Content operation supervision system of intensive website platform

Abstract

The invention discloses a content operation supervision system of an intensive website platform, which relates to the technical field of content supervision and comprises an intention modeling module for generating an identity inertia chain by recording a traceable access path of a user and extracting an access starting point, a jump structure and a drop point from the identity inertia chain to form an access intention baseline; the context alignment module is used for performing structured arrangement on the current browsing path of the user corresponding to the access intention baseline to obtain a content context chain, and marking the disjunction position of the content context chain and the access intention baseline as a context fault chain; and the risk propagation module tracks all reachable content nodes downstream for the context fault chain, and constructs a fault propagation graph. Double recognition is performed through the structure jump sequence and the page semantic consistency, so that abnormal behaviors such as jump direction reversal, hierarchical jump, short-period rollback and theme mutation can be positioned, and misjudgment is remarkably reduced.

Description

Technical Field

[0001] This invention relates to the field of content supervision technology, and in particular to an integrated content operation and supervision system for website platforms. Background Technology

[0002] As the content structure of centralized website platforms becomes increasingly sophisticated, navigation paths, recommended content links, and resource loading links between pages exhibit complex characteristics of multiple levels, triggers, and redirects. Traditional content operation and supervision primarily focus on static analysis of user interaction logs and page access records, identifying risky content tags, abnormal access frequencies, or sensitive access areas to achieve content supervision. In existing business practices, behavioral trajectory analysis, path graph modeling, and user profiling methods are also widely adopted to extract preference directions from user access behavior, supporting content recommendation, access quality assessment, and the tracing of violations.

[0003] However, in actual operation, user access paths are often affected by factors such as mixed interaction paths, topic branching, and the intervention of recommended entry points. Although the access process remains within the website system, it may have deviated from the user's true intention trajectory. Therefore, the focus of supervision is no longer just the risky content itself, but rather when, where, and how the access process deviates from the user's intended path. Traditional supervision focuses on page-level detection, content recognition, or rule triggering, making it difficult to identify access path breaks, contextual disconnects, and cascading propagation trends. It also cannot restore the user's intended path in real time when abnormal access occurs. Figure 1 The context path. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a content operation and supervision system for an integrated website platform to solve the problem of difficulty in accurately identifying deviations in access path intent and performing context recovery.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides an integrated content operation and supervision system for website platforms, comprising:

[0008] The intent modeling module generates an identity inertial chain by recording the user's traceable access path, and extracts the access start point, jump structure and landing point from it to form the access intent baseline;

[0009] The context alignment module organizes the user's current browsing path corresponding to the access intent baseline in a structured manner to obtain the content context chain, and marks the position where the content context chain is disconnected from the access intent baseline as the context break chain.

[0010] The risk propagation module tracks all reachable content nodes downstream of the context fault chain, constructs a fault propagation graph, and automatically aggregates the set of content nodes continuously pointed to by the same fault chain according to the guidance path of the context fault chain to generate a risk path set.

[0011] The path recovery module determines the upstream fault node in the fault propagation graph based on the risk path set, blocks access to the content pointed to by the fault node, and switches the access path to the contextual complete path corresponding to the access intent baseline.

[0012] As a preferred embodiment of the content operation and supervision system for the centralized website platform described in this invention, the specific steps for generating an identity inertial chain by recording the user's traceable access path are as follows:

[0013] When a user successfully logs in, the user's session identifier is obtained, and the page access behavior in the session is recorded. Each page jump is organized into access path fragments and strung together in chronological order to form a complete access path.

[0014] The complete access paths formed by the same user in multiple sessions are aggregated and associated with the user's identity identifier to obtain the identity inertial chain.

[0015] As a preferred embodiment of the content operation and supervision system for the integrated website platform described in this invention, the extraction of the access starting point, redirection structure, and landing point constitutes an access intent baseline. The specific steps are as follows:

[0016] Statistical analysis is performed on each access path in the identity inertia chain, and access paths that have reached a preset access count threshold and have a stable page jump order are selected as a set of stable access paths.

[0017] Extract the starting page identifier, page jump order, and ending page identifier of each access path from the set of stable access paths, and combine them according to the access order to generate an access intent baseline.

[0018] As a preferred embodiment of the content operation and supervision system for the centralized website platform described in this invention, the specific steps for structuring and organizing the user's current browsing path corresponding to the baseline of access intent are as follows:

[0019] Record the page identifiers visited sequentially, the page entry positions, and the interactive controls that trigger page navigation in the user's current session to form a record of the current browsing path;

[0020] Extract adjacent page pairs from the current browsing path record, associate them with the interactive control that triggered the jump and the page entry location, and form a jump relationship entry;

[0021] The jump relationship entries are compared with path segments in the access intent baseline that have the same start and end points, and then linked together according to the access time order and page level to construct a content context chain.

[0022] As a preferred embodiment of the content operation and supervision system for the integrated website platform described in this invention, the specific steps for marking the location where the content context chain and the access intent baseline are disconnected as a context break chain are as follows:

[0023] Compare the jump order in the content context chain with the corresponding jump order in the access intent baseline, mark the jump positions with opposite jump directions, jump levels crossed, and repeated backtracking, and form a set of abnormal order positions;

[0024] The page topic tags at each jump location in the content context chain are compared with the page topic tags at the same location in the access intent baseline. Jump locations with inconsistent topics, low relevance, and abrupt topic changes are marked to form a set of semantically abnormal locations.

[0025] The set of sequential abnormal locations and the set of semantic abnormal locations are merged, and adjacent abnormal locations are merged according to the access time to generate the corresponding contextual fault chain.

[0026] As a preferred embodiment of the content operation and supervision system for the integrated website platform described in this invention, the specific steps for constructing the fault propagation graph are as follows:

[0027] Starting from each abnormal position in the context fault chain, the downstream page identifiers and resource identifiers directly connected to the abnormal position are read from the website content management records to form a set of downstream candidate nodes.

[0028] Read the external links, internal jump relationships, and resource reference relationships of each node in the downstream candidate node set, add them to the extended set, and form a hierarchical connection relationship with the downstream candidate node set;

[0029] All nodes in the extended set are organized according to the access level. The connection relationship between the abnormal position in the context fault chain and the reachable nodes at each level is represented as directed edges, and organized into a fault propagation graph.

[0030] As a preferred embodiment of the content operation and supervision system of the integrated website platform described in this invention, the extended set is obtained by expanding the downstream candidate node set step by step through external links, internal jump relationships and resource reference relationships.

[0031] As a preferred embodiment of the content operation and supervision system for the centralized website platform described in this invention, the specific steps for automatically aggregating the set of content nodes continuously pointed to by the same fault chain based on the guiding path of the context fault chain are as follows:

[0032] In the fault propagation map, all nodes directly connected to the same context fault chain are selected and sorted from near to far according to the access level to form a candidate node sequence;

[0033] Identify groups of nodes that are continuously pointed to by the same abnormal location in the candidate node sequence, treat each group of nodes as a complete risk path, and record the starting node, each intermediate node and the ending node on the risk path.

[0034] All risk paths are classified according to context break chain identifier and session identifier, and multiple risk paths belonging to the same context break chain are grouped into a path set;

[0035] Associate all risk paths in each path set with their corresponding context fault chains to generate a risk path set.

[0036] As a preferred embodiment of the content operation and supervision system for the centralized website platform described in this invention, the steps for blocking access to content pointed to by fault nodes and switching the access path to the complete contextual path corresponding to the access intent baseline are as follows:

[0037] Each risk path in each risk path set is analyzed according to the access level, the abnormal location closest to the access origin is identified, and the corresponding node is determined as the upstream fault node of the risk path set.

[0038] Register the page identifier and jump control information associated with the upstream fault node as the object to be blocked, set the page jump operation corresponding to the object to be blocked to a prohibited state, and monitor page jump requests in the session;

[0039] When it is detected that a user access reaches the object to be blocked, query the context complete node in the access intent baseline that matches the current identity and is consistent with the current access topic, and replace the access target node that originally pointed to the object to be blocked with the current context complete node.

[0040] Align the replaced access target node with the subsequent jump order in the access intent baseline to form a context-complete access path.

[0041] As a preferred embodiment of the content operation and supervision system for the centralized website platform described in this invention, the specific steps for analyzing each risk path in each risk path set according to access level are as follows:

[0042] Read the access order of the starting node, each intermediate node and the ending node from each risk path in the risk path set, map it to the access level number, and form a node sequence;

[0043] In the node sequence, based on the abnormal location markers corresponding to the context fault chain, the set of nodes corresponding to the abnormal location is determined, forming an abnormal node sequence;

[0044] In the sequence of abnormal nodes, the abnormal node with the smallest access level number is found in ascending order of access level number and is determined as the upstream fault node.

[0045] The beneficial effects of this invention are as follows: The identity inertia chain formed through multiple user sessions reflects the user's stable access path structure, making deviation judgment individualized and temporally continuous; the context fault chain simultaneously performs dual identification based on structural jump order and page semantic consistency, enabling the location of abnormal behaviors such as jump direction reversal, hierarchical jumps, short-cycle backtracking, and topic abrupt changes, significantly reducing false positives; by recursively tracing reachable nodes downstream through the fault propagation graph, the potential cascading propagation range of abnormal access to the content system can be identified, not only discovering the deviation behavior itself but also revealing content segments that may be affected; through the context path recovery mechanism that blocks the upstream fault node and aligns with the access intent baseline, access is not interrupted but returns to the user's true intent path, achieving risk isolation + uninterrupted access intervention, significantly improving the user experience. Attached Figure Description

[0046] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a schematic diagram of a content operation and supervision system for an integrated website platform.

[0048] Figure 2 Build a flowchart for the baseline of user access intent.

[0049] Figure 3 This is a flowchart of context tomography detection and labeling.

[0050] Figure 4 A flowchart for risk propagation analysis and path generation. Detailed Implementation

[0051] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0052] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0053] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0054] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides an integrated website platform content operation and supervision system, comprising the following steps:

[0055] The intent modeling module generates an identity inertial chain by recording the user's traceable access path, and extracts the access start point, jump structure and landing point from it to form the access intent baseline.

[0056] After a user completes a normal login, the system obtains the user's corresponding session identifier, continuously records page access behavior for the session, and records each process of jumping from the source page to the target page as an access event containing the source page identifier, the target page identifier, the identifier of the interactive control that triggered the jump, and a timestamp. The access events are sorted according to the timestamp order, and the source page identifier in adjacent access events is paired with the target page identifier of the previous access event to form a complete access path from the login start point to the exit end point.

[0057] The complete access paths obtained by the same user in multiple sessions are aggregated, and each complete access path is associated with the corresponding user identity identifier to form an access path set for a single user. The multiple complete access paths in the access path set are arranged in the order of session time as an access node sequence to form an identity inertial chain covering multiple sessions.

[0058] For each complete access path in the identity inertial chain, statistics are compiled, and the number of accesses is taken as the frequency of occurrence of the complete access path; for users whose identity is... In this case, the number of visits for each complete access path is calculated using the following expression:

[0059] ;

[0060] in, Indicates user identity as At that time, the complete access path Number of visits, Indicates user identity as The number of sessions indicates that the user's identity is... At that time, the first Different complete access path styles, Indicates the path style index. Indicates the user's identification number. Represents the session index. Indicates user identity as At that time, in the The complete access path formed in a session For indicator functions, when the first The complete access path in each session and the complete access path The value is 1 if the page navigation order is consistent, otherwise the value is 0.

[0061] Each complete access path in the identity inertia chain is filtered, and complete access paths that have reached a preset access count threshold and whose page jump order remains consistent across multiple sessions are selected as the stable access path set. This ensures that the stable access path set only includes access paths that appear repeatedly in the identity inertia chain and have a stable jump structure.

[0062] It should be noted that the preset access count threshold is set based on historical access statistics. For example, if the statistics show that a certain access path under the same user appears no less than 3 times in the past month and accounts for more than 10% of the user's total number of sessions, it is considered to be stable. Therefore, the preset access count threshold can be set to 3 times for example to ensure that the stable access path set only retains access paths that are repetitive and representative.

[0063] Each access path in the stable access path set is parsed. The starting page identifier corresponding to the first page is extracted from each access path. The intermediate page identifiers are read sequentially from the access path and recorded as the page jump order according to their appearance order. The landing page identifier is extracted from the last page of the access path. The starting page identifier, the page jump order, and the landing page identifier are combined into an access intent chain unit according to the access order.

[0064] All access intent chain units in the stable access path set are summarized, and all access intent chain units are organized into an access intent baseline according to the combination of the starting page identifier, the order of page jumps, and the ending page identifier.

[0065] The context alignment module structures the user's current browsing path corresponding to the access intent baseline to obtain the content context chain, and marks the locations where the content context chain is disconnected from the access intent baseline as context break chains.

[0066] Once the baseline of the access intent is obtained, the user's current browsing path is continuously recorded during the user's new session. A session identifier is assigned to the current session, and the page identifier, page entry position, and interactive control identifier that triggers the page jump are recorded in the order of access. These are then compiled into a current browsing path record, which fully reflects the access sequence from the login start point to the end of the current session.

[0067] In the current browsing path record, page access entries arranged in chronological order are paired up. The identifier of the previous destination page in each pair of adjacent page access entries is regarded as the source page identifier, and the identifier of the next destination page is regarded as the target page identifier. At the same time, the entry position of the next destination page and the interactive control identifier that triggers the page jump between the two entries are retained. The source page identifier, target page identifier, page entry position and interactive control identifier are combined into a jump relationship entry.

[0068] For each jump relationship entry, the source page identifier and target page identifier are read. In the access intent baseline, the path segment whose starting page identifier matches the source page identifier and whose ending page identifier matches the target page identifier is found. The jump relationship entries are matched one-to-one with the matched access intent baseline path segments. The corresponding jump relationship entries are arranged in the order of access time and page hierarchy to form a content context chain organized in the access order.

[0069] In the content context chain, the jump order of each jump location is compared with the jump order of the corresponding location in the access intent baseline according to the link order. When the jump direction of the current session is opposite to the jump direction of the access intent baseline, the jump of the current session crosses the expected page level in the access intent baseline, or the current session exhibits repeated back-down behavior that does not appear in the access intent baseline within a short period of time, the corresponding jump location is marked as an abnormal sequence location, and all abnormal sequence locations are collected to form an abnormal sequence location set.

[0070] It should be noted that the expected page hierarchy is determined based on the hierarchical order in which pages in the access intent baseline jump from the upper navigation page to the lower content page. For example, if the access intent baseline is recorded as jumping from the channel homepage to the list page and then to the details page, then jumping directly from the channel homepage to the details page in the current session is considered to cross the expected page hierarchy.

[0071] Each jump location in the content context chain is assigned a page topic tag (obtained by extracting page title and body keywords using TF-IDF and mapping them to a pre-maintained thesaurus). The page topic tag corresponding to the jump location in the current session is compared with the page topic tag corresponding to the same location in the access intent baseline. When the two page topic tags are inconsistent, the corresponding topic similarity is lower than a preset similarity threshold, or the page topic shows a significant abrupt change at adjacent jump locations, the jump location is marked as a semantically abnormal location, and all semantically abnormal locations are collected to form a semantically abnormal location set.

[0072] To further explain, a significant mutation is determined by statistically analyzing the changes in topic similarity between historical normal and abnormal visits. The threshold for determining a significant mutation is the point where the absolute value of the difference in topic similarity between two adjacent jump positions is between 0.4 and 0.6. Therefore, when the absolute value of the difference in topic similarity between two adjacent jump positions is greater than or equal to the threshold for determining a significant mutation, it is considered that there is a significant mutation in the page topic at the adjacent jump position.

[0073] It should be noted that during the construction of the semantic anomaly location set, the consistency between page topic tags is quantified using a topic similarity function. For example, for the first topic in the content context chain... For each redirect location, the topic similarity is calculated using the following expression:

[0074] ;

[0075] in, The first in the content context chain The topic similarity of each jump location This indicates that the current browsing path is at the [location missing]. The page theme tags of the redirect location, The page's topic tag at the corresponding location indicates the baseline of the visitor's intent. This represents a similarity function used to measure the semantic similarity of topic tags between two pages.

[0076] when When the similarity is below the preset threshold, the first... Each jump location is included in the semantic anomaly location set.

[0077] It should be noted that the preset similarity threshold is set based on the distribution of topic similarity between historical normal access and abnormal access. For example, if statistics show that the topic similarity of normal access is mostly higher than 0.7 and the topic similarity of abnormal access is mostly lower than 0.5, the preset similarity threshold is set to 0.5 to distinguish between jump positions with basically the same page topic and jump positions with obviously different page topics.

[0078] The set of sequential abnormal positions and the set of semantic abnormal positions are merged. Jump positions that are adjacent or close to each other on the timeline and belong to either the set of sequential abnormal positions or the set of semantic abnormal positions are merged, and consecutive abnormal jump positions are divided into several abnormal segments; for the first... Define a jump location and an exception indication function:

[0079] ;

[0080] in, The first in the content context chain The result of the indicator for whether the jump position is an abnormal position, when the value is 1, indicates that the jump position is an abnormal position. The jump position belongs to the merged set of abnormal positions, and a value of 0 indicates that the jump position is the first jump position. The jump location was not identified as an abnormal location. Represents the set of positions with out-of-order sequences. Represents the set of semantically abnormal locations. This indicates the set union operation.

[0081] By satisfying all The jump positions are merged in time order, and each consecutive abnormal jump position and its corresponding upstream normal jump position are recorded as a context fault segment. All context fault segments are arranged according to the access order of the content context chain to obtain the corresponding context fault chain.

[0082] Preferably, by utilizing the access intent baseline formed based on identity inertia chains, the direction, level, and back-back behavior of each page jump in the current session are compared one by one to form a set of sequential abnormal positions. This makes abnormal jumps no longer dependent on static rules but closely follow the user's own stable access habits. By assigning page topic tags to each jump position and combining topic similarity functions and preset similarity thresholds, topic deviation and topic mutation positions are grouped into a set of semantic abnormal positions, enabling fine-grained identification of seemingly normal jumps that semantically deviate significantly from the access intent baseline. By using an anomaly indicator function, the set of sequential abnormal positions and the set of semantic abnormal positions are unified and merged into contextual fault fragments and arranged into a contextual fault chain. This not only accurately depicts where and how the access path deviates from the user's original access intent, but also provides a precise fault starting point for subsequent fault propagation analysis and risk path blocking, taking into account personalization, low false alarms, and traceability.

[0083] The risk propagation module tracks all reachable content nodes downstream of the context fault chain, constructs a fault propagation graph, and automatically aggregates the set of content nodes continuously pointed to by the same fault chain according to the guidance path of the context fault chain to generate a risk path set.

[0084] Each abnormal position in the context fault chain is regarded as the starting point. The downstream page identifier and resource identifier directly connected to the abnormal position are read from the website content management record. The downstream page identifier and resource identifier corresponding to each abnormal position are organized into downstream candidate nodes. All downstream candidate nodes corresponding to the abnormal positions are summarized to form a set of downstream candidate nodes for the context fault chain.

[0085] For each downstream candidate node in the downstream candidate node set, read the external links (hyperlinks in the page content pointing to external domains or third-party site pages of the centralized website platform, used to jump from the current page to external pages), internal jump relationships (inter-page jump relationships between pages within the same centralized website platform through interactive controls such as site menus, buttons, tags, and pagination controls), and resource reference relationships (reference relationships of static or dynamic resource files such as images, videos, scripts, style files, and embedded components during the page display process, used to load and present related resource content in the current page). Add the page identifiers and resource identifiers that can be accessed through external links, internal jump relationships, and resource reference relationships as new downstream nodes to the extended set, and record the new downstream nodes and their corresponding source nodes as hierarchical connection relationships according to the order of access, so that the extended set expands the content nodes level by level along the access direction based on the downstream candidate node set.

[0086] It should be noted that the extended set is obtained by expanding the downstream candidate node set step by step through external links, internal jump relationships, and resource reference relationships.

[0087] When the extended set is constructed, all nodes in the extended set are organized according to the access level of the access path relative to the anomaly position in the context fault chain. The anomaly position in the context fault chain is taken as the starting level node, and the reachable page identifier and resource identifier in each level are recorded as the corresponding level node. The connection relationship between the anomaly position in the context fault chain and the reachable nodes in each access level is represented as directed edges, resulting in a fault propagation graph composed of a set of nodes and a set of directed edges.

[0088] In the fault propagation graph, for each context fault chain, all content nodes directly connected to the current context fault chain are selected, the abnormal position corresponding to the context fault chain is regarded as the starting node, and the downstream nodes that have directed edge connections with the starting node are arranged in order of access level from near to far to form a candidate node sequence. The access level number and source abnormal position identifier of each node are retained in the candidate node sequence.

[0089] In the candidate node sequence, the node groups continuously pointed to by the same abnormal location are identified based on the source abnormal location identifier. For each node group continuously pointed to by the same abnormal location, the starting node, each intermediate node and the ending node in the node group are recorded in order of access level from small to large. The directed connection relationship corresponding to the node group is regarded as a complete risk path, and the corresponding context fault chain identifier and session identifier are associated with each risk path to form a risk path list for a single context fault chain.

[0090] All risk paths in the risk path list are classified according to the context break chain identifier and the session identifier. For multiple risk paths belonging to the same context break chain, the corresponding risk paths are aggregated into a set of paths. Each set of paths is associated with the corresponding context break chain, and the set of paths corresponding to all context break chains is jointly represented as a risk path set.

[0091] The path recovery module determines the upstream fault node in the fault propagation graph based on the risk path set, blocks access to the content pointed to by the fault node, and switches the access path to the contextual complete path corresponding to the access intent baseline.

[0092] For each risk path in the risk path set, the starting node, each intermediate node, and the ending node are read sequentially according to the access order. The access order of each node in the risk path is mapped to the access level number, forming a node sequence with access level numbers. Based on the abnormal position marker in the context fault chain associated with the node sequence, the set of nodes belonging to abnormal positions in the node sequence is determined, forming an abnormal node sequence.

[0093] In the sequence of abnormal nodes, the abnormal node with the smallest access level number is searched in ascending order of access level number. The abnormal node with the smallest access level number is determined as the upstream fault node corresponding to the current risk path, and the corresponding upstream fault node identifier is recorded for each risk path in the risk path set.

[0094] Read the page identifier and jump control information associated with each upstream fault node, register the page identifier and jump control information as objects to be blocked, configure the page jump operation corresponding to the object to be blocked to a prohibited state, and monitor all page jump requests during the user session. When it is detected that the page identifier and jump control information carried in the page jump request match any object to be blocked, the page jump request is determined to be a jump request that hits the object to be blocked and the original page jump is blocked.

[0095] When blocking the execution of the original page redirect, the user identity and current access topic information corresponding to the current user session are read (determined by the page topic tag of the current page). Access intent chain units that match the current user identity and whose page topic matches the current access topic are retrieved in the access intent baseline. The page corresponding to the position with the access level closest to the current access level is selected from the retrieved access intent chain units as the context complete node. The target page identifier in the page redirection request of the page to be blocked is replaced with the page identifier corresponding to the context complete node, so that the access target of the current session is switched from the page pointed to by the upstream fault node to the context complete node in the access intent baseline.

[0096] After the access target replacement is completed, the context complete node is used as the new access starting point. The access path of the current session is updated according to the subsequent page jump order corresponding to the context complete node in the access intent baseline. The access path composed of the context complete node and the subsequent page jump order is recorded as the context complete access path.

[0097] In summary, this invention utilizes the identity inertia chain formed through multiple user sessions to reflect the user's stable access path structure, enabling deviation judgment to be individualized and temporally continuous. The contextual fault chain simultaneously performs dual identification based on structural jump order and page semantic consistency, allowing the identification of abnormal behaviors such as jump direction reversal, hierarchical jumps, short-cycle backtracking, and topic abrupt changes, significantly reducing false positives. By recursively tracing reachable nodes downstream through the fault propagation graph, the potential cascading propagation range of abnormal access to the content system can be identified, not only detecting the deviation behavior itself but also discerning potentially affected content segments. Through the contextual path recovery mechanism that blocks the upstream fault node and aligns with the access intent baseline, access is uninterrupted but returns to the user's true intent path, achieving risk isolation + uninterrupted access intervention, significantly improving the user experience.

[0098] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A content operation and supervision system for an integrated website platform, characterized in that: include, The intent modeling module generates an identity inertial chain by recording the user's traceable access path, and extracts the access start point, jump structure and landing point from it to form the access intent baseline; The context alignment module organizes the user's current browsing path corresponding to the access intent baseline in a structured manner to obtain the content context chain, and marks the position where the content context chain is disconnected from the access intent baseline as the context break chain. The risk propagation module tracks all reachable content nodes downstream of the context fault chain, constructs a fault propagation graph, and automatically aggregates the set of content nodes continuously pointed to by the same fault chain according to the guidance path of the context fault chain to generate a risk path set. The path recovery module determines the upstream fault node in the fault propagation graph based on the risk path set, blocks access to the content pointed to by the fault node, and switches the access path to the contextual complete path corresponding to the access intent baseline.

2. The content operation and supervision system for an integrated website platform as described in claim 1, characterized in that: The specific steps for generating an identity inertial chain by recording the user's traceable access path are as follows. When a user successfully logs in, the user's session identifier is obtained, and the page access behavior in the session is recorded. Each page jump is organized into access path fragments and strung together in chronological order to form a complete access path. The complete access paths formed by the same user in multiple sessions are aggregated and associated with the user's identity identifier to obtain the identity inertial chain.

3. The content operation and supervision system for an integrated website platform as described in claim 1, characterized in that: The extraction of the access start point, jump structure, and landing point constitutes the access intent baseline. The specific steps are as follows. Statistical analysis is performed on each access path in the identity inertia chain, and access paths that have reached a preset access count threshold and have a stable page jump order are selected as a set of stable access paths. Extract the starting page identifier, page jump order, and ending page identifier of each access path from the set of stable access paths, and combine them according to the access order to generate an access intent baseline.

4. The content operation and supervision system for an integrated website platform as described in claim 1, characterized in that: The specific steps for structuring and organizing the user's current browsing path corresponding to the baseline of access intent are as follows. Record the page identifiers visited sequentially, the page entry positions, and the interactive controls that trigger page navigation in the user's current session to form a record of the current browsing path; Extract adjacent page pairs from the current browsing path record, associate them with the interactive control that triggered the jump and the page entry location, and form a jump relationship entry; The jump relationship entries are compared with path segments in the access intent baseline that have the same start and end points, and then linked together according to the access time order and page level to construct a content context chain.

5. The content operation and supervision system for an integrated website platform as described in claim 1, characterized in that: The specific steps for marking the locations where the content context chain and the access intent baseline become disconnected as context break chains are as follows. Compare the jump order in the content context chain with the corresponding jump order in the access intent baseline, mark the jump positions with opposite jump directions, jump levels crossed, and repeated backtracking, and form a set of abnormal order positions; The page topic tags at each jump location in the content context chain are compared with the page topic tags at the same location in the access intent baseline. Jump locations with inconsistent topics, low relevance, and abrupt topic changes are marked to form a set of semantically abnormal locations. The set of sequential abnormal locations and the set of semantic abnormal locations are merged, and adjacent abnormal locations are merged according to the access time to generate the corresponding contextual fault chain.

6. The content operation and supervision system for an integrated website platform as described in claim 1, characterized in that: The specific steps for constructing the fault propagation map are as follows: Starting from each abnormal position in the context break chain, the downstream page identifiers and resource identifiers directly connected to the abnormal position are read from the website content management records to form a set of downstream candidate nodes. Read the external links, internal jump relationships, and resource reference relationships of each node in the downstream candidate node set, add them to the extended set, and form a hierarchical connection relationship with the downstream candidate node set; All nodes in the extended set are organized according to the access level. The connection relationship between the abnormal position in the context fault chain and the reachable nodes at each level is represented as directed edges and organized into a fault propagation graph.

7. The content operation and supervision system for an integrated website platform as described in claim 6, characterized in that: The extended set is obtained by expanding the downstream candidate node set step by step through external links, internal jump relationships, and resource reference relationships.

8. The content operation and supervision system for an integrated website platform as described in claim 1, characterized in that: The automatic aggregation of content nodes continuously pointed to by the same fault chain based on the guiding path of the context fault chain involves the following steps. In the fault propagation map, all nodes directly connected to the same context fault chain are selected and sorted from near to far according to the access level to form a candidate node sequence; Identify groups of nodes that are continuously pointed to by the same abnormal location in the candidate node sequence, treat each group of nodes as a complete risk path, and record the starting node, each intermediate node and the ending node on the risk path. All risk paths are classified according to context break chain identifier and session identifier, and multiple risk paths belonging to the same context break chain are grouped into a path set; Associate all risk paths in each path set with their corresponding context fault chains to generate a risk path set.

9. The content operation and supervision system for an integrated website platform as described in claim 1, characterized in that: The steps to block access to the content pointed to by the fault node and switch the access path to the complete contextual path corresponding to the access intent baseline are as follows: Each risk path in each risk path set is analyzed according to the access level, the abnormal location closest to the access origin is identified, and the corresponding node is determined as the upstream fault node of the risk path set. Register the page identifier and jump control information associated with the upstream fault node as the object to be blocked, set the page jump operation corresponding to the object to be blocked to a prohibited state, and monitor page jump requests in the session; When it is detected that a user access reaches the object to be blocked, query the context complete node in the access intent baseline that matches the current identity and is consistent with the current access topic, and replace the access target node that originally pointed to the object to be blocked with the current context complete node. Align the replaced access target node with the subsequent jump order in the access intent baseline to form a context-complete access path.

10. The content operation and supervision system for an integrated website platform as described in claim 9, characterized in that: The specific steps for analyzing each risk path in each risk path set according to access level are as follows. Read the access order of the starting node, each intermediate node and the ending node of each risk path in the risk path set, map it to the access level number, and form a node sequence; In the node sequence, based on the abnormal location markers corresponding to the context fault chain, the set of nodes corresponding to the abnormal location is determined, forming an abnormal node sequence; In the sequence of abnormal nodes, the abnormal node with the smallest access level number is found in ascending order of access level number and is determined as the upstream fault node.

Images