Intelligent diagnosis method, device and equipment for waybills distributed wrongly and storage medium

Abstract

The invention relates to the technical field of logistics, in particular to an intelligent diagnosis method, device and equipment for waybills distributed wrongly and a storage medium. The intelligent diagnosis method for the waybills distributed wrongly comprises the following steps of obtaining waybill distribution data of a consumption waybill distribution center, performing calculation according to the waybill distribution data by adopting a plurality of prediction algorithms, obtaining a plurality of calculation process data and calculation results, and storing the calculation process data and the calculation results into a database; detecting the order distribution accuracy rate and the recognition rate of the website, and if the order distribution accuracy rate or the recognition rate of the website is lower than a set critical value, extracting corresponding calculation process data and calculation results in the database and summarizing the calculation process data and the calculation results; and according to a preset rule, carrying out statistics and analysis on the summarized result to obtain an analysis result, and outputting the analysis result. According to the method, big data can be analyzed and counted, the misclassification data of the website can be classified and counted according to the early warning indexes, misclassification reasons are automatically analyzed, and the website can effectively and quickly intervene.

Description

Technical Field

[0001] The present invention relates to the field of logistics technology, and in particular to an intelligent diagnosis method, device, equipment and storage medium for misassigned waybills. Background Art

[0002] In the process of parcel delivery and order splitting, for the sake of management, multi-segment codes are usually used to identify parcel delivery orders. Currently, logistics companies are able to quickly count the order splitting accuracy of each segment in the multi-segment code through the system. However, although there is order splitting accuracy data, there is a lack of effective analysis methods. Due to the complex logic of the order splitting engine, the cost of purely manual analysis is high and inefficient.

[0003] Therefore, there is an urgent need to systematically sort out the corresponding data, monitor and analyze the output results of the existing order splitting algorithm, and display the final results on the page content through business logic sorting, so that the business party can more quickly and accurately understand the source of the error, intervene in the order splitting in time, correct the order splitting, and quickly solve the problems arising from automatic sorting. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the object of the present invention is to provide an intelligent diagnosis method, device, equipment and storage medium for allocating erroneous waybills, aiming to solve the technical problem in the prior art that the calculation error data of the logistics system order splitting algorithm is difficult to effectively analyze.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] The first aspect of the present invention provides an intelligent diagnosis method for misassigned waybills, comprising the following steps: obtaining order data from a consumer order distribution center, using a variety of prediction algorithms, performing calculations based on the order distribution data to obtain a plurality of calculation process data and calculation results, and storing the calculation process data and calculation results in a database; obtaining outlet data, detecting the order distribution accuracy and recognition rate of the outlets, and if the order distribution accuracy or recognition rate of the outlets is lower than a set critical value, extracting the corresponding calculation process data and calculation results from the database and summarizing them; obtaining summary results, performing statistics and analysis on the summary results according to pre-set rules, obtaining analysis results, and outputting the analysis results.

[0007] Optionally, in a first implementation method of the first aspect of the present invention, the order data of the consumer order distribution center is obtained, multiple prediction algorithms are used, calculations are performed based on the order data to obtain multiple calculation process data and calculation results, and the calculation process data and calculation results are stored in a database, specifically including: obtaining the order data of the consumer order distribution center, extracting the order number and address data in the order data; using a text parsing algorithm to extract keywords in the address data; using multiple prediction algorithms to perform calculations based on the keywords to obtain multiple calculation process data and calculation results, and storing the order number, keywords, calculation process data and calculation results in a database.

[0008] Optionally, in a second implementation of the first aspect of the present invention, the text parsing algorithm is used to extract keywords from the address data, specifically including: preprocessing the address data, removing irrelevant characters and stop words, and unifying the formats of all characters to obtain standardized data; using a word segmentation tool to split the standardized data to obtain split results; using a machine learning model to identify the split results to obtain multiple keywords, classifying and extracting the identified keywords, and organizing them into a structured format.

[0009] Optionally, in a third implementation of the first aspect of the present invention, the use of multiple prediction algorithms, calculations are performed according to the keywords, multiple calculation process data and calculation results are obtained, and the keywords, calculation process data and calculation results are stored in a database, specifically including: constructing multiple prediction algorithms, the prediction algorithms including: a delivery range algorithm, a keyword manual specification algorithm, a map fence algorithm, and a historical library calculation algorithm; using multiple prediction algorithms, respectively performing calculations according to the keywords, to obtain multiple calculation process data and calculation results; classifying and sorting the multiple calculation process data and calculation results according to the algorithm type, and storing the classified data in the database.

[0010] Optionally, in a fourth implementation method of the first aspect of the present invention, the acquiring of the network point data, detecting the order splitting accuracy and recognition rate of the network point, if the order splitting accuracy or recognition rate of the network point is lower than a set critical value, extracting the corresponding calculation process data and calculation results in the database and summarizing them, specifically includes: setting a first critical value for the order splitting accuracy and a second critical value for the recognition rate, respectively, acquiring the network point data regularly, and comparing the order splitting accuracy and recognition rate of the network point data with the first critical value and the second critical value, respectively; if the order splitting accuracy is lower than the first critical value, extracting the network point number in the corresponding network point data; if the recognition rate is lower than the second critical value, extracting the network point number in the corresponding network point data; summarizing all the extracted network point numbers, removing the same network point numbers, extracting the corresponding calculation process data and calculation results in the database according to the network point numbers and summarizing them.

[0011] Optionally, in a fifth implementation of the first aspect of the present invention, the obtaining of summary results, performing statistics and analysis on the summary results according to pre-set rules, obtaining analysis results, and outputting the analysis results specifically include: formulating statistical analysis rules, the statistical analysis rules including: the display form of predicted data and actual data, and an explanation template for analysis results; obtaining summary results, performing statistics and analysis on the summary results according to pre-set statistical analysis rules, obtaining analysis results; visualizing and outputting the analysis results, the analysis results including: comparison results of predicted outlets and receipt outlets of various prediction algorithms, comparison results of predicted three-segment codes and receipt three-segment codes of various prediction algorithms, and error details of prediction algorithms.

[0012] Optionally, in a sixth implementation method of the first aspect of the present invention, the analysis results are visualized and output, and the analysis results include: comparison results of prediction points and receipt points of various prediction algorithms, comparison results of prediction three-segment codes and receipt three-segment codes of various prediction algorithms, and error details of the prediction algorithms, specifically including: comparing the prediction points of various prediction algorithms with the receipt points, and marking the prediction algorithms with accurate predictions and incorrect predictions respectively to obtain a first comparison result; comparing the prediction three-segment codes of various prediction algorithms with the receipt three-segment codes, marking the erroneous parts of the prediction three-segment codes, and obtaining a second comparison result; tracing the prediction algorithm with the prediction error according to the summary results, marking the data causing the prediction error, and forming an error detail result; visualizing the first comparison result, the second comparison result and the error detail result, and outputting the visualization result to generate a display page.

[0013] The second aspect of the present invention provides an intelligent diagnostic device for allocating incorrect waybills, including: a calculation module, used to obtain the order distribution data of the consumer order distribution center, adopt a variety of prediction algorithms, calculate according to the order distribution data to obtain a plurality of calculation process data and calculation results, and store the calculation process data and calculation results in a database; an extraction module, used to obtain the outlet data, detect the order distribution accuracy and recognition rate of the outlet, if the order distribution accuracy or recognition rate of the outlet is lower than the set critical value, extract the corresponding calculation process data and calculation results in the database and summarize them; an analysis module, used to obtain the summary results, statistics and analyze the summary results according to pre-set rules, obtain the analysis results, and output the analysis results.

[0014] Optionally, in a first implementation method of the second aspect of the present invention, the calculation module includes: an extraction sub-module, used to obtain the order data of the consumer order distribution center, and extract the order number and address data in the order distribution data; a parsing sub-module, used to use a text parsing algorithm to extract keywords in the address data; a prediction sub-module, used to use multiple prediction algorithms, perform calculations based on the keywords, obtain multiple calculation process data and calculation results, and store the order number, keywords, calculation process data and calculation results in a database.

[0015] Optionally, in a second implementation of the second aspect of the present invention, the parsing submodule includes: a preprocessing unit, used to preprocess the address data, remove irrelevant characters and stop words, and unify the formats of all characters to obtain standardized data; a splitting unit, used to split the standardized data using a word segmentation tool to obtain a splitting result; an identification unit, used to identify the splitting result using a machine learning model to obtain multiple keywords, classify and extract the identified keywords, and organize them into a structured format.

[0016] Optionally, in a third implementation method of the second aspect of the present invention, the prediction submodule includes: a construction unit, used to construct multiple prediction algorithms, the prediction algorithms including: a delivery range algorithm, a keyword manual specification algorithm, a map fence algorithm, and a historical library calculation algorithm; a calculation unit, used to adopt a plurality of prediction algorithms, perform calculations according to the keywords respectively, and obtain multiple calculation process data and calculation results; a classification unit, used to classify and organize the multiple calculation process data and calculation results according to the algorithm type, and store the classified data in a database.

[0017] Optionally, in a fourth implementation method of the second aspect of the present invention, the extraction module includes: a setting unit, used to respectively set a first critical value for the order splitting accuracy and a second critical value for the recognition rate, regularly obtain network point data, and compare the order splitting accuracy and the recognition rate of the network point data with the first critical value and the second critical value, respectively; a first extraction unit, used to extract the network point number in the corresponding network point data if the order splitting accuracy is lower than the first critical value; a second extraction unit, used to extract the network point number in the corresponding network point data if the recognition rate is lower than the second critical value; a summarizing unit, used to summarize all extracted network point numbers, remove the same network point numbers, extract the corresponding calculation process data and calculation results in the database according to the network point number, and summarize them.

[0018] Optionally, in the fifth implementation method of the second aspect of the present invention, the analysis module includes: a formulation sub-module for formulating statistical analysis rules, and the statistical analysis rules include: the display form of predicted data and actual data, and the description template of the analysis results; an analysis sub-module for obtaining summary results, and performing statistics and analysis on the summary results according to pre-set statistical analysis rules to obtain analysis results; a visualization sub-module for visualizing and outputting the analysis results, and the analysis results include: the comparison results of the predicted outlets and the receipt outlets of various prediction algorithms, the comparison results of the predicted three-segment codes and the receipt three-segment codes of various prediction algorithms, and the error details of the prediction algorithm.

[0019] Optionally, in a sixth implementation method of the second aspect of the present invention, the visualization sub-module includes: a first comparison unit, used to compare the prediction points of various prediction algorithms with the receipt points, and mark the prediction algorithms with accurate predictions and incorrect predictions respectively to obtain a first comparison result; a second comparison unit, used to compare the prediction three-segment codes of various prediction algorithms with the receipt three-segment codes, and mark the erroneous parts of the prediction three-segment codes to obtain a second comparison result; a tracing unit, used to trace the prediction algorithm with prediction errors based on the summary results, mark the data that causes the prediction errors, and form an error detail result; a generation unit, used to visualize the first comparison result, the second comparison result and the error detail result, and output the visualization processing result to generate a display page.

[0020] A third aspect of the present invention provides an intelligent diagnostic device for allocating erroneous waybills, comprising a memory and at least one processor, wherein the memory stores computer-readable instructions; the at least one processor calls the computer-readable instructions in the memory to execute the various steps of the intelligent diagnostic method for allocating erroneous waybills as described above.

[0021] A fourth aspect of the present invention provides a computer-readable storage medium having computer-readable instructions stored thereon, which, when executed by a processor, implement the various steps of the intelligent diagnosis method for allocating erroneous waybills as described above.

[0022] Beneficial effect: The present invention provides an intelligent diagnosis method for allocating incorrect waybills, which first obtains the order data of the consumer order distribution center, adopts multiple prediction algorithms, and performs calculations based on the order data to obtain multiple calculation process data and calculation results, and stores the calculation process data and calculation results in a database, so that subsequent statistical analysis has more data basis and can better find potential causes of errors in the system; then, by acquiring the branch data, the order distribution accuracy and recognition rate of the branch are detected, if the order distribution accuracy or recognition rate of the branch is lower than the set critical value, the corresponding calculation process data and calculation results in the database are extracted and summarized, so that the data can be better compared with the predicted data later; finally, by obtaining the summary results, according to pre-set rules, the summary results are counted and analyzed to obtain analysis results, and the analysis results are output, and by comparing the predicted data of multiple prediction algorithms with the actual data, the prediction accuracy of various prediction algorithms is systematically compared, so that the logistics company can intuitively see the statistics of order distribution errors and the analysis results of the causes of errors, so that the order distribution algorithm can be optimized in a targeted manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A first flow chart of the intelligent diagnosis method for allocating wrong waybills provided in an embodiment of the present invention;

[0024] Figure 2 A second flow chart of the intelligent diagnosis method for allocating wrong waybills provided in an embodiment of the present invention;

[0025] Figure 3 A third flow chart of the intelligent diagnosis method for allocating wrong waybills provided in an embodiment of the present invention;

[0026] Figure 4 A fourth flow chart of the intelligent diagnosis method for allocating wrong waybills provided in an embodiment of the present invention;

[0027] Figure 5 A fifth flow chart of the intelligent diagnosis method for allocating wrong waybills provided in an embodiment of the present invention;

[0028] Figure 6 A sixth flow chart of the intelligent diagnosis method for allocating wrong waybills provided in an embodiment of the present invention;

[0029] Figure 7 A seventh flow chart of the intelligent diagnosis method for allocating wrong waybills provided in an embodiment of the present invention;

[0030] Figure 8 A schematic diagram of the structure of an intelligent diagnostic device for misassigned waybills provided in an embodiment of the present invention;

[0031] Fig. 9Another structural schematic diagram of the intelligent diagnosis device for misassigned waybills provided in an embodiment of the present invention;

[0032] Fig.10 A schematic diagram of the structure of an intelligent diagnostic device for allocating incorrect waybills provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The present invention provides an intelligent diagnosis method, device, equipment and storage medium for allocating wrong waybills. The present invention first obtains the order data of the consumer order distribution center, uses multiple prediction algorithms to calculate the order data, obtains multiple calculation process data and calculation results, so that it can subsequently compare which prediction algorithm has better accuracy and the error conditions of various algorithms, and can better find the error cause of the system algorithm; then, by obtaining the outlet data, the order accuracy and recognition rate of the outlet are detected. If the order accuracy or recognition rate of the outlet is lower than the set critical value, the corresponding calculation process data and calculation results in the database are extracted and summarized, so that the actual data can be better compared with the predicted data; finally, by obtaining the summary results, according to the pre-set rules, the summary results are counted and analyzed, the analysis results are obtained, and the analysis results are output. By comparing the predicted data of multiple prediction algorithms with the actual data, it can show why the order distribution algorithm or system calculation error, the abnormal cause and other contents, so that the management personnel can correct the error in time, optimize the order distribution algorithm or system, and reduce the error from the root.

[0034] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0035] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 The first embodiment of the intelligent diagnosis method for allocating wrong waybills in the embodiment of the present invention includes:

[0036] S101. Obtain the order data of the consumer order center, use multiple prediction algorithms, calculate according to the order data to obtain multiple calculation process data and calculation results, and store the calculation process data and calculation results in the database;

[0037] Specifically, you can use a consumption KAFKA topic with 32 partitions, corresponding to 32 instances of the consumer group. You can put data into a storage table with date as the dimension every day, and store 30 days of data in total. You will record the keywords parsed by many algorithms based on the address of the order number and the prediction results for easy statistics and aggregation. When the data falls into the storage table, the system will use many complex and sophisticated algorithms to deeply process the data and store it in the clickhouse database.

[0038] S102. Obtaining the data of the outlets, detecting the accuracy and recognition rate of the order splitting of the outlets, if the accuracy or recognition rate of the order splitting of the outlets is lower than the set critical value, extracting the corresponding calculation process data and calculation results from the database and summarizing them;

[0039] Specifically, you can use XXL-JOB scheduled tasks to read the accuracy and recognition rate configuration of tens of thousands of outlets. XXL-JOB reads the accuracy and recognition rate configuration data preset for each outlet from a huge number of tens of thousands of outlet information bases. These configuration data are like precise rulers to measure whether the performance of each outlet in the order distribution business is up to standard. For those outlets that meet the conditions for summary statistics, the system will quickly call a special summary calculation interface. For example, by writing efficient SQL query statements, the data of the corresponding outlets can be accurately pulled from the clickhouse database.

[0040] S103. Obtain summary results, perform statistics and analysis on the summary results according to pre-set rules, obtain analysis results, and output the analysis results.

[0041] After the data is pulled, the complex business logic inside the program begins to play a role, just like an experienced statistician, carefully summarizing and statistically analyzing the data. For example, it will classify and organize the data according to different algorithms, and calculate the accuracy contribution of each algorithm in different outlets, the fluctuation of recognition rate, etc. After a series of rigorous calculations, the final summary statistical results are returned to the user, and these results are written into a special summary table at the same time. This summary table is like a stage for data display, providing a data source for subsequent front-end page query and display, so that business personnel can intuitively understand the order distribution business status of each outlet and order distribution error analysis.

[0042] See also Figure 2 The second embodiment of the intelligent diagnosis method for allocating wrong waybills in the embodiment of the present invention includes:

[0043] S201. Obtain the order data from the consumer order center and extract the order number and address data from the order data;

[0044] S202. Use a text parsing algorithm to extract keywords from the address data;

[0045] Based on the address information corresponding to the order number, use advanced text parsing technology to extract key address keywords, and then multi-dimensional analysis and calculation can be performed based on these keywords to finally obtain the sub-order prediction result.

[0046] S203. Use multiple prediction algorithms to calculate according to the keywords, obtain multiple calculation process data and calculation results, and store the order number, keywords, calculation process data and calculation results in the database.

[0047] The original order number, parsed keywords, applied algorithm steps, and final calculation results are all important data details that need to be recorded in detail and stored in the database. These records not only help to preserve the integrity of the data, but also provide a solid and comprehensive data foundation for subsequent data statistics and summary work, enabling data analysts to mine valuable information from this rich data treasure trove, such as sub-order trends and algorithm accuracy evaluation.

[0048] Please refer to Figure 3 , the third embodiment of the intelligent diagnosis method for misallocated waybills in the embodiments of the present invention includes:

[0049] S301. Preprocess the address data, remove irrelevant characters and stop words, and unify the format of all characters to obtain standardized data;

[0050] Specifically, punctuation marks, special characters, spaces, etc. need to be deleted. Only letters, numbers, and necessary delimiters (such as spaces between streets, cities, and states) are retained. The removal of stop words can be completed by creating a stop word list, which contains common meaningless words (such as "of", "is", "in", etc.). By traversing the address data, the words matching the stop word list can be removed. In addition, by converting all characters to lowercase, problems caused by inconsistent case can be avoided.

[0051] S302. Use a word segmentation tool to split the standardized data to obtain a split result;

[0052] An appropriate word segmentation tool can be selected according to the language type. For example, jieba can be used for Chinese.

[0053] S303. Use a machine learning model to identify the split result, obtain multiple keywords, classify and extract the identified keywords, and organize them into a structured format.

[0054] Machine learning models can automatically process large amounts of text data without the need for manual screening and labeling, which greatly improves work efficiency. During use, new data can be continuously added, allowing machine learning models to update their knowledge base through incremental learning or online learning to adapt to new contexts and keywords. Specifically, machine learning models can be built based on convolutional neural networks (CNN), recurrent neural networks (RNN) and their variants (such as LSTM, GRU), and Transformers. The resulting models can learn and capture complex features in the data, thereby more accurately identifying keywords.

[0055] See also Figure 4 The fourth embodiment of the intelligent diagnosis method for allocating wrong waybills in the embodiment of the present invention includes:

[0056] S401. Construct multiple prediction algorithms, including: delivery range algorithm, keyword manual designation algorithm, map fence algorithm, and history library calculation algorithm;

[0057] S402. Using a variety of prediction algorithms, respectively, according to the keywords to calculate, to obtain a plurality of calculation process data and calculation results;

[0058] There are many algorithms that can be used to predict and calculate parcel splitting. In addition to the various prediction algorithms mentioned above, there are also district-county-only algorithms, township-only algorithms, etc. Different types of algorithms have different calculation logics, so the calculation results obtained will also be different. The daily forecasted order volume data of logistics companies can reach tens of millions, which will amplify the algorithm differences and errors that are usually difficult to find. By using multiple prediction algorithms for calculation, detailed calculation process data and calculation results can be obtained, forming sufficient sample data.

[0059] S403. Classify and sort multiple calculation process data and calculation results according to algorithm types, and store the classified data in a database.

[0060] For example, the calculation process data and calculation results obtained by the delivery range algorithm are stored, and the data of other types of algorithms are also stored separately. By classifying and effectively organizing the data, the subsequent summary analysis results can be more accurate.

[0061] See also Figure 5 The fifth embodiment of the intelligent diagnosis method for allocating wrong waybills in the embodiment of the present invention includes:

[0062] S501. Set a first critical value for the accuracy of the order and a second critical value for the recognition rate, obtain the network data regularly, and compare the accuracy of the order and the recognition rate of the network data with the first critical value and the second critical value respectively;

[0063] The order splitting accuracy rate is a measure of the proportion of orders correctly assigned to the corresponding delivery areas or personnel when processing orders. The recognition rate refers to the proportion of accurate recognition of order information or customer identity by the outlets when processing orders or customers. Both indicators are crucial to ensuring the accuracy of order processing. In order to improve the efficiency of order splitting, logistics companies generally use intelligent order splitting algorithms to handle the order splitting of packages. Although this method is efficient, there is still a small probability of order splitting errors. If the causes of such errors are eliminated one by one manually, it will increase a lot of labor costs and is not feasible. The use of intelligent diagnosis can well find the causes and solve them.

[0064] S502. If the order splitting accuracy is lower than the first critical value, extract the corresponding branch number from the branch data;

[0065] S503. If the recognition rate is lower than the second critical value, extract the dot number in the corresponding dot data;

[0066] S504. Summarize all extracted network point numbers, remove the same network point numbers, extract the corresponding calculation process data and calculation results in the database according to the network point numbers, and summarize them.

[0067] The branch numbers extracted by the above rules may be repeated. For example, when the branch's order splitting accuracy and recognition rate meet the extraction requirements, the branch number of the branch will be repeated. By removing the repeated branch numbers, the system's computational load can be reduced, and no repeated content will appear in the summarized analysis results.

[0068] See also Figure 6 The sixth embodiment of the intelligent diagnosis method for allocating wrong waybills in the embodiment of the present invention includes:

[0069] S601. Formulate statistical analysis rules, the statistical analysis rules including: display forms of forecast data and actual data, and description templates of analysis results;

[0070] When outputting the analysis results, it is necessary to consider the display format so that users and those who need the analysis results can understand them more easily and find the desired content. Specifically, the comparison results of the predicted data and the actual data can be displayed in different types of charts to more intuitively see the prediction accuracy and error points of different prediction algorithms. In addition, when outputting the analysis results, it is necessary to explain the results so that business managers can understand the analysis results more clearly. The expression format of the analysis results can also be achieved by setting the description template in advance.

[0071] S602. Obtain summary results, perform statistics and analysis on the summary results according to pre-set statistical analysis rules, and obtain analysis results;

[0072] S603. Visualize and output the analysis results, including: comparison results of predicted outlets and receipt outlets of various prediction algorithms, comparison results of predicted three-segment codes and receipt three-segment codes of various prediction algorithms, and error details of prediction algorithms.

[0073] Specifically, the outlets and three-segment codes of incorrectly sorted orders can be displayed through the development page. By displaying the summary table data, the outlet data related to abnormal receipt can be presented to the business party in a clear and intuitive manner. The displayed data content can include predicted outlet information, which allows business personnel to clearly understand the outlets to which the original system predicts that the package should be delivered; the predicted three-segment code can further refine and display the predicted delivery area information. In addition, by comparing the actual receipt outlets and the receipt three-segment code, business personnel can quickly find the differences in the order sorting errors.

[0074] See also Figure 7 The seventh embodiment of the intelligent diagnosis method for allocating wrong waybills in the embodiment of the present invention includes:

[0075] S701. Compare the prediction points of various prediction algorithms with the receipt points, mark the prediction algorithms with accurate predictions and incorrect predictions, and obtain the first comparison result;

[0076] S702. Compare the predicted three-segment code of various prediction algorithms with the signed three-segment code, mark the erroneous part of the predicted three-segment code, and obtain the second comparison result;

[0077] S703. Based on the summary results, the prediction algorithm with the prediction error is traced back, the data causing the prediction error is marked, and the error detailed results are formed;

[0078] By marking the different parts, users can quickly find the differences when viewing the analysis results. Specifically, when visualizing the marked content, you can use background color or change the font color to show the differences.

[0079] S704. Perform visualization processing on the first comparison result, the second comparison result and the error detail result, and output the visualization processing result to generate a display page.

[0080] In this embodiment, in addition to displaying the comparison results between the predicted outlets and the receipt outlets and the predicted results between the predicted three-segment code and the receipt three-segment code, the system will also display in detail the algorithm details that lead to incorrect order sorting.

[0081] For example, the details can show that the error in the delivery range algorithm may be caused by deviation, or it may be a misjudgment of the keyword manual assignment algorithm. In addition, the analysis results generated by the system will also deeply analyze the causes of the error, such as fuzzy matching of address keywords, untimely updates of map data, or interference from historical data. In addition to the above reasons, keywords are often the key clues to order splitting errors. These reasons can be traced back with the help of calculation process data and calculation results to find the most likely influencing factors, thereby helping the business side to optimize the influencing factors at the source.

[0082] The above describes the intelligent diagnosis method for misassigning waybills in the embodiment of the present invention. The following describes the intelligent diagnosis device for misassigning waybills in the embodiment of the present invention. Figure 8 , an embodiment of the intelligent diagnosis device for allocating wrong waybills in the embodiment of the present invention includes:

[0083] The calculation module 10 is used to obtain the order data of the consumption order distribution center, use multiple prediction algorithms, perform calculations based on the order data to obtain multiple calculation process data and calculation results, and store the calculation process data and calculation results in a database;

[0084] The extraction module 20 is used to obtain the network point data, detect the order separation accuracy and recognition rate of the network point, and if the order separation accuracy or recognition rate of the network point is lower than the set critical value, extract the corresponding calculation process data and calculation results in the database and summarize them;

[0085] The analysis module 30 is used to obtain the summary results, perform statistics and analysis on the summary results according to pre-set rules, obtain analysis results, and output the analysis results.

[0086] See also Fig. 9 , an embodiment of the intelligent diagnosis device for allocating wrong waybills in the embodiment of the present invention includes:

[0087] The calculation module 10 is used to obtain the order data of the consumption order distribution center, use multiple prediction algorithms, perform calculations based on the order data to obtain multiple calculation process data and calculation results, and store the calculation process data and calculation results in a database;

[0088] The extraction module 20 is used to obtain the network point data, detect the order separation accuracy and recognition rate of the network point, and if the order separation accuracy or recognition rate of the network point is lower than the set critical value, extract the corresponding calculation process data and calculation results in the database and summarize them;

[0089] The analysis module 30 is used to obtain the summary results, perform statistics and analysis on the summary results according to pre-set rules, obtain analysis results, and output the analysis results;

[0090] In this embodiment, the calculation module 10 includes:

[0091] Extraction submodule 11, used to obtain the order data of the consumption order center, and extract the order number and address data in the order data;

[0092] The parsing submodule 12 is used to extract keywords from the address data using a text parsing algorithm;

[0093] The prediction submodule 13 is used to use a plurality of prediction algorithms to perform calculations according to the keywords, obtain a plurality of calculation process data and calculation results, and store the order number, keyword, calculation process data and calculation results in a database;

[0094] In this embodiment, the parsing submodule 12 includes:

[0095] A preprocessing unit 121 is used to preprocess the address data, remove irrelevant characters and stop words, and unify the formats of all characters to obtain standardized data;

[0096] A splitting unit 122 is used to split the standardized data using a word segmentation tool to obtain a splitting result;

[0097] The recognition unit 123 is used to recognize the split result by using a machine learning model to obtain a plurality of keywords, classify and extract the recognized keywords, and organize them into a structured format;

[0098] In this embodiment, the prediction submodule 13 includes:

[0099] A construction unit 131 is used to construct a plurality of prediction algorithms, wherein the prediction algorithms include: a delivery range algorithm, a keyword manual designation algorithm, a map fence algorithm, and a history library calculation algorithm;

[0100] A calculation unit 132, used to use a plurality of prediction algorithms to perform calculations according to the keywords respectively, to obtain a plurality of calculation process data and calculation results;

[0101] A classification unit 133 is used to classify and sort multiple calculation process data and calculation results according to the algorithm type, and store the classified data in a database;

[0102] In this embodiment, the extraction module 20 includes:

[0103] A setting unit 21 is used to set a first critical value for the order splitting accuracy rate and a second critical value for the recognition rate, respectively, to periodically obtain network point data, and to compare the order splitting accuracy rate and the recognition rate of the network point data with the first critical value and the second critical value, respectively;

[0104] A first extraction unit 22, for extracting a network point number from the corresponding network point data if the order separation accuracy is lower than a first critical value;

[0105] A second extraction unit 23, for extracting a dot number from the corresponding dot data if the recognition rate is lower than a second critical value;

[0106] A summarizing unit 24 is used to summarize all extracted network point numbers, remove the same network point numbers, extract the corresponding calculation process data and calculation results in the database according to the network point numbers, and summarize them;

[0107] In this embodiment, the analysis module 30 includes:

[0108] A formulation submodule 31 is used to formulate statistical analysis rules, wherein the statistical analysis rules include: a display format of predicted data and actual data, and a description template of analysis results;

[0109] The analysis submodule 32 is used to obtain the summary results, and to perform statistics and analysis on the summary results according to the pre-set statistical analysis rules to obtain the analysis results;

[0110] The visualization submodule 33 is used to visualize and output the analysis results, including: the comparison results of the predicted network points and the receipt network points of various prediction algorithms, the comparison results of the predicted three-segment codes and the receipt three-segment codes of various prediction algorithms, and the error details of the prediction algorithms;

[0111] In this embodiment, the visualization submodule 33 includes:

[0112] The first comparison unit 331 is used to compare the prediction points of various prediction algorithms with the signing points, mark the prediction algorithms with accurate predictions and wrong predictions respectively, and obtain a first comparison result;

[0113] The second comparison unit 332 is used to compare the predicted three-segment code of various prediction algorithms with the received three-segment code, mark the part with errors in the predicted three-segment code, and obtain a second comparison result;

[0114] The tracing unit 333 is used to trace the prediction algorithm of the prediction error according to the summary result, mark the data causing the prediction error, and form an error detailed result;

[0115] The generating unit 334 is used to perform visualization processing on the first comparison result, the second comparison result and the error detail result, and output the visualization processing result to generate a display page.

[0116] The intelligent diagnostic device for allocating incorrect waybills of the present invention first obtains the order splitting data, and then uses a variety of prediction algorithms to predict and calculate the obtained data, and stores the calculation process data and calculation results in a database, and then further obtains the actual branch data, and compares the actual branch data with the results of the various prediction algorithms to obtain detailed analysis results. The analysis result content may include: predicted branch company, prediction three-segment code, algorithm rules of prediction results, summary content affecting the order quantity, solution processing details, etc. Finally, the analysis results are output, such as inserted into the MYSQL database business table, for the front-end WEB page to query these summary data, so that the business management party can clearly understand the possible reasons for the calculation errors caused by the order splitting algorithm of the existing system, and thereby solve the corresponding problems in a targeted manner and correct the prediction results.

[0117] The above is a detailed description of the intelligent diagnosis device for allocating erroneous waybills in an embodiment of the present invention from the perspective of modular functional entities. The following is a detailed description of the intelligent diagnosis device for allocating erroneous waybills in an embodiment of the present invention from the perspective of hardware processing.

[0118] Fig.10 A schematic diagram of the structure of an intelligent diagnostic device for misallocating waybills provided in an embodiment of the present invention, wherein the intelligent diagnostic device 900 for misallocating waybills may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 910 (for example, one or more processors) and a memory 920, and one or more storage media 930 (for example, one or more mass storage devices) storing application programs 933 or data 932. Among them, the memory 920 and the storage medium 930 may be temporary storage or permanent storage. The program stored in the storage medium 930 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the intelligent diagnostic device 900 for misallocating waybills. Furthermore, the processor 910 may be configured to communicate with the storage medium 930, and execute a series of instruction operations in the storage medium 930 on the intelligent diagnostic device 900 for misallocating waybills, so as to implement the steps of the intelligent diagnostic method for misallocating waybills provided in the above-mentioned method embodiments.

[0119] The intelligent diagnostic device 900 for allocating wrong waybills may also include one or more power supplies 940, one or more wired or wireless network interfaces 950, one or more input and output interfaces 960, and / or one or more operating systems 931, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. It will be appreciated by those skilled in the art that Fig.10The structure of the intelligent diagnostic device for misassigning waybills shown does not constitute a limitation on the intelligent diagnostic device for misassigning waybills, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0120] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of an intelligent diagnosis method for allocating erroneous waybills.

[0121] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the equipment or device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0122] If the integrated unit is implemented in the form of 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 is essentially 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. The computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0123] It is understandable that those skilled in the art can make equivalent substitutions or changes based on the technical solution and inventive concept of the present invention, and all these changes or substitutions should fall within the protection scope of the claims attached to the present invention.

Claims

1. An intelligent diagnosis method for misassigned waybills, characterized in that: The steps include: Obtain the order data of the consumer order distribution center, use multiple prediction algorithms, perform calculations based on the order data to obtain multiple calculation process data and calculation results, and store the calculation process data and calculation results in a database; Obtaining the data of outlets, testing the order splitting accuracy and recognition rate of the outlets, if the order splitting accuracy or recognition rate of the outlets is lower than the set critical value, extracting the corresponding calculation process data and calculation results from the database and summarizing them; Obtain summary results, perform statistics and analysis on the summary results according to pre-set rules, obtain analysis results, and output the analysis results.

2. The intelligent diagnosis method for misassigned waybills according to claim 1, characterized in that: The obtaining of the order data of the consumer order distribution center, using a plurality of prediction algorithms, performing calculations based on the order data to obtain a plurality of calculation process data and calculation results, and storing the calculation process data and calculation results in a database, specifically includes: Obtain the order data from the consumer order center and extract the order number and address data from the order data; Use text parsing algorithms to extract keywords from address data; A plurality of prediction algorithms are used to perform calculations according to the keywords to obtain a plurality of calculation process data and calculation results, and the order number, keywords, calculation process data and calculation results are stored in a database.

3. The intelligent diagnosis method for misassigned waybills according to claim 2, characterized in that: The text parsing algorithm is used to extract keywords from the address data, specifically including: Preprocess the address data, remove irrelevant characters and stop words, and unify the format of all characters to obtain standardized data; Use word segmentation tools to split the standardized data and obtain the split results; A machine learning model is used to identify the split results, obtain multiple keywords, classify and extract the identified keywords, and organize them into a structured format.

4. The intelligent diagnosis method for misassigned waybills according to claim 2, characterized in that: The method adopts multiple prediction algorithms, performs calculations according to the keywords, obtains multiple calculation process data and calculation results, and stores the keywords, calculation process data and calculation results in a database, specifically including: Constructing multiple prediction algorithms, including: delivery range algorithm, keyword manual designation algorithm, map fence algorithm, and history library calculation algorithm; Using a variety of prediction algorithms, respectively performing calculations according to the keywords, to obtain a plurality of calculation process data and calculation results; Multiple calculation process data and calculation results are classified and sorted according to the algorithm type, and the classified data is stored in the database.

5. The intelligent diagnosis method for misassigned waybills according to claim 1, characterized in that: The method of acquiring the network point data, detecting the order splitting accuracy and recognition rate of the network point, and if the order splitting accuracy or recognition rate of the network point is lower than the set critical value, extracting the corresponding calculation process data and calculation results in the database and summarizing them, specifically includes: A first critical value for the order splitting accuracy rate and a second critical value for the recognition rate are respectively set, and the network point data are regularly acquired, and the order splitting accuracy rate and the recognition rate of the network point data are respectively compared with the first critical value and the second critical value; If the order splitting accuracy is lower than the first critical value, extract the branch number in the corresponding branch data; If the recognition rate is lower than the second critical value, extracting the dot number in the corresponding dot data; Summarize all extracted network point numbers, remove the same network point numbers, extract the corresponding calculation process data and calculation results in the database according to the network point numbers and summarize them.

6. The intelligent diagnosis method for misassigned waybills according to claim 1, characterized in that: The obtaining of the summary results, performing statistics and analysis on the summary results according to a preset rule, obtaining analysis results, and outputting the analysis results, specifically includes: Formulate statistical analysis rules, including: display forms of forecast data and actual data, and explanation templates of analysis results; Obtain summary results, and perform statistics and analysis on the summary results according to pre-set statistical analysis rules to obtain analysis results; The analysis results are visualized and output, and the analysis results include: comparison results of predicted outlets and receipt outlets of various prediction algorithms, comparison results of predicted three-segment codes and receipt three-segment codes of various prediction algorithms, and error details of prediction algorithms.

7. The intelligent diagnosis method for misassigned waybills according to claim 6, characterized in that: The analysis results are visualized and output, and the analysis results include: the comparison results of the predicted network points and the receipt network points of various prediction algorithms, the comparison results of the predicted three-segment codes and the receipt three-segment codes of various prediction algorithms, and the error details of the prediction algorithm, specifically including: Compare the prediction points of various prediction algorithms with the receipt points, mark the prediction algorithms with accurate predictions and wrong predictions respectively, and obtain the first comparison result; Compare the predicted three-segment codes of various prediction algorithms with the received three-segment codes, mark the erroneous parts of the predicted three-segment codes, and obtain a second comparison result; Based on the summary results, the prediction algorithm with prediction errors is traced, the data causing the prediction errors are marked, and detailed error results are formed; The first comparison result, the second comparison result and the error detail result are visualized, and the visualization processing result is output to generate a display page.

8. An intelligent diagnostic device for misassigned waybills, characterized in that: include: A calculation module is used to obtain the order data of the consumer order distribution center, use multiple prediction algorithms, perform calculations based on the order data to obtain multiple calculation process data and calculation results, and store the calculation process data and calculation results in a database; The extraction module is used to obtain the network point data, detect the order separation accuracy and recognition rate of the network point, and if the order separation accuracy or recognition rate of the network point is lower than the set critical value, extract the corresponding calculation process data and calculation results in the database and summarize them; The analysis module is used to obtain the summary results, perform statistics and analysis on the summary results according to pre-set rules, obtain the analysis results, and output the analysis results.

9. An intelligent diagnostic device for misassigned waybills, characterized in that: comprising a memory and at least one processor, wherein the memory has computer-readable instructions stored therein; The at least one processor calls the computer-readable instructions in the memory to execute the various steps of the intelligent diagnosis method for allocating incorrect waybills as described in any one of claims 1-7.

10. A computer-readable storage medium having computer-readable instructions stored thereon, characterized in that: When the computer-readable instructions are executed by a processor, the steps of the intelligent diagnosis method for allocating erroneous waybills as described in any one of claims 1-7 are implemented.

Images