Drill-down analysis method, device and equipment based on user movement line

Abstract

The embodiment of the invention discloses a drill-down analysis method, device and equipment based on a user movement line. The scheme comprises the following steps: for a current service, analyzing a service log of a service execution user, and extracting to obtain a corresponding first user behavior chain; performing service semantic association on the first user behavior chain and the current service through a plurality of preset service dimensions to obtain a second user behavior chain; aiming at the second user behavior chain, extracting a corresponding key parameter, extracting a corresponding business loss logic feature according to the key parameter, and determining a target user with business loss in the business execution users according to the business loss logic feature; and performing drill-down analysis based on the first user behavior chain corresponding to the target user, and extracting to obtain a corresponding path feature.

Description

Technical Field

[0001] This specification relates to the field of Internet technology, and in particular to methods, apparatus and equipment for drill-down analysis based on user movement. Background Technology

[0002] With the development of computer and internet technology, many businesses can now be conducted online.

[0003] In the execution of online business, drill-down analysis of conversion rates at key points in the user journey can pinpoint churn between these points, allowing for monitoring the destination of unconverted users and ultimately improving product quality and user experience. Drill-down analysis is a data analysis technique whose core characteristic is the use of hierarchical decomposition to achieve fine-grained data exploration, revealing the underlying reasons behind aggregated data.

[0004] In the traditional approach, drill-down analysis can be performed using two methods.

[0005] Firstly, a static analysis scheme based on a predefined funnel is used to pre-set fixed paths, thereby enabling drill-down analysis based on corresponding basic metrics. Taking e-commerce as an example, this fixed path can be set as a three-step process: user registration → product page → product payment. The drop-off between user nodes along this fixed path can be analyzed using basic data such as page views (PV) and unique visitors (UV).

[0006] Secondly, the no-code behavior analysis solution automatically generates click and page jump events within a preset number of steps through DOM parsing, thereby generating a conversion funnel and Sankey diagram. To avoid overly complex analysis, the number of steps analyzed is typically set to 3-5.

[0007] Therefore, a drill-down analysis solution that can incorporate business semantics is needed. Summary of the Invention

[0008] This specification provides one or more embodiments of a drill-down analysis method, apparatus, device, and storage medium based on user movement to solve the following technical problem: the need for a drill-down analysis scheme that can incorporate business semantics.

[0009] To solve the above-mentioned technical problems, one or more embodiments of this specification are implemented as follows: This specification provides one or more embodiments of a drill-down analysis method based on user movement, including: For the current business, the business logs of the users executing the business are parsed to extract the corresponding first user behavior chain; The first user behavior chain is semantically associated with the current business through multiple preset business dimensions to obtain the second user behavior chain; For the second user behavior chain, the corresponding key parameters are extracted, and the corresponding business churn logic features are extracted based on the key parameters. Based on the business churn logic features, the target users who have experienced business churn are identified among the users who performed the business. Drill-down analysis is performed based on the first user behavior chain corresponding to the target user to extract the corresponding path features.

[0010] This specification provides one or more embodiments of a drilling analysis device based on user movement, comprising: The behavior chain extraction module analyzes the business logs of the user executing the business for the current business and extracts the corresponding first user behavior chain. The business semantic association module associates the first user behavior chain with the current business through multiple preset business dimensions to obtain the second user behavior chain; The target user determination module extracts the corresponding key parameters for the second user behavior chain, extracts the corresponding business churn logic features based on the key parameters, and determines the target users with business churn among the business execution users based on the business churn logic features. The path feature extraction module performs drill-down analysis based on the first user behavior chain corresponding to the target user to extract the corresponding path features.

[0011] This specification provides one or more embodiments of a user-trajectory-based drilling analysis device, comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to: For the current business, the business logs of the users executing the business are parsed to extract the corresponding first user behavior chain; The first user behavior chain is semantically associated with the current business through multiple preset business dimensions to obtain the second user behavior chain; For the second user behavior chain, the corresponding key parameters are extracted, and the corresponding business churn logic features are extracted based on the key parameters. Based on the business churn logic features, the target users who have experienced business churn are identified among the users who performed the business. Drill-down analysis is performed based on the first user behavior chain corresponding to the target user to extract the corresponding path features.

[0012] This specification provides one or more embodiments of a non-volatile computer storage medium storing computer-executable instructions, wherein the computer-executable instructions are configured as follows: For the current business, the business logs of the users executing the business are parsed to extract the corresponding first user behavior chain; The first user behavior chain is semantically associated with the current business through multiple preset business dimensions to obtain the second user behavior chain; For the second user behavior chain, the corresponding key parameters are extracted, and the corresponding business churn logic features are extracted based on the key parameters. Based on the business churn logic features, the target users who have experienced business churn are identified among the users who performed the business. Drill-down analysis is performed based on the first user behavior chain corresponding to the target user to extract the corresponding path features.

[0013] The above-described at least one technical solution adopted in one or more embodiments of this specification can achieve the following beneficial effects: By dynamically extracting user behavior chains from business logs, the limitations of predefined paths are avoided, and any user journey node can be flexibly covered, including abnormal jumps and cross-platform behaviors (such as the connection between H5 and mini-programs), to achieve full path analysis.

[0014] By introducing multiple business dimensions for semantic association, user behavior is closely integrated with business logic, enabling analysis to no longer rely on basic metrics but to delve into the business context and improve the business relevance of the analysis.

[0015] By extracting business churn logic features through key parameters, we can accurately identify target users, avoid missing long paths due to step truncation in no-code tracking solutions, and ensure complete exploration of complex user behaviors. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating a user-trajectory-based drill-down analysis method provided for one or more embodiments of this specification; Figure 2 This is a flowchart illustrating the target user determination process in one or more embodiments of this specification. Figure 3 This is a flowchart illustrating a drill-down analysis process in one or more embodiments of this specification. Figure 4 A schematic diagram of a drilling analysis device based on user movement path provided for one or more embodiments of this specification; Figure 5 This is a schematic diagram of a drilling analysis device based on user movement, provided for one or more embodiments of this specification. Detailed Implementation

[0018] This specification provides a drilling analysis method, apparatus, device, and storage medium based on user movement.

[0019] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0020] In traditional static analysis schemes based on predefined funnels, drill-down analysis is only performed within the fixed path due to the pre-defined fixed path settings. This results in a fixed path, making it difficult to drill down into nodes outside the path. Furthermore, it relies on basic metrics such as UV / PV and lacks a connection to business semantics. For example, the fixed path might be business execution page A → business execution page B → end page, but if a user accidentally clicks on an ad during business execution, causing an abnormal redirect, the path becomes business execution page A → ad page → ad product page, making it impossible to detect this path. Additionally, it's difficult to correlate behavioral data from different platforms (e.g., mini-program platforms, H5 page platforms), making it difficult to perform analysis actions on paths like business execution page (H5) → payment page (mini-program).

[0021] For code-free behavior analysis solutions, because they have a set number of path steps, nodes exceeding that number of steps are truncated, making it difficult to perform analysis actions. Furthermore, they are difficult to bind to business semantics, resulting in difficulties in performing business semantic-level analysis.

[0022] Based on this Figure 1This document presents a flowchart illustrating a user-trajectory-based drill-down analysis method provided in one or more embodiments. This method can be applied to various business domains, such as e-commerce, online business operations and maintenance, instant messaging, gaming, and government services. The process can be executed by computing devices specific to the domain (e.g., intelligent customer service servers or intelligent mobile terminals for e-commerce). Certain input parameters or intermediate results within the process can be manually adjusted to improve accuracy.

[0023] Figure 1 The process may include the following steps: S102: For the current business, parse the business logs of the user executing the business and extract the corresponding first user behavior chain.

[0024] The current business is an online business, such as e-commerce (including goods trading, online communication, etc.), online education (including course learning, online examination, etc.), online medical services (including online registration, online consultation, etc.).

[0025] The user executing the business task is the user performing the current business task. The business log records various operational information of the user during the business task execution process, such as page visits, click events, and data input. The business log can be obtained by extracting it from the user's original logs within the platform. By parsing the business log, the user's operation sequence and behavioral trajectory can be identified, thereby extracting the first user behavior chain, which reflects the user's original behavioral path during the business task execution process.

[0026] Taking e-commerce as an example, the first user behavior chain records a series of actions a user takes, from entering the platform homepage, browsing product category pages, clicking to enter a specific product details page, adding products to the shopping cart, and finally making a checkout payment. By analyzing these business logs, an example of the first user behavior chain can be extracted as: Platform Homepage → Product Category Page → Product Details Page → Shopping Cart → Checkout Payment Page. Of course, based on actual needs, the first user behavior chain can be further refined by adding component, control, and button-level actions. In this case, the section "Platform Homepage → Product Category Page → Product Details Page" can become "Platform Homepage → Click the product category entry button → Product Category Page → Click the specific product link → Product Details Page".

[0027] S104: By using multiple preset business dimensions, the first user behavior chain is semantically associated with the current business to obtain the second user behavior chain.

[0028] The first user behavior chain is directly extracted and parsed from the business logs. At this point, it does not involve any business semantic descriptions. Based on this, it is semantically associated with multiple preset business dimensions to give it business semantic content.

[0029] Specifically, business semantic association refers to associating each node in the first user behavior chain with business meaning within the business domain and context of the current business, based on its simple grammatical or structural features.

[0030] Figure 2 This is a flowchart illustrating the process of determining a target user in an application scenario, provided for one or more embodiments of this specification. Taking e-commerce as an example, if the current business corresponds to e-commerce, then the preset multiple business dimensions include the business instance dimension corresponding to the e-commerce business, the business channel dimension, and the e-commerce product information summary dimension.

[0031] The business instance dimension is used to describe the nodes that a user may reach in the business process, as well as the description and encoding of each node. Therefore, the business instance dimension can be reflected through the product line instance node dimension table and the structured path modeling attribute fields of the instance node dimension table.

[0032] The product line instance node dimension table defines the node metadata of specific instances within a business product line, describing the nodes a user may reach in the business process. This includes attributes such as a unique node identifier, node business name and description, node's hierarchical position in the business process, node type (e.g., entry node, conversion node, churn node), and the definition of flow relationships between nodes. For example, in e-commerce, the product line instance node dimension table corresponding to a certain node might include the following: Node ID: CHECKOUT_PAGE, Node Name: Cashier Page, Node Type: Key Conversion Node, Belonging Product Line: V1 Product Line, Expected Next Node: [PAYMENT_PAGE, CART_PAGE], Business Importance: High.

[0033] The Structured Path Modeling Attribute (SPMA) field in the instance node dimension table is a structured path modeling attribute field that provides a unique semantic code containing hierarchical information for each node, enabling node location and pattern recognition in path analysis. For example, for the node: Product Payment Page, SPMA: ECOM.CHECKOUT.PAYMENT.CONFIRM.BUTTON (corresponding to the e-commerce domain, checkout process, payment module, confirmation page, and button component).

[0034] The business channel dimension describes the various cooperative channels corresponding to the e-commerce business, which can be reflected in the application partner type information dimension table. The application partner type information dimension table can include the partner's unique identifier, partner type classification (e.g., direct traffic, search engines, social media, advertising networks, etc.), channel quality rating, cooperative business model, target user group characteristics, etc.

[0035] E-commerce product information aggregation dimensions are used for daily business health monitoring, trend analysis, and performance evaluation, and can be reflected in the application summary table. The application summary table can include user scale indicators (including daily active users, weekly active users, monthly active users, etc.), traffic quality indicators (including number of sessions, page views, bounce rate, etc.), user engagement indicators (including average session duration, visit depth, etc.), etc.

[0036] By using business data corresponding to multiple business dimensions, the semantically related information in the first user behavior chain is updated. For each node in the first user behavior chain, not all nodes have semantic relationships. In this case, the nodes with semantic relationships can be selected for updating, and when performing semantic association, only the corresponding business dimension needs to be selected for association.

[0037] Specifically, each node in the first user behavior chain is identified. Typically, each node corresponds to a single page in the first user behavior chain.

[0038] Based on the page type corresponding to each node, select nodes that match the preset type, and use a preset association rule base to select the corresponding business dimension for them, updating the information contained in the node. The page type can be preset and can be determined by the page title, page ID, etc., corresponding to the node. For example, for e-commerce, page types can include the platform homepage, product detail page, payment page, etc.

[0039] The preset types are also pre-set. For different businesses, the more important page types are selected and associated with the corresponding business semantics to facilitate the subsequent drill-down analysis process.

[0040] The association rule base marks whether different page types have mandatory or optional association business dimensions. Mandatory association means that when a page type appears, the corresponding business dimension is forcibly associated with it, while optional association means that the business dimension is associated only after the node meets certain conditions (such as the user stays for a long time or the user performs a preset action on the page).

[0041] For nodes that do not conform to the preset type, the importance score of the node is determined based on its sequence position in the first user behavior chain and the page type it corresponds to. Generally speaking, nodes at the ends of the sequence are more important. Weights can be assigned to different sequence positions and different page types, and the importance score of each node can be calculated using these two dimensions.

[0042] For nodes with importance scores higher than the preset score, the corresponding business dimension is selected based on a preset association rule base, and the information contained in the node is updated. For nodes with high importance, even if they are not of the preset type, business semantic association is still performed. For the remaining nodes with low importance, no semantic association is required, thereby reducing the introduction of noise.

[0043] Taking e-commerce as an example. Assume the first user behavior chain is: platform homepage → product category page → product details page → shopping cart → checkout / payment page. The product category page, product details page, and checkout / payment page are preset types and require semantic association. The platform homepage is considered a node requiring semantic association after importance scoring. Therefore, the second user behavior chain after semantic association could be: entering the platform homepage through social channel A → selecting the product category page of a currently high-conversion-rate product → deeply browsing the product details page → shopping cart → payment method obstacles leading to unpaid checkout / payment page.

[0044] Specifically, based on the business dimensions selected from the association rule base: Social channel A can be identified through the interface corresponding to the business channel dimension; high-conversion products can be identified through the instance node dimension table to determine the corresponding product type, and the conversion rate of that product type can be determined through the records in the e-commerce product information summary dimension; deep browsing can be located through the SPMA field to locate the product details browsing behavior, thereby determining the dwell time and whether to conduct a deep evaluation; the shopping cart node has not undergone semantic expansion; payment method obstacles can be identified through the business channel dimension to determine the corresponding payment method partner, and the SPMA field can be used to determine whether the payment was successful.

[0045] S106: For the second user behavior chain, extract the corresponding key parameters, extract the corresponding business churn logic features based on the key parameters, and determine the target users with business churn among the business execution users based on the business churn logic features.

[0046] Key parameters are primarily used to determine the target user, while business churn refers to a user's failure to achieve the goals expected by the business owner during the execution of the current business. For example, in the case of an e-commerce business, business churn could be due to a failed purchase; in the case of an online healthcare business, business churn could be due to a user failing to complete an online consultation or being dissatisfied with the consultation results.

[0047] Specifically, such as Figure 2 As shown, when determining key parameters, the second user behavior chain is grouped into windows to obtain one or more user behavior sub-chains. The main function of window grouping is to divide the second user behavior chain into smaller, targeted sub-chains, facilitating detailed analysis of user behavior.

[0048] The window grouping method can be based on various factors such as time, business stage, and behavior type. Taking e-commerce as an example, users can be grouped according to the time period when they enter the platform or the maximum number of nodes. In this case, the time difference between each user behavior sub-chain is less than a preset time difference, or the number of nodes in each user behavior sub-chain is less than a preset number. When a user browses multiple products in a behavior chain, they can also be grouped according to different products. In this case, each user behavior sub-chain corresponds to the same product.

[0049] Once the user behavior sub-chains are obtained, a corresponding identifier can be set for each sub-chain. This identifier can be set based on a grouping method. For example, grouping based on time period or maximum number of nodes can be set according to its sequence in the original second user behavior chain; grouping based on product can be set according to product identifiers.

[0050] For each user behavior subchain, the corresponding maximum instance index is extracted. The maximum instance index describes the maximum position of the nodes contained in the user behavior subchain, and can be used to determine the nodes contained in the user behavior subchain.

[0051] For each node in the sequence corresponding to the largest instance index, extract its corresponding business channel information, as well as the corresponding node duration information and node completion information.

[0052] Business channel information primarily refers to the channel source of a node. This channel can include entry channels for the business, other channels for business execution, etc. For example, in e-commerce, it indicates which channels users use to enter the product page and which channels they use to complete payment transactions. This information can be extracted through semantic associations added to the second user behavior chain. Typically, only some nodes can have their corresponding business channel information extracted.

[0053] Node duration information indicates the time a user spends at each node, which can be obtained based on the time difference between entering and leaving the node.

[0054] Node completion information can be categorized into multiple sub-business dimensions. For example, firstly, based on the user ID and timestamp, the user's historical transactions are linked (e.g., the previous independent session determines the last executed transaction, which is then considered a historical transaction), and it is determined whether the user has completed that previous transaction. The completion rate of the historical transaction is then used as a sub-business dimension within the node completion information. Next, within the user behavior sub-chain corresponding to the current transaction, it is determined whether the user has completed the corresponding node, thus using the completion rate of the current transaction as a business dimension within the node completion information. Simultaneously, it can also be determined whether there are any return actions or task switching actions within the user behavior sub-chain; that is, whether there is an action of returning to the previous node from the current node or switching from the current transaction to other tasks (e.g., not performing a payment action on the settlement and payment page, but instead entering a customer service page or an activity page).

[0055] At this point, for the target user, the corresponding business churn logic features are extracted based on business channel information, node duration information, and node completion information. Among them, the sub-business dimension of node completion information includes the completion degree of the user's historical business, the completion degree of the current business, the node return action and task switching action in the current business.

[0056] This content is used as a business churn logic feature to identify target users among those performing business operations who have experienced business churn and meet pre-defined analysis requirements. For example, based on the completion status of the current business based on node completion information, users with business churn are identified. Then, users are filtered based on business channel information (e.g., the current analysis only analyzes specific payment channels). Furthermore, based on the user's node duration information, it is determined which nodes the user spent too much time on. And based on the nodes corresponding to node return actions and task switching actions, it is determined which nodes the user performed actions that caused churn (e.g., the current analysis only analyzes churn on product detail pages and payment pages). Thus, the target users with business churn who meet the pre-defined analysis requirements are finally obtained.

[0057] S108: Drill down the analysis based on the first user behavior chain corresponding to the target user to extract the corresponding path features.

[0058] Figure 3This is a flowchart illustrating drill-down analysis in one or more embodiments of this specification, illustrating a specific application scenario. Drill-down analysis based on the first user behavior chain corresponding to the target user allows for a more detailed analysis of each link in the first user behavior chain, thereby uncovering hidden path features.

[0059] Based on the user identifier corresponding to the target user, cross-platform behavior association is performed on the first user behavior chain. In traditional solutions, it is usually difficult to effectively associate user cross-platform behaviors, which limits the nodes to be analyzed. This solution addresses this limitation by using cross-platform behavior association.

[0060] Specifically, based on the user identifier corresponding to the target user (such as user login ID, user fingerprint, etc.), the first user behavior chain corresponding to the target user on each terminal (such as smartphone, personal computer, tablet, etc.) is determined.

[0061] In the first user behavior chain corresponding to each terminal, the first user behavior chain with a difference between timestamps lower than a preset difference is selected. Each node in this chain is used as the smallest splicing unit, and the chains are sorted and spliced ​​according to the time sequence to obtain the intermediate user behavior chain. The timestamp refers to the time when the user accesses the first user behavior chain. The timestamp corresponding to each first user behavior chain can be either the timestamp of the first node or the timestamp of the last node.

[0062] The preset difference can be set based on requirements, for example, between 5 and 10 minutes. This assumes that within this timeframe, the first user behavior chain executed by the user on different terminal devices is likely to belong to the same service, allowing for cross-device behavior association. In this case, the nodes in each first user behavior chain are separated and used as the smallest splicing unit for assembly. The splicing at this stage primarily considers chronological order to restore the original sequence of user actions.

[0063] Business logic validation is performed on the intermediate user behavior chain. Business logic refers to the pre-defined logical relationships within the current business logic. For example, when a product details page first appears, it should follow the product category page of its major product category; the payment page should follow the shopping cart or product details page, etc. Since the intermediate user behavior chain is obtained by breaking down the original first user behavior chain into nodes and then reassembling it, some objective factors may exist (such as an anomaly in the time system of a certain terminal, or different actual users logging into the same user-identified account, etc.), leading to business logic anomalies. Therefore, business logic validation is necessary.

[0064] If there are no business logic anomalies, it can be directly used as the first user behavior chain after cross-platform behavior association.

[0065] If a business logic anomaly exists, intent recognition is performed on the first user behavior chain corresponding to each terminal to obtain the business intent for each terminal. Intent recognition refers to determining the user's purpose in executing the first user behavior chain based on the page type and sequence of each node. This can be achieved through a pre-set intent rule base. For example, in e-commerce, when the first user behavior chain is "Product A details page → shopping cart → payment page", the intent is the intention to purchase product A; when the first user behavior chain is "Product A details page → product category page → product B details page", the intent is the intention to compare prices of product A and product B; when the first user behavior chain is "platform homepage → user service page → customer service page", the intent is the intention to provide service support, etc.

[0066] If the business intents corresponding to each terminal match, then the first user behavior chain corresponding to each terminal is used as the smallest splicing unit, sorted and spliced ​​according to time sequence to obtain the first user behavior chain after cross-terminal behavior association. Intent matching refers to the existence of a corresponding relationship between two intents. For example, there is a matching relationship between the intent to buy and the intent to compare prices for the same product, a matching relationship between the intent to purchase a product and the intent to participate in promotional activities, and a matching relationship between the intent to return or exchange a product and the intent to provide service support, etc. This intent matching relationship can be pre-set through a mapping table.

[0067] For each first user behavior chain that matches intent, it can be considered as the operation of the same actual user. More likely, it is a business logic exception caused by an anomaly in the time system. In this case, instead of using nodes as the smallest splicing unit, the first user behavior chain is used directly as the smallest splicing unit to prevent the occurrence of business logic exceptions after splicing nodes.

[0068] For the first user behavior chains with mismatched intentions, they can be considered as different actual users performing different business operations using the same user's account, and therefore cross-platform behavior association is no longer required.

[0069] Filtering the first user behavior chain reveals key nodes. Key nodes are those that may have a significant impact on the target user's business churn process. For example, they might be nodes corresponding to back actions or task switching actions, or nodes with low business completion rates. Alternatively, filtering can be based on the importance of each node's page type to identify key nodes.

[0070] Based on key nodes, construct the behavior sequence corresponding to the target user. This behavior sequence can be directly obtained by concatenating key nodes. However, this direct concatenation method tends to result in an excessively long behavior sequence with limited information, which is detrimental to the subsequent generation of path features.

[0071] Based on this, key nodes are segmented using multiple pre-defined window sorting dimensions to obtain multiple independent analysis units. Window sorting dimensions refer to the dimensions used when segmenting windows and sorting within windows. For example, they may include the page type of the node, the type of the current business, and the degree of correlation with the user's historical business (e.g., in e-commerce, this could be reflected in whether the purchase was for the same or similar products). Using these dimensions, key nodes that are more related to each other are selected and segmented into the same analysis unit to facilitate drill-down analysis within the more relevant analysis unit.

[0072] For each analysis unit, the behavior sequences are sorted according to time series to obtain the corresponding behavior sequences. Based on the sequence order and page type of the key nodes, corresponding sequence numbers are generated. While generating behavior sequences based solely on time series is possible, it's difficult to clearly show the relationships between different key nodes. Therefore, generating sequence numbers based on page type allows for quick identification of the business relationships between different nodes. For example, assuming a single analysis unit is: Platform Homepage → Product Search Page → Product A Details Page → Product Search Page → Product B Details Page → Payment Page, the sequence numbers generated solely based on sequence order would be: Node 1 → Node 2 → … → Node 6. However, if page type is added, the sequence numbers become: Node 1 (Platform Homepage 1) → Node 2 (Product Search Page 1) → Node 3 (Product Details Page 1) → Node 4 (Product Search Page 2) → Node 5 (Product Details Page 2) → Node 6 (Payment Page 1). This clearly shows the types of different nodes, facilitating subsequent data processing.

[0073] By analyzing behavioral sequences, corresponding path features are extracted and visualized. Once the behavioral sequences are obtained, the corresponding path features can be extracted. These path features reflect the specific path taken by the target user during this business execution process. Developers and operations personnel can analyze these path features to improve the business and enhance the user experience.

[0074] When extracting path features, key nodes within the behavioral sequence are used to extract their corresponding page features. Page features are extracted based on controls on the page, such as buttons, input boxes, dropdown menus, and images. These page features can include the control's type, position, size, color, and text content. For example, on a product details page, the size of the product image, the color of the price tag, and the position of the "add to cart" button can be extracted. Simultaneously, layout features can also be extracted, including the overall page structure and the distribution of different areas. For example, determining whether the page uses a top-bottom, left-right, or mixed layout.

[0075] Based on page features and the target user's operation records, structured path features for each control within the key nodes are extracted. These structured path features reflect the user's operational characteristics towards different controls within a single page, and can be achieved through SPM (Site Layout Processing) point marking. For example, on a product details page, if a user clicks on a product image, views the price tag, or clicks the "add to cart" button, the order and position of these actions can be recorded using SPM point marking, thus forming the structured path features of that page.

[0076] In addition, the time interval between operations can be considered when extracting structured path features. For example, how long after a user clicks the "Add to Cart" button will they click the "Checkout" button? This time interval can also be included as part of the structured path features.

[0077] Of course, the extracted page features and structured path features may differ for different types of key nodes. For example, on a payment page, in addition to extracting regular control features, it is also necessary to focus on the selection of payment methods and the click status of the payment button.

[0078] Based on the sequence number corresponding to each key node in the behavior sequence, the structured path features are combined to obtain the corresponding path features. By combining the global dimension of the nodes and the local dimension of the space, complete and accurate path features can be obtained. These features not only describe the user's access order to different nodes, but also reveal the detailed access order of controls within each node, which is beneficial for developers to analyze and utilize path features.

[0079] Based on this, when performing visualization, the path characteristics are standardized using front-end languages. For example, JSON is used for structural standardization, including standardized field naming, standardized data format, and establishment of hierarchical relationships.

[0080] Based on the standardized path characteristics, corresponding visualization charts are generated. For example, a Sankey diagram can be used. A Sankey diagram is a visualization chart composed of nodes and traffic, which can clearly show the direction and magnitude of traffic between key nodes. In a Sankey diagram, nodes represent key nodes, and traffic represents the behavioral path of target users between different key nodes. Through a Sankey diagram, one can intuitively see the flow of target users from one key node to another during business execution, as well as the traffic magnitude at each key node, so that developers can quickly analyze the reasons for user churn.

[0081] Of course, besides Sankey diagrams, other types of visualization charts can be used, such as flowcharts and tree diagrams. Flowcharts can clearly show the entire process of business execution, presenting key nodes and their relationships in a flow format, making it easier for developers to grasp the overall execution process. Tree diagrams, on the other hand, can display key nodes according to a hierarchical structure, highlighting the hierarchical relationships between different key nodes, which helps developers to deeply analyze the structure and logic of the business.

[0082] We conduct statistical analysis on each target user in the current business and add the obtained traffic data to the nodes of a visualization chart. Sankey diagrams can present the traffic data of target users at each key node in a visual way, using lines of varying thickness and color intensity. For example, darker colors represent higher traffic. This allows developers to quickly determine which key nodes in the business process have higher traffic and which have lower traffic by observing the changes in line thickness and color.

[0083] By dynamically extracting user behavior chains from business logs, the limitations of predefined paths are avoided, and any user journey node can be flexibly covered, including abnormal jumps and cross-platform behaviors (such as the connection between H5 and mini-programs), to achieve full path analysis.

[0084] By introducing multiple business dimensions for semantic association, user behavior is closely integrated with business logic, enabling analysis to no longer rely on basic metrics but to delve into the business context and improve the business relevance of the analysis.

[0085] By extracting business churn logic features through key parameters, we can accurately identify target users, avoid missing long paths due to step truncation in no-code tracking solutions, and ensure complete exploration of complex user behaviors.

[0086] Based on the same idea, one or more embodiments of this specification also provide apparatus and devices corresponding to the above methods, such as... Figure 4 , Figure 5 As shown.

[0087] Figure 4A schematic diagram of a drilling analysis device based on user movement path provided for one or more embodiments of this specification, the device comprising: The behavior chain extraction module 402, for the current business, parses the business logs of the user executing the business and extracts the corresponding first user behavior chain; The business semantic association module 404 associates the first user behavior chain with the current business through multiple preset business dimensions to obtain the second user behavior chain. The target user determination module 406 extracts the corresponding key parameters for the second user behavior chain, extracts the corresponding business churn logic features based on the key parameters, and determines the target users with business churn among the business execution users based on the business churn logic features. The path feature extraction module 408 performs drill-down analysis based on the first user behavior chain corresponding to the target user to extract the corresponding path features.

[0088] Optionally, if the current business corresponds to an e-commerce business, the business semantic association module 404 determines that the preset multiple business dimensions include the business instance dimension, business channel dimension, and e-commerce product information summary dimension corresponding to the e-commerce business. The relevant information with semantic association in the first user behavior chain is updated using the business data corresponding to the multiple business dimensions.

[0089] Optionally, the business semantic association module 404 determines each node in the first user behavior chain; Based on the page type corresponding to each node, select nodes that match the preset type, select the corresponding business dimension for them through the preset association rule library, and update the information contained in the node; For nodes that do not conform to the preset type, the importance score of the node is determined based on the node's sequence position in the first user behavior chain and the page type corresponding to the node. For nodes whose importance score is higher than the preset score, the corresponding business dimension is selected for them through the preset association rule base, and the information contained in the node is updated.

[0090] Optionally, the target user determination module 406 performs window grouping on the second user behavior chain to obtain one or more user behavior sub-chains; For the user behavior subchain, the corresponding maximum instance index is extracted; For each node in the sequence corresponding to the maximum instance index, extract its corresponding business channel information, as well as the corresponding node duration information and node completion information.

[0091] Optionally, the target user determination module 406 extracts corresponding business churn logic features for the target user based on the business channel information, the node duration information, and the node completion information; wherein, the sub-business dimension of the node completion information includes the completion degree of the user's historical business, the completion degree of the current business, the node return action and task switching action in the current business; Based on the business churn logic characteristics, target users who have experienced business churn and meet the preset analysis requirements are identified among the users performing the business.

[0092] Optionally, the path feature extraction module 408 performs cross-end behavior association on the first user behavior chain based on the user identifier corresponding to the target user; Filter the first user behavior chain to obtain key nodes; Based on the key nodes, construct the behavior sequence corresponding to the target user; The corresponding path features are extracted from the behavioral sequence and then visualized and mapped.

[0093] Optionally, the path feature extraction module 408 determines the first user behavior chain corresponding to the target user on each terminal based on the user identifier corresponding to the target user; In the first user behavior chain corresponding to each terminal, select the first user behavior chain where the difference between timestamps is lower than the preset difference, take each node in it as the smallest splicing unit, sort and splice them according to the time sequence to obtain the intermediate user behavior chain. Perform business logic verification on the aforementioned intermediate user behavior chain; If there is a business logic anomaly, then perform intent recognition on the first user behavior chain corresponding to each terminal to obtain the business intent corresponding to each terminal. If the business intents corresponding to each terminal match, then the first user behavior chain corresponding to each terminal is used as the smallest splicing unit, and the chain is sorted and spliced ​​according to the time sequence to obtain the first user behavior chain after cross-terminal behavior association.

[0094] Optionally, the path feature extraction module 408 segments the key nodes using multiple pre-set window sorting dimensions to obtain multiple independent analysis units; For the analysis unit, sorting is performed according to the time series to obtain the corresponding behavior sequence, and a corresponding sequence number is generated based on the sequence order and page type of the key node.

[0095] Optionally, the path feature extraction module 408 extracts the corresponding page features for key nodes contained in the behavior sequence; Based on the page features and the target user's operation records, extract the structured path features of the target user for each control in the key nodes; Based on the sequence number corresponding to each key node in the behavior sequence, the structured path features are combined to obtain the corresponding path features.

[0096] Optionally, the path feature extraction module 408 standardizes the path features using a front-end language. Based on the standardized path features, generate corresponding visualization charts; Statistics are collected on each target user in the current business, and the collected traffic data is added to each node in the visualization chart.

[0097] Figure 5 A schematic diagram of a drilling analysis device based on user movement, provided for one or more embodiments of this specification, the device comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to: For the current business, the business logs of the users executing the business are parsed to extract the corresponding first user behavior chain; The first user behavior chain is semantically associated with the current business through multiple preset business dimensions to obtain the second user behavior chain; For the second user behavior chain, the corresponding key parameters are extracted, and the corresponding business churn logic features are extracted based on the key parameters. Based on the business churn logic features, the target users who have experienced business churn are identified among the users who performed the business. Drill-down analysis is performed based on the first user behavior chain corresponding to the target user to extract the corresponding path features.

[0098] Based on the same idea, one or more embodiments of this specification also provide a non-volatile computer storage medium corresponding to the above method, storing computer-executable instructions, wherein the computer-executable instructions are configured as follows: For the current business, the business logs of the users executing the business are parsed to extract the corresponding first user behavior chain; The first user behavior chain is semantically associated with the current business through multiple preset business dimensions to obtain the second user behavior chain; For the second user behavior chain, the corresponding key parameters are extracted, and the corresponding business churn logic features are extracted based on the key parameters. Based on the business churn logic features, the target users who have experienced business churn are identified among the users who performed the business. Drill-down analysis is performed based on the first user behavior chain corresponding to the target user to extract the corresponding path features.

[0099] In the 1990s, improvements to a technology could be clearly distinguished as either hardware improvements (e.g., improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the methodology). However, with technological advancements, many methodological improvements today can be considered direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved methodology into the hardware circuit. Therefore, it cannot be said that a methodological improvement cannot be implemented using hardware physical modules. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit whose logic function is determined by the user programming the device. Designers can program and "integrate" a digital system onto a PLD themselves, without needing chip manufacturers to design and manufacture dedicated integrated circuit chips. Furthermore, nowadays, instead of manually manufacturing integrated circuit chips, this programming is mostly implemented using "logic compiler" software. Similar to the software compiler used in program development, the original code before compilation must also be written in a specific programming language, called a Hardware Description Language (HDL). There are many HDLs, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, and RHDL (Ruby Hardware Description Language). Currently, the most commonly used are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also understand that by simply performing some logic programming on the method flow using one of these hardware description languages ​​and programming it into an integrated circuit, the hardware circuit implementing the logical method flow can be easily obtained.

[0100] The controller can be implemented in any suitable manner. For example, it can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code form, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included therein for implementing various functions can also be considered as structures within the hardware component. Alternatively, the means for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0101] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0102] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware components.

[0103] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0104] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0105] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0106] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0107] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0108] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0109] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0110] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0111] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0112] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0113] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0114] The above description is merely one or more embodiments of this specification and is not intended to limit this specification. Various modifications and variations can be made to the one or more embodiments of this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of this specification.

Claims

1. A drill-down analysis method based on user movement, comprising: For the current business, the business logs of the users executing the business are parsed to extract the corresponding first user behavior chain; The first user behavior chain is semantically associated with the current business through multiple preset business dimensions to obtain the second user behavior chain; For the second user behavior chain, the corresponding key parameters are extracted, and the corresponding business churn logic features are extracted based on the key parameters. Based on the business churn logic features, the target users who have experienced business churn are identified among the users who performed the business. Drill-down analysis is performed based on the first user behavior chain corresponding to the target user to extract the corresponding path features.

2. The method as described in claim 1, wherein the first user behavior chain is semantically associated with the current business through multiple preset business dimensions to obtain a second user behavior chain, specifically including: If the current business corresponds to an e-commerce business, then the preset multiple business dimensions are determined to include the business instance dimension, business channel dimension, and e-commerce product information summary dimension corresponding to the e-commerce business. The relevant information with semantic association in the first user behavior chain is updated using the business data corresponding to the multiple business dimensions.

3. The method as described in claim 2, wherein updating the semantically related information in the first user behavior chain using the business data corresponding to the multiple business dimensions specifically includes: Identify each node in the first user behavior chain; Based on the page type corresponding to each node, select nodes that match the preset type, select the corresponding business dimension for them through the preset association rule library, and update the information contained in the node; For nodes that do not conform to the preset type, the importance score of the node is determined based on the node's sequence position in the first user behavior chain and the page type corresponding to the node. For nodes whose importance score is higher than the preset score, the corresponding business dimension is selected for them through the preset association rule base, and the information contained in the node is updated.

4. The method as described in claim 1, specifically extracting the corresponding key parameters for the second user behavior chain, includes: The second user behavior chain is grouped into windows to obtain one or more user behavior sub-chains. For the user behavior subchain, the corresponding maximum instance index is extracted; For each node in the sequence corresponding to the maximum instance index, extract its corresponding business channel information, as well as the corresponding node duration information and node completion information.

5. The method as described in claim 4, wherein the corresponding business churn logic features are extracted based on the key parameters, and target users experiencing business churn are identified among the business execution users based on the business churn logic features, specifically including: For the target user, corresponding business churn logic features are extracted based on the business channel information, the node duration information, and the node completion information; wherein, the sub-business dimension of the node completion information includes the completion degree of the user's historical business, the completion degree of the current business, the node return action and task switching action in the current business; Based on the business churn logic characteristics, target users who have experienced business churn and meet the preset analysis requirements are identified among the users performing the business.

6. The method as described in claim 1, wherein drill-down analysis is performed based on the first user behavior chain corresponding to the target user to extract the corresponding path features, specifically including: Based on the user identifier corresponding to the target user, cross-platform behavior association is performed on the first user behavior chain; Filter the first user behavior chain to obtain key nodes; Based on the key nodes, construct the behavior sequence corresponding to the target user; The corresponding path features are extracted from the behavioral sequence and then visualized and mapped.

7. The method as described in claim 6, wherein cross-platform behavior association is performed on the first user behavior chain based on the user identifier corresponding to the target user, specifically including: Based on the user identifier corresponding to the target user, determine the first user behavior chain corresponding to the target user on each terminal; In the first user behavior chain corresponding to each terminal, select the first user behavior chain where the difference between timestamps is lower than the preset difference, take each node in it as the smallest splicing unit, sort and splice them according to the time sequence to obtain the intermediate user behavior chain. Perform business logic verification on the aforementioned intermediate user behavior chain; If there is a business logic anomaly, then perform intent recognition on the first user behavior chain corresponding to each terminal to obtain the business intent corresponding to each terminal. If the business intents corresponding to each terminal match, then the first user behavior chain corresponding to each terminal is used as the smallest splicing unit, and the chain is sorted and spliced ​​according to the time sequence to obtain the first user behavior chain after cross-terminal behavior association.

8. The method as described in claim 6, wherein constructing the behavior sequence corresponding to the target user based on the key nodes specifically includes: By using multiple pre-set window sorting dimensions, the key nodes are segmented to obtain multiple independent analysis units; For the analysis unit, sort it according to the time series to obtain the corresponding behavior sequence, and generate the corresponding sequence number based on the sequence order and page type of the key node.

9. The method as described in claim 6, wherein extracting corresponding path features from the behavior sequence specifically includes: For the key nodes contained in the behavior sequence, extract their corresponding page features; Based on the page features and the target user's operation records, extract the structured path features of the target user for each control in the key nodes; Based on the sequence number corresponding to each key node in the behavior sequence, the structured path features are combined to obtain the corresponding path features.

10. The method as described in claim 6, specifically including: performing visualization mapping, The path characteristics are standardized using a front-end language. Based on the standardized path features, generate corresponding visualization charts; Statistics are collected on each target user in the current business, and the collected traffic data is added to each node in the visualization chart.

11. A drilling analysis device based on user movement, comprising: The behavior chain extraction module analyzes the business logs of the user executing the business for the current business and extracts the corresponding first user behavior chain. The business semantic association module associates the first user behavior chain with the current business through multiple preset business dimensions to obtain the second user behavior chain; The target user determination module extracts the corresponding key parameters for the second user behavior chain, extracts the corresponding business churn logic features based on the key parameters, and determines the target users with business churn among the business execution users based on the business churn logic features. The path feature extraction module performs drill-down analysis based on the first user behavior chain corresponding to the target user to extract the corresponding path features.

12. The apparatus of claim 11, wherein the business semantic association module, if the current business corresponds to an e-commerce business, determines that the preset multiple business dimensions include the business instance dimension, business channel dimension, and e-commerce product information summary dimension corresponding to the e-commerce business; The relevant information with semantic association in the first user behavior chain is updated using the business data corresponding to the multiple business dimensions.

13. The apparatus of claim 12, wherein the business semantic association module determines each node in the first user behavior chain; Based on the page type corresponding to each node, select nodes that match the preset type, select the corresponding business dimension for them through the preset association rule library, and update the information contained in the node; For nodes that do not conform to the preset type, the importance score of the node is determined based on the node's sequence position in the first user behavior chain and the page type corresponding to the node. For nodes whose importance score is higher than the preset score, the corresponding business dimension is selected for them through the preset association rule base, and the information contained in the node is updated.

14. The apparatus of claim 11, wherein the target user determination module performs window grouping on the second user behavior chain to obtain one or more user behavior sub-chains; For the user behavior subchain, the corresponding maximum instance index is extracted; For each node in the sequence corresponding to the maximum instance index, extract its corresponding business channel information, as well as the corresponding node duration information and node completion information.

15. The apparatus of claim 14, wherein the target user determination module, for the target user, extracts corresponding business churn logic features based on the business channel information, the node duration information, and the node completion information; wherein, The sub-business dimensions of the node completion information include the completion rates of the user's historical business, the completion rate of the current business, the node return actions and task switching actions in the current business; Based on the business churn logic characteristics, target users who have experienced business churn and meet the preset analysis requirements are identified among the users performing the business.

16. The apparatus of claim 11, wherein the path feature extraction module performs cross-terminal behavior association on the first user behavior chain based on the user identifier corresponding to the target user; Filter the first user behavior chain to obtain key nodes; Based on the key nodes, construct the behavior sequence corresponding to the target user; The corresponding path features are extracted from the behavioral sequence and then visualized and mapped.

17. The apparatus of claim 16, wherein the path feature extraction module determines, based on the user identifier corresponding to the target user, the first user behavior chain corresponding to the target user on each terminal; In the first user behavior chain corresponding to each terminal, select the first user behavior chain where the difference between timestamps is lower than the preset difference, take each node in it as the smallest splicing unit, sort and splice them according to the time sequence to obtain the intermediate user behavior chain. Perform business logic verification on the aforementioned intermediate user behavior chain; If there is a business logic anomaly, then perform intent recognition on the first user behavior chain corresponding to each terminal to obtain the business intent corresponding to each terminal. If the business intents corresponding to each terminal match, then the first user behavior chain corresponding to each terminal is used as the smallest splicing unit, and the chain is sorted and spliced ​​according to the time sequence to obtain the first user behavior chain after cross-terminal behavior association.

18. The apparatus of claim 16, wherein the path feature extraction module segments the key node by pre-setting multiple window sorting dimensions to obtain multiple independent analysis units; For the analysis unit, sort it according to the time series to obtain the corresponding behavior sequence, and generate the corresponding sequence number based on the sequence order and page type of the key node.

19. The apparatus of claim 16, wherein the path feature extraction module extracts the corresponding page features for key nodes contained in the behavior sequence; Based on the page features and the target user's operation records, extract the structured path features of the target user for each control in the key nodes; Based on the sequence number corresponding to each key node in the behavior sequence, the structured path features are combined to obtain the corresponding path features.

20. The apparatus of claim 16, wherein the path feature extraction module standardizes the path features using a front-end language; Based on the standardized path features, generate corresponding visualization charts; Statistics are collected on each target user in the current business, and the collected traffic data is added to each node in the visualization chart.

21. A drilling analysis device based on user movement, comprising: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to: For the current business, the business logs of the users executing the business are parsed to extract the corresponding first user behavior chain; The first user behavior chain is semantically associated with the current business through multiple preset business dimensions to obtain the second user behavior chain; For the second user behavior chain, the corresponding key parameters are extracted, and the corresponding business churn logic features are extracted based on the key parameters. Based on the business churn logic features, the target users who have experienced business churn are identified among the users who performed the business. Drill-down analysis is performed based on the first user behavior chain corresponding to the target user to extract the corresponding path features.

Images