Logistics data analysis method, device, equipment and medium

Abstract

This invention provides a logistics data analysis method, apparatus, device, and medium, comprising: determining the corresponding basic data permission range based on user role information; obtaining user preference features based on user historical query behavior; dynamically calculating the optimal permission range within the basic data permission range using a dynamic permission recommendation model based on user preference features, business roles, and query intent; obtaining the query date range input by the user; optimizing the user-input query date range using a time window optimization model according to the current query intent to obtain an optimized query time range; obtaining corresponding logistics data based on the optimal permission range and the optimized query time range; and calculating intent indicators based on the logistics data and the current query intent.

Description

Technical Field

[0001] This invention relates to the field of logistics management technology, specifically to logistics data analysis methods, apparatus, equipment and media, and more specifically to logistics data analysis methods, apparatus, equipment and media based on permission adaptation and deep learning. Background Technology

[0002] Data analytics is crucial for optimizing business operations in the logistics industry. However, traditional logistics data analytics systems suffer from numerous technical shortcomings, failing to meet the industry's sophisticated operational needs. Firstly, access control relies on a simple role-based model, failing to adapt to the multi-level organizational structure of logistics companies (headquarters, regional offices, provincial headquarters, and distribution centers). This prevents dynamic adjustments to data access based on user business scope and query needs, resulting in insufficient data security and targeted analysis. Secondly, query time ranges are often fixed, lacking intelligent validation and optimization mechanisms for user-input dates. This makes it easy to obtain invalid data exceeding storage expiration dates or incomplete data inconsistent with business patterns, impacting data analysis accuracy and user experience. Thirdly, core logistics indicators such as on-time delivery rates, misclassification, loss, and damage are analyzed in isolation, lacking the ability to mine temporal correlations between multiple indicators. This hinders trend prediction and potential risk identification, and lacks automated anomaly detection and root cause analysis, making it difficult to extract valuable business insights from the data and providing effective support for logistics companies' operational decisions. Summary of the Invention

[0003] The main objective of this invention is to solve the technical problems in the prior art, such as the inability to dynamically adjust the data access scope according to the user's actual business scope and query needs, the lack of intelligent verification and optimization mechanisms for the query date entered by the user, and the easy acquisition of invalid data that exceeds the data storage period or incomplete data that does not conform to business rules.

[0004] The first aspect of this invention provides a logistics data analysis method, comprising: Determine the corresponding basic data permission scope based on user role information; Based on users’ historical query behavior, user preference features are obtained. Based on user preference features, business roles and query intent, the optimal permission range within the basic data permission range is dynamically calculated through a dynamic permission recommendation model. Obtain the query date range input by the user, and optimize the query date range input by the user using a time window optimization model based on the current query intent to obtain the optimized query time range; Match the corresponding logistics data based on the optimal permission range and the optimized query time range; Intent metrics are obtained by calculating intent metrics based on logistics data and current query intent.

[0005] Optionally, in a first implementation of the first aspect of the present invention, determining the corresponding basic data permission scope based on user role information includes: Based on the multi-level organizational structure of logistics enterprises, a role-level-basic data permission scope mapping system is constructed to form a basic permission mapping rule base; user role information is obtained, and the corresponding basic data permission scope is obtained based on the basic permission mapping rule base.

[0006] Optionally, in a second implementation of the first aspect of the present invention, the step of obtaining user preference features based on historical user query behavior, and dynamically calculating the optimal permission range within the basic data permission range based on user preference features, business roles, and query intent through a dynamic permission recommendation model, includes: Based on the user's business role, and combined with the business scope and data access responsibilities of each role in the logistics industry knowledge graph, a candidate data pool matching the business role is selected within the basic data permissions of the user role. The logistics industry knowledge graph is a semantic knowledge network built for logistics data analysis permission management scenarios. It is a standardized knowledge system that integrates logistics industry business rules, organizational structure, and data responsibilities. It establishes relationships between entities based on multi-level organizational roles, business scopes, data types, and data access responsibilities within the logistics industry. Obtain users' historical query behavior and generate user preference feature vectors based on the users' historical query behavior, including indicator preferences, range preferences, granularity preferences, and time preferences; Obtain the user's current query content, identify the user's query intent based on the current query content, and convert the current query intent into a standardized intent feature vector consistent with the dimensions of the user's preference feature vector; The user preference feature vector is fused with the standardized intent feature vector to generate a fused feature vector; Based on the candidate data pool, the optimal permission range is determined by further filtering through collaborative filtering algorithm according to the fused feature vector.

[0007] Optionally, in a third implementation of the first aspect of the present invention, the step of obtaining the query date range input by the user and optimizing the query date range input by the user using a time window optimization model based on the current query intent to obtain an optimized query time range includes: Obtain the user query date range, and verify the obtained user query date range based on the preset earliest query time and latest query time to obtain the verified query date range; Based on the user's query intent, the corresponding business periodicity pattern is matched, and the boundary of the verified query date range is optimized based on the matched business periodicity pattern to obtain the optimized query date range; The daily data quality score within the optimized query date range is calculated based on preset requirements. If the daily data quality score is less than the preset value, the current daily data is considered not to meet the preset requirements. If the completeness rate of the daily data that meets the preset requirements is greater than or equal to the preset value, the daily data that does not meet the preset requirements is repaired, and the repaired diary data is retained after the repair is successful. If repair is not possible, the optimal time range is selected based on the data quality score higher than the preset value and the longest continuous duration. If the duration of the selected continuous time range is less than the preset minimum analysis cycle, the dates are extended forward / backward within the compliance verification boundary to supplement high-quality data that meets the preset requirements until the minimum analysis cycle is met. If the extension still fails to meet the requirements, fragmented data reference results and alternative query time ranges that meet the preset requirements are pushed to the user, and finally, the optimized query time range that meets the analysis requirements is determined.

[0008] Optionally, in a fourth implementation of the first aspect of the present invention, the step of validating the obtained user query date range based on a preset earliest query time and latest query time includes: If the start time of the user's query date range is earlier than the preset earliest query time, the start time in the current user's query date range will be changed to the preset earliest query time; if the end time of the user's query date range is later than the preset latest query time, the end time in the current user's query date range will be changed to the preset latest query time.

[0009] Optionally, in a fifth implementation of the first aspect of the present invention, the method further includes: Construct a logistics industry indicator system that meets preset requirements, including: timeliness achievement rate, misclassification, loss and damage indicators; The association rules between various indicators in a logistics industry indicator system that meets preset requirements are analyzed by using a multi-indicator association deep learning model; wherein, the multi-indicator association deep learning model uses an LSTM network to process time series data of multiple indicators and learn the temporal association rules between indicators. Based on the intent index, a heatmap of the correlation between the indices associated with the intent index is generated according to the correlation rules between the indices obtained from the analysis. Based on the correlation heatmap, trend prediction of correlation indicators is performed to identify potential risks.

[0010] Optionally, in a sixth implementation of the first aspect of the present invention, the method further includes: Anomaly detection and attribution analysis are performed on the matched logistics data using anomaly detection and root cause analysis models. The anomaly detection and root cause analysis model includes: identifying abnormal data points using the isolated forest algorithm based on matched logistics data; and then mining root causes using the decision tree root cause analysis algorithm combined with association rules and / or knowledge graphs.

[0011] A second aspect of the present invention provides a logistics data analysis device, comprising: The basic data permission scope acquisition module is used to determine the corresponding basic data permission scope based on user role information; The optimal permission range determination module is used to obtain user preference features based on users' historical query behavior, and dynamically calculate the optimal permission range within the basic data permission range based on user preference features, business roles, and query intent through a dynamic permission recommendation model. The query time range optimization module is used to obtain the query date range input by the user, and optimize the query date range input by the user using a time window optimization model based on the current query intent to obtain the optimized query time range; The logistics data acquisition module is used to match the corresponding logistics data based on the optimal permission range and the optimized query time range. The intent index calculation module is used to calculate intent indexes based on logistics data and the current query intent.

[0012] Optionally, in a first implementation of the second aspect of the present invention, the basic data permission scope acquisition module includes: Based on the multi-level organizational structure of logistics enterprises, a role-level-basic data permission scope mapping system is constructed to form a basic permission mapping rule base; user role information is obtained, and the corresponding basic data permission scope is obtained based on the basic permission mapping rule base.

[0013] Optionally, in a second implementation of the second aspect of the present invention, the optimal permission range determination module includes: The candidate data pool acquisition submodule is used to filter out candidate data pools that match the user's business role within the user's basic data permissions, based on the user's business role and the business scope and data access responsibilities of each role in the logistics industry knowledge graph. The logistics industry knowledge graph is a semantic knowledge network built for logistics data analysis permission management scenarios. It is a standardized knowledge system that integrates logistics industry business rules, organizational structure, and data responsibilities. It establishes relationships between entities based on multi-level organizational roles, business scopes, data types, and data access responsibilities within the logistics industry. The user preference feature vector construction submodule is used to obtain users' historical query behavior and generate user preference feature vectors including indicator preference, range preference, granularity preference and time preference based on users' historical query behavior; The standardized intent feature vector construction submodule is used to obtain the user's current query content, identify the user's query intent based on the current query content, and convert the current query intent into a standardized intent feature vector with the same dimension as the user's preference feature vector. The feature vector fusion generation submodule is used to fuse the user preference feature vector with the standardized intent feature vector to generate a fused feature vector; The permission range determination submodule is used to further filter candidates based on the candidate data pool using a collaborative filtering algorithm according to the fused feature vector, and finally determine the optimal permission range.

[0014] Optionally, in a third implementation of the second aspect of the present invention, the query time range optimization module includes: The validation submodule is used to obtain the user's query date range, and to validate the obtained user query date range according to the preset earliest query time and latest query time to obtain the validated query date range. The optimization submodule is used to match the corresponding business periodicity pattern according to the user's query intent, and to perform boundary optimization on the verified query date range based on the matched business periodicity pattern to obtain the optimized query date range; The scoring and filtering submodule is used to calculate the daily data quality score within the optimized query date range according to preset requirements. When the daily data quality score is less than the preset value, the current daily data is considered not to meet the preset requirements. If the completeness rate of the daily data that meets the preset requirements is greater than or equal to the preset value, the daily data that does not meet the preset requirements is repaired, and the repaired diary data is retained after the repair is successful. If repair is not possible, the optimal time range is selected as the data quality score that is higher than the preset value and has the longest continuous duration. If the duration of the selected continuous time range is less than the preset minimum analysis cycle, the date range is extended forward / forward within the compliance verification boundary to supplement high-quality data that meets the preset requirements until the minimum analysis cycle is met. If the extension still fails to meet the requirements, fragmented data reference results and alternative query time ranges that meet the preset requirements are pushed to the user, and finally, the optimized query time range that meets the analysis requirements is determined.

[0015] Optionally, in a fourth implementation of the second aspect of the present invention, the verification submodule includes: If the start time of the user's query date range is earlier than the preset earliest query time, the start time in the current user's query date range will be changed to the preset earliest query time; if the end time of the user's query date range is later than the preset latest query time, the end time in the current user's query date range will be changed to the preset latest query time.

[0016] Optionally, in a fifth implementation of the second aspect of the present invention, the apparatus further includes: a prediction module, configured to predict the trend of indicators associated with intent indicators through a multi-indicator association deep learning model, and identify potential risks; The multi-index association deep learning model includes: Construct a logistics industry indicator system that meets preset requirements, including: timeliness achievement rate, misclassification, loss and damage indicators; An LSTM network is used to process time series data of multiple indicators in a logistics industry indicator system that meets preset requirements, and to obtain the temporal correlation rules between the indicators. Based on the intent index, a heatmap of the correlation between the indices associated with the intent index is generated according to the correlation rules between the indices obtained from the analysis. Based on the correlation heatmap, trend prediction of correlation indicators is performed to identify potential risks.

[0017] Optionally, in a sixth implementation of the first aspect of the present invention, the apparatus further includes: an anomaly attribution analysis module, used to perform anomaly detection and attribution analysis on the matched logistics data through an anomaly detection and root cause analysis model; The anomaly detection and root cause analysis model includes: identifying abnormal data points using the isolated forest algorithm based on matched logistics data; and then performing root cause mining by combining the decision tree root cause analysis algorithm with association rules and / or knowledge graphs.

[0018] A third aspect of the present invention provides an electronic device, the electronic device comprising a memory and at least one processor, wherein the memory stores instructions; The at least one processor invokes the instructions in the memory to cause the electronic device to perform the various steps of the logistics data analysis method as described above.

[0019] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions, characterized in that the instructions, when executed by a processor, implement the various steps of the logistics data analysis method described above.

[0020] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention breaks away from the traditional, simple role-based permission model. Based on a role-hierarchy-basic data permission mapping system, it defines compliant basic permission boundaries from the bottom up, adhering to the "principle of least privilege" to ensure that core logistics data is not accessed without authorization, thus significantly improving data security. At the same time, relying on a dynamic permission recommendation model, it integrates user business roles, historical query preferences, and current query intent to achieve personalized and precise filtering of data permissions within the basic permissions. This allows users to obtain only data highly relevant to their work needs, effectively avoiding redundant data interference, improving the targeting and efficiency of data access, and solving the problems of insufficient data security and analysis targeting in existing technologies. 2. This invention employs a three-pronged progressive approach—compliance verification, business cycle optimization, and data quality stratification—to optimize the user-input query date range using a time window optimization model. This approach avoids invalid queries that exceed data storage expiration dates from the outset, ensuring that the time range aligns with the temporal patterns of logistics operations. Simultaneously, through refined design including data repair, compliance boundary completion, and multiple result pushes, it maximizes the use of effective data while guaranteeing data quality and temporal continuity. This avoids analytical distortion caused by incomplete data or time gaps, solving the problems of fixed query time ranges and easy acquisition of invalid / incomplete data in existing technologies. This lays a high-quality time-dimensional data foundation for subsequent data analysis. 3. This invention constructs a core indicator system for the logistics industry, including timeliness achievement rate, misclassification, loss, and damage. It uses an LSTM network to process multi-indicator time series data, accurately mining the temporal correlation rules between indicators, and visualizing the correlation relationships through a correlation heatmap. Simultaneously, it completes trend prediction of related indicators based on the correlation rules, and automatically identifies potential operational risks by combining preset risk thresholds. This achieves an upgrade from "single indicator calculation" to "multi-indicator correlation mining" and from "post-event analysis" to "pre-event warning," solving the problems of existing technologies lacking multi-indicator time series correlation mining capabilities and being unable to achieve trend prediction and potential risk identification. It can help logistics companies avoid operational losses in advance and improve operational stability. Attached Figure Description

[0021] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a first flowchart of a logistics data analysis method provided in an embodiment of the present invention.

[0022] Figure 2 The second flowchart of the logistics data analysis method provided in the embodiment of the present invention.

[0023] Figure 3 The third flowchart of the logistics data analysis method provided in the embodiments of the present invention.

[0024] Figure 4 This is a schematic diagram of a logistics data analysis device provided in an embodiment of the present invention.

[0025] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0026] This invention provides a logistics data analysis method, apparatus, device, and medium, comprising: determining the corresponding basic data permission range based on user role information; obtaining user preference features based on user historical query behavior; dynamically calculating the optimal permission range within the basic data permission range using a dynamic permission recommendation model based on user preference features, business roles, and query intent; obtaining the query date range input by the user; optimizing the user-input query date range using a time window optimization model according to the current query intent to obtain an optimized query time range; obtaining corresponding logistics data based on the optimal permission range and the optimized query time range; and calculating intent indicators based on the logistics data and the current query intent. This invention solves the technical problems in the prior art where the data access range cannot be dynamically adjusted according to the user's actual business scope and query needs, where there is a lack of intelligent verification and optimization mechanisms for user-input query dates, and where invalid data exceeding the data storage period or incomplete data that does not conform to business rules is easily obtained.

[0027] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0028] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 The first embodiment of the logistics data analysis method in this invention includes: 101. Determine the corresponding basic data permission scope based on user role information; In this embodiment, a role-level-basic data permission scope mapping system is constructed based on the multi-level organizational structure of the logistics enterprise to form a basic permission mapping rule base; user role information is obtained, and the corresponding basic data permission scope is obtained based on the basic permission mapping rule base.

[0029] This embodiment achieves precise matching between user roles and basic data permission scopes, laying a compliant and adaptable permission foundation for subsequent dynamic calculation of the optimal permission scope.

[0030] 102. Based on users' historical query behavior, obtain user preference features. Based on user preference features, business roles, and query intent, dynamically calculate the optimal permission range within the basic data permission range through a dynamic permission recommendation model. In this embodiment, based on the user's business role and combined with the business scope and data access responsibilities of each role in the logistics industry knowledge graph, a candidate data pool matching the business role is selected within the basic data permissions of the user role. The logistics industry knowledge graph is a semantic knowledge network built for logistics data analysis permission management scenarios. It is a standardized knowledge system that integrates logistics industry business rules, organizational structure, and data responsibilities. It establishes relationships between entities based on multi-level organizational roles, business scopes, data types, and data access responsibilities within the logistics industry. Obtain users' historical query behavior and generate user preference feature vectors based on the users' historical query behavior, including indicator preferences, range preferences, granularity preferences, and time preferences; Obtain the user's current query content, identify the user's query intent based on the current query content, and convert the current query intent into a standardized intent feature vector consistent with the dimensions of the user's preference feature vector; The user preference feature vector is fused with the standardized intent feature vector to generate a fused feature vector; Based on the candidate data pool, the optimal permission range is determined by further filtering through collaborative filtering algorithm according to the fused feature vector.

[0031] This embodiment builds upon the aforementioned defined scope of basic data permissions, relies on a dynamic permission recommendation model, and integrates user business roles, historical query preferences, and current query intent. Figure 3 Within the core dimensions, data permissions are refined and personalized within the basic permission boundaries to ultimately determine the optimal permission range.

[0032] 103. Obtain the query date range input by the user, and optimize the query date range input by the user using the time window optimization model based on the current query intent to obtain the optimized query time range; In this embodiment, the user query date range is obtained, and the obtained user query date range is verified according to the preset earliest query time and the preset latest query time. If the start time of the user query date range is earlier than the preset earliest query time, the start time in the current user query date range is changed to the preset earliest query time; if the end time of the user query date range is later than the preset latest query time, the end time in the current user query date range is changed to the preset latest query time, thus obtaining the verified query date range. Based on the user's query intent, the corresponding business periodicity pattern is matched, and the boundary of the verified query date range is optimized based on the matched business periodicity pattern to obtain the optimized query date range; The daily data quality score within the optimized query date range is calculated based on preset requirements. If the daily data quality score is less than the preset value, the current daily data is considered not to meet the preset requirements. If the completeness rate of the daily data that meets the preset requirements is greater than or equal to the preset value, the daily data that does not meet the preset requirements is repaired, and the repaired diary data is retained after the repair is successful. If repair is not possible, the optimal time range is selected based on the data quality score higher than the preset value and the longest continuous duration. If the duration of the selected continuous time range is less than the preset minimum analysis cycle, the dates are extended forward / backward within the compliance verification boundary to supplement high-quality data that meets the preset requirements until the minimum analysis cycle is met. If the extension still fails to meet the requirements, fragmented data reference results and alternative query time ranges that meet the preset requirements are pushed to the user, and finally, the optimized query time range that meets the analysis requirements is determined.

[0033] This embodiment is the core step in time-dimensional data filtering and optimization in logistics data analysis methods. Building upon the aforementioned optimal permission range, it focuses on the time dimension of user queries. Relying on a time window optimization model, it intelligently optimizes the original query date range input by the user through a three-tiered progressive process of compliance verification, business optimization, and data quality filtering. This addresses the technical shortcomings of traditional logistics data analysis systems, such as fixed query time ranges and the ease with which invalid / incomplete data can be obtained. It ensures that the subsequently matched logistics data is compliant in the time dimension, conforms to business patterns, and meets data quality standards, laying a data foundation for accurate logistics data analysis in the time dimension.

[0034] 104. Match the corresponding logistics data based on the optimal permission range and the optimized query time range; 105. Based on logistics data and the current query intent, calculate the intent index to obtain the intent index.

[0035] This embodiment addresses the shortcomings of traditional logistics data analysis systems, such as rigid access control, lack of intelligent optimization of query time ranges, and limited indicator analysis. First, it matches basic data access ranges to user roles based on the multi-level organizational structure of logistics enterprises. Then, it integrates users' historical query preferences and current query intent, using a dynamic access recommendation model to determine the optimal access range within the basic permissions. Simultaneously, it utilizes a time window optimization model to perform compliance checks, business cycle optimization, and data quality filtering on the user-input query date range. Finally, it combines the optimal access range with logistics data and calculates intent indicators, achieving dynamic and refined control of logistics data permissions and intelligent optimization of query time ranges. This provides precise data support for the refined operation and scientific decision-making of logistics enterprises.

[0036] Please see Figure 2 The second embodiment of the logistics data analysis method in this invention includes: 201. Determine the corresponding basic data permission scope based on user role information; In this embodiment, a role-level-basic data permission scope mapping system is first constructed to form a basic permission mapping rule library. Specifically, the organizational structure of logistics enterprises is first standardized and decomposed to clarify the business jurisdiction and data access boundaries of each level. Typical levels include headquarters level, regional level, provincial level, and distribution center level. The business jurisdiction of each level is covered in a hierarchical manner of "headquarters unified management → regional management → provincial implementation → distribution execution". The data access scope is also gradually refined and localized as the level goes down.

[0037] Based on the business functions of each level, standard business roles are defined to adapt to logistics operations. These roles are strongly linked to their respective levels, ensuring that the responsibilities and authority of each role match the business scope of its level. For example: Headquarters level: Chief Logistics Operations Officer, Data Analyst, Strategic Planner; Regional level: Regional Operations Manager, Regional Data Specialist; Provincial level: Provincial head, provincial operations manager; Distribution center level: Distribution manager, distribution data entry clerk, on-site operations specialist.

[0038] Define the basic data permission scope for each role. This scope refers to the minimum compliant boundary of logistics data that a role can access. The definition is based on the business jurisdiction of the role's level and the job responsibilities, while adhering to the "principle of least privilege" to ensure that roles can only access the basic data necessary to complete their duties, thus guaranteeing logistics data security. The structured correspondence between organizational hierarchy, business roles, and basic data permission scopes is stored in a structured manner, forming a basic permission mapping rule base. This rule base is a configurable and updatable structured database, supporting real-time modifications based on organizational structure adjustments and role / functional changes, ensuring flexibility in permission control.

[0039] Obtain user role information and retrieve the corresponding basic data permission range based on the basic permission mapping rule library; Specifically, when a user logs into the logistics data analysis system and initiates a data query request, the system automatically extracts core role information from the user's account information, including the user's organizational level and assigned standard business role. At the same time, the system verifies the validity of the role information to confirm that the role has been registered in the basic permission mapping rule base. If it is an unregistered non-standard role, a permission exception reminder will be triggered, and the administrator must complete the role registration before permission matching can be performed.

[0040] The extracted and verified user role information is used as the search criteria to perform a precise search in the basic permission mapping rule library, matching the unique basic data permission range corresponding to the role.

[0041] Once the matching is complete, the corresponding basic data permission range is loaded into the user's query session in real time. This serves as the basic permission boundary for the user's current and subsequent queries. All of the user's logistics data query operations cannot exceed this basic range, ensuring the compliance and security of data access from the ground up. Simultaneously, this basic permission range will serve as the core basis for dynamically calculating the optimal permission range in subsequent steps. The optimal permission range will be finely filtered within this basic range, without exceeding the boundaries of the basic permissions.

[0042] 202. Based on users' historical query behavior, obtain user preference features. Based on user preference features, business roles, and query intent, dynamically calculate the optimal permission range within the basic data permission range through a dynamic permission recommendation model. In this embodiment, the basic data permission scope is used as an inviolable boundary, and the role association information retrieved from the logistics industry knowledge graph is used as the filtering condition. All logistics data within the basic permission scope is precisely filtered, and only data that highly matches the work responsibilities and business scope of the user's business role is retained to form a candidate data pool. The logistics industry knowledge graph is a semantic knowledge network built for logistics data analysis permission management. It has pre-stored the entity association relationships of multi-level organizational roles, business scopes, data types, and data access responsibilities in the logistics industry, as well as the core work scenarios and data requirement matching rules of each role.

[0043] The system automatically collects all historical query behavior data from users in the logistics data analysis system, including but not limited to the type of indicators used in historical queries, the geographical scope of queries, the data granularity of queries, the time period preferences for queries, and high-frequency operations in historical queries. Simultaneously, the collected raw behavioral data is cleaned to remove invalid data such as erroneous operations, duplicate queries, and test queries, ensuring the accuracy of preference extraction. Based on the business characteristics of logistics data analysis, user preferences are divided into four core dimensions: indicator preference, scope preference, granularity preference, and time preference. Each dimension undergoes standardized quantification, transforming non-numerical preferences into calculable values. The quantified values ​​of the four dimensions are combined in a fixed order to form a user preference feature vector. If a user is a new user with no historical query behavior, the standard preference vector of the corresponding business role in the logistics industry knowledge graph is used as the initial value.

[0044] The system collects the user's current query command, including entered keywords, selected query conditions, and clicked function modules. Natural Language Processing (NLP) technology is used to perform semantic recognition and intent parsing on the query content, accurately extracting the core of the user's current query needs, including key information such as target metrics, target geographic scope, target data granularity, and target time range. The parsed current query intent is then numerically processed according to dimensions and quantification standards completely consistent with the user preference feature vector. The four dimensions of the quantified current query intent are combined in the same fixed order as the user preference feature vector to form a standardized intent feature vector. Its dimensions, numerical range, and quantification standards are completely consistent with the user preference feature vector, providing a prerequisite for subsequent feature vector fusion. The user preference feature vector and the standardized intent feature vector are then fused to generate a fused feature vector. Based on the candidate data pool, the optimal permission range is determined by a collaborative filtering algorithm according to the fused feature vector.

[0045] 203. Obtain the query date range input by the user, and optimize the query date range input by the user using the time window optimization model based on the current query intent to obtain the optimized query time range; In this embodiment, the original query date range entered by the user in the logistics data analysis system is collected, including the start and end times of the query. Users can submit query requests in various ways, such as manually entering the date, selecting the time range, and quick selection. All input formats are uniformly converted into a standardized timestamp format. The logistics data analysis system pre-configures two core time boundary rules: the earliest query time and the latest query time. These rules are set based on the logistics company's data storage strategy, business data retention period, and industry regulatory requirements, and can be flexibly adjusted by the administrator according to the company's actual needs. The standardized original user query date range is compared one by one with the preset time boundaries, and time ranges exceeding the boundaries are automatically adjusted according to hard correction rules to finally obtain the verified query date range. The correction rules are as follows: If the user's query start time is earlier than the preset earliest query time, the query start time will be automatically corrected to the preset earliest query time. If the user's query end time is later than the preset latest query time, the query end time will be automatically corrected to the preset latest query time; If the start time entered by the user is later than the end time, an input error reminder will be triggered, guiding the user to re-enter a valid time range; If the user's original query date range is completely within the preset time boundary, the original time range will be directly retained as the result after verification.

[0046] A pre-built database of periodic patterns in logistics operations is constructed within the logistics data analysis system. This database is based on the business characteristics of the logistics industry, historical operational data, and business analysis needs corresponding to different query intents. It covers typical time-cycle patterns and business scenario correlation patterns in the logistics industry, and its core components include: General cyclical patterns: weekly patterns (e.g., logistics order volume is higher on weekends than on weekdays), monthly patterns (e.g., order volume surges at the end of the month), quarterly patterns (e.g., a surge in order volume during major e-commerce promotions), and holiday patterns (e.g., peaks and troughs in logistics business during Spring Festival, 618, and Double 11). Intent-based periodic patterns: Match specific business analysis cycles to different query intents. For example, querying "timeliness achievement rate" is adapted to an analysis cycle based on natural weeks / days, querying "monthly operation summary" is adapted to a complete natural month cycle, querying "logistics performance during major promotions" is adapted to a continuous cycle before and after major promotions, and querying "daily operation efficiency of distribution centers" is adapted to a fine cycle of 24 hours / natural days.

[0047] By using the user's current query intent as the search keyword, the system accurately retrieves uniquely matching business periodic patterns from the pattern database, which serve as the basis for optimizing the time range boundaries.

[0048] Boundary optimization rules based on business cycle patterns: Based on the validated query date range, and following the principles of "aligning with the business analysis cycle, ensuring data integrity, and not exceeding compliance boundaries," the start / end boundaries of the time range are flexibly adjusted. The core optimization logic is to complete the time boundaries to match the full nodes of the business cycle, avoiding incomplete time ranges such as "half a cycle" or "incomplete business periods." Specific optimization methods are illustrated below: If the query intent is "analyze the timeliness achievement rate of the East China distribution center this week", and the validated time range is March 4, 2026 - March 7, 2026 (Wednesday to Saturday), and the matching business cycle is a natural week (Monday to Sunday), then the start time will be optimized to March 2, 2026 (Monday), and the end time will remain March 8, 2026 (Friday), completing the start node of the natural week; If the query intent is "analyze the monthly misclassification rate in February", and the validated time range is February 5, 2026 - February 28, 2026, and the matching business cycle is a complete calendar month, then the start time will be optimized to February 1, 2026 to form a complete monthly time range. If the query intent is "analyze the damage rate of logistics during the 618 promotion", and the matching business cycle is a dedicated cycle of 3 days before the promotion to 7 days after the promotion, then the verified time range will be completed with the full nodes of that dedicated cycle.

[0049] After boundary optimization, an optimized query date range is generated. This range is not only completely within the compliant time boundaries verified in the first stage, but also highly consistent with the business periodicity corresponding to the user's query intent, ensuring both compliance and business adaptability of the time range. If the verified time range completely matches the business periodicity, it is directly retained as the optimized result.

[0050] According to the preset logistics data quality evaluation index system, each day within the optimized query date range is given an independent comprehensive data quality score, with higher scores indicating better data quality for that day.

[0051] If the daily data quality score is greater than or equal to the preset value, it is judged as qualified daily data that meets the preset requirements; If the daily data quality score is less than the preset value, it is judged as unqualified daily data that does not meet the preset requirements.

[0052] To determine whether the completeness rate of qualified daily data meets the standard, calculate the proportion of qualified daily data within the optimized date range and compare it with the preset completeness rate threshold. The core purpose of this step is to determine the impact of unqualified data. If the completeness rate of qualified data is ≥ the preset threshold, it means that the unqualified data is a small number of scattered points and has repair value. If the completeness rate of qualified data is < the preset threshold, it means that the proportion of unqualified data is too high and has no repair significance. When non-compliant daily data can be repaired, time series interpolation is used to repair a small number of scattered non-compliant daily data points, maximizing the preservation of the integrity of the original query time range and avoiding time breaks caused by a single non-compliant data point. The quality score of the repaired daily data is recalculated. If the score is greater than or equal to the preset value, the repair is considered successful, and the repaired daily data is retained. In this case, the optimized date range is a range that is fully compliant and time-continuous, and is directly used as the final optimized query time range. If the score after repair is less than the preset value, it is determined that it cannot be repaired, and the longest continuous compliant time range is selected. When there is no point in repairing, select the optimal time range that meets the data quality standards and has the longest continuous duration. When unqualified data cannot be repaired, the optimized query date range is continuously traversed and judged to mark all sub-time intervals with acceptable data quality and continuous time. Then, the longest sub-interval is extracted as the initial optimal time range. The duration of the initial optimal time range is compared with the preset minimum analysis cycle to ensure that the final output time range has practical data analysis value and to avoid statistically meaningless analysis results due to excessively short duration. If the initial optimal duration is greater than or equal to the preset minimum analysis period: this range meets the analysis requirements and can be directly used as a candidate query time range after optimization. If the initial optimal duration is less than the preset minimum analysis cycle, then within the time boundary of compliance verification, the system will automatically traverse in the direction before / after the initial optimal time range to supplement daily data with qualified data quality until the duration of the supplemented time range meets the preset minimum analysis cycle, thus forming the complete optimal time range. When a continuous range of compliant time that meets the minimum analysis cycle cannot be found within the compliance boundaries, forced filtering is no longer implemented. Instead, two types of compliant results are pushed to the user, balancing data utilization and user query experience, and avoiding query failures due to data issues. Fragmented data reference results: Outputs the single indicator calculation results of all fragmented daily data that meet the data quality standards, and clearly marks "Time is not continuous, only for single indicator reference, does not support trend analysis or correlation analysis" in the results, to provide users with basic data reference; Compliant alternative query time range: Retrieve the recent time range that meets the data quality standards, is continuous in time, and fits the original query intent and business pattern, and push it to the user as an alternative solution, while simultaneously displaying an explanation of the business relevance of the alternative range.

[0053] 204. Match the corresponding logistics data based on the optimal permission range and the optimized query time range; 205. Based on logistics data and the current query intent, calculate the intent index to obtain the intent index.

[0054] Please see Figure 3 The third embodiment of the logistics data analysis method in this invention includes: 301. Determine the corresponding basic data permission scope based on user role information; 302. Based on users' historical query behavior, obtain user preference features. Based on user preference features, business roles, and query intent, dynamically calculate the optimal permission range within the basic data permission range through a dynamic permission recommendation model. 303. Obtain the query date range input by the user, and optimize the query date range input by the user using the time window optimization model based on the current query intent to obtain the optimized query time range; 304. Match the corresponding logistics data based on the optimal permission range and the optimized query time range; 305. Based on logistics data and the current query intent, calculate the intent index to obtain the intent index; 306. Predict the trends of indicators associated with intent indicators and identify potential risks through a multi-indicator association deep learning model; In this embodiment, a logistics industry indicator system that meets preset requirements is constructed, including: timeliness achievement rate, misclassification, loss and damage indicators; An LSTM network is used to process time series data of multiple indicators in a logistics industry indicator system that meets preset requirements, and to obtain the temporal correlation rules between the indicators. Based on the intent index, a heatmap of the correlation between the indices associated with the intent index is generated according to the correlation rules between the indices obtained from the analysis. Based on the correlation heatmap, trend prediction of correlation indicators is performed to identify potential risks.

[0055] 307. Perform anomaly detection and attribution analysis on the matched logistics data using anomaly detection and root cause analysis models; In this embodiment, the anomaly detection and root cause analysis model includes: identifying abnormal data points using the isolated forest algorithm based on the matched logistics data; and then performing root cause mining by combining the decision tree root cause analysis algorithm with association rules and / or knowledge graphs.

[0056] This embodiment breaks through the limitations of traditional logistics data analysis's single-dimensional indicator calculation. Relying on deep learning and machine learning algorithms, it achieves in-depth correlation mining, trend prediction, anomaly location, and root cause tracing of core logistics indicators. It upgrades from "data calculation" to "data insight and risk warning," providing logistics companies with accurate trend references and problem solutions for operational decisions, and realizing intelligent and in-depth logistics data analysis.

[0057] The logistics data analysis method in the embodiments of the present invention has been described above. The logistics data analysis device in the embodiments of the present invention will be described below. Please refer to [link / reference]. Figure 4 One embodiment of the logistics data analysis device in this invention includes: The basic data permission scope acquisition module 401 is used to determine the corresponding basic data permission scope based on user role information. In this embodiment, the basic data permission range acquisition module 401 includes: Based on the multi-level organizational structure of logistics enterprises, a role-level-basic data permission scope mapping system is constructed to form a basic permission mapping rule base; user role information is obtained, and the corresponding basic data permission scope is obtained based on the basic permission mapping rule base.

[0058] The optimal permission range determination module 402 is used to obtain user preference features based on the user's historical query behavior, and dynamically calculate the optimal permission range within the basic data permission range based on the user preference features, business roles and query intent through a dynamic permission recommendation model. In this embodiment, the optimal permission range determination module 402 includes: The candidate data pool acquisition submodule 4021 is used to filter out candidate data pools that match the user's business role within the user's basic data permission scope, based on the user's business role and the business scope and data access responsibilities of each role in the logistics industry knowledge graph. The logistics industry knowledge graph is a semantic knowledge network built for logistics data analysis permission management scenarios. It is a standardized knowledge system that integrates logistics industry business rules, organizational structure, and data responsibilities. It establishes relationships between entities based on multi-level organizational roles, business scopes, data types, and data access responsibilities within the logistics industry. The user preference feature vector construction submodule 4022 is used to obtain users' historical query behavior and generate user preference feature vectors including indicator preference, range preference, granularity preference and time preference based on users' historical query behavior; The standardized intent feature vector construction submodule 4023 is used to obtain the user's current query content, identify the user's query intent based on the user's current query content, and convert the current query intent into a standardized intent feature vector with the same dimension as the user's preference feature vector. The feature vector fusion generation submodule 4024 is used to fuse the user preference feature vector with the standardized intent feature vector to generate a fused feature vector; The permission range determination submodule 4025 is used to further filter based on the candidate data pool and the fused feature vector through a collaborative filtering algorithm to finally determine the optimal permission range.

[0059] The query time range optimization module 403 is used to obtain the query date range input by the user, and optimize the query date range input by the user using a time window optimization model according to the current query intent to obtain the optimized query time range; In this embodiment, the query time range optimization module 403 includes: The verification submodule 4031 is used to obtain the user's query date range and verify the obtained user query date range according to the preset earliest query time and latest query time. If the start time of the user's query date range is earlier than the preset earliest query time, the start time in the current user query date range is changed to the preset earliest query time; if the end time of the user query date range is later than the preset latest query time, the end time in the current user query date range is changed to the preset latest query time, thus obtaining the verified query date range. The optimization submodule 4032 is used to match the corresponding business periodicity pattern according to the user's query intent, and to perform boundary optimization on the verified query date range based on the matched business periodicity pattern to obtain the optimized query date range; The scoring and filtering submodule 4033 is used to calculate the daily data quality score within the optimized query date range according to preset requirements. When the daily data quality score is less than the preset value, the current daily data is considered not to meet the preset requirements. If the completeness rate of the daily data that meets the preset requirements is greater than or equal to the preset value, the daily data that does not meet the preset requirements is repaired, and the repaired diary data is retained after the repair is successful. If repair is not possible, the data quality score that is higher than the preset value and has the longest continuous duration is selected as the optimal time range. If the duration of the selected continuous time range is less than the preset minimum analysis cycle, the dates are extended forward / backward within the compliance verification boundary to supplement high-quality data that meets the preset requirements until the minimum analysis cycle is met. If the extension still fails to meet the requirements, fragmented data reference results and alternative query time ranges that meet the preset requirements are pushed to the user, and finally the optimized query time range that meets the analysis requirements is determined.

[0060] The logistics data acquisition module 404 is used to match the corresponding logistics data based on the optimal permission range and the optimized query time range. The intent index calculation module 405 is used to calculate intent indexes based on logistics data and the current query intent.

[0061] The prediction module 406 is used to predict the trend of indicators associated with intent indicators and identify potential risks through a multi-indicator association deep learning model. In this embodiment, the multi-index correlation deep learning model includes: Construct a logistics industry indicator system that meets preset requirements, including: timeliness achievement rate, misclassification, loss and damage indicators; An LSTM network is used to process time series data of multiple indicators in a logistics industry indicator system that meets preset requirements, and to obtain the temporal correlation rules between the indicators. Based on the intent index, a heatmap of the correlation between the indices associated with the intent index is generated according to the correlation rules between the indices obtained from the analysis. Based on the correlation heatmap, trend prediction of correlation indicators is performed to identify potential risks.

[0062] The anomaly attribution analysis module 407 is used to perform anomaly detection and attribution analysis on the matched logistics data through anomaly detection and root cause analysis models. In this embodiment, the anomaly detection and root cause analysis model includes: identifying abnormal data points using the isolated forest algorithm based on the matched logistics data; and then performing root cause mining by combining the decision tree root cause analysis algorithm with association rules and / or knowledge graphs.

[0063] above Figure 4 The logistics data analysis device in this embodiment of the invention will be described in detail from the perspective of modular functional entities. The electronic device in this embodiment of the invention will be described in detail from the perspective of hardware processing.

[0064] Figure 5 This is a schematic diagram of the structure of an electronic device 700 provided in an embodiment of the present invention. The electronic device 700 can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 710 (e.g., one or more processors) and a memory 720, and one or more storage media 730 (e.g., one or more mass storage devices) for storing application programs 733 or data 732. The memory 720 and storage media 730 can be temporary or persistent storage. The program stored in the storage media 730 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the electronic device 700. Furthermore, the processor 710 may be configured to communicate with the storage media 730 and execute the series of instruction operations in the storage media 730 on the electronic device 700.

[0065] Electronic device 700 may also include one or more power supplies 740, one or more wired or wireless network interfaces 750, one or more input / output interfaces 750, and / or one or more operating systems 731, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 5 The illustrated electronic device structure does not constitute a limitation on electronic devices and may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0066] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of a logistics data analysis method.

[0067] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system, device, or unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0068] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0069] The above-described 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 the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A logistics data analysis method, characterized in that, include: Determine the corresponding basic data permission scope based on user role information; Based on users’ historical query behavior, user preference features are obtained. Based on user preference features, business roles and query intent, the optimal permission range within the basic data permission range is dynamically calculated through a dynamic permission recommendation model. Obtain the query date range input by the user, and optimize the query date range input by the user using a time window optimization model based on the current query intent to obtain the optimized query time range; Match the corresponding logistics data based on the optimal permission range and the optimized query time range; Intent metrics are obtained by calculating intent metrics based on logistics data and current query intent.

2. The logistics data analysis method according to claim 1, characterized in that, The process of determining the corresponding basic data permission scope based on user role information includes: Based on the multi-level organizational structure of logistics enterprises, a role-level-basic data permission scope mapping system is constructed to form a basic permission mapping rule base; user role information is obtained, and the corresponding basic data permission scope is obtained based on the basic permission mapping rule base.

3. The logistics data analysis method according to claim 1, characterized in that, The process of obtaining user preference features based on historical user query behavior, and then dynamically calculating the optimal permission range within the basic data permission range based on user preference features, business roles, and query intent using a dynamic permission recommendation model, includes: Based on the user's business role, and combined with the business scope and data access responsibilities of each role in the logistics industry knowledge graph, a candidate data pool matching the business role is selected within the basic data permissions of the user role. The logistics industry knowledge graph is a semantic knowledge network built for logistics data analysis permission management scenarios. It is a standardized knowledge system that integrates logistics industry business rules, organizational structure, and data responsibilities. It establishes relationships between entities based on multi-level organizational roles, business scopes, data types, and data access responsibilities within the logistics industry. Obtain users' historical query behavior and generate user preference feature vectors based on the users' historical query behavior, including indicator preferences, range preferences, granularity preferences, and time preferences; Obtain the user's current query content, identify the user's query intent based on the current query content, and convert the current query intent into a standardized intent feature vector consistent with the dimensions of the user's preference feature vector; The user preference feature vector is fused with the standardized intent feature vector to generate a fused feature vector; Based on the candidate data pool, the optimal permission range is determined by further filtering through collaborative filtering algorithm according to the fused feature vector.

4. The logistics data analysis method according to claim 1, characterized in that, The process of obtaining the user-input query date range and optimizing it using a time window optimization model based on the current query intent to obtain an optimized query time range includes: Obtain the user query date range, and verify the obtained user query date range based on the preset earliest query time and latest query time to obtain the verified query date range; Based on the user's query intent, the corresponding business periodicity pattern is matched, and the boundary of the verified query date range is optimized based on the matched business periodicity pattern to obtain the optimized query date range; The daily data quality score within the optimized query date range is calculated based on preset requirements. If the daily data quality score is less than the preset value, the current daily data is considered not to meet the preset requirements. If the completeness rate of the daily data that meets the preset requirements is greater than or equal to the preset value, the daily data that does not meet the preset requirements is repaired, and the repaired diary data is retained after the repair is successful. If repair is not possible, the optimal time range is selected based on the data quality score higher than the preset value and the longest continuous duration. If the duration of the selected continuous time range is less than the preset minimum analysis cycle, the dates are extended forward / backward within the compliance verification boundary to supplement high-quality data that meets the preset requirements until the minimum analysis cycle is met. If the extension still fails to meet the requirements, fragmented data reference results and alternative query time ranges that meet the preset requirements are pushed to the user, and finally, the optimized query time range that meets the analysis requirements is determined.

5. The logistics data analysis method according to claim 1, characterized in that, The step of validating the obtained user query date range based on the preset earliest and latest query times includes: If the start time of the user's query date range is earlier than the preset earliest query time, the start time in the current user's query date range will be changed to the preset earliest query time; if the end time of the user's query date range is later than the preset latest query time, the end time in the current user's query date range will be changed to the preset latest query time.

6. The logistics data analysis method according to claim 1, characterized in that, The method further includes: Construct a logistics industry indicator system that meets preset requirements, including: timeliness achievement rate, misclassification, loss and damage indicators; The association rules between various indicators in a logistics industry indicator system that meets preset requirements are analyzed by using a multi-indicator association deep learning model; wherein, the multi-indicator association deep learning model uses an LSTM network to process time series data of multiple indicators and learn the temporal association rules between indicators. Based on the intent index, a heatmap of the correlation between the indices associated with the intent index is generated according to the correlation rules between the indices obtained from the analysis. Based on the correlation heatmap, trend prediction of correlation indicators is performed to identify potential risks.

7. The logistics data analysis method according to claim 1, characterized in that, The method further includes: Anomaly detection and attribution analysis are performed on the matched logistics data using anomaly detection and root cause analysis models. The anomaly detection and root cause analysis model includes: identifying abnormal data points using the isolated forest algorithm based on matched logistics data; and then mining root causes using the decision tree root cause analysis algorithm combined with association rules and / or knowledge graphs.

8. A logistics data analysis device, characterized in that, include: The basic data permission scope acquisition module is used to determine the corresponding basic data permission scope based on user role information; The optimal permission range determination module is used to obtain user preference features based on users' historical query behavior, and dynamically calculate the optimal permission range within the basic data permission range based on user preference features, business roles, and query intent through a dynamic permission recommendation model. The query time range optimization module is used to obtain the query date range input by the user, and optimize the query date range input by the user using a time window optimization model based on the current query intent to obtain the optimized query time range; The logistics data acquisition module is used to match the corresponding logistics data based on the optimal permission range and the optimized query time range. The intent index calculation module is used to calculate intent indexes based on logistics data and the current query intent.

9. An electronic device comprising a memory and at least one processor, wherein the memory stores instructions; The at least one processor invokes the instructions in the memory to cause the electronic device to perform the various steps of the logistics data analysis method as described in any one of claims 1-7.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions are executed by the processor, they implement the various steps of the logistics data analysis method as described in any one of claims 1-7.

Images