WO2011069255A1 - Procédé et système permettant une catégorisation d'adresse url efficace et exhaustive - Google Patents

Procédé et système permettant une catégorisation d'adresse url efficace et exhaustive Download PDF

Info

Publication number
WO2011069255A1
WO2011069255A1 PCT/CA2010/001952 CA2010001952W WO2011069255A1 WO 2011069255 A1 WO2011069255 A1 WO 2011069255A1 CA 2010001952 W CA2010001952 W CA 2010001952W WO 2011069255 A1 WO2011069255 A1 WO 2011069255A1
Authority
WO
WIPO (PCT)
Prior art keywords
url
categorization engine
web page
urls
semantic
Prior art date
Application number
PCT/CA2010/001952
Other languages
English (en)
Inventor
Olivier Mirandette
Marc Tremblay
Eric Melin
Original Assignee
Neuralitic Systems
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Neuralitic Systems filed Critical Neuralitic Systems
Priority to GB1209172.4A priority Critical patent/GB2488274A/en
Priority to US13/515,079 priority patent/US8935390B2/en
Priority to CA2769946A priority patent/CA2769946C/fr
Publication of WO2011069255A1 publication Critical patent/WO2011069255A1/fr

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/50—Network services
    • H04L67/535—Tracking the activity of the user
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/955—Retrieval from the web using information identifiers, e.g. uniform resource locators [URL]

Definitions

  • a method and system are disclosed to categorize URLs (Uniform Resource Locators) of web pages accessed by multiple users over an IP (Internet Protocol) based data network.
  • the method and system collect, by means of at least one monitoring probe, real time data from IP data traffic occurring on the IP based data network, and extract parameters from the collected real time data, the parameters including an URL of a web page.
  • the URL is processed by a rule based categorization engine, to associate a matching category to the URL of the web page.
  • the URL is transferred to a semantic based categorization engine.
  • a matching category is associated to the transferred URL by the semantic based categorization engine, based on a semantic analysis of the textual content extracted from the web page associated to the URL.
  • a method and system are disclosed, wherein the matching category associated to a transferred URL by the semantic based categorization engine is transferred back to the rule based categorization engine.
  • a set of text based rules of the rule based categorization engine is automatically updated, in order to acquire the capability to associate the matching category to the transferred URL.
  • a method and system are disclosed, wherein the URLs for which no matching category has been inferred by the rule based categorization engine over a determined period of time are memorized; and only the N URLs with the highest occurrence over the determined period of time are transferred to the semantic based categorization engine; where N is a pre-defined number of URLs.
  • a method and system are disclosed, wherein additional parameters related to an access to a web page are extracted, by the at least one monitoring probe, from the collected real time data.
  • An analysis from a Business Intelligence perspective of the matching categories associated to the URLs is performed by an analytic system; the analysis taking into consideration the additional parameters.
  • Figure 1 illustrates a rule based URL categorization engine; according to a non- restrictive illustrative embodiment
  • Figure 2 illustrates a system for efficient and exhaustive URL categorization, according to a non-restrictive illustrative embodiment
  • Figure 3 illustrates a method for efficient and exhaustive URL categorization, according to a non-restrictive illustrative embodiment
  • Figure 4 illustrates a system for efficient and exhaustive URL categorization in the context of a mobile network, according to a non-restrictive illustrative embodiment.
  • IP based data networks including fixed broadband networks and mobile networks
  • fixed and mobile service providers have an increasing need to understand their subscriber's habits and behaviours, in terms of consumption of IP based data services, to identify means to differentiate from the competition.
  • various tools are used to analyze these IP based data services.
  • the extreme diversity of available web pages complicates this analysis.
  • One means to address this complexity is to classify the web pages among a pre-defined list of categories, and to perform the analysis from the perspective of these categories, rather than from the perspective of the web pages themselves.
  • One technology to classify web pages consists in using a set of rules to associate the Uniform Resource Locator (URL) of a web page to one among a pre-defined list of categories.
  • This type of technology referred to as a rule based categorization engine, is usually effective in terms of performance, but does not scale easily to the diversity of available web sites and pages. It can only categorize a limited subset of web pages (those covered by the set of rules), and needs manual upgrade to extend its set of rules.
  • An object of the present is therefore to provide a method and system for efficient and exhaustive URL categorization.
  • the present method is adapted for categorizing URLs of web pages accessed by users over an IP (Internet Protocol) based data network. For doing so, the method collects, by means of at least one monitoring probe, real time data from IP data traffic occurring on the IP based data network; and extracts from the collected real time data parameters related to a web page, the parameters including an URL of the web page. Then, the method processes the URL with a rule based categorization engine, to associate a matching category to the URL of the web page; the matching category being inferred from a pre-defined list of categories. When no matching category is inferred, the method transfers the URL of the web page to a semantic based categorization engine.
  • IP Internet Protocol
  • the method processes the transferred URL with the semantic based categorization engine.
  • the processing consists in: extracting textual content from the web page associated to the URL; performing a semantic analysis of the textual content; and associating a matching category to the transferred URL of the web page based on the semantic analysis of the textual content extracted from the web page, the matching category being inferred from a pre-defined list of categories.
  • the present system is adapted for categorizing URLs of web pages accessed by users over an IP based data network. For doing so, the system comprises at least one monitoring probe, for collecting real time data from IP data traffic occurring on the IP based data network; and for extracting from the collected real time data parameters related to a web page, the parameters including an URL of the web page.
  • the system also comprises a rule based categorization engine for processing the URL, and to associate a matching category to the URL of the web page, the matching category being inferred from a pre-defined list of categories.
  • the system further comprises a semantic based categorization engine, for further processing the URL of the web page, when no matching category is inferred by the rule based categorization engine.
  • the further processing consists in: extracting textual content from the web page associated to the URL; performing a semantic analysis of the textual content; and associating a matching category to the URL of the web page based on the semantic analysis of the textual content extracted from the web page, the matching category being inferred from a predefined list of categories.
  • a complementary mechanism is implemented as follows: the matching category associated to a transferred URL by the semantic based categorization engine is transferred back to the rule based categorization engine; and a set of text based rules of the rule based categorization engine is automatically updated to acquire the capability to associate the matching category to the transferred URL.
  • a complementary mechanism is implemented, to memorize the URLs which cannot be categorized by the rule based categorization engine, and to transmit at regular intervals only a subset of these URLs to the semantic based categorization engine. The subset consists of a pre-defined number of URLs, which have the highest occurrence on the data network, and cannot be categorized by the rule based categorization engine.
  • FIG. 1 The context of Figure 1 is an IP based data network offering IP based data services to end users.
  • This network may be a fixed broadband network relying on a variety of access technologies, including cable, DSL (Digital Subscriber Line), fiber optic.
  • the network may be a mobile network, for instance a cellular network or a WIMAX (Worldwide Interoperability for Microwave Access) network, or a combination of both fixed broadband and mobile networks.
  • Subscribers use various types of devices to access data services via the IP based data network.
  • Mobile devices 10 e.g. mobile phones, tablets, etc
  • computers 20 e.g. fixed like a Personal Computer (PC), nomadic like a laptop, etc
  • PC Personal Computer
  • web browsing represents a significant part of the activities of the subscribers.
  • the web pages 40 accessed during the web browsing sessions may be part of a dedicated web portal belonging to the network Operator, or may alternatively be part of the Internet at large.
  • Network Operators are increasingly interested in measuring and analyzing the activities of their subscribers in terms of IP based data services consumption.
  • the network Operator makes use of this information directly, to develop targeted value added services offerings or to customize its data plan offerings.
  • the network Operator shares this information with third party content providers to adapt the content offerings.
  • the network Operator shares this information with on-line advertising companies, to generate personalized advertisements targeting clusters of subscribers with similar interests, or even targeting individual subscribers.
  • Other scenarios based on the ecosystem involving network Operators, content providers, and advertisers may be generated at will.
  • measuring and analyzing the subscriber's activities related to web browsing consists in monitoring the web pages 40 accessed by the subscribers of the Operator, and assigning a category to each of these web pages.
  • the category is selected among a predefined list of categories representing the diversity of the contents hosted on Internet web pages. Examples of such categories include, among others: sports, news, finance, music and video...
  • a hierarchy of categories can be built to better represent the variety of thematic available on the Internet. For instance, sports may be divided into the sub-categories soccer, tennis, swimming... Although there is no limitation on the number of layers in the hierarchy of categories, a main category and a single sub-category is considered effective in terms of classification.
  • a web page is allocated the category sport / tennis when its content is more closely related to this specific category.
  • a classification of the web pages accessed by subscribers based on a limited number of categories / sub-categories is a practical source of information to analyze the trends and behaviors of these subscribers related to web browsing. From a business analyst perspective, it is complicated to deal with statistics directly related to the URLs of the web pages accessed by subscribers, since there is almost infinity of such URLs. Grouping these URLs within a limited list of categories (typically a few dozen categories, possibly organized in main and sub categories) reduces the size of the data to be analyzed to a size that is manageable from a human being perspective.
  • the different steps involved in the aforementioned method are the following: capturing the URLs of the web pages accessed by the subscribers in real time. Then, assigning a category to each URL. And finally, performing an analysis from a Business Intelligence perspective on metrics associated to the categories (e.g. analyzing the percentage of web pages viewed for each category).
  • the real time capture of the URLs of the web pages 40 accessed by the subscribers relies on a monitoring probe 50, based for example on DPI (Deep Packet Inspection).
  • the DPI technology is well known in the art for allowing the real time capture of live IP traffic 30, and for allowing the processing of the captured traffic along various layers of the OSI (Open System Interconnection) model to extract relevant information.
  • the capture of the live IP traffic 30 is performed at one or several carefully selected location(s) in the IP based data network infrastructure of the Operator. Figure 4 will illustrate this point of capture in the case of a cellular network.
  • the monitoring probe 50 has the capability to identify the IP based data traffic corresponding to an access to a specific web page by a subscriber, and to extract the URL corresponding to this web page.
  • the DPI based monitoring probe identifies the protocol used for web browsing: HTTP (Hyper Text Transfer Protocol). It then extracts the URL of a specific web page from the HTTP layer of the IP packets exchanged during the corresponding web browsing session. Additional parameters are also collected by the DPI based monitoring probe 50, to add contextual information to the URLs. For instance, timestamps are added to record the time at which a web page has been accessed.
  • HTTP Hyper Text Transfer Protocol
  • a unique identifier of the specific subscriber (or alternatively, a unique identifier of the device 10 or 20 used by this specific subscriber) who accessed a web page is extracted from the live IP traffic 30. This type of information is useful to count the number of unique subscribers who accessed a web page (or a category in the context of the present method and system). Associating a unique identifier of a subscriber to a specific IP data flow corresponding to a web browsing session of this subscriber is usually not trivial. However, this type of association has been described in details in the art related to DPI technology, and is also technically feasible.
  • a unique identifier of the subscriber is the International Mobile Subscriber Identity (IMSI) or the Mobile Subscriber ISDN (MSISDN). Both the IMSI and MSISDN are extracted from UMTS specific IP based control traffic, and associated to a specific data session including HTTP based IP traffic.
  • the parameters 55 extracted by the monitoring probe 50 are transmitted to an analytic system 60. These parameters include the URLs of all the web pages 40 accessed via the Operator data network, along with additional contextual information (like for example the timestamps and the subscriber's unique identifiers mentioned previously). From an operational point of view, the transmission usually occurs at regular intervals: for example, every fifteen minutes, or every hour, or every day.
  • the interval depends on the capabilities and performances of the monitoring probe 50, and of the analytic system 60.
  • the analytic system 60 usually includes a high performance database (not shown) to store the parameters transmitted by the monitoring probe 50.
  • the parameters 55 are transformed and mapped to an optimized data model for efficient storage.
  • the analytic system 60 also includes a dedicated analytic engine, to perform Business Intelligence oriented processing of the stored data. For instance, historic data over a period of reference are used to track trends and behaviors related to subscriber's consumption of IP based data services.
  • the analytic system 60 is capable of analyzing different types of data services.
  • the parameters 55 transmitted by the monitoring probe 50 include information related to these various types of data services. However, in the following, we will focus on web browsing analytics and URL categorization.
  • the engine uses an optimized algorithm to associate a category to the URL, chosen among a list of pre-defined categories.
  • the URL category 74 is transferred back to the analytic system 60.
  • the fact that a web page 40 corresponding to the URL 72 has been accessed is recorded in the analytic system (usually in a high performance database as already mentioned) as follows: the URL category 74 is recorded, along with contextual parameters such as the timestamp and a unique identifier for the subscriber who accessed the web page.
  • the full URL 72 may be memorized as well.
  • a URL is composed of a host, a path, a file and optional parameters (in the form of key / value pairs).
  • the host represents an entire web site belonging to an organization.
  • the path represents the localization of a specific web page in the web site.
  • the file represents the name of a file where the content of a web page is stored.
  • the optional parameters allow interactions with the referenced web page. For instance, the web page www.vahoo.ca/sports/tennis/coupe rogers/results.html is broken down as follows: www.yahoo.ca is the host, sports/tennis/coupe_rogers is the path, and results.html is the file.
  • a single web site, represented by its host may include diversified content belonging to several categories.
  • the web page www.vahoo.ca/sports/tennis/coupe rogers/results.html could be associated to the category sport / tennis. While the web page www.yahoo.ca/news/tsx.html could be associated to the category finance / stock markets.
  • viewing of these two web pages is memorized in the analytic system 60 as viewing content from the categories sport / tennis and finance / stock markets, respectively.
  • FIG. 1 Another point is the localization of the rule based categorization engine 70.
  • Figure 1 it has been represented as a standalone entity interfacing with the analytic system 60.
  • the rule based categorization engine may be integrated to either the monitoring probe 50 or to the analytic system 60.
  • rule based categorization engine 70 relies on a rule engine, which is a technology well known in the art of artificial intelligence.
  • a possible embodiment of a rule engine in the context of the present method and system is described as follows.
  • Text based rules are included in an input file.
  • Each line in the input file represents a text based rule and consists of a single URL or an aggregation of URLs, followed by the associated category. Aggregation of URLs is represented using specific symbols.
  • the symbol * can be used as an aggregator as follows: www.yahoo.ca/sports/* means every URL starting with www.yahoo.ca/sports/.
  • the symbols * and - can be used as aggregators as follows: www.vahoo.ca/sports/* - /tennis/ means every URL starting with www.yahoo.ca/sports/ except those starting with www.yahoo.ca/sports/tennis.
  • the aggregation rules for the URLs allow for the allocation of a category to multiple web pages via a single line in the input file.
  • the input file is interpreted by the rule engine 70 and converted to metadata in a format specific to the rule engine.
  • the rule engine 70 consists in software instructions executed on either a standard or a dedicated hardware platform. When presented with a URL, the rule engine 70 executes the software instructions to interpret the metadata in order to find the associated category.
  • the format of the metadata and the set of software instructions are both optimized, in order to find the category associated to a URL in a minimum amount of time.
  • all the information contained in the text based rules of the input file is present in the metadata, but represented in a format that is efficiently interpreted by the software instructions.
  • the rule based categorization engine 70 may interpret a combination of the hosts, paths and files that constitute the URLs. In particular, this may allow a better granularity and flexibility in the definition of the text based rules in the input file.
  • the main advantage of using a rule engine 70 for the categorization of the URLs is that it relies on optimized mechanisms that improve the performances of the categorization process, compared to alternative technologies.
  • the main drawback is that the effective range of URLs that can be categorized is limited to the set of text based rules defined in the input file: only the URLs represented by the text based rules can be categorized.
  • the aggregation rules allow for the representation of a large number of URLs via a single rule, but still, the variety of existing URLs (for instance considering the whole Internet) is so large, that only a subset of these existing URLs can be categorized via this mechanism.
  • Another drawback is that the list of available rules is static by nature, preventing the implementation of a dynamic mechanism to automatically upgrade the list of rules to follow the evolution of the web pages 40 accessed by subscribers.
  • the semantic based categorization engine 100 retrieves the web page content 05 of the URL 102 associated to the web page 40, and extracts the textual content from the web page content 105 (web pages usually also contain multimedia content in the form of images, videos, etc; which are not exploited by the present method and system). For this purpose, the semantic based categorization engine 100 connects to the related web site, downloads the web page 40 corresponding to the URL 102, and performs the extraction of the textual content from the web page content.
  • the semantic based categorization engine 100 performs a semantic analysis of the textual content, and infers the best matching category for this textual content.
  • the category is selected among a pre-defined list of categories; preferably the same list that is used by the rule based categorization engine 70.
  • the semantic based categorization engine 100 is a technology well known in the art of artificial intelligence.
  • a semantic based categorization engine has the capability to analyze a text from a semantic perspective, to recognize specific patterns and associate a meaning to these patterns.
  • a critical component of the semantic based categorization engine 100 is the ontology: it is the metadata that contains the background knowledge and the rules necessary to perform the semantic based interpretation of the text. The following describes how the semantic based categorization engine is trained and becomes operational.
  • the ontology is created and the semantic based categorization engine is trained with a sample population, to learn to associate a textual input with the proper output (usually an interpretation or categorization of the textual input).
  • the semantic based categorization engine 100 is capable of determining the proper output when presented with any new textual input. Additionally, a percentage of confidence can also be provided, representing the level of confidence that the inferred output is the best matching one for the submitted textual input.
  • the training process of the semantic based categorization engine 100 usually relies on one among a variety of Artificial Intelligence technologies.
  • the training period consists first in generating the appropriate ontology, which is closely related to the pre- defined list of URL categories. Then, it consists in training the semantic based categorization engine 100 with a selection of URLs for each pre-defined URL category. The URLs are carefully selected by a specialized staff, including linguists, based on criteria granting the relevance of the textual content associated to an URL with the matching URL category. The semantic based categorization engine 100 is trained to associate the textual content of the URLs of the training sample with their matching URL categories.
  • a validation phase is performed.
  • a new sample of URLs is selected, for which the category is known in advance.
  • the sample of URLs is presented to the semantic based categorization engine 100, and the inferred category for each URL is compared with the expected result.
  • a fine tuning of the semantic based categorization engine 100 is performed when inaccurate URL categorization occurs.
  • the validation sample is carefully selected by the specialized staff, to assess the robustness of the semantic based categorization process. For example, ambiguous textual content that may be interpreted as belonging to more than a single URL category is selected on purpose, to validate that the expected URL category is selected by the semantic based categorization engine 100.
  • the URL category 104 inferred by the semantic based categorization engine 100 is transmitted to the analytic system 60.
  • This URL category is exploited by the analytic system 60 in a similar manner as the one described in the context of Figure 1 , for the URL categories 74 transmitted by the rule based categorization engine 70.
  • the URL category 104 is transmitted to the rule based categorization engine 70, as part of an enhancement process 106 of the categorization capabilities of the rule based categorization engine 70.
  • a new text based rule for the URL 102 and its URL category 104 is created in the input file of the rule based categorization engine 70.
  • the rule based categorization engine 70 acquires the capability to categorize the URL 102 with the proper URL category 104.
  • an existing rule may be modified when applicable, to add the capability to categorize the URL 102 to the existing rule.
  • the semantic based categorization engine 100 has a limited capacity in terms of the range of URLs that it is capable of categorizing in a fixed amount of time.
  • a reasonable order of magnitude with existing technologies allows for the categorization of the textual content of a few web pages per second.
  • Increasing the web page categorization capacity means significantly increasing the underlying hardware platform capacity. This may have a prohibitive effect on the cost and scalability.
  • the following mechanism is implemented to address this issue.
  • the rule based categorization engine 70 Over a given period of time (for example every hour or every day), all the URLs that are not categorized by the rule based categorization engine 70 are memorized and counted. This mechanism of memorizing and counting may be implemented in the analytic system 60. Then, the top N (for example top 100 or top 1000) URLs not categorized 102 over the considered period of time are transmitted to the semantic based categorization engine 100. Finally, the enhancement process 106 is applied as described previously. With this mechanism, the rule based categorization engine 70 acquires the capacity to categorize 100 (or 1000) new critical URLs every hour (or day).
  • the same considerations apply as for the rule based categorization engine 70.
  • the rule based categorization engine may be integrated to the analytic system 60 or may be integrated with the rule based categorization engine 70.
  • a semantic based categorization engine 100 has been used.
  • an alternative Artificial Intelligence based technology can be used without changing the scope of the present method and system.
  • the only constraint on the Artificial Intelligence based technology is its capability to categorize any URL with a high percentage of accuracy.
  • the extensive coverage in terms of URL categorization has a cost in terms of performances.
  • such a technology is a complement of the high performance rule based categorization engine 70, and cannot replace it.
  • a mobile network 410 is considered in Figure 4. Examples of such mobile networks include cellular networks implementing one of the following standards: Universal Mobile Telecommunication System (UMTS), Code Division Multiple Access (CDMA), LTE (Long Term Evolution). Worldwide Interoperability for Microwave Access (WIMAX) networks are another type of mobile networks that are also relevant to the present method and system.
  • the mobile network 410 comprises a mobile data network 420, to transport the data traffic generated by the data services provided by the mobile Operator. Such data services consist, among others, in web browsing, messaging, multimedia delivery, on-line gaming, and business applications.
  • the Internet Protocol is the underlying networking protocol used in mobile data networks, in the case of any type of cellular network, as well as for WIMAX networks.
  • the mobile data network 420 is the interface between the mobile subscribers engaged in a data session, represented by their mobile devices 400, and external networks, like the Internet 430. Any IP based data session in which a mobile device 400 is engaged, goes through the mobile data network 420.
  • Different types of external networks can be accessed via the mobile data network 420: the Internet 430, an ISP network (not represented in Figure 4), and also services and applications network infrastructures operated by the Mobile Operator (not represented in Figure 4), like the IP Multimedia Subsystem (IMS). More specifically, Figure 4 represents a Universal Mobile Telecommunication System (UMTS) network, as an example of a mobile network.
  • the Gateway GPRS Support Node (GGSN) 422 is a key equipment for the transport of the IP data traffic in the UMTS mobile data network 420.
  • the GGSN 422 is connected through its Gi interface 426 to the external networks, including the Internet 430. It is connected through its Gn interface 424 to the rest of the UMTS packet switched core network.
  • the packet switched core network is the name of the mobile data network 420 for a UMTS network.
  • a monitoring probe 450 uses the Gn interface 424, or the Gi interface 426, as a point of capture for the IP data traffic circulating in the mobile data network 420.
  • the role of the monitoring probe 450 is to extract relevant parameters from the IP data traffic (using DPI technologies), and to transmit these parameters to the analytics system 460.
  • the Gn interface 424 and Gi interface 426 are shown as the points of capture for the IP data traffic circulating in the mobile data network 420, other points of capture could alternately be selected.
  • the parameters include the URLs of the web pages 440 accessed from the mobile devices 400 of the subscribers through the mobile data network 420.
  • the rule based categorization engine 470 and the semantic based categorization engine 480 implement the mechanisms described previously (and represented in Figures 1 , 2, and 3), to perform the categorization of the URLs received by the analytic system 460.
  • the categorized URLs in conjunction with other parameters (timestamps, unique identifiers of the subscribers accessing the URLs) transmitted by the monitoring probe 450, are analyzed by the analytic system 460 from a Business Intelligence perspective. This type of analysis is known in the art as web analytics.
  • the following metrics may be calculated by the analytic system 460: total number of accesses to web pages which URLs have been associated to a specific URL category, over a determined period of time; total number of unique subscribers accessing web pages which URLs have been associated to a specific category, over a determined period of time.
  • web pages 440 on Figure 4 have been represented as being part of the Internet 430, mobile Operators usually have their own web portal, which is hosted in the mobile data network 420.
  • the categorization may be applied to both the Operator web portal, and to the Internet, or may focus on only one of these two.
  • Figure 4 represents an embodiment of the present system in the context of a UMTS network, it can be applied to another type of mobile network, or to a fixed broadband network, as well.
  • the main difference is the localization of the monitoring probe 450, which is located at a point of convergence for all the IP based data traffic.

Landscapes

  • Engineering & Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

La présente invention se rapporte à un procédé et à un système qui permettent une catégorisation d'adresses URL (localisateurs de ressources uniformes) des pages web auxquelles de multiples utilisateurs ont accès sur un réseau de données IP (protocole Internet). Le procédé et le système collectent des données en temps réel provenant du trafic de données IP se trouvant sur le réseau de données IP, et extraient des paramètres des données collectées en temps réel, les paramètres comprenant une adresse URL d'une page web. L'adresse URL est traitée par un moteur de catégorisation basée sur des règles afin d'associer une catégorie concordante à l'adresse URL de la page web. Lorsqu'aucune catégorie concordante n'est déduite, l'adresse URL est transférée à un moteur de catégorisation sémantique. Une catégorie concordante est associée à l'adresse URL transférée par le moteur de catégorisation sémantique sur la base d'une analyse sémantique du contenu textuel extrait de la page web associée à l'adresse URL.
PCT/CA2010/001952 2009-12-11 2010-12-08 Procédé et système permettant une catégorisation d'adresse url efficace et exhaustive WO2011069255A1 (fr)

Priority Applications (3)

Application Number Priority Date Filing Date Title
GB1209172.4A GB2488274A (en) 2009-12-11 2010-12-08 A method and system for efficient and exhaustive url categorization
US13/515,079 US8935390B2 (en) 2009-12-11 2010-12-08 Method and system for efficient and exhaustive URL categorization
CA2769946A CA2769946C (fr) 2009-12-11 2010-12-08 Procede et systeme permettant une categorisation d'adresse url efficace et exhaustive

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US28560409P 2009-12-11 2009-12-11
US61/285,604 2009-12-11

Publications (1)

Publication Number Publication Date
WO2011069255A1 true WO2011069255A1 (fr) 2011-06-16

Family

ID=44145066

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CA2010/001952 WO2011069255A1 (fr) 2009-12-11 2010-12-08 Procédé et système permettant une catégorisation d'adresse url efficace et exhaustive

Country Status (4)

Country Link
US (1) US8935390B2 (fr)
CA (1) CA2769946C (fr)
GB (1) GB2488274A (fr)
WO (1) WO2011069255A1 (fr)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10079876B1 (en) 2014-09-30 2018-09-18 Palo Alto Networks, Inc. Mobile URL categorization
US11580163B2 (en) 2019-08-16 2023-02-14 Palo Alto Networks, Inc. Key-value storage for URL categorization
US11748433B2 (en) 2019-08-16 2023-09-05 Palo Alto Networks, Inc. Communicating URL categorization information

Families Citing this family (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9251239B1 (en) * 2008-05-15 2016-02-02 Salesforce.Com, Inc. System, method and computer program product for applying a public tag to information
US8805838B1 (en) * 2009-12-22 2014-08-12 Amazon Technologies, Inc. Systems and methods for automatic item classification
US9424356B2 (en) 2010-12-09 2016-08-23 Microsoft Technology Licensing, Llc Updating a search index using reported browser history data
US20120310941A1 (en) * 2011-06-02 2012-12-06 Kindsight, Inc. System and method for web-based content categorization
US8677346B1 (en) 2011-09-27 2014-03-18 Symantec Corporation Providing installer package information to a user
CN103092857A (zh) * 2011-11-01 2013-05-08 腾讯科技（深圳）有限公司 历史记录的整理方法和装置
US10122722B2 (en) * 2013-06-20 2018-11-06 Hewlett Packard Enterprise Development Lp Resource classification using resource requests
US9239737B2 (en) 2013-11-15 2016-01-19 Sap Se Concise resource addressing
US10073918B2 (en) 2014-08-12 2018-09-11 Entit Software Llc Classifying URLs
US10225158B1 (en) * 2014-12-22 2019-03-05 EMC IP Holding Company LLC Policy based system management
KR102455232B1 (ko) 2015-06-23 2022-10-17 삼성전자 주식회사 콘텍스트 기반 탭 관리를 위한 방법 및 전자 장치
US20170032019A1 (en) * 2015-07-30 2017-02-02 Anthony I. Lopez, JR. System and Method for the Rating of Categorized Content on a Website (URL) through a Device where all Content Originates from a Structured Content Management System
US10432490B2 (en) * 2015-07-31 2019-10-01 Cisco Technology, Inc. Monitoring single content page application transitions
CN107045501A (zh) * 2016-02-05 2017-08-15 腾讯科技（深圳）有限公司 一种网页访问方法和装置
US10445393B2 (en) 2016-05-31 2019-10-15 Bootstrap Collective LLC System and method of creating and processing semantic URL
CN106970984B (zh) * 2017-03-29 2020-11-06 杭州迪普科技股份有限公司 一种url过滤库更新方法及装置
CN109995889B (zh) * 2018-01-02 2022-02-25 中国移动通信有限公司研究院 映射关系表的更新方法、装置、网关设备及存储介质
US10866704B1 (en) * 2019-06-04 2020-12-15 Microsoft Technology Licensing, Llc Intelligent web browser segregation
US11451569B1 (en) * 2021-04-17 2022-09-20 Arbor Networks, Inc. File extraction from network data to artifact store files and file reconstruction
US11848839B2 (en) * 2021-09-30 2023-12-19 Forescout Technologies, Inc. Entity attribute designation based on logic programming

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070081471A1 (en) * 2005-10-06 2007-04-12 Alcatel Usa Sourcing, L.P. Apparatus and method for analyzing packet data streams
US20080040389A1 (en) * 2006-08-04 2008-02-14 Yahoo! Inc. Landing page identification, tagging and host matching for a mobile application
US20080162448A1 (en) * 2006-12-28 2008-07-03 International Business Machines Corporation Method for tracking syntactic properties of a url
US20080189307A1 (en) * 2007-02-01 2008-08-07 Ayyappan Sankaran Method for categorizing content published on internet
US20080209057A1 (en) * 2006-09-28 2008-08-28 Paul Martini System and Method for Improved Internet Content Filtering
US7421498B2 (en) * 2003-08-25 2008-09-02 Microsoft Corporation Method and system for URL based filtering of electronic communications and web pages
WO2008124947A1 (fr) * 2007-04-16 2008-10-23 Neuralitic Systems Procédé et système de filtrage d'un trafic ip dans des réseaux ip mobiles
US7590716B2 (en) * 2004-09-09 2009-09-15 Websense Uk Limited System, method and apparatus for use in monitoring or controlling internet access
US20090240638A1 (en) * 2008-03-19 2009-09-24 Yahoo! Inc. Syntactic and/or semantic analysis of uniform resource identifiers

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7376729B2 (en) * 2000-03-23 2008-05-20 Freebit Co., Ltd. System for internet connections, method for calculating connection fees for network connection services, billing system for network connection services, and system for network connection management
US20010032258A1 (en) * 2000-03-31 2001-10-18 Kabushiki Kaisha Free Bit.Com System for internet connections, system for providing internet user information, method for providing internet user preference information, and method for distributing digital contents using the internet
US7062475B1 (en) * 2000-05-30 2006-06-13 Alberti Anemometer Llc Personalized multi-service computer environment
US20030046311A1 (en) * 2001-06-19 2003-03-06 Ryan Baidya Dynamic search engine and database
US7194464B2 (en) * 2001-12-07 2007-03-20 Websense, Inc. System and method for adapting an internet filter
US7693863B2 (en) * 2004-12-20 2010-04-06 Claria Corporation Method and device for publishing cross-network user behavioral data
US20090182818A1 (en) * 2008-01-11 2009-07-16 Fortinet, Inc. A Delaware Corporation Heuristic detection of probable misspelled addresses in electronic communications
US20090234876A1 (en) * 2008-03-14 2009-09-17 Timothy Schigel Systems and methods for content sharing

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7421498B2 (en) * 2003-08-25 2008-09-02 Microsoft Corporation Method and system for URL based filtering of electronic communications and web pages
US7590716B2 (en) * 2004-09-09 2009-09-15 Websense Uk Limited System, method and apparatus for use in monitoring or controlling internet access
US20070081471A1 (en) * 2005-10-06 2007-04-12 Alcatel Usa Sourcing, L.P. Apparatus and method for analyzing packet data streams
US20080040389A1 (en) * 2006-08-04 2008-02-14 Yahoo! Inc. Landing page identification, tagging and host matching for a mobile application
US20080209057A1 (en) * 2006-09-28 2008-08-28 Paul Martini System and Method for Improved Internet Content Filtering
US20080162448A1 (en) * 2006-12-28 2008-07-03 International Business Machines Corporation Method for tracking syntactic properties of a url
US20080189307A1 (en) * 2007-02-01 2008-08-07 Ayyappan Sankaran Method for categorizing content published on internet
WO2008124947A1 (fr) * 2007-04-16 2008-10-23 Neuralitic Systems Procédé et système de filtrage d'un trafic ip dans des réseaux ip mobiles
US20090240638A1 (en) * 2008-03-19 2009-09-24 Yahoo! Inc. Syntactic and/or semantic analysis of uniform resource identifiers

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
"Proceedings of ACM Fourteenth Conference on Information and Knowledge Management CIKM'05, Bremen, Germany, pages 325-326, 31 October-5 November 2005", 31 October 2005, article KAN, M. ET AL.: "Fast Webpage Classification Using URL Features" *
"Proceedings of the 2006 IEEE/WIC/ACM International Conference on Web Intelligence WT06, Hong Kong, pages 277-283, 18-22 December 2006", 18 December 2006, article ZHANG, J. ET AL.: "The Role of URLs in Objectionable Web Content Categorization" *
"Proceedings of the Fifth International Conference on Semantics, Knowledge and Grid, SKG 2009, Zhuhai, China, pages 270-273, 12-14 October 2009", 12 October 2009, article LI, C. ET AL.: "Classifier-Guided Topical Crawler: A Novel Method of Automatically Labeling the Positive URLs" *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10079876B1 (en) 2014-09-30 2018-09-18 Palo Alto Networks, Inc. Mobile URL categorization
US10554736B2 (en) 2014-09-30 2020-02-04 Palo Alto Networks, Inc. Mobile URL categorization
US11580163B2 (en) 2019-08-16 2023-02-14 Palo Alto Networks, Inc. Key-value storage for URL categorization
US11748433B2 (en) 2019-08-16 2023-09-05 Palo Alto Networks, Inc. Communicating URL categorization information
US11983220B2 (en) 2019-08-16 2024-05-14 Palo Alto Networks, Inc. Key-value storage for URL categorization

Also Published As

Publication number Publication date
CA2769946C (fr) 2013-02-05
CA2769946A1 (fr) 2011-06-16
US8935390B2 (en) 2015-01-13
GB2488274A (en) 2012-08-22
US20120271941A1 (en) 2012-10-25
GB201209172D0 (en) 2012-07-04

Similar Documents

Publication Publication Date Title
CA2769946C (fr) Procede et systeme permettant une categorisation d'adresse url efficace et exhaustive
US20120317151A1 (en) Model-Based Method for Managing Information Derived From Network Traffic
US20130066814A1 (en) System and Method for Automated Classification of Web pages and Domains
US10530671B2 (en) Methods, systems, and computer readable media for generating and using a web page classification model
CN103218431B (zh) 一种能识别网页信息自动采集的系统
US20140223575A1 (en) Privacy protection in recommendation services
US20130066875A1 (en) Method for Segmenting Users of Mobile Internet
US20130064109A1 (en) Analyzing Internet Traffic by Extrapolating Socio-Demographic Information from a Panel
US20140304653A1 (en) Method For Generating Rules and Parameters for Assessing Relevance of Information Derived From Internet Traffic
CN106230809B (zh) 一种基于url的移动互联网舆情监测方法及系统
Fang et al. Fine-grained HTTP web traffic analysis based on large-scale mobile datasets
Laiche et al. QoE‐aware traffic monitoring based on user behavior in video streaming services
Liu et al. Request dependency graph: A model for web usage mining in large-scale web of things
US9973950B2 (en) Technique for data traffic analysis
US9069763B2 (en) Services management application integrating social media and automated infrastructure monitoring
Ribeiro et al. Youniverse: Large-scale channel and video metadata from english-speaking youtube
US20130064108A1 (en) System and Method for Relating Internet Usage with Mobile Equipment
US20120215793A1 (en) Method and system for matching segment profiles to a device identified by a privacy-compliant identifier
US10027774B2 (en) Information on navigation behavior of web page users
JP2013011952A (ja) 潜在クラス分析装置、潜在クラス分析方法及びプログラム
Husin et al. Discovering users navigation of online newspaper using Markov model
KR102381132B1 (ko) 세션 정보 저장 및 렌더링을 이용한 세션 리플레이 서비스 제공 방법
Vassio et al. Data Analysis and Modelling of Users’ Behavior on the Web
Yuan et al. Mobile phone recommendation based on phone interest
Lei et al. A Systematic Literature Review on Relationship Between Internet Usage Behavior and Internet QoS in Campus

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 10835340

Country of ref document: EP

Kind code of ref document: A1

DPE1 Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101)
WWE Wipo information: entry into national phase

Ref document number: 2769946

Country of ref document: CA

ENP Entry into the national phase

Ref document number: 1209172

Country of ref document: GB

Kind code of ref document: A

Free format text: PCT FILING DATE = 20101208

WWE Wipo information: entry into national phase

Ref document number: 1209172.4

Country of ref document: GB

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 13515079

Country of ref document: US

122 Ep: pct application non-entry in european phase

Ref document number: 10835340

Country of ref document: EP

Kind code of ref document: A1