diagram.

4 is a flow chart of a procedure for generating a user credential score.

Figure 5 is a flow diagram of another procedure for generating a user credential score.

Figure 6 is a flow diagram showing the ranking of search results using user credential scores.

The same reference symbols are used in the drawings to the

100. . . System for ranking search results for searching for content generated by users

102. . . network

104a. . . User device

106. . . Web server

108. . . Search server

110. . . Interactive processing server

112. . . Repository/user-generated content of user-generated content

114. . . Search engine

116. . . Search index repository

118. . . Result ranking module

120. . . Correlation score module

122. . . Content Analysis Module / Content Analyzer

124. . . Interactive data repository

126. . . User profile repository

128. . . Article parser

130. . . Quality weighting module

132. . . Social drawing module

134. . . Voucher scoring module

136. . . User ranking module

138. . . Voucher data repository

Claims (21)

A computer implemented method for analyzing quality of content generated by a user, comprising: identifying, by a computer operation, a plurality of interactions between a plurality of users including content items generated by a plurality of users via an electronic network, each The interaction is between a pair of users; assigning a weighting factor to each interaction, wherein the weighting factor represents one of the qualities of the interaction; and is generated by a computer operation for each of the plurality of users a user credential score, wherein the user credential scores are based on the weighting factors of each of the plurality of interactions; storing the user credential scores associated with a user identifier on a computer Reading the storage device; assigning one of the content relevance measurements for each of the plurality of user-generated content items; and measuring the score based on the user credentials and the assigned content relevance by the following steps To rank the plurality of user-generated content items: normalizing the user credential scores associated with the plurality of user-generated content items and the content relevance Measurement; and the user credential such scores have been normalized to the sum of the normalized correlation for the content of the measured quantity of these projects are combined. The method of claim 1, further comprising: receiving a search query; and The search query is responded to by the operation of a processor to identify the plurality of user generated content items. The method of claim 2, wherein the measure of the content relevance that has been assigned for each item is assigned to the search query based on the relevance of each item. The method of claim 1, further comprising: generating a user activity map by operation of the system, the user activity map identifying a link between the users based on the plurality of interactions between the users; determining by operation of the system One authority of each user is scored, wherein the authority score of the user is based on the contribution score of the user to which the user is linked in the user activity map; and each operation is determined by the operation of the system One of the user's contribution scores, wherein the contribution score is based on the authority score of the user to which the user is linked in the user activity map. The method of claim 4, wherein the authority scores and the contribution scores are generated by repeating a repeated update procedure until the repeated update procedure reaches a predetermined convergence threshold. The method of claim 1, wherein the plurality of user interactions comprise user generated content on at least one of a question and answer website, a bulletin board website, a blog, or a social networking website. The method of claim 1, wherein the weighting factor is derived by combining a plurality of quality factors, at least one of the quality factors being selected from the group consisting of: one user content item and another user One associated with a previous content item, a content item relative to it One of his content items is originality, a content item corresponds to a coverage of one of the non-generic terms in the content item, a richness of the content item, and a timeliness of a content item. The method of claim 1, further comprising rewarding the user based on the user credential scores. The method of claim 1, wherein the plurality of interactions comprise interactions between users in an electronic community. A system for ranking content generated by a user or a user, comprising: at least one server adapted to receive and publish content generated by a user for accessing across a network; at least one storage device storing the user And the at least one processor configured to: identify an interaction between user pairs associated with the stored user-generated content; objectively measure based on one of the qualities of each interaction to generate the interaction a weighting factor; generating a user credential score for each user based on the identified interactions and the weighting factors of the interactions; and scoring based on the user credential and content items generated for the plurality of users One of each of the content relevance measures to rank at least one of the user or user generated content, wherein the ranking is generated by the following steps: normalizing the content items associated with the plurality of user generated content items User voucher scoring and measurement of the relevance of the content; and scoring the normalized user credential with the normalized content The correlation measures are combined. The system of claim 10, wherein the at least one processor is configured to generate each of the authority scores and an associated contribution score by repeatedly updating based on the identified interactions and the weighting factors One of the user's user credentials is scored. The system of claim 11, wherein the at least one processor is further configured to: receive a search query; identify a plurality of user-generated content items responsive to the search query; and generate a set based on the ranking of the content items Search results. The system of claim 12, wherein the at least one processor is configured to measure a correlation associated with a content item generated by each user and a user associated with each user generated content item A weighted combination of the user credential scores ranks the content items generated by the identified users. The system of claim 10, wherein the objective measure of the quality of the interaction is derived from a combination of one of the following factors: a correlation of a content item of one user with a previous content item associated with one of the other users Sexuality, the originality of a content item relative to one of the other content items, and one of the non-generic terms in the content item. The system of claim 10, wherein one of the interactions comprises a first user's electronic response to electronic information posted by a second user, wherein the weighting factor for the particular interaction is Among At least one of: one of the first user's electronic response to the electronic information posted by the second user, one of the relative non-universal information in the electronic response, or one of the electronic responses Relatively original. An electronic system for analyzing the quality of content, the system comprising: at least one server adapted to receive and publish content for access across a network; at least one storage device that stores user generated content; And means for generating an objective measure of the quality of each of the plurality of interactions to generate a weighting factor for the interaction, wherein each interaction occurs with a pair of users associated with the stored user generated content Means for generating a score based on the user's voucher for each user based on the weighting factors of the interactions; and for correlating content based on the user credential scores and content items generated for the plurality of users A measure of the content of the user-generated content, wherein the ranking is generated by normalizing the user voucher scores associated with the plurality of user-generated content items and the amount of relevance of the content. And combining the normalized user account scores with the normalized measurements of the content of the content for the items. The system of claim 16, further comprising means for generating a search result for searching for one of the user-generated content based on a combination of the scores of the user voucher scores and the items for the user-generated content . An article comprising a computer readable storage medium storing instructions operable to cause one or more processors to: identify a user comprising a plurality of users via an electronic network a plurality of interactions of content items, each interaction being between a pair of users; assigning a weighting factor to each interaction, wherein the weighting factor represents one of the qualities of the interaction; generating a user for each of the plurality of users Voucher scoring, wherein the user voucher scoring is based on the weighting factor of each of the plurality of interactions; storing the user voucher scores on a computer readable storage device; generating for a plurality of users a content item, assigning one of the content relevance measurements; and ranking the plurality of user-generated content items based on the measurement of the user credential scores and the assigned content relevance by the following steps: normalization and The user account scores associated with the plurality of user-generated content items and the measurement of the content relevance; and the user credentials that will be normalized The correlation between the amount of already formalized for the content of these projects be combined measure. The article of claim 18, wherein the computer readable storage medium further stores instructions operable to cause the one or more processors to perform the following additional actions: receiving a search query; Identifying the plurality of user-generated content items in response to the search query. The item of claim 19, wherein the measure of the content relevance that has been assigned for each item is assigned to the search query based on the relevance of each item. The article of claim 18, wherein the computer readable storage medium further stores instructions operable to cause the one or more processors to perform an additional action of: generating a link between the users based on the plurality of interactions between the users a user activity map; determining one authority rating of each user, wherein the authority score of the user is based on the contribution score of the user to which the user is linked in the user activity map; and determining each One of the user's contribution scores, wherein the contribution score is based on the authority score of the user to which the user is linked in the user activity map.