Website detection method and device, electronic equipment and storage medium
Through the knowledge graph and tag dissemination algorithm, combined with web page content and URL association relationship, the recall and accuracy of URL detection are improved, the problem of insufficient URL detection capabilities in the existing technology is solved, and more comprehensive risk URL recognition is achieved.
Patent Information
- Application Number
- CN202410985940.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-22
- Publication Date
- 2024-11-15
AI Technical Summary
The existing URL detection method is based on web page content and domain name information, and has weak recall capabilities and cannot effectively detect potential risky URLs. Especially web pages with strong content confrontation or risky links are missing, and only detect them for the current URL, ignoring external links. Potential risks in URLs.
Using knowledge graph technology and tag propagation algorithm, select candidate propagation nodes and tags through entity nodes, edge relationships and propagation configuration parameters, improve the recall capability of risk websites, combine web page content for risk detection, introduce correlation relationship dimensions, and improve detection accuracy and coverage.
It significantly improves the recall and accuracy of risk websites, reduces the missed detection rate, identifies URLs that seem normal but contain risk external links, expands the detection range, and achieves more comprehensive security protection.
Smart Images

Figure CN118972107A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the Internet field, to the field of network security technology, and specifically to a URL detection method, device, electronic device, and storage medium. Background Art
[0002] Risky websites such as malicious websites, fraudulent activities, and privacy leaks have become a huge hidden danger in cyberspace, seriously threatening the rights and interests of the majority of netizens. Internet companies have the responsibility and obligation to actively fulfill their social responsibilities, strengthen their identification capabilities, and ensure that they can detect these risky websites in a timely and accurate manner.
[0003] The URL detection in related technologies is more based on the webpage content and its own domain name information to determine whether the URL has risks, and the recall ability is relatively weak. It is very important to improve the detection ability of risky URLs. Summary of the invention
[0004] The present disclosure provides a URL detection method, device, electronic device and storage medium.
[0005] According to one aspect of the present disclosure, a method for detecting a website is provided, comprising:
[0006] Obtain the entity node corresponding to the URL to be detected in the URL relationship knowledge graph as the monitoring node in the label propagation algorithm;
[0007] Selecting a candidate propagation node for the listening node from the entity nodes according to the edge relationship between different entity nodes and the first propagation relationship in the propagation configuration parameter;
[0008] Determine a candidate label of the listening node according to the second label of the candidate propagation node and the second propagation relationship between the candidate propagation node and the listening node;
[0009] The detection result of the to-be-detected website is determined according to the candidate tags of the monitoring node.
[0010] According to one aspect of the present disclosure, a URL detection device is provided, comprising:
[0011] The monitoring node module is used to obtain the entity node corresponding to the URL to be detected in the URL relationship knowledge graph as the monitoring node in the label propagation algorithm;
[0012] A propagation node module, configured to select a candidate propagation node for the listening node from the entity nodes according to the edge relationship between different entity nodes and the first propagation relationship in the propagation configuration parameter;
[0013] a label propagation module, configured to determine a candidate label of the listening node according to the second label of the candidate propagation node and a second propagation relationship between the candidate propagation node and the listening node;
[0014] The detection result module is used to determine the detection result of the URL to be detected according to the candidate label of the monitoring node.
[0015] According to another aspect of the present disclosure, there is provided an electronic device, the electronic device comprising:
[0016] at least one processor; and
[0017] a memory communicatively connected to the at least one processor; wherein,
[0018] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method provided by any embodiment of the present disclosure.
[0019] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable a computer to execute the method provided by any embodiment of the present disclosure.
[0020] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, wherein when the computer program is executed by a processor, the method provided according to any embodiment of the present disclosure is implemented.
[0021] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] Figure 1 is a flow chart of a URL detection method provided according to an embodiment of the present disclosure;
[0023] Figure 2a is a flow chart of another URL detection method provided according to an embodiment of the present disclosure;
[0024] Figure 2b is an example diagram of a candidate propagation node provided according to an embodiment of the present disclosure;
[0025] Figure 3a is a flow chart of another URL detection method provided according to an embodiment of the present disclosure;
[0026] Figure 3b is a specific example diagram of a node in a label propagation algorithm provided according to an embodiment of the present disclosure;
[0027] Figure 4 is a structural schematic diagram of a URL detection device provided according to an embodiment of the present disclosure;
[0028] Figure 5 It is a block diagram of an electronic device used to implement the URL detection method of the embodiment of the present disclosure. DETAILED DESCRIPTION
[0029] The detection of risky URLs in related technologies mainly includes: First, judging the risk of a URL based on the content of the web page. Obtain the web page content of the URL, match it with keyword rules, or use text classification models such as convolutional neural networks to judge the type of web page content, so as to deal with the risk. Second, perform risky domain name detection. For the domain name dimension, obtain domain name information features such as domain name string, IP address, and behavior, design rules for domain name information features to calculate the risk score of the domain name, so as to assess the risk level of the domain name.
[0030] However, the first method relies on web page content to judge URL risk. The detection method based on the single "point" of web page content itself has defects such as missed detection, undetectable, and insufficient perception. Regarding the missed detection defect, due to the strictness of the detection rules, some risky web pages that do not reach the detection threshold or are seriously disguised in content are missed. These URLs may escape because of the strong confrontation of the content or deliberate evasion of the detection rules; Regarding the undetectable defect, there are a large number of web pages that seem normal in content but actually contain risky links. Although the content itself has no obvious risk, there are huge risks hidden in the links; Regarding the insufficient perception defect, the existing URL detection system only detects the snapshot of the current URL (Uniform Resource Locator), but ignores other external link URLs that may be contained in the snapshot, and will not continue to detect external links. These external links may potentially contain huge risks. The second risk domain name detection can only detect at the domain name level, with a narrow application scope, and the features used only contain direct information about the domain name itself.
[0031] This solution provides a new method for detecting risky websites based on knowledge graph technology and label propagation algorithm. The label propagation algorithm can select LPA (Label Propagation Algorithm) algorithm, SLPA (Speaker-Listener Label Propagation Algorithm, speaker listener label propagation) algorithm, or BMLPA (Balance Multi-label Propagation Algorithm, balanced multi-label propagation) algorithm, etc. This solution utilizes the high efficiency of knowledge graph in relational calculation, and combines the label propagation algorithm to detect risky websites according to the association relationship between different nodes. On the basis of risk detection based on web page content, the dimension of association relationship is introduced, which improves the recall ability of potential risky websites. The following embodiment is explained by taking the selection of SLPA algorithm as an example.
[0032] Figure 1 This is a flow chart of a URL detection method provided according to an embodiment of the present disclosure. The method is applicable to the case of detecting risky URLs. The method can be executed by a URL detection device, which can be implemented in software and / or hardware and can be integrated into an electronic device.
[0033] like Figure 1 As shown, the URL detection method of this embodiment may include:
[0034] S101, obtaining the entity node corresponding to the URL to be detected in the URL relationship knowledge graph as a monitoring node in the label propagation algorithm;
[0035] S102, selecting a candidate propagation node for the listening node from the entity nodes according to the edge relationship between different entity nodes and the first propagation relationship in the propagation configuration parameter;
[0036] S103, determining a candidate label of the listening node according to the second label of the candidate propagation node and the second propagation relationship between the candidate propagation node and the listening node;
[0037] S104: Determine a detection result of the to-be-detected website according to the candidate tags of the monitoring node.
[0038] In the disclosed embodiments, the web page content of a URL can be obtained, and risk detection can be performed on the web page content to obtain the risk attributes of the URL; based on the knowledge graph technology, the URL is used as an entity node, and the risk attributes of the URL are used as labels. The information of multiple data sources is integrated to obtain the edge relationships between different URLs, and a URL relationship knowledge graph is constructed based on the entity nodes, labels, and edge relationships.
[0039] The propagation configuration parameters are used to configure the propagation node (Speaker) in the SLPA algorithm, including the first propagation relationship that can propagate labels, that is, the effective propagation relationship, which can be configured according to the target type that needs to be identified in the actual application. If the edge relationship between other nodes and the listening node belongs to the first propagation relationship, the other nodes can propagate labels to the listening node, that is, the other nodes can be candidate propagation nodes for the listening node; if the edge relationship between other nodes and other nodes does not belong to the first propagation relationship, the other nodes cannot propagate labels to the listening node. In a network with many nodes and complex relationships, by configuring the first propagation relationship in the propagation configuration parameters and selecting candidate propagation nodes for the listening node according to the first propagation relationship, unnecessary or redundant label propagation can be reduced, thereby improving the efficiency and accuracy of label propagation. Moreover, allowing users to configure the first propagation relationship according to specific application scenarios and requirements can enhance the flexibility of label propagation, thereby adapting to different types of network structures and community division requirements.
[0040] Each candidate propagation node can propagate the label to the listening node according to its own second label and the second propagation relationship between itself and the listening node, and according to a certain propagation strategy, obtain the candidate label obtained by the listening node from the candidate propagation node; the candidate labels obtained by the listening node in different iterations of the label propagation algorithm can be processed to obtain the final target label of the listening node, and the target label of the listening node is used as the risk attribute of the corresponding website to be detected, that is, the risk type and risk level of the website to be detected are obtained. According to the edge relationship between different entity nodes in the website relationship knowledge graph, the first propagation relationship configured in the label propagation algorithm selects the candidate propagation node that can propagate the label for the listening node, and according to the second label of the candidate propagation node and the second propagation relationship between the candidate propagation node and the listening node, the candidate label propagated by the candidate propagation node to the listening node is determined, so that the listening node updates its own label, and the label updated by the listening node is used as the detection result of the website to be detected. By introducing the edge relationship between different websites and the first propagation relationship in the propagation configuration parameters for label propagation based on the webpage content, a new label propagation method is realized, which can improve the recall rate and accuracy of risk website detection.
[0041] The technical solution provided by the embodiment of the present disclosure selects a candidate propagation node that can propagate a label for a listening node based on the edge relationship between different entity nodes in the URL relationship knowledge graph and the first propagation relationship configured in the label propagation algorithm, and determines the candidate label that the candidate propagation node propagates to the listening node based on the second label of the candidate propagation node and the second propagation relationship between the candidate propagation node and the listening node, so that the listening node updates its own label, and uses the label updated by the listening node as the detection result of the URL to be detected, thereby improving the recall rate and accuracy of risky URL detection.
[0042] Figure 2a is a flow chart of another URL detection method provided according to an embodiment of the present disclosure. Figure 2a The URL detection method of this embodiment may include:
[0043] S201, obtaining the entity node corresponding to the URL to be detected in the URL relationship knowledge graph as a monitoring node in the label propagation algorithm;
[0044] S202, acquiring neighbor nodes of a listening node and edge relationships between the listening node and the neighbor nodes from the URL relationship knowledge graph;
[0045] S203, if the edge relationship between the listening node and the neighboring node successfully matches any first propagation relationship in the propagation configuration parameter, taking the neighboring node as the candidate propagation node;
[0046] S204, determining a candidate label of the listening node according to the second label of the candidate propagation node and the second propagation relationship between the candidate propagation node and the listening node;
[0047] S205: Determine a detection result of the to-be-detected website according to the candidate label of the monitoring node.
[0048] In the URL relationship knowledge graph, the listening node is connected to the neighboring node, and the neighboring nodes of the listening node and the edge relationship between the listening node and the neighboring node can be obtained from the URL relationship knowledge graph; the edge relationship is used to characterize the switching relationship between the URLs corresponding to the listening node and the neighboring node. For each neighboring node of the listening node, the edge relationship between the listening node and the neighboring node can be matched with each valid first propagation relationship in the propagation configuration parameter; if the corresponding edge relationship successfully matches any first propagation relationship, the neighboring node is a candidate propagation node; if the corresponding edge relationship fails to match each first propagation relationship, the neighboring node does not propagate the label to the listening node.
[0049] Figure 2b is an example diagram of a candidate propagation node provided according to an embodiment of the present disclosure. Figure 2b, the listening node L has neighbor nodes A, B, C, and D, and the edge relationships between the listening node L and neighbor nodes A, B, and D all have first propagation relationships that match successfully, but the edge relationship between the listening node L and neighbor node C fails to match each first propagation relationship. Therefore, neighbor nodes A, B, and D are candidate propagation nodes for the listening node L, and neighbor node C does not propagate labels to the listening node L. By configuring the first propagation relationship in the propagation configuration parameters and selecting candidate propagation nodes for the listening node according to the first propagation relationship, unnecessary or redundant label information propagation can be reduced, thereby improving the efficiency and accuracy of label propagation.
[0050] In an optional implementation, the first propagation relationship includes a propagation type and a propagation direction, the propagation type includes at least one of the following: nesting, jumping, or external linking; the propagation direction is reverse propagation.
[0051] By using nesting (iframe), jump, external link and other URL association relationships to determine whether a URL has potential risks, the recall rate of malicious URLs is greatly improved; in addition, other URL association relationships can be used for label propagation, such as the relationship of shared IP addresses. The first propagation relationship includes not only effective propagation types, but also effective propagation directions, and reverse propagation can be used as an effective propagation direction. Through research, it is found that risky websites may try to disguise themselves by linking to well-known or trusted secure websites to increase their credibility or evade detection; the probability of secure websites linking to risky websites is low. By using reverse propagation as an effective propagation direction and shielding forward propagation, it is possible to prevent risky URLs from propagating their own risk labels to secure URLs, thereby contaminating secure URLs and reducing the false call rate of secure URLs.
[0052] The technical solution provided by the embodiment of the present disclosure matches the propagation type and propagation direction between the listening node and the neighboring nodes with each first propagation relationship in the propagation configuration parameters, and uses the neighboring nodes with successfully matched relationships as candidate propagation nodes; and, based on the propagation type and propagation direction between the listening node and each candidate propagation node, label propagation is performed, thereby further improving the efficiency and accuracy of label propagation.
[0053] In an optional implementation, the URL relationship knowledge graph is constructed in the following manner: obtaining the web page content of each URL; based on a text classification model, performing risk identification on the web page content to obtain risk attributes of the URL; wherein the risk attributes include risk type and / or risk level; taking the URL as an entity node, the association relationship between different URLs as an edge relationship, and taking the risk attribute as a label of the entity node to obtain the URL relationship knowledge graph.
[0054] Before constructing the URL relationship knowledge graph, the risk attributes of the URL can be constructed. Exemplarily, the webpage content of each URL is obtained, and the risk attributes of the webpage content are identified based on the text classification model to obtain the risk attributes of the URL. For each risk type, it can also be finely divided into high, medium and low risk levels to obtain a risk attribute library of the URL, that is, a risk label library of the URL, which is used to provide seed labels for label propagation of the label propagation algorithm. The text classification network includes but is not limited to using a text convolutional neural network (textCNN).
[0055] In the process of constructing the URL relationship knowledge graph, information from multiple data sources can be integrated, and the entity nodes, edge relationships and attributes of the knowledge graph can be designed using knowledge modeling technology. Entity nodes may include URLs of various levels such as primary domain names, subdomain names, and URLs (Uniform Resource Locators). Exemplarily, each original URL in the URL database is obtained, and each original address is parsed to determine whether each original address has URLs of various levels such as primary domain names, subdomain names, and URLs. If any level is missing, the level is supplemented. Take the following URL as an example: www.zhuyuming.com.cn. / sheet / ABCDE? tab=123&_t=45678. Through hierarchical supplementation, the main domain name: zhuyuming.com.cn; subdomain name: www.zhuyuming.com.cn can be obtained. By using URLs of each level as entity nodes, the coverage rate of entity nodes is improved. In addition, we obtain nesting, jump or external link relationships between different URLs as edge relationships, write entity nodes and edge relationships into the graph database based on entity relationship extraction technology, and use the risk attributes of URLs as labels of entity nodes to obtain a URL relationship knowledge graph. When other URL relationships such as shared IP addresses are used for label propagation, the edge relationship can also be a shared IP address.
[0056] By characterizing the content dimension of the URL, the risk attributes of the URL are obtained. On this basis, entity nodes, edge relationships and labels are designed, and the URL relationship knowledge graph is constructed using entity relationship extraction technology. Moreover, by taking the main domain name, subdomain name, and URL of different levels as entity nodes, it is convenient to perform risk detection on URLs at each level in the future, thereby improving the coverage of risk URL detection. By introducing edge relationships such as nesting, jumping, and external links between URLs, it is convenient to combine various edge relationships for label propagation, thereby further improving the recall capability of risky URLs.
[0057] In an optional implementation, after selecting a candidate propagation node for the listening node from the entity node, the method further includes: initializing labels for the listening node and each of the candidate propagation nodes in the label propagation algorithm using the risk attribute of the entity node.
[0058] By using the risk attributes of the entity nodes to initialize the labels of each listening node and each candidate propagation node in the label propagation algorithm, the initial labels of each listening node and each candidate propagation node are obtained, so that the initial labels are accurate and reliable. By using the risk attributes as the seeds of label propagation, the efficiency of label propagation can be improved.
[0059] In an optional implementation manner, after determining the detection result of the to-be-detected website, the method further includes: if any risky website belongs to a preset whitelist website, determining that the detection result of the website is set to be risk-free.
[0060] In the disclosed embodiment, a list of known safe and trustworthy websites can also be regularly updated and maintained as whitelisted websites. The whitelisted websites are used to filter websites with risky labels obtained through the label propagation algorithm. If any website with a risky address belongs to the whitelisted website, the detection result of the website is set to risk-free. By using the whitelisted website to filter the websites that are mistakenly recalled by the label propagation algorithm, the accuracy of risky website detection is further improved.
[0061] Figure 3a is a flow chart of another URL detection method provided according to an embodiment of the present disclosure. Figure 3a The URL detection method of this embodiment may include:
[0062] S301, obtaining the entity node corresponding to the URL to be detected in the URL relationship knowledge graph as a monitoring node in the label propagation algorithm;
[0063] S302, selecting a candidate propagation node for the listening node from the entity nodes according to the edge relationship between different entity nodes and the first propagation relationship in the propagation configuration parameter;
[0064] S303, based on the first propagation label, propagation condition and output policy label in each propagation strategy, as well as the second propagation relationship between the candidate propagation node and the listening node, and the second label of each candidate propagation node, determine whether the propagation strategy propagates the policy label to the listening node;
[0065] S304, determining a candidate label of the listening node according to the policy label obtained by the listening node;
[0066] S305: Determine a detection result of the to-be-detected website according to the candidate label of the monitoring node.
[0067] The propagation strategy is used to configure the listening node (Listener) in the label propagation algorithm, and may include at least one first propagation label, a propagation condition, and an output policy label. The propagation condition is used to determine whether the propagation strategy propagates the corresponding policy label to the listening node. Exemplarily, it can be determined whether the propagation condition of the propagation strategy is met based on the first propagation label in each propagation strategy, the second label of each candidate propagation node, and the second propagation relationship between the candidate propagation node and the listening node; if the propagation condition of the propagation strategy is met, the propagation strategy propagates its own policy label to the listening node; otherwise, it does not propagate.
[0068] It should be noted that there may be multiple propagation strategies, and the first propagation labels, propagation conditions, and output policy labels in various propagation strategies can be flexibly configured. For any propagation strategy, the policy label output by the propagation strategy may belong to the corresponding first propagation label, or it may be a new type of label other than the first propagation label. Exemplarily, for a propagation strategy with a first propagation label {a, b, c}, its policy label may be d; for a propagation strategy with a first propagation label {a, b}, its policy label may be e. By configuring the output of the propagation strategy to be a new type of policy label in addition to the first propagation label, the label type can be dynamically expanded, so that the label propagation algorithm can not only propagate existing labels, but also propagate new types of labels, and the risk types of the URLs are more comprehensive; moreover, by propagating labels in units of propagation strategies, compared with propagating in units of labels, the scope of label propagation is wider and the scalability is stronger.
[0069] In an optional implementation, the propagation condition is an importance threshold; the first propagation label in each propagation strategy, the propagation condition and the output strategy label, as well as the second propagation relationship between the candidate propagation node and the listening node, the second label of each candidate propagation node, determines whether the propagation strategy propagates the strategy label to the listening node, including: matching the second label of each candidate propagation node with the first propagation label in the propagation strategy to obtain a target propagation node that hits the propagation strategy; determining the importance of the propagation strategy based on the second propagation relationship between the target propagation node and the listening node and the importance of the first propagation relationship in the propagation configuration parameters; if the importance of the propagation strategy is greater than the importance threshold, the propagation strategy propagates the strategy label to the listening node.
[0070] The propagation configuration parameters include not only the effective first propagation relationship, but also the importance of the first propagation relationship. For each candidate propagation node of the listening node, the second label of the candidate propagation node can be matched with the first propagation label in the propagation strategy. If the second label successfully matches any first propagation label, the candidate propagation node is the target propagation node that hits the propagation strategy; otherwise, the propagation strategy is not hit. For each target propagation node that hits the propagation strategy, the importance of the target propagation node is determined based on the second propagation relationship between the target propagation node and the listening node, and the importance of the first propagation relationship; the importance of each target propagation node that hits the propagation strategy is statistically analyzed to obtain the importance of the propagation strategy; if the importance of the propagation strategy is greater than the importance threshold of the propagation strategy, the propagation strategy propagates the output policy label to the listening node; otherwise, the propagation strategy does not output the policy label.
[0071] By setting different importances for various first propagation relationships and performing label propagation in units of propagation strategies, the flexibility of label propagation is further improved compared to using the same importance for each propagation node and performing label propagation in units of labels. By flexibly configuring the propagation strategy and propagation configuration parameters according to risk detection requirements, the matching degree between risky URL detection and risk detection requirements can be further improved. Exemplarily, the importance of the first propagation relationship "nesting and reverse" and the first propagation relationship "jump and reverse" can both be a first value, and the importance of the first propagation relationship "external link + reverse" can be a second value, and the first value can be greater than the second value. The propagation condition in the propagation strategy can be an importance threshold. By flexibly setting the importance of different first propagation relationships, the flexibility of label propagation can be further improved.
[0072] In an optional implementation, the candidate label of the listening node is determined based on the policy label obtained by the listening node, including: if the listening node obtains policy labels from at least two propagation strategies, then according to the priority of the at least two propagation strategies, each of the policy labels is screened to obtain the candidate label of the listening node.
[0073] In the case of multiple propagation strategies, the propagation strategies also have configuration information such as strategy name and priority. If at least two propagation strategies propagate corresponding strategy labels to the listening node, the strategy labels with relatively low priorities can be filtered according to the priorities of each propagation strategy, and the strategy labels with relatively high priorities can be retained to obtain candidate labels for the listening node, further improving the accuracy of label propagation.
[0074] The following example provides a specific example of label propagation, taking the case where the propagation strategy parameters may include the following first propagation relationships: "Nested and Reverse", "Jump and Reverse", and "External Link + Reverse", and the importance of each first propagation relationship is 2, 2, and 1, respectively. Propagation strategy 1 has the following first propagation labels {a, b, c}, the importance threshold is 4, and the strategy label to be output is d; propagation strategy 2 has the following first propagation labels {a, b}, the importance threshold is 3, and the strategy label to be output is e, and the priority of propagation strategy 1 is higher than that of propagation strategy 2. Figure 3b is a specific example diagram of a node in a label propagation algorithm provided according to an embodiment of the present disclosure, with reference to Figure 3b , the listening node L has candidate propagation nodes S1, S2, S3, and S4. The second labels of the candidate propagation nodes S1, S2, S3, and S4 are a, a, b, and c respectively. The second propagation relationships between the candidate propagation nodes S1, S2, S3, and S4 and the listening node L are "nested and reverse", "external link + reverse", "jump and reverse", and "external link + reverse" as examples.
[0075] For propagation strategy 1, the second labels of candidate propagation nodes S1, S2, S3, and S4 all hit the first propagation label of propagation strategy 1, so they are all target propagation nodes of propagation strategy 1; based on the importance of the second propagation relationship and the first propagation relationship, the importance of target propagation nodes S1, S2, S3, and S4 are 2, 1, 2, and 1, respectively, so the importance of propagation strategy 1 is 6. Since the importance of propagation strategy 1 is greater than the importance threshold of propagation strategy 1, propagation strategy 1 outputs strategy label d to the listening node. For propagation strategy 2, the second labels of candidate propagation nodes S1, S2, and S3 all hit the first propagation label of propagation strategy 2, while the second label of candidate propagation node S4 does not hit propagation strategy 2, so candidate propagation nodes S1, S2, and S3 are target propagation nodes of propagation strategy 2; based on the importance of the second propagation relationship and the first propagation relationship, the importance of target propagation nodes S1, S2, and S3 are 2, 1, and 2, respectively, so the importance of propagation strategy 2 is 5. Since the importance of propagation strategy 2 is greater than the importance threshold of propagation strategy 2, propagation strategy 2 outputs strategy label e to the listening node. Considering that propagation strategy 1 has a higher priority than propagation strategy 2, the candidate label of the listening node is d.
[0076] In an optional implementation, the detection result of the URL to be detected is determined based on the candidate tags of the listening node, including: obtaining the candidate tags of the listening node in different iteration rounds and the propagation frequency of the candidate tags; obtaining the risk priority of each candidate tag; screening the candidate tags according to the propagation frequency and / or risk priority of each candidate tag, and using the remaining candidate tags as the detection result of the URL to be detected.
[0077] Exemplarily, a memory may be maintained for each listening node to store the candidate labels received during the iteration of the label propagation algorithm and their corresponding propagation frequencies. When the label propagation algorithm converges or the traversal reaches the maximum number of iterations, the candidate labels of the listening node may be screened according to the propagation frequency of each candidate label, and the candidate label with the highest propagation frequency may be used as the target label of the listening node. In addition, a risk priority may be set for each candidate label, and the propagation frequency and frequency threshold of the candidate label may be compared to retain the candidate label with a propagation frequency greater than the frequency threshold; according to the risk priority of each retained candidate label, the candidate label with the highest risk priority may be used as the target label of the listening node. The target label of the listening node is used as the risk attribute of the corresponding website to be detected to obtain the risk type and risk level of the website to be detected. The target label of the listening node is obtained by screening each candidate label according to the propagation frequency and risk priority of each candidate label, thereby further improving the accuracy of website detection.
[0078] The technical solution provided by the embodiment of the present disclosure utilizes the label propagation algorithm to configure propagation configuration parameters and propagation strategies to propagate risk labels, monitors and intercepts risky URLs based on updated label information and combined with whitelist filtering, implements a new risky URL detection method using edge relationships, and improves the recall rate and accuracy of risky URL detection.
[0079] Figure 4 1 is a schematic diagram of a URL detection device provided according to an embodiment of the present disclosure. The device is suitable for detecting risky URLs. The device can be implemented in software and / or hardware and can be integrated into electronic devices. Figure 4 As shown, the URL detection device 400 of this embodiment may include:
[0080] A monitoring node module 410 is used to obtain the entity node corresponding to the URL to be detected in the URL relationship knowledge graph as a monitoring node in the label propagation algorithm;
[0081] A propagation node module 420, configured to select a candidate propagation node for the listening node from the entity nodes according to the edge relationship between different entity nodes and the first propagation relationship in the propagation configuration parameter;
[0082] The label propagation module 430 is used to determine the candidate label of the listening node according to the second label of the candidate propagation node and the second propagation relationship between the candidate propagation node and the listening node;
[0083] The detection result module 440 is used to determine the detection result of the to-be-detected website according to the candidate tags of the monitoring node.
[0084] In an optional implementation, the propagation node module 430 includes:
[0085] An edge relationship unit, used for obtaining neighbor nodes of a listening node and edge relationships between the listening node and the neighbor nodes from the URL relationship knowledge graph;
[0086] The propagation node unit is used to use the neighbor node as the candidate propagation node if the edge relationship between the monitoring node and the neighbor node successfully matches any first propagation relationship in the propagation configuration parameters.
[0087] In an optional implementation, the first propagation relationship includes a propagation type and a propagation direction, the propagation type includes at least one of the following: nesting, jumping, or external linking; the propagation direction is reverse propagation.
[0088] In an optional implementation, the label propagation module 430 includes:
[0089] A strategy propagation unit, configured to determine whether the propagation strategy propagates the strategy label to the listening node based on the first propagation label, the propagation condition and the output strategy label in each propagation strategy, the second propagation relationship between the candidate propagation node and the listening node, and the second label of each candidate propagation node;
[0090] The candidate label unit is used to determine the candidate label of the monitoring node according to the policy label obtained by the monitoring node.
[0091] In an optional implementation manner, the propagation condition is an importance threshold; and the strategy propagation unit includes:
[0092] A label matching subunit, used for matching the second label of each candidate propagation node with the first propagation label in the propagation strategy to obtain a target propagation node that hits the propagation strategy;
[0093] An importance subunit, configured to determine the importance of the propagation strategy according to the second propagation relationship between the target propagation node and the monitoring node and the importance of the first propagation relationship in the propagation configuration parameter;
[0094] The strategy propagation subunit is used to propagate the strategy label to the monitoring node if the importance of the propagation strategy is greater than the importance threshold.
[0095] In an optional implementation manner, the candidate tag unit is specifically used to:
[0096] If the listening node obtains policy labels from at least two propagation strategies, the policy labels are screened according to the priorities of the at least two propagation strategies to obtain a candidate label for the listening node.
[0097] In an optional implementation, the detection result module 440 includes:
[0098] A propagation frequency unit, used to obtain candidate tags and propagation frequencies of candidate tags of the monitoring node in different iteration rounds;
[0099] A priority unit is used to obtain the risk priority of each candidate label;
[0100] The detection result unit is used to screen the candidate tags according to the propagation frequency and / or risk priority of each candidate tag, and use the remaining candidate tags as the detection result of the website to be detected.
[0101] In an optional implementation, the URL relationship knowledge graph is constructed in the following manner:
[0102] Get the webpage content of each URL;
[0103] Based on the text classification model, risk identification is performed on the webpage content to obtain the risk attribute of the website; wherein the risk attribute includes risk type and / or risk level;
[0104] The URLs are taken as entity nodes, the association relationships between different URLs are taken as edge relationships, and the risk attributes are taken as labels of the entity nodes to obtain the URL relationship knowledge graph.
[0105] In an optional implementation manner, the above-mentioned website detection device 400 further includes:
[0106] The initialization module is used to use the risk attribute of the entity node to initialize the label of the monitoring node and each candidate propagation node in the label propagation algorithm.
[0107] In an optional implementation manner, the above-mentioned website detection device 400 further includes:
[0108] The risk-free module is used to determine that the detection result of any risky URL is set to risk-free if it belongs to a preset whitelist URL.
[0109] The technical solution of the disclosed embodiment detects risky URLs based on the knowledge graph technology and label propagation algorithm, combined with the edge relationship between URLs, on the basis of web page content detection, and makes up for the shortcomings of risk detection based on web page content in the following three aspects: 1) Reduce the missed detection rate, and recall the web pages that have risky content but have not been detected through edge relationships, thereby reducing missed detections; 2) Reduce blind spots, and detect URL addresses that seem to have normal content but actually contain risky external links, thereby reducing blind spots that cannot be detected in the content dimension; 3) Expand the perception surface, improve the detection ability of risky external links contained in web page addresses, and be able to identify potential risks that are not perceived in the content dimension, and achieve more comprehensive security protection. The above solution provides important technical support for network security work.
[0110] In the technical solution disclosed herein, the acquisition, storage and application of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.
[0111] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.
[0112] Figure 5 It is a block diagram of an electronic device used to implement the URL detection method of the embodiment of the present disclosure. Figure 5 Schematic block diagram of an example electronic device 500 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as a laptop.
[0113] Computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. Electronic devices may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.
[0114] like Figure 5 As shown, the electronic device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic device 500 can also be stored. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.
[0115] Multiple components in the electronic device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the electronic device 500 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.
[0116] The computing unit 501 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), an image processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 501 performs the various methods and processes described above, such as the URL detection method. For example, in some embodiments, the URL detection method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 508. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 500 via ROM 502 and / or communication unit 509. When the computer program is loaded into RAM 503 and executed by the computing unit 501, one or more steps of the URL detection method described above may be performed. Alternatively, in other embodiments, the computing unit 501 may be configured to perform the URL detection method in any other appropriate manner (e.g., by means of firmware).
[0117] Figure 5 A schematic block diagram of an example electronic device 500 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.
[0118] like Figure 5As shown, the electronic device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic device 500 can also be stored. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.
[0119] Multiple components in the electronic device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the electronic device 500 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.
[0120] The computing unit 501 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), an image processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 501 performs the various methods and processes described above, such as the URL detection method. For example, in some embodiments, the URL detection method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 508. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 500 via ROM 502 and / or communication unit 509. When the computer program is loaded into RAM 503 and executed by the computing unit 501, one or more steps of the URL detection method described above may be performed. Alternatively, in other embodiments, the computing unit 501 may be configured to perform the URL detection method in any other appropriate manner (e.g., by means of firmware).
[0121] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
[0122] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.
[0123] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0124] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, audio input, or tactile input).
[0125] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communications network). Examples of communications networks include: a local area network (LAN), a wide area network (WAN), and the Internet.
[0126] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.
[0127] Artificial intelligence is a discipline that studies how to use computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.). It includes both hardware-level and software-level technologies. Artificial intelligence hardware technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, and big data processing; artificial intelligence software technologies mainly include computer vision technology, audio recognition technology, natural language processing technology, machine learning / deep learning technology, big data processing technology, knowledge graph technology, and other major directions.
[0128] Cloud computing refers to a technology system that uses network access to elastically scalable shared physical or virtual resource pools. Resources can include servers, operating systems, networks, software, applications, and storage devices, and can be deployed and managed on demand and in a self-service manner. Cloud computing technology can provide efficient and powerful data processing capabilities for technical applications such as artificial intelligence and blockchain, as well as model training.
[0129] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.
[0130] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.
Claims
1. A URL detection method, comprising: Obtain the entity node corresponding to the URL to be detected in the URL relationship knowledge graph as the monitoring node in the label propagation algorithm; Selecting a candidate propagation node for the listening node from the entity nodes according to the edge relationship between different entity nodes and the first propagation relationship in the propagation configuration parameter; Determine a candidate label of the listening node according to the second label of the candidate propagation node and the second propagation relationship between the candidate propagation node and the listening node; The detection result of the to-be-detected website is determined according to the candidate tags of the monitoring node.
2. The method according to claim 1, wherein: The selecting a candidate propagation node for the listening node from the entity nodes according to the edge relationship between different entity nodes and the first propagation relationship in the propagation configuration parameter comprises: Obtaining neighbor nodes of a listening node and edge relationships between the listening node and the neighbor nodes from the URL relationship knowledge graph; If the edge relationship between the listening node and the neighboring node successfully matches any first propagation relationship in the propagation configuration parameters, the neighboring node is used as the candidate propagation node.
3. The method according to claim 2, wherein: The first propagation relationship includes a propagation type and a propagation direction, the propagation type includes at least one of the following: nesting, jumping or external link; the propagation direction is reverse propagation.
4. The method according to claim 1, wherein: The determining, according to the second label of the candidate propagation node and the second propagation relationship between the candidate propagation node and the listening node, the candidate label of the listening node comprises: Based on the first propagation label, the propagation condition and the outputted policy label in each propagation strategy, as well as the second propagation relationship between the candidate propagation node and the listening node, and the second label of each candidate propagation node, determining whether the propagation strategy propagates the policy label to the listening node; According to the policy label obtained by the monitoring node, a candidate label of the monitoring node is determined.
5. The method according to claim 4, wherein: The propagation condition is an importance threshold; the determining whether the propagation strategy propagates the policy label to the listening node based on the first propagation label, the propagation condition and the output policy label in each propagation strategy, the second propagation relationship between the candidate propagation node and the listening node, and the second label of each candidate propagation node includes: Matching the second label of each candidate propagation node with the first propagation label in the propagation strategy to obtain a target propagation node that hits the propagation strategy; Determine the importance of the propagation strategy according to the second propagation relationship between the target propagation node and the listening node and the importance of the first propagation relationship in the propagation configuration parameter; If the importance of the propagation strategy is greater than the importance threshold, the propagation strategy propagates the strategy label to the listening node.
6. The method according to claim 4, wherein: The step of determining the candidate label of the listening node according to the policy label obtained by the listening node includes: If the listening node obtains policy labels from at least two propagation strategies, the policy labels are screened according to the priorities of the at least two propagation strategies to obtain a candidate label for the listening node.
7. The method according to claim 6, wherein: The step of determining the detection result of the to-be-detected website according to the candidate tag of the monitoring node includes: Obtaining candidate tags and propagation frequencies of candidate tags of the monitoring node in different iteration rounds; Get the risk priority of each candidate label; The candidate tags are screened according to the propagation frequency and / or risk priority of each candidate tag, and the remaining candidate tags are used as the detection result of the website to be detected.
8. The method according to any one of claims 1 to 7, wherein: The URL relationship knowledge graph is constructed in the following way: Get the webpage content of each URL; Based on the text classification model, risk identification is performed on the webpage content to obtain the risk attribute of the website; wherein the risk attribute includes risk type and / or risk level; The URLs are taken as entity nodes, the association relationships between different URLs are taken as edge relationships, and the risk attributes are taken as labels of the entity nodes to obtain the URL relationship knowledge graph.
9. The method according to claim 8, after selecting a candidate propagation node for the listening node from the entity node, further comprising: The risk attribute of the entity node is used to initialize the labels of the monitoring node and each of the candidate propagation nodes in the label propagation algorithm.
10. The method according to claim 1, characterized in that After determining the detection result of the to-be-detected website, the method further includes: If any risky URL belongs to the preset whitelist URL, the detection result of the URL is determined to be set to no risk.
11. A website detection device, comprising: The monitoring node module is used to obtain the entity node corresponding to the URL to be detected in the URL relationship knowledge graph as the monitoring node in the label propagation algorithm; A propagation node module, configured to select a candidate propagation node for the listening node from the entity nodes according to the edge relationship between different entity nodes and the first propagation relationship in the propagation configuration parameter; a label propagation module, configured to determine a candidate label of the listening node according to the second label of the candidate propagation node and a second propagation relationship between the candidate propagation node and the listening node; The detection result module is used to determine the detection result of the URL to be detected according to the candidate label of the monitoring node.
12. The device according to claim 11, wherein: The propagation node module includes: An edge relationship unit, used for acquiring neighbor nodes of a listening node and edge relationships between the listening node and the neighbor nodes from the URL relationship knowledge graph; The propagation node unit is used to use the neighbor node as the candidate propagation node if the edge relationship between the monitoring node and the neighbor node successfully matches any first propagation relationship in the propagation configuration parameters.
13. The device according to claim 12, wherein: The first propagation relationship includes a propagation type and a propagation direction, the propagation type includes at least one of the following: nesting, jumping or external link; the propagation direction is reverse propagation.
14. The device according to claim 11, wherein: The label propagation module includes: A strategy propagation unit, configured to determine whether the propagation strategy propagates the strategy label to the listening node based on the first propagation label, the propagation condition and the output strategy label in each propagation strategy, the second propagation relationship between the candidate propagation node and the listening node, and the second label of each candidate propagation node; The candidate label unit is used to determine the candidate label of the monitoring node according to the policy label obtained by the monitoring node.
15. The device according to claim 14, wherein: The propagation condition is an importance threshold; The strategy propagation unit comprises: A label matching subunit, used for matching the second label of each candidate propagation node with the first propagation label in the propagation strategy to obtain a target propagation node that hits the propagation strategy; An importance subunit, configured to determine the importance of the propagation strategy according to the second propagation relationship between the target propagation node and the monitoring node and the importance of the first propagation relationship in the propagation configuration parameter; The strategy propagation subunit is used to propagate the strategy label to the monitoring node if the importance of the strategy is greater than the importance threshold.
16. The device according to claim 14, wherein: The candidate label unit is specifically used for: If the listening node obtains policy labels from at least two propagation strategies, the policy labels are screened according to the priorities of the at least two propagation strategies to obtain a candidate label for the listening node.
17. The device according to claim 16, wherein: The detection result module includes: A propagation frequency unit, used to obtain candidate tags and propagation frequencies of candidate tags of the monitoring node in different iteration rounds; A priority unit, used to obtain the risk priority of each candidate label; The detection result unit is used to screen the candidate tags according to the propagation frequency and / or risk priority of each candidate tag, and use the remaining candidate tags as the detection result of the website to be detected.
18. The device according to any one of claims 11 to 17, wherein: The URL relationship knowledge graph is constructed in the following way: Get the webpage content of each URL; Based on the text classification model, risk identification is performed on the webpage content to obtain the risk attribute of the website; wherein the risk attribute includes risk type and / or risk level; The URLs are taken as entity nodes, the association relationships between different URLs are taken as edge relationships, and the risk attributes are taken as labels of the entity nodes to obtain the URL relationship knowledge graph.
19. The apparatus according to claim 18, further comprising: The initialization module is used to use the risk attribute of the entity node to initialize the label of the monitoring node and each candidate propagation node in the label propagation algorithm.
20. The device according to claim 11, characterized in that The device also includes: The risk-free module is used to determine that the detection result of any risky URL is set to risk-free if it belongs to a preset whitelist URL.
21. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 10.
22. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to make a computer execute the method according to any one of claims 1-10.
23. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 10.
