A logistics service data analysis method and device

Abstract

The application discloses a kind of logistics business data analysis method and device, it is related to computer technology field.The specific embodiment of the method includes: receiving the natural language description information corresponding to the logistics business data input by user, identifies the target index information and target dimension information corresponding to the logistics business data from natural language description information;Based on target index information and target dimension information, the attribution analysis path corresponding to the logistics business data is constructed, and the numerical change data corresponding to target index information is inquired according to attribution analysis path;Based on attribution analysis path and numerical change data, the index change influence factor corresponding to the logistics business data is determined.This embodiment automatically locates the index information and dimension information corresponding to the logistics business data according to natural language description information, and automatically determines index change influence factor through the attribution analysis path constructed, without manual analysis or configuration analysis template, improves analysis efficiency and accuracy, reduces development difficulty.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a method and apparatus for analyzing logistics business data. Background Technology

[0002] In logistics business analysis scenarios, it is necessary to analyze logistics business data to determine the reasons for changes in indicators, such as revenue and costs, and then develop targeted measures to improve business conditions, such as increasing revenue or reducing costs. Currently, logistics business data analysis is mainly carried out by manually writing queries or configuring relevant analysis templates. These methods involve a large workload of manual analysis, resulting in slow analysis speed and poor accuracy. Furthermore, they require the configuration of numerous analysis templates, making development difficult. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide a method and apparatus for analyzing logistics business data, which can automatically locate the corresponding indicator information and dimension information of logistics business data based on the natural language description information input by the user, and automatically determine the influencing factors of indicator changes through the constructed attribution analysis path, without the need for manual analysis or configuration of analysis templates, thereby improving analysis efficiency and accuracy and reducing development difficulty.

[0004] To achieve the above objectives, according to one aspect of the present invention, a method for analyzing logistics business data is provided, comprising:

[0005] Receive natural language description information corresponding to logistics business data input by users, and identify target indicator information and target dimension information corresponding to logistics business data from the natural language description information;

[0006] Based on the target indicator information and target dimension information, construct the attribution analysis path corresponding to the logistics business data, and query the numerical change data corresponding to the target indicator information according to the attribution analysis path;

[0007] Based on attribution analysis paths and numerical change data, the factors influencing the changes in indicators corresponding to logistics business data are identified.

[0008] Optionally, target indicator information and target dimension information corresponding to logistics business data are identified from natural language description information, including:

[0009] Entity recognition processing is performed on natural language description information to obtain initial indicator and dimensional information corresponding to logistics business data;

[0010] The similarity between the initial indicator information and dimension information and the pre-configured indicator information and dimension information is calculated respectively. Based on the similarity calculation results, the target indicator information and target dimension information corresponding to the logistics business data are determined.

[0011] Optionally, the similarity between the initial indicator information and dimension information and the pre-configured indicator information and dimension information is calculated, including:

[0012] The initial indicator information and dimension information are encoded into indicator vectors and dimension vectors, respectively;

[0013] Semantic similarity is calculated based on the indicator vector and dimension vector and the pre-configured indicator information and dimension information to obtain the similarity calculation result.

[0014] Optionally, an attribution analysis path corresponding to the logistics business data is constructed based on the target indicator information and target dimension information, including:

[0015] In response to a hit in a pre-configured attribution formula, an attribution analysis path corresponding to the logistics business data is constructed based on the pre-configured attribution formula.

[0016] In response to a missed attribution formula, the target dimension information is removed from the pre-configured full dimension set to obtain a dimension set to be drilled down. The dimension values ​​in the dimension set to be drilled down are traversed, and the target attribution formula is generated based on the traversal results. The attribution analysis path corresponding to the logistics business data is constructed based on the target attribution formula.

[0017] Optionally, based on the attribution analysis path and numerical change data, the influencing factors of the changes in the corresponding indicators of logistics business data are determined, including:

[0018] Based on the numerical change data, determine the contribution of each path node in the attribution analysis path to the change of the target indicator;

[0019] Construct a list of influencing factors based on the contribution of changes to each path node;

[0020] Identify the factors influencing changes in logistics business data based on the list of influencing factors.

[0021] Optionally, after determining the influencing factors of changes in indicators corresponding to logistics business data based on attribution analysis paths and numerical change data, the process also includes:

[0022] Generate a visualized dimensional tree structure diagram based on the factors influencing indicator changes; or, input the factors influencing indicator changes and natural language description information into a pre-trained language model to obtain a data analysis text report.

[0023] Present users with a dimensional tree structure diagram or a data analysis text report.

[0024] According to a second aspect of the present invention, an apparatus for analyzing logistics business data is provided, comprising:

[0025] The receiving module is used to receive natural language description information corresponding to the logistics business data input by the user, and to identify the target indicator information and target dimension information corresponding to the logistics business data from the natural language description information.

[0026] The module is used to construct the attribution analysis path corresponding to the logistics business data based on the target indicator information and target dimension information, and to query the numerical change data corresponding to the target indicator information according to the attribution analysis path;

[0027] The determination module is used to identify the influencing factors of changes in indicators corresponding to logistics business data based on attribution analysis paths and numerical change data.

[0028] According to a third aspect of the present invention, an electronic device is provided, comprising:

[0029] One or more processors;

[0030] Memory, used to store one or more programs.

[0031] When one or more programs are executed by one or more processors, the one or more processors implement the methods of any of the above embodiments.

[0032] According to a fourth aspect of the present invention, a computer-readable medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method of any of the above embodiments.

[0033] According to a fifth aspect of the present invention, a computer program product is provided, including a computer program that, when executed by a processor, implements the method of any of the above embodiments.

[0034] One embodiment of the above invention has the following advantages or beneficial effects: by receiving natural language description information corresponding to logistics business data input by the user, the target indicator information and target dimension information corresponding to the logistics business data are identified from the natural language description information; based on the target indicator information and target dimension information, an attribution analysis path corresponding to the logistics business data is constructed, and the numerical change data corresponding to the target indicator information is queried according to the attribution analysis path; based on the attribution analysis path and the numerical change data, the influencing factors of the indicator changes corresponding to the logistics business data are determined; thereby, the indicator information and dimension information corresponding to the logistics business data can be automatically located according to the natural language description information input by the user, and the influencing factors of indicator changes can be automatically determined through the constructed attribution analysis path, without the need for manual analysis or configuration of analysis templates, thus improving analysis efficiency and accuracy and reducing development difficulty.

[0035] The further effects of the aforementioned unconventional alternative methods will be explained below in conjunction with specific implementation methods. Attached Figure Description

[0036] The accompanying drawings are provided to better understand the invention and are not intended to unduly limit the scope of the invention. Wherein:

[0037] Figure 1 This is a schematic diagram of the main process of the logistics business data analysis method according to an embodiment of the present invention;

[0038] Figure 2 This is a schematic diagram of the main flow of a logistics business data analysis method according to a preferred embodiment of the present invention;

[0039] Figure 3 This is a schematic diagram of the main modules of a logistics business data analysis device according to an embodiment of the present invention;

[0040] Figure 4 This is an exemplary system architecture diagram in which embodiments of the present invention can be applied;

[0041] Figure 5 This is a schematic diagram of the structure of a computer system suitable for implementing terminal devices or servers of the present invention. Detailed Implementation

[0042] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0043] It should be noted that the acquisition, storage, and application of personal information involved in the embodiments of the present invention comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0044] In logistics business analysis scenarios, it is necessary to analyze logistics business data to determine the reasons for changes in indicators, such as revenue and costs, and then develop targeted measures to improve business conditions, such as increasing revenue or reducing costs. Currently, logistics business data analysis is mainly carried out by manually writing queries or configuring relevant analysis templates. These methods involve a large workload of manual analysis, resulting in slow analysis speed and poor accuracy. Furthermore, they require the configuration of numerous analysis templates, making development difficult.

[0045] In view of this, according to one aspect of the present invention, a method for analyzing logistics business data is provided.

[0046] Figure 1This is a schematic diagram of the main flow of a logistics business data analysis method according to an embodiment of the present invention. Figure 1 As shown, the method for analyzing logistics business data according to an embodiment of the present invention includes the following steps S101 to S103.

[0047] Step S101: Receive natural language description information corresponding to the logistics business data input by the user, and identify the target indicator information and target dimension information corresponding to the logistics business data from the natural language description information.

[0048] Logistics business data refers to structured data used to measure the operational status and efficiency of logistics operations, including but not limited to order fulfillment data, capacity scheduling data, transportation timeliness data, warehouse turnover data, and exception handling records. Natural language description information refers to the textual expression of analytical needs or business questions input by users in natural language form, such as "Why did the trunk transportation delay rate in East China increase this month?" or "Investigate the reasons for the changes in picking efficiency of a certain warehouse over the past week." This type of information can be submitted through input boxes or dialog boxes in the user interface. Specifically, based on the graphical user interface (GUI) text input method, users manually type text information in the input area provided by the system, and the system receives this text information as natural language description information; alternatively, based on speech recognition technology, the user's spoken input is converted into parsable text information and used as natural language description information.

[0049] After receiving natural language description information, the system identifies target metrics and target dimensions relevant to logistics business data. Target metrics refer to key business indicators that users are interested in and whose numerical changes need to be analyzed, such as business revenue, costs, order volume, delivery on-time rate, and order cancellation rate. Target dimensions refer to categorized fields that define the calculation scope of the target metrics, such as region, time period, logistics link, and warehouse number. When identifying this information, keyword matching and part-of-speech tagging techniques in the natural language processing model can be used to extract predefined indicator names and dimension fields from the natural language description information. Alternatively, a trained intent recognition and entity extraction model can be used to perform semantic parsing of the natural language description information to determine the target metrics and target dimensions. Through these methods, the user's natural language requirements can be structured into explicit analytical tasks.

[0050] Step S102: Construct the attribution analysis path corresponding to the logistics business data based on the target indicator information and target dimension information, and query the numerical change data corresponding to the target indicator information according to the attribution analysis path.

[0051] Attribution analysis paths are logical deduction paths used to analyze the reasons for changes in target indicator values. These paths include multiple path nodes, each corresponding to a set of detailed dimensions or feature conditions of logistics business data. The path nodes form hierarchical or parallel relationships, used to break down the numerical changes of the target indicator under different dimensional combinations layer by layer. For example, for the target indicator of declining delivery on-time rate, the attribution analysis path can include multiple path nodes such as different regions under the regional dimension, different days under the time dimension, and different transportation types under the transportation mode dimension, used to analyze in turn which specific dimensional combinations caused significant fluctuations in the indicator. Constructing attribution analysis paths corresponding to logistics business data can be based on predefined indicator dimension mapping relationships. That is, based on the identified target indicator information and target dimension information, preset rules are invoked to determine the dimensional combinations that can be used for attribution and their priority order, thereby constructing the attribution analysis path structure. Alternatively, a data-driven path generation mechanism can be used to automatically recommend path nodes and hierarchical structures by analyzing the changing characteristics of various indicators in historical logistics business data and their correlation with different dimensions, achieving dynamic adaptation of attribution analysis path construction to business scenarios.

[0052] Numerical change data refers to the changes between historical and current statistical values ​​of a target indicator across different time periods or combinations of dimensions. This can be expressed as a numerical increase or decrease, and can be further converted into absolute changes or rates of change. Specifically, numerical change data can be queried using a multidimensional data analysis model. By combining the dimensions of each path node in a pre-constructed attribution analysis path, the statistical values ​​of the target indicator at each path node can be retrieved from a database or multidimensional dataset, and their changes can be calculated. Alternatively, streaming data processing and time-series comparison mechanisms can be used to acquire the indicator data stream at each path node in real time and compare it with historical benchmark values ​​within a set time window to obtain the numerical change data of the target indicator.

[0053] Step S103: Based on the attribution analysis path and numerical change data, determine the influencing factors of the indicator changes corresponding to the logistics business data.

[0054] Factors influencing indicator changes refer to key dimensional elements or business characteristics that cause numerical changes in indicators. These factors accurately reflect the magnitude of the impact of a particular business process, business condition, or external environment on indicator fluctuations. For example, in analyzing an increase in delivery delay rates, possible factors influencing indicator changes include abnormal traffic conditions in a specific area, a decrease in the processing capacity of a certain type of transportation, or a sudden surge in order volume during a specific time period. Specifically, determining the factors influencing indicator changes corresponding to logistics business data can be based on ranking the contribution of changes in path nodes. This involves normalizing the magnitude of numerical changes corresponding to each path node in the attribution analysis path to assess the degree of contribution of that path node to the overall target indicator change, and then using the dimensions corresponding to path nodes with higher contribution as the influencing factors of indicator changes. Alternatively, multi-factor regression modeling or factor decomposition algorithms can be used. By using the numerical changes under path nodes as input variables, a mathematical model can be constructed between explanatory variables and the target indicator, quantifying the influence coefficient of each dimension on indicator changes, and using the dimension features with high significance as the influencing factors of indicator changes.

[0055] This embodiment receives natural language descriptions of logistics business data input by the user, identifies target indicator information and target dimension information corresponding to the logistics business data from the natural language descriptions, constructs an attribution analysis path for the logistics business data based on the target indicator information and target dimension information, and queries the numerical change data corresponding to the target indicator information according to the attribution analysis path; and determines the influencing factors of the indicator changes corresponding to the logistics business data based on the attribution analysis path and the numerical change data. Thus, it can automatically locate the indicator information and dimension information corresponding to the logistics business data based on the natural language descriptions input by the user, and automatically determine the influencing factors of indicator changes through the constructed attribution analysis path, without the need for manual analysis or configuration of analysis templates, improving analysis efficiency and accuracy, and reducing development difficulty.

[0056] Optionally, the target indicator information and target dimension information corresponding to the logistics business data are identified from the natural language description information, including: performing entity recognition processing on the natural language description information to obtain the initial indicator information and dimension information corresponding to the logistics business data; calculating the similarity between the initial indicator information and dimension information and the pre-configured indicator information and dimension information respectively; and determining the target indicator information and target dimension information corresponding to the logistics business data based on the similarity calculation results.

[0057] In this embodiment, entity recognition processing is performed on the natural language description information input by the user to identify the initial indicator information and initial dimension information involved. Specifically, based on a pre-trained natural language processing model, such as a named entity recognition model or a domain-specific word matching algorithm, keywords in the natural language description information are matched and extracted with candidate indicators and dimensions. The extracted initial indicator information and initial dimension information are then compared with a pre-configured set of indicator information and dimension information to calculate similarity, for example, using word vector cosine similarity, edit distance, or semantic embedding comparison, to evaluate the semantic or expressive closeness between the two. Based on the similarity score, the most matching indicator information and dimension information are selected as the target indicator information and target dimension information corresponding to the logistics business data. In addition to the above methods, recognition can also be achieved through rule matching, such as pattern matching recognition of natural language description information based on a predefined keyword rule table. This embodiment can improve the system's ability to understand user natural language input, making the identified indicators and dimensions more consistent with the user's true intentions, thereby providing an accurate and reliable structured input foundation for subsequent attribution analysis and improving the intelligence and automation of the data analysis process.

[0058] Optionally, the similarity between the initial indicator information and dimension information and the pre-configured indicator information and dimension information is calculated, including: encoding the initial indicator information and dimension information into indicator vectors and dimension vectors respectively; and performing semantic similarity calculation based on the indicator vectors and dimension vectors and the pre-configured indicator information and dimension information to obtain the similarity calculation result.

[0059] In this implementation, initial indicator and dimension information are encoded into corresponding indicator and dimension vectors using a natural language processing model. For example, a word embedding model or a custom semantic encoding model can be employed to transform text information into low- or high-dimensional semantic vector representations. The encoded indicator and dimension vectors are then compared one-to-one with pre-configured standard indicator and dimension information vectors. Semantic similarity calculation methods, such as cosine similarity, Manhattan distance, or Euclidean distance, are used to calculate the similarity between the two. Based on the calculation results, the match with the highest similarity can be selected as the final target indicator and target dimension information, thus completing the accurate conversion from natural language to structured data. Besides using vector semantic similarity calculation, rule-based fuzzy matching methods or semantic analysis methods combined with syntax tree structures can also be used. These methods are suitable for scenarios where a semantic model cannot be trained, enhancing the system's coverage and fault tolerance. This embodiment can accurately identify the indicator and dimensional information corresponding to the user's true intent in the case of the diversity and ambiguity of natural language expressions. It provides structured and highly matched basic data for subsequent attribution analysis path construction and numerical change data query, thereby improving the system's ability to understand business data and the level of intelligent automatic analysis.

[0060] Optionally, constructing an attribution analysis path for logistics business data based on target indicator information and target dimension information includes: in response to hitting a pre-configured attribution formula, constructing an attribution analysis path for logistics business data based on the pre-configured attribution formula; in response to not hitting a pre-configured attribution formula, deleting the target dimension information from the pre-configured full dimension set to obtain a dimension set to be drilled down, traversing the dimension values ​​in the dimension set to be drilled down, generating a target attribution formula based on the traversal results, and constructing an attribution analysis path for logistics business data based on the target attribution formula.

[0061] The system determines whether the target metric and target dimension information match a pre-configured attribution formula. These pre-configured attribution formulas are logical rules pre-defined based on historical business analysis experience and domain knowledge, covering common attribution analysis needs. If an existing attribution formula is matched, the corresponding attribution analysis path is directly generated using that formula, eliminating the need for additional path construction and ensuring efficiency and rule reusability. If no pre-configured attribution formula is matched, dynamic construction logic is initiated. Specifically, the current target dimension information is removed from the pre-configured full dimension set, resulting in a set of dimensions to be drilled down into. This set represents dimensions that may be further subdivided or supplemented for analysis. Each dimension item in this set is sequentially traversed, and a new attribution formula is constructed based on the correlation between each dimension item, its value, and the target metric. These target attribution formulas can be generated in the form of rule trees, dimension path chains, or vector paths, and serve as path nodes to form the final attribution analysis path, representing a hierarchical attribution relationship chain from high-level to detailed dimensions. This embodiment implements an attribution path generation mechanism that combines static rules with dynamic construction. It can improve processing efficiency by relying on existing knowledge, and can also quickly build adaptive attribution analysis paths when facing novel or marginal business needs, thereby improving the system's versatility and the accuracy of attribution analysis.

[0062] Optionally, the factors influencing the changes in indicators corresponding to logistics business data are determined based on the attribution analysis path and numerical change data, including: determining the contribution of each path node in the attribution analysis path to the change of the target indicator based on the numerical change data; constructing a list of influencing factors based on the contribution of each path node; and determining the factors influencing the changes in indicators corresponding to logistics business data based on the list of influencing factors.

[0063] After constructing the attribution analysis path corresponding to the logistics business data and obtaining the corresponding numerical change data, a deeper analysis of the changes in the target indicators is conducted to identify the core influencing factors leading to these changes. Specifically, firstly, based on the numerical change data corresponding to each path node in the attribution analysis path, the specific contribution of each path node to the change in the target indicator is calculated. Path nodes represent specific dimension values ​​under a certain business dimension, such as "destination = Shanghai" or "transportation mode = air freight," and their corresponding numerical change data reflects the fluctuation of the target indicators (such as logistics timeliness, transportation costs, etc.) under that business dimension. By comparing the changes in indicators under different time periods or different dimension conditions, the proportion of each path node in the overall change can be quantitatively assessed. After obtaining the change contribution of all path nodes, these path nodes are sorted from high to low according to their contribution, constructing a list of influencing factors. This list of influencing factors clearly identifies the factors that have the greatest impact on the change in the target indicator and their degree of influence. Based on the path nodes in the list of influencing factors and their corresponding contribution, the final influencing factors of the indicator change are determined. These factors can be the main transportation modes that lead to increased logistics costs, or the key destination distribution that leads to decreased logistics timeliness. In addition to the static evaluation method based on contribution mentioned above, a model-driven approach can also be used to identify the influencing factors of indicator changes. For example, regression analysis models or causal inference methods based on graph neural networks can be used to uncover potential causal relationships from the complex correlations between path nodes and indicators. This approach can systematically and quantitatively identify the hidden driving factors of indicator changes in logistics business data, achieving efficient tracing from dimensional paths to core influencing factors. This not only improves the accuracy of attribution analysis but also provides users with evidence-based and targeted optimization directions.

[0064] Optionally, after determining the influencing factors of the indicator changes corresponding to the logistics business data based on the attribution analysis path and numerical change data, the process further includes: generating a visualized dimensional tree structure diagram based on the influencing factors of indicator changes; or, inputting the influencing factors of indicator changes and natural language description information into a pre-trained language model to obtain a data analysis text report; and displaying the dimensional tree structure diagram or data analysis text report to the user.

[0065] After identifying the influencing factors of changes in logistics business data indicators, the analysis results are output in a structured and visualized manner to facilitate user understanding and decision-making. Specifically, based on the identified influencing factors, a visualized dimensional tree structure is constructed. This dimensional tree structure uses the target indicator as the root node and expands sequentially according to the dimensional hierarchy, with each path node with significant contribution as a branch of the tree structure, clearly showing the influence path and degree of contribution of each dimension value on the change of the target indicator. This graphical representation allows users to quickly grasp the structural logic of overall business changes and key issue areas. Alternatively, the influencing factors of indicator changes and the user's original natural language description information can be used as input and submitted to a pre-trained language model, such as a large language model based on the Transformer structure. Through joint understanding of influencing factors and user intent, a complete data analysis text report is automatically generated. This data analysis text report can be automatically written from multiple dimensions such as business background, indicator change trends, key dimension analysis, possible causes, and optimization suggestions, enabling non-professional users to quickly obtain highly readable and interpretable data analysis conclusions. The generated dimensional tree structure diagram or data analysis text report can be presented to users in a combined text and graphics format to enhance their analysis experience and understanding efficiency. In addition to the above methods, various output formats such as email and embedding in business dashboards can also be used to meet the business application needs of multiple scenarios. This implementation method can provide users with intuitive, clear, and business-valuable feedback on the results of complex data attribution analysis processes, lowering the data interpretation threshold and improving the visibility and action guidance of analysis conclusions.

[0066] Figure 2 This is a schematic diagram of the main flow of a logistics business data analysis method according to a preferred embodiment of the present invention. Figure 2 As shown, this embodiment provides a method for attribution analysis of logistics business data based on natural language input, where the logistics business data can be logistics operation data. In this embodiment, the natural language input part of the logistics operation data analysis problem provides a text box for user input. Users can directly enter the natural language text of the analysis problem in this text box, and the system uses it as subsequent input data. For example, a user might enter, "We need to analyze the revenue of the next-day delivery product line in Beijing and Hebei Province in order to assess the operational status of different product lines in various regions." After receiving such text, it is used as the starting point for analysis.

[0067] The logistics business data attribution metrics and dimension identification section parses the natural language text. The system pre-maintains all data metric names and dimensions, along with their possible values. For example, metrics include revenue, cost, and order volume; dimensions include province / region and product line; and province / region values ​​include Beijing and Hebei. First, the system calls a pre-trained language model, combined with prompts, to identify the metrics, dimensions, and dimension values ​​in the natural language text. For the identified text entities, since they may not perfectly match the standard fields in the metadata, the system vectorizes the natural language text into a 1536-dimensional vector and calculates similarity with all known metric and dimension value vectors to find the best-matching metric name and dimension value, thus establishing entity connections.

[0068] After obtaining the indicators and dimensions, the logistics operation data automatically drills down through the dimensions, and the data query section generates corresponding attribution analysis paths. Specifically, one approach is to use pre-configured drill-down formulas containing multiplication and division, such as Revenue = Number of Orders × Average Revenue per Order, which the system can directly use. Another approach is to automatically generate drill-down formulas, suitable for accumulative indicators. When certain dimensions are not mentioned in the natural language text, these dimensions can be defined as drill-down dimensions, and all their dimension values ​​are iterated to construct a complete accumulation formula. For example, when analyzing "express delivery revenue" without mentioning the "province / region" dimension, formulas like "Express delivery revenue = Beijing region express delivery revenue + Hebei province express delivery revenue + Shanghai region express delivery revenue" are automatically generated. Subsequently, using the dimension values ​​and indicator names in the generated formulas, the current and base period values ​​are retrieved from the database, generally at a daily granularity. The current value is yesterday's data, and the base period value is yesterday's or historical data.

[0069] After completing the data query, the logistics operation factor analysis section breaks down each formula into a left-hand side item (e.g., revenue) and multiple right-hand side items (e.g., revenue from different provinces / regions), recording the current value and the base period value for each. Combining the formula structure and the numerical values ​​of each item, the following impact scoring formula is used to calculate the contribution of each right-hand side item to the changes in the left-hand side indicators:

[0070] Influence_score(x i ) = f(x j '',x i '|i ≠ j) - f(x) j '',x i ''|i is not equal to j)

[0071] In this formula, the term on the left side is denoted as y, the current value is y', and the base period value is y''; the terms on the right side of the formula are (x1,…,x m The current value is denoted as (x1',...,x). mThe base period value is denoted as (x1'',…,xm''), where m is a positive integer; the formula is y=fi(x1,…,xm''). m ); i and j are positive integers. f is a function represented on the right side of the equals sign in the drill-down formula. For example, there are two formulas: Revenue = Number of Units × Average Revenue per Unit; Revenue = Beijing Revenue + Hebei Revenue. f1(Number of Units, Average Revenue per Unit) = Number of Units × Average Revenue per Unit; f2(Beijing Revenue, Hebei Revenue) = Beijing Revenue + Hebei Revenue.

[0072] When the overall indicator shows an upward trend, the system sorts the factors from largest to smallest contribution; when it shows a downward trend, it sorts them from smallest to largest contribution, ultimately selecting the top few contributing factors and generating a standard analysis text. For example, if y'-y''>=0, then it sorts according to Influence_score(x i Sort all items from largest to smallest and extract the top q factor items. If y'-y''<0, then sort according to Influence_score(x i Sort all items from smallest to largest and extract the top q factor items. q is an integer that can be set manually. If the total number of factor items is less than q, then q is equal to the total number of factor items. The top q factor items that have the greatest impact on y are obtained to generate text R.

[0073] After completing the attribution analysis, the logistics operations analysis attribution results display section will visualize or present the results in natural language. First, the system can generate a dimensional tree structure diagram, corresponding to the aforementioned drill-down tree, displaying key dimensional items in descending order of factor contribution, helping users quickly locate key business impact points. Second, the system can also generate a natural language analysis report. This report is generated by a language model, with input consisting of the aforementioned standard analysis text, drill-down formulas, and the user's original natural language input. The final output is a structurally complete, semantically fluent, and stylistically consistent data analysis explanation text.

[0074] Traditional manual data analysis takes minutes to hours, while this embodiment can output analysis results in seconds. Compared with template-based analysis methods that require users to manually select indicators and produce fixed and inflexible output results, this embodiment is based on natural language input, which is flexible and intelligent. It can automatically complete the entire process from problem understanding to attribution analysis to graphic and text output, which greatly improves the efficiency of logistics business data analysis and user experience.

[0075] According to a second aspect of the present invention, an apparatus for analyzing logistics business data is provided.

[0076] Figure 3 This is a schematic diagram of the main modules of a logistics business data analysis device according to an embodiment of the present invention. Figure 3As shown, the logistics business data analysis device 300 includes:

[0077] The receiving module 301 is used to receive natural language description information corresponding to logistics business data input by the user, and to identify target indicator information and target dimension information corresponding to the logistics business data from the natural language description information.

[0078] The construction module 302 is used to construct the attribution analysis path corresponding to the logistics business data based on the target indicator information and target dimension information, and to query the numerical change data corresponding to the target indicator information according to the attribution analysis path;

[0079] The determination module 303 is used to determine the influencing factors of the changes in indicators corresponding to logistics business data based on the attribution analysis path and numerical change data.

[0080] Optionally, the receiving module 301 is also used for:

[0081] Entity recognition processing is performed on natural language description information to obtain initial indicator and dimensional information corresponding to logistics business data;

[0082] The similarity between the initial indicator information and dimension information and the pre-configured indicator information and dimension information is calculated respectively. Based on the similarity calculation results, the target indicator information and target dimension information corresponding to the logistics business data are determined.

[0083] Optionally, the receiving module 301 is also used for:

[0084] The initial indicator information and dimension information are encoded into indicator vectors and dimension vectors, respectively;

[0085] Semantic similarity is calculated based on the indicator vector and dimension vector and the pre-configured indicator information and dimension information to obtain the similarity calculation result.

[0086] Optionally, the building module 302 is also used for:

[0087] In response to a hit in a pre-configured attribution formula, an attribution analysis path corresponding to the logistics business data is constructed based on the pre-configured attribution formula.

[0088] In response to a missed attribution formula, the target dimension information is removed from the pre-configured full dimension set to obtain a dimension set to be drilled down. The dimension values ​​in the dimension set to be drilled down are traversed, and the target attribution formula is generated based on the traversal results. The attribution analysis path corresponding to the logistics business data is constructed based on the target attribution formula.

[0089] Optionally, the determining module 303 is also used for:

[0090] Based on the numerical change data, determine the contribution of each path node in the attribution analysis path to the change of the target indicator;

[0091] Construct a list of influencing factors based on the contribution of changes to each path node;

[0092] Identify the factors influencing changes in logistics business data based on the list of influencing factors.

[0093] Optionally, the analysis device 300 also includes a display module, which is used for:

[0094] Generate a visualized dimensional tree structure diagram based on the factors influencing indicator changes; or, input the factors influencing indicator changes and natural language description information into a pre-trained language model to obtain a data analysis text report.

[0095] Present users with a dimensional tree structure diagram or a data analysis text report.

[0096] It should be noted that the specific implementation details of the logistics business data analysis device of the present invention have been described in detail in the above logistics business data analysis method, so the content will not be repeated here.

[0097] According to a third aspect of the present invention, an electronic device is provided, comprising: one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method provided in the first aspect of the present invention.

[0098] According to a fourth aspect of the present invention, a computer-readable medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method provided in the first aspect of the present invention.

[0099] According to a fifth aspect of the present invention, a computer program product is provided, including a computer program that, when executed by a processor, implements the method provided in the first aspect of the present invention.

[0100] Figure 4 An exemplary system architecture 400 is shown, which can be applied to the analysis method or apparatus for logistics business data according to embodiments of the present invention.

[0101] like Figure 4 As shown, system architecture 400 may include terminal devices 401, 402, and 403, a network 404, and a server 405. Network 404 serves as the medium for providing communication links between terminal devices 401, 402, and 403 and server 405. Network 404 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0102] Users can use terminal devices 401, 402, and 403 to interact with server 405 via network 404 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 401, 402, and 403, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social media platform software, etc. (for example only).

[0103] Terminal devices 401, 402, and 403 can be various electronic devices with displays that support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0104] Server 405 can be a server providing various services, such as a backend management server supporting shopping websites browsed by users using terminal devices 401, 402, and 403 (for example only). The backend management server can analyze and process received data such as logistics business data analysis requests, and feed back the factors affecting changes in indicators (such as processing results - for example only) to the terminal devices.

[0105] It should be noted that the logistics business data analysis method provided in this embodiment of the invention is generally run by server 405, and correspondingly, the logistics business data analysis device is generally set in server 405.

[0106] It should be understood that Figure 4 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0107] The following is for reference. Figure 5 It shows a schematic diagram of the structure of a computer system 500 suitable for implementing a terminal device of the present invention. Figure 5 The terminal device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0108] like Figure 5 As shown, the computer system 500 includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 502 or programs loaded from storage section 508 into random access memory (RAM) 503. The RAM 503 also stores various programs and data required for the operation of the system 500. The CPU 501, ROM 502, and RAM 503 are interconnected via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0109] The following components are connected to I / O interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. Drive 510 is also connected to I / O interface 505 as needed. Removable media 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 510 as needed so that computer programs read from them can be installed into storage section 508 as needed.

[0110] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 509, and / or installed from removable medium 511. When the computer program is run by the central processing unit (CPU) 501, it performs the functions defined above in the system of this invention.

[0111] It should be noted that the computer-readable medium shown in this invention can be a computer-readable signal medium, a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit programs for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0112] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more operable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually operate substantially in parallel, and they may sometimes operate in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0113] The modules described in the embodiments of the present invention can be implemented in software or hardware. The described modules can also be housed in a processor; for example, a processor can be described as including a receiving module, a building module, and a determining module. The names of these modules do not necessarily limit the module itself; for example, the determining module can also be described as "a module for determining the influencing factors of indicator changes corresponding to logistics business data based on attribution analysis paths and numerical change data."

[0114] In another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or it may exist independently and not assembled into the device. The computer-readable medium carries one or more programs, which, when executed by the device, cause the device to include: receiving natural language description information corresponding to logistics business data input by a user; identifying target indicator information and target dimension information corresponding to the logistics business data from the natural language description information; constructing an attribution analysis path corresponding to the logistics business data based on the target indicator information and target dimension information; querying numerical change data corresponding to the target indicator information according to the attribution analysis path; and determining the influencing factors of indicator changes corresponding to the logistics business data based on the attribution analysis path and the numerical change data.

[0115] The computer program product provided in this embodiment of the invention includes a computer program that, when executed by a processor, implements the logistics business data analysis method in this embodiment of the invention.

[0116] According to the technical solution of the present invention, the following advantages or beneficial effects are achieved: by receiving natural language description information corresponding to logistics business data input by the user, the target indicator information and target dimension information corresponding to the logistics business data are identified from the natural language description information; based on the target indicator information and target dimension information, an attribution analysis path corresponding to the logistics business data is constructed, and the numerical change data corresponding to the target indicator information is queried according to the attribution analysis path; based on the attribution analysis path and the numerical change data, the influencing factors of the indicator changes corresponding to the logistics business data are determined; thereby, the indicator information and dimension information corresponding to the logistics business data can be automatically located according to the natural language description information input by the user, and the influencing factors of indicator changes can be automatically determined through the constructed attribution analysis path, without the need for manual analysis or configuration of analysis templates, thus improving analysis efficiency and accuracy and reducing development difficulty.

[0117] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

[0118] It should be noted that the acquisition, storage, and application of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

Claims

1. A method for analyzing logistics business data, characterized in that, include: Receive natural language description information corresponding to logistics business data input by the user, and identify target indicator information and target dimension information corresponding to the logistics business data from the natural language description information; Based on the target indicator information and target dimension information, construct the attribution analysis path corresponding to the logistics business data, and query the numerical change data corresponding to the target indicator information according to the attribution analysis path; Based on the attribution analysis path and the numerical change data, the influencing factors of the indicator changes corresponding to the logistics business data are determined.

2. The method according to claim 1, characterized in that, Identifying target indicator information and target dimension information corresponding to the logistics business data from the natural language description information includes: The natural language description information is processed by entity recognition to obtain the initial indicator information and dimension information corresponding to the logistics business data; The similarity between the initial indicator information and dimension information and the pre-configured indicator information and dimension information is calculated respectively, and the target indicator information and target dimension information corresponding to the logistics business data are determined based on the similarity calculation results.

3. The method according to claim 2, characterized in that, Calculate the similarity between the initial indicator information and dimension information and the pre-configured indicator information and dimension information, including: The initial indicator information and dimension information are encoded into indicator vectors and dimension vectors, respectively; The semantic similarity is calculated based on the indicator vector and dimension vector and the pre-configured indicator information and dimension information to obtain the similarity calculation result.

4. The method according to claim 1, characterized in that, Based on the target indicator information and target dimension information, an attribution analysis path is constructed for the logistics business data, including: In response to a hit of a pre-configured attribution formula, an attribution analysis path corresponding to the logistics business data is constructed based on the pre-configured attribution formula. In response to a missed attribution formula, the target dimension information is removed from the pre-configured full dimension set to obtain a dimension set to be drilled down. The dimension values ​​in the dimension set to be drilled down are traversed, and a target attribution formula is generated based on the traversal results. The attribution analysis path corresponding to the logistics business data is constructed based on the target attribution formula.

5. The method according to claim 1, characterized in that, Based on the attribution analysis path and the numerical change data, the influencing factors of the indicator changes corresponding to the logistics business data are determined, including: Based on the numerical change data, determine the contribution of each path node in the attribution analysis path to the change of the target indicator; Construct a list of influencing factors based on the contribution of changes to each path node; The factors influencing the changes in the indicators corresponding to the logistics business data are determined based on the list of influencing factors.

6. The method according to claim 1, characterized in that, After determining the influencing factors of the indicator changes corresponding to the logistics business data based on the attribution analysis path and the numerical change data, the method further includes: A visualized dimensional tree structure diagram is generated based on the factors influencing the changes in the indicators; or, the factors influencing the changes in the indicators and the natural language description information are input into a pre-trained language model to obtain a data analysis text report. The dimensional tree structure diagram or the data analysis text report is displayed to the user.

7. A device for analyzing logistics business data, characterized in that, include: The receiving module is used to receive natural language description information corresponding to logistics business data input by the user, and to identify target indicator information and target dimension information corresponding to the logistics business data from the natural language description information. The construction module is used to construct the attribution analysis path corresponding to the logistics business data based on the target indicator information and target dimension information, and to query the numerical change data corresponding to the target indicator information according to the attribution analysis path; The determination module is used to determine the influencing factors of the indicator changes corresponding to the logistics business data based on the attribution analysis path and numerical change data.

8. An electronic device, characterized in that, include: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-6.

9. A computer-readable medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-6.

Images