Abstract: Human interaction with a webpage may be determined by processing an event stream generated by the client device during the webpage interaction. A classification server receives the event stream and compares components of the event stream, including components of an event header message, with prerecorded datasets. The datasets include prerecorded event streams having a known interaction type. Training clients may be provided for generating the prerecorded datasets.

Abstract: A system and method to detect and prevent non-human interaction between a client and a web server invokes an effect to change the event generation behavior at the client. Subsequent event streams from the client to the server are analyzed to determine whether the event streams contain events corresponding to expected reactions of a human operator at the client to the effect. Indications of non-human behavior may invoke more direct human testing, for example using a dynamic CAPTCHA application, or may cause a termination of the client/URL interaction.

Abstract: When a user clicks a widget, such as a button, to invoke a widget behavior in a web page, an overlay module may delay the invoking of the widget behavior while executing a stimulatory effect. The stimulatory effect may be a visual effect, such as a gradual fill of the button, that stimulates the user to provide a re-click. A re-click can cause the stimulatory effect to conclude and to invoke the widget behavior. The re-click event, based on human impatience, may be determined to be an indicator of a human user.

Abstract: During an interaction with a web page, user interface events are recorded and augmented with page layout data from the document object model. An event stream is formed with the page layout data and communicated to an event server. The event server processes the event stream to determine a location at which the events were generated, which can be used to predict an area of a user's attention.

Abstract: A web page that includes content form fields may be modified to include an event observer module and an authored content module. The authored content module adds a hidden “events observed” field to the form fields. Events generated during the authoring of content by a user are recorded by the event observer module. When the content is submitted from a client browser to the web server, the events generated during the authoring of the content are added to the events observed field and submitted with the content. The web server uses the events to determine a DOM of the web page and compare the observed DOM with a stored DOM for that web page and that particular interaction. The page structure may be optionally modified by the web server to enhance the analysis of the DOM comparison. The web server analysis facilitates detection of non-human content submission at a client browser.

Abstract: Unobserved user interface events may be detected based on upon previously recorded data streams of events. The prior data streams are obtained by recording events from users who have browsers that support both event capturing and bubbling. When users with browsers that support only event bubbling interact with the page, the data stream is augmented by inferring unobserved events based on similarity to other event stream records.

Abstract: In a web crawler, a URL selection module selects URLs for pages to be downloaded. The URL selection module accesses an interaction data store that stores interaction data for web pages, including interaction data that indicates human interactions with the pages. To reduce the effects of link farms, the URL selection module filters the URLs to select only those URLs that have human interaction histories and provides the selected URLs to a download module for web page downloading.

Abstract: An indexing system for documents such as web pages divides a document into elements, such as document object model elements. User attention data from prior interactions with the document are analyzed to determine those elements of a document that satisfy a threshold requirement of user attention. Elements meeting the user attention threshold requirement are added to a set of indexable content for the document. Furthermore, document sections are determined based on attention data and each section is indexed separately. Indexing is per section and based only on the indexable content, thereby enhancing the index relevance, increasing the efficiency of search engines and reducing spamdexing.

Abstract: During an interaction with a web page, user interface events are recorded and augmented with page layout data from the document object model. An event stream is formed with the page layout data and communicated to an event server. The event server processes the event stream to determine a location at which the events were generated and determines attention data for each event stream that indicates where a user's attention was directed during an interaction. The attention data for a plurality of event streams is aggregated to determine common or popular areas of interest of the web page.

Abstract: Selection of web page advertisements for placement into a web page may use attention data that ranks the elements of the web page according to the attention the elements receive from users during interactions with the web page. The content of ads can be matched to the content of the most highly ranked elements of the web page rather than the web page as a whole. Alternatively, ads can be selected that historically produce a preferred pattern of behavior such as an attention shift from a highly ranked element of the web page to the ad.

Abstract: A web page that includes content form fields may be modified to include an event observer module and an authored content module. Events generated during the authoring of content by a user are recorded by the event observer module and sent to an event server with an InteractionID. The authored content module inserts hidden fields into the form fields that are updated with the InteractionID when content is submitted to the web server. The web server provides the InteractionID in a bind request to the event server. The event server binds the content to the events used to create the content in response to the request.