Intelligent matching method, device and equipment for destination website information and storage medium

Abstract

The invention relates to the technical field of information processing, in particular to a target website information intelligent matching method, device and equipment and a storage medium, and the method comprises the steps: inputting real-time order data into a website matching model, outputting to obtain information of a plurality of prepared target websites, and obtaining feature variable information in real time; the score of each website in the preliminary destination website information is calculated according to the feature variable information and the weight information in a weighted average manner, the optimal destination website information is generated based on the website with the highest score, the optimal destination website information is sent to the client, change factors are comprehensively considered, the flexibility is improved, the optimal destination website can be accurately predicted, and the accuracy of the optimal destination website is improved. The accuracy of parcel delivery is improved, overload of a website or delivery delay is prevented, the dependence on human resources is reduced, the labor cost is reduced, the automation degree is improved, the parcel sorting and matching efficiency is improved, the order can be responded in real time, and the customer satisfaction degree is improved.

Description

Technical Field

[0001] The present invention relates to the field of information processing technology, and in particular to a method, device, equipment and storage medium for intelligently matching destination network point information. Background Art

[0002] In the express logistics industry, accurately matching and selecting destination outlets is an important step to ensure that packages can be delivered quickly and safely. Last-mile delivery refers to optimizing the last-mile delivery route and time by accurately matching destination outlets. Existing outlet matching methods mainly rely on manual allocation, with a low degree of automation, which reduces the efficiency of package sorting and matching and is prone to errors. They cannot respond to orders in real time, especially when processing large-scale data, which reduces customer satisfaction. Or they match based on simple rules (such as geographical distance), but such simple matching rules do not comprehensively consider other complex influencing factors such as traffic conditions, weather conditions, and load conditions. They only rely on pre-collected static data rather than real-time data, which is inflexible and cannot accurately predict the optimal destination outlet, reducing the accuracy of package delivery, and easily causing some outlets to be overloaded or delayed. Summary of the invention

[0003] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a method, device, equipment and storage medium for intelligently matching destination network information that comprehensively considers changing factors, improves flexibility, can accurately predict the optimal destination network point, improves the accuracy of package delivery, prevents network point overload or delivery delays, reduces dependence on human resources, reduces labor costs, improves the degree of automation, improves the efficiency of package sorting and matching, can respond to orders in real time, and improves customer satisfaction.

[0004] The first aspect of the present invention provides a method for intelligent matching of destination network point information, comprising: collecting historical order data, extracting a key feature set from the historical order data, the key feature set including delivery network points, delivery time and customer preferences; constructing a network point matching model based on the key feature set, and receiving real-time order data sent by a client; inputting the real-time order data into the network point matching model, outputting a plurality of preliminary destination network point information, and acquiring feature variable information in real time, the feature variable information including the traffic conditions, weather conditions and load conditions of each network point; querying the weight information corresponding to the traffic conditions, the weather conditions and the load conditions, respectively, and calculating the score of each network point in the preliminary destination network point information by weighted average according to the feature variable information and the weight information; generating the best destination network point information based on the network point with the highest score, and sending the best destination network point information to the client.

[0005] Optionally, in a first implementation method of the first aspect of the present invention, the collecting of historical order data and the extraction of key feature sets in the historical order data, wherein the key feature sets include delivery outlets, delivery time and customer preferences, include: collecting historical order data, parsing the historical order data to obtain historical order parsed data; removing duplicate data in the historical order parsed data to obtain historical order deduplicated data; using regular expressions to correct erroneous data in the historical order deduplicated data to obtain historical order corrected data; filling missing values ​​in the historical order corrected data to obtain historical order filled data; based on a preset target standard format, formatting the historical order filled data to obtain historical order standard data; extracting key feature sets in the historical order standard data, wherein the key feature sets include delivery outlets, delivery time and customer preferences.

[0006] Optionally, in a second implementation method of the first aspect of the present invention, the branch matching model is constructed according to the key feature set, and real-time order data sent by the client is received, including: dividing the key feature set according to a preset split ratio to obtain a training set, a validation set, and a test set; selecting a basic model, using a random forest learning algorithm and training the basic model through the training set to obtain a branch matching model; respectively tuning and evaluating the branch matching model through the validation set and the test set; and receiving real-time order data sent by the client.

[0007] Optionally, in a third implementation method of the first aspect of the present invention, the real-time order data is input into the network point matching model, and multiple prepared destination network point information is outputted, and characteristic variable information is acquired in real time, and the characteristic variable information includes the traffic conditions, weather conditions and load conditions of each network point. The method comprises: inputting the real-time order data into the network point matching model, and multiple prepared destination network point information is outputted; extracting the order destination information in the real-time order data; identifying the delivery range of the order destination information; and acquiring characteristic variable information in real time according to the delivery range, and the characteristic variable information includes the traffic conditions, weather conditions and load conditions of each network point.

[0008] Optionally, in a fourth implementation method of the first aspect of the present invention, the querying of weight information respectively corresponding to the traffic condition, the weather condition, and the load condition, and calculating the score of each outlet in the preliminary destination outlet information by weighted average according to the characteristic variable information and the weight information, includes: calling a preset weight configuration list; querying the weight information respectively corresponding to the traffic condition, the weather condition, and the load condition from the weight configuration list; and calculating the score of each outlet in the preliminary destination outlet information by weighted average according to the characteristic variable information and the weight information.

[0009] Optionally, in a fifth implementation method of the first aspect of the present invention, the best destination outlet information is generated based on the outlet with the highest score, and the best destination outlet information is sent to the client, including: sorting each outlet in the preliminary destination outlet information according to the score to obtain sorting information; determining the best outlet based on the sorting information, and identifying the outlet name, outlet address, outlet number and outlet contact information corresponding to the best outlet; generating the best destination outlet information based on the outlet name, the outlet address, the outlet number and the outlet contact information; and sending the best destination outlet information to the client.

[0010] Optionally, in a sixth implementation method of the first aspect of the present invention, after the best destination outlet information is generated based on the outlet with the highest score and the best destination outlet information is sent to the client, it also includes: creating a listener and using the listener to listen to the actual delivery outlet information; when the listener listens to the actual delivery outlet information, obtaining the current actual delivery outlet information; determining whether the actual delivery outlet information is consistent with the best destination outlet information; if not, generating feedback information and uploading the feedback information to the blockchain.

[0011] The second aspect of the present invention provides an intelligent matching device for destination network point information, including: a collection and extraction module, used to collect historical order data, and extract key feature sets from the historical order data, wherein the key feature sets include delivery network points, delivery time and customer preferences; a construction and receiving module, used to construct a network point matching model according to the key feature sets, and receive real-time order data sent by a client; an input acquisition module, used to input the real-time order data into the network point matching model, output a plurality of preliminary destination network point information, and obtain feature variable information in real time, wherein the feature variable information includes the traffic conditions, weather conditions and load conditions of each network point; a query and calculation module, used to query the weight information corresponding to the traffic conditions, the weather conditions and the load conditions, respectively, and calculate the score of each network point in the preliminary destination network point information by weighted average according to the feature variable information and the weight information; a generation and sending module, used to generate the best destination network point information based on the network point with the highest score, and send the best destination network point information to the client.

[0012] Optionally, in a first implementation method of the second aspect of the present invention, the collection and extraction module includes: a collection and parsing unit, which is used to collect historical order data, parse the historical order data, and obtain historical order parsed data; a removal unit, which is used to remove duplicate data in the historical order parsed data, and obtain historical order deduplication data; a correction unit, which is used to correct erroneous data in the historical order deduplication data using a regular expression, and obtain historical order corrected data; a filling unit, which is used to fill missing values ​​in the historical order corrected data, and obtain historical order filled data; a conversion unit, which is used to convert the format of the historical order filled data based on a preset target standard format, and obtain historical order standard data; a first extraction unit, which is used to extract a key feature set in the historical order standard data, and the key feature set includes delivery outlets, delivery time and customer preferences.

[0013] Optionally, in a second implementation method of the second aspect of the present invention, the construction receiving module includes: a division unit, used to divide the key feature set according to a preset splitting ratio to obtain a training set, a verification set and a test set; a training selection unit, used to select a basic model, and use a random forest learning algorithm to train the basic model through the training set to obtain a network matching model; a tuning and evaluation unit, used to tune and evaluate the network matching model through the verification set and the test set respectively; a receiving unit, used to receive real-time order data sent by the client.

[0014] Optionally, in a third implementation of the second aspect of the present invention, the input acquisition module includes: an input unit, used to input the real-time order data into the network point matching model, and output a plurality of preliminary destination network point information; a second extraction unit, used to extract the order destination information from the real-time order data; an identification unit, used to identify the delivery range of the order destination information; an acquisition unit, used to acquire characteristic variable information in real time according to the delivery range, and the characteristic variable information includes traffic conditions, weather conditions and load conditions of each network point.

[0015] Optionally, in a fourth implementation of the second aspect of the present invention, the query and calculation module includes: a calling unit, used to call a preset weight configuration list; a query unit, used to query the weight information corresponding to the traffic conditions, the weather conditions and the load conditions from the weight configuration list; and a calculation unit, used to calculate the score of each network point in the preliminary destination network point information by weighted average based on the characteristic variable information and the weight information.

[0016] Optionally, in a fifth implementation of the second aspect of the present invention, the generation and sending module includes: a sorting unit, used to sort each outlet in the prepared destination outlet information according to the score to obtain sorting information; a determination and identification unit, used to determine the best outlet based on the sorting information, and identify the outlet name, outlet address, outlet number and outlet contact information corresponding to the best outlet; a generation unit, used to generate the best destination outlet information based on the outlet name, the outlet address, the outlet number and the outlet contact information; and a sending unit, used to send the best destination outlet information to the client.

[0017] Optionally, in a sixth implementation method of the second aspect of the present invention, it also includes: creating a monitoring module, used to create a listener, and using the listener to monitor the actual delivery point information; an acquisition module, used to acquire the current actual delivery point information when the listener listens to the actual delivery point information; a judgment module, used to judge whether the actual delivery point information is consistent with the optimal destination point information; and an upload module, used to generate feedback information if not, and upload the feedback information to the blockchain.

[0018] The third aspect of the present invention provides a destination network point information intelligent matching device, the destination network point information intelligent matching device comprising: a memory and at least one processor, the memory storing instructions; at least one of the processors calls the instructions in the memory to enable the destination network point information intelligent matching device to execute each step of the destination network point information intelligent matching method described in any one of the above items.

[0019] A fourth aspect of the present invention provides a computer-readable storage medium having instructions stored thereon, and when the instructions are executed by a processor, the various steps of the method for intelligently matching destination network point information described in any one of the above items are implemented.

[0020] In the technical solution of the present invention, real-time order data is input into the network point matching model, multiple prepared destination network point information is output, characteristic variable information is obtained in real time, and the score of each network point in the prepared destination network point information is calculated by weighted average according to the characteristic variable information and weight information. The best destination network point information is generated based on the network point with the highest score, and the best destination network point information is sent to the client. The changing factors are comprehensively considered to improve flexibility, accurately predict the best destination network point, improve the accuracy of package delivery, prevent network overload or delivery delays, reduce dependence on human resources, reduce labor costs, improve the degree of automation, improve the efficiency of package sorting and matching, and can respond to orders in real time to improve customer satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1A first flow chart of a method for intelligently matching destination network point information provided by an embodiment of the present invention;

[0022] Figure 2 A second flow chart of the method for intelligently matching destination network point information provided by an embodiment of the present invention;

[0023] Figure 3 A third flow chart of the method for intelligently matching destination network point information provided by an embodiment of the present invention;

[0024] Figure 4 A fourth flow chart of the method for intelligently matching destination network point information provided by an embodiment of the present invention;

[0025] Figure 5 A schematic diagram of a structure of a device for intelligently matching destination network point information provided by an embodiment of the present invention;

[0026] Figure 6 Another structural diagram of the device for intelligently matching destination network point information provided by an embodiment of the present invention;

[0027] Figure 7 A schematic diagram of the structure of a destination network point information intelligent matching device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The present invention provides a method, device, equipment and storage medium for intelligent matching of destination network point information, which comprehensively considers changing factors, improves flexibility, can accurately predict the optimal destination network point, improves the accuracy of package delivery, prevents network point overload or delivery delays, reduces dependence on human resources, reduces labor costs, improves the degree of automation, improves the efficiency of package sorting and matching, can respond to orders in real time, and improves customer satisfaction.

[0029] 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.

[0030] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1, an embodiment of the method for intelligently matching destination network point information in an embodiment of the present invention includes:

[0031] 101. Collect historical order data and extract key feature sets from the historical order data, where the key feature sets include delivery outlets, delivery time, and customer preferences;

[0032] In this embodiment, historical order data is collected from different data sources such as e-commerce platform data, logistics system data or customer feedback data. The historical order data includes basic order information, logistics information, user behavior data and environmental factors. The key feature set in the historical order data is extracted. The key feature set includes delivery outlets, delivery time and customer preferences. Data aggregation is performed based on various dimensions of each order (such as outlets, time, and customers) to obtain statistical features such as average delivery time and customer ratings. Geographic Information System (GIS) technology is used to process the geographical location information of the outlets, and the delivery path, delivery distance, etc. are calculated in combination with map data. The impact of order time-related features (such as peak season, off-season, etc.) on delivery timeliness is analyzed. Decision trees, random forests and other methods are used to perform feature selection to find features that are strongly correlated with delivery efficiency and customer preferences.

[0033] 102. Build a network point matching model based on the key feature set and receive real-time order data sent by the client;

[0034] In this embodiment, the key feature set is divided into a training set, a validation set, and a test set. The training set is used to train the model, and the validation set is used to evaluate the accuracy of the model. Hyperparameter tuning is performed to ensure that the model has good generalization capabilities. The final performance of the model is evaluated on the test set, and real-time order data sent by the client is received through an API or message queue system.

[0035] 103. Input the real-time order data into the network point matching model, output a plurality of preliminary destination network point information, and obtain characteristic variable information in real time, wherein the characteristic variable information includes the traffic conditions, weather conditions, and load conditions of each network point;

[0036] In this embodiment, features related to network point matching (such as delivery address, timeliness requirements, customer preferences, etc.) are extracted from real-time order data, and the extracted features are input into a trained network point matching model. The model will predict multiple suitable preliminary destination network point information based on the rules in the historical data. After the output of the order matching model, it is necessary to obtain some external feature variables in real time to further optimize the decision to ensure that the order can be delivered at the most suitable network point, that is, to obtain feature variable information in real time, and the feature variable information includes the traffic conditions, weather conditions and load conditions of each network point.

[0037] 104. Query the weight information corresponding to the traffic conditions, weather conditions, and load conditions, and calculate the score of each network point in the prepared destination network point information by weighted average according to the characteristic variable information and the weight information;

[0038] In this embodiment, a weight is assigned to each feature variable to indicate the importance of the feature in network point selection. For each prepared destination network point, its specific values ​​need to be calculated based on traffic conditions, weather conditions, and load conditions. For each network point, the comprehensive score of the network point is calculated based on the weight information of the above three features and their scores.

[0039] 105. Generate optimal destination network point information based on the network point with the highest score, and send the optimal destination network point information to the client;

[0040] In this embodiment, based on the comprehensive score of each outlet, all outlets are sorted from high to low according to the score, and the outlet with the highest score, that is, the first outlet after sorting, is selected as the best destination outlet. After selecting the outlet with the highest score, the next task is to generate detailed information of the outlet, including the outlet name, outlet address and outlet contact information.

[0041] In an embodiment of the present invention, real-time order data is input into an outlet matching model, multiple prepared destination outlet information is output, characteristic variable information is acquired in real time, scores of each outlet in the prepared destination outlet information are calculated by weighted average based on the characteristic variable information and weight information, optimal destination outlet information is generated based on the outlet with the highest score, the optimal destination outlet information is sent to the client, and variable factors are comprehensively considered to improve flexibility, accurately predict the optimal destination outlet, improve the accuracy of package delivery, prevent outlet overload or delivery delays, reduce dependence on human resources, reduce labor costs, improve the degree of automation, improve the efficiency of package sorting and matching, respond to orders in real time, and improve customer satisfaction.

[0042] See also Figure 2 The second embodiment of the method for intelligently matching destination network point information in the embodiment of the present invention includes:

[0043] 201. Collect historical order data, parse the historical order data, and obtain historical order parsing data;

[0044] In this embodiment, historical order data is collected from different data sources such as e-commerce platform data, logistics system data or customer feedback data. The historical order data includes basic order information, logistics information, user behavior data and environmental factors. Each field in the historical order data is parsed to obtain historical order parsed data.

[0045] 202. Remove duplicate data from historical order parsing data to obtain deduplicated historical order data;

[0046] In this embodiment, the order ID is a unique identifier for each order. If the same order ID appears multiple times in the historical order parsing data, it means that it is duplicate data. If all fields of two records are exactly the same, they are considered to be duplicates and can be directly removed to finally obtain the deduplicated data of historical orders.

[0047] 203. Correct the erroneous data in the historical order deduplication data using a regular expression to obtain corrected historical order data;

[0048] In this embodiment, the type of erroneous data is understood, the fields that need to be corrected are determined, and regular expressions are used to correct the erroneous data in the historical order deduplication data to obtain the corrected historical order data. After correcting the data using regular expressions, a quality check is performed on the corrected data to ensure that the correction does not introduce new errors and that all corrected records comply with business logic, for example, the quantity of goods should not be negative, the payment status should be consistent with the order amount, etc.

[0049] 204. Fill missing values ​​in the historical order correction data to obtain historical order filling data;

[0050] In this embodiment, it is clear which fields have missing values. For some fields with predefined default values ​​or regular values, fixed values ​​can be used to fill them. If the missing data is time-series, forward filling or backward filling methods can be used. For numeric fields, if the missing data is random and the specific value cannot be inferred, the mean, median or mode of the field can be used to fill it. If there is a certain relationship between the missing values ​​and other fields, the missing values ​​can be predicted by machine learning methods (such as regression analysis), or interpolation methods can be used to fill the missing values ​​based on other related fields. If the missing values ​​seriously affect the data quality and there is no effective way to fill them, you can choose to delete the records containing missing values ​​(that is, delete the entire row of data).

[0051] 205. Based on the preset target standard format, the historical order filling data is formatted and the historical order standard data is obtained;

[0052] In this embodiment, based on a preset target standard format, the historical order filling data is format converted, and then the historical order filling data is converted into a standard format to obtain the historical order standard data.

[0053] 206. Extract key feature sets from the historical order standard data, where the key feature sets include delivery outlets, delivery time, and customer preferences;

[0054] In this embodiment, a key feature set is extracted from the historical order standard data, and the key feature set includes delivery outlets, delivery time and customer preferences. If there are fields related to delivery outlets in the historical order standard data, such as "delivery address", "delivery outlet", "delivery area" and other fields are directly extracted, and the delivery time is directly calculated based on the "order creation time" and "order completion time". Customer preference refers to the personalized choice displayed by the customer when placing an order. The method of extracting customer preference usually depends on multiple data fields in the historical orders, and it may be necessary to combine the information in multiple orders for analysis. If there is a delivery method field in the order, the customer's preferred delivery method can be extracted. If there is a payment method field, the customer's payment preference can be extracted.

[0055] 207. Divide the key feature set according to a preset split ratio to obtain a training set, a validation set, and a test set;

[0056] In this embodiment, the key feature set is divided according to a preset split ratio to obtain a training set, a validation set and a test set. For example, the training set accounts for 70%, the validation set accounts for 15%, and the test set accounts for 15%.

[0057] 208. Select a basic model, use a random forest learning algorithm and train the basic model through a training set to obtain a network point matching model;

[0058] In this embodiment, a basic model is selected, a random forest classifier is imported, and the model is initialized. The training set is passed into the model for training. During the training process, the random forest learns the relationship between different features from the data and builds multiple decision trees based on these relationships. After the training is completed, a network point matching model is obtained.

[0059] 209. Use the validation set and test set to tune and evaluate the network point matching model respectively;

[0060] In this embodiment, after training is completed, the validation set is used to evaluate the performance of the current model. If the performance of the validation set is not as expected, the performance can be optimized by adjusting the hyperparameters of the model, and the tuned model is continuously evaluated on the validation set until the performance of the model is stable and meets the expected generalization ability. After the validation set is tuned, the best model is selected and predicted on the test set to evaluate the performance of the model on the test set, mainly to test whether the model has good generalization ability, and to evaluate multiple indicators, including precision, recall rate, F1 value, etc.

[0061] 210. Receive real-time order data sent by the client;

[0062] In this embodiment, the real-time order data sent by the client is received through the message queue, and the real-time order data is cleansed.

[0063] In the embodiment of the present invention, multiple data cleanings are performed to ensure the high accuracy of the historical order data used, avoid inaccurate or invalid model training due to data problems, and convert the data into a standardized format to meet the preset target standards, facilitate subsequent model processing and analysis, and improve the compatibility and maintainability of the system. By extracting key feature sets and focusing on important variables related to business goals, redundant information is reduced to make model training more targeted and effective. The data is reasonably divided into training sets, validation sets, and test sets to ensure the scientificity and accuracy of model evaluation and avoid overfitting problems. Through feedback from validation sets and test sets, the branch matching model is continuously optimized to improve the prediction accuracy and practicality of the model. By receiving real-time order data, it is ensured that the system can dynamically respond to customer needs, thereby improving the real-time and flexibility of the system.

[0064] See also Figure 3 The third embodiment of the method for intelligently matching destination network point information in the embodiment of the present invention includes:

[0065] 301. Input the real-time order data into the network point matching model, and output information of multiple prepared destination network points;

[0066] In this embodiment, features related to branch matching (such as delivery address, timeliness requirements, customer preferences, etc.) are extracted from real-time order data, and the extracted features are input into a trained branch matching model. The model will output a list of candidate branches based on the patterns in historical data. The list of candidate branches includes information on multiple prepared destination branches.

[0067] 302. Extracting order destination information from real-time order data;

[0068] In this embodiment, order destination information is extracted from real-time order data. The order destination information refers to the delivery address specified by the user, that is, the final location of the order delivery. The order data usually includes multiple fields, and the destination information is nested in these fields. By parsing the original order data structure, the destination information can be extracted from other irrelevant data.

[0069] 303. Identify the delivery range of the order destination information;

[0070] In this embodiment, the maximum delivery distance from the distribution center or warehouse to the destination is set. For example, assuming that the distribution center is located in the center of a city, the delivery range may be specified as within 10 kilometers of the city or within a specific area.

[0071] 304. Acquire characteristic variable information in real time according to the delivery range, the characteristic variable information including traffic conditions, weather conditions and load conditions of each outlet;

[0072] In this embodiment, real-time characteristic variable information within the delivery range is obtained, and the characteristic variable information includes the traffic conditions, weather conditions and load conditions of each outlet. Methods for obtaining traffic information include traffic data API, GPS monitoring, road sensors and cameras. Weather conditions can be obtained from weather data API, and load conditions can be obtained through the outlet system.

[0073] 305. Calling the preset weight configuration list;

[0074] In this embodiment, a preset weight configuration list is called. The weight configuration list refers to a pre-set parameter set, which usually includes a series of adjustable weight values. These weights are used to influence the decision-making of the system in different scenarios.

[0075] 306. Query the weight information corresponding to the traffic condition, weather condition and load condition from the weight configuration list;

[0076] In this embodiment, the weight configuration list is a list containing multiple parameters (such as traffic conditions, weather conditions, load conditions, etc.) and the weight values ​​corresponding to these parameters. The weight information corresponding to the traffic conditions, weather conditions and load conditions are queried from the weight configuration list. These weight values ​​are used to quantify the influence of each factor in the decision-making process.

[0077] 307. Calculate the score of each network point in the preparation destination network point information by weighted average according to the characteristic variable information and the weight information;

[0078] In this embodiment, characteristic variables are collected and sorted for different network points. Characteristic variables refer to specific factors considered when evaluating each network point. The characteristic variables of each network point are quantifiable indicators that reflect its performance or status. The score of each network point in the preparation destination network point information is calculated by weighted average based on the characteristic variable information and weight information. For example, for network point A, traffic conditions: characteristic value is 0.8, weight is 0.5, weather conditions: characteristic value is 0.7, weight is 0.3, load conditions: characteristic value is 0.6, weight is 0.2, then the score of network point A can be calculated as follows: Network point A score = (0.8×0.5)+(0.7×0.3)+(0.6×0.2)=0.4+0.21+0.12=0.73.

[0079] In the embodiment of the present invention, by inputting real-time order data into the network point matching model, the accuracy and diversity of the destination network point selection are ensured, and a variety of possible matching schemes are provided for subsequent optimization. The order destination information is extracted and combined with the real-time characteristics such as delivery range, traffic, weather and load, to provide comprehensive background information, so that the network point selection is more scientific and refined. By acquiring various characteristic variables in real time, it is possible to flexibly respond to environmental changes, such as traffic or weather emergencies, and effectively improve the delivery efficiency and timeliness. Through the weight configuration list and weighted average calculation, various factors can be optimized according to the actual situation to ensure that the final selected network point is balanced and optimized in multiple dimensions, thereby improving the accuracy and rationality of the decision. The preset weight configuration list is used so that the model can adjust the weight distribution according to different needs to adapt to different scenarios and changing conditions.

[0080] See also Figure 4 The fourth embodiment of the method for intelligently matching destination network point information in the embodiment of the present invention includes:

[0081] 401. Sort each outlet in the preliminary destination outlet information according to the score to obtain sorting information;

[0082] In this embodiment, the network points in the preliminary destination network point information are sorted according to the scores obtained by weighted average calculation to obtain the sorting information. The network points may be arranged in descending order to obtain the sorting information.

[0083] 402. Determine the best outlet based on the ranking information, and identify the outlet name, outlet address, outlet number, and outlet contact information corresponding to the best outlet;

[0084] In this embodiment, the best outlet is determined based on the ranking information, and the outlet name, outlet address, outlet number and outlet contact information of the best outlet are extracted by querying a database, a system interface or an existing data table.

[0085] 403. Generate optimal destination outlet information according to outlet name, outlet address, outlet number and outlet contact information;

[0086] In this embodiment, the outlet name is the identification name of the outlet, which is usually a unique identifier of the outlet and is used to distinguish different outlets. The outlet address is the specific geographical location of the outlet, indicating the street, district, and other address information of the outlet. The outlet number is a unique number identification for each outlet, which can help the system or personnel quickly identify the outlet. The outlet contact information is the telephone, fax or other contact information of the outlet, which is used to communicate or contact the outlet. The extracted outlet name, outlet address, outlet number and outlet contact are combined into a complete output to generate the best destination outlet information.

[0087] 404. Send the best destination network point information to the client;

[0088] In this embodiment, the client receiving method is determined. If the client interacts with the back-end system through an API, the best destination branch information can be returned through an HTTP request. For mobile terminals or clients based on push services, the branch information can be sent to the client through message push. If the client receives information through email or text messages, the branch information can be formatted into text and sent via email or text messages. If the client has a graphical user interface (GUI), the branch information can be directly displayed through the front-end page.

[0089] 405. Create a listener and use the listener to monitor the actual delivery point information;

[0090] In this embodiment, a listener is created and subscribed to or bound to a data source, database table, message queue, etc. to ensure that changes in branch information can be received. Once the branch information in the data source (such as a database table, message queue) changes, the listener will immediately capture the event.

[0091] 406. When the listener listens to the actual delivery point information, the current actual delivery point information is obtained;

[0092] In this embodiment, after the listener detects a change event, it will query the database for the latest data of the relevant outlets. The specific query method can be to obtain the complete information of a certain outlet based on the primary key (such as the outlet ID), or to obtain all outlet data that meet certain conditions through conditional query (such as all outlets in a certain area). If the system uses a caching mechanism (such as Redis, Memcached, etc.), the listener may first query the data in the cache. If the cache misses, it will query the data from the database. This can improve query efficiency and reduce the access pressure on the database. If the outlet data is stored in an external service (such as the system of a third-party logistics service provider), the listener may need to obtain the latest information of the current outlet through API calls.

[0093] 407. Determine whether the actual delivery outlet information is consistent with the optimal destination outlet information;

[0094] In this embodiment, it is determined whether the actual delivery point information is consistent with the optimal destination point information. Specifically, the actual delivery point information is compared with the various fields of the optimal destination point information, including but not limited to the following contents: checking whether the addresses of the two are consistent, checking whether the unique identifiers of the outlets (such as ID) are the same, and comparing whether the contact information (such as telephone, email, etc.) are consistent.

[0095] 408. If not, generate feedback information and upload the feedback information to the blockchain;

[0096] In this embodiment, if the actual delivery point information is inconsistent with the optimal destination point information, feedback information is generated. The main function of the feedback information is to record the inconsistent event and upload the feedback information to the blockchain.

[0097] In the embodiment of the present invention, by sorting the prepared outlets, it is possible to ensure that the outlets finally selected have the best performance in multiple dimensions, thereby improving the distribution efficiency and accuracy. The actual delivery outlets are monitored in real time through listeners. Once deviations occur, timely feedback and adjustments can be made to reduce the probability of manual intervention and errors. The feedback information is uploaded to the blockchain to ensure the transparency, immutability and traceability of the entire distribution process, which helps to improve the trust and security of the system. By comparing the actual delivery outlet information with the scheduled outlet information, deviations can be discovered and corrected in a timely manner, further optimizing the distribution process and system decisions, and improving overall efficiency.

[0098] The above describes the intelligent matching method of the destination network point information in the embodiment of the present invention. The following describes the intelligent matching device of the destination network point information in the embodiment of the present invention. Figure 5 In one embodiment of the present invention, an intelligent matching device for destination network point information includes:

[0099] The collection and extraction module 501 is used to collect historical order data and extract key feature sets from the historical order data, where the key feature sets include delivery outlets, delivery time and customer preferences;

[0100] A receiving module 502 is constructed to construct a network point matching model according to the key feature set and receive real-time order data sent by the client;

[0101] An input acquisition module 503 is used to input the real-time order data into the network point matching model, output a plurality of preliminary destination network point information, and acquire characteristic variable information in real time, wherein the characteristic variable information includes the traffic conditions, weather conditions, and load conditions of each network point;

[0102] A query calculation module 504 is used to query the weight information corresponding to the traffic conditions, weather conditions and load conditions, and calculate the score of each network point in the prepared destination network point information by weighted average according to the characteristic variable information and the weight information;

[0103] The generating and sending module 505 is used to generate the best destination network point information based on the network point with the highest score, and send the best destination network point information to the client.

[0104] In this embodiment, real-time order data is input into the branch matching model, and multiple prepared destination branch information is output, characteristic variable information is obtained in real time, and the score of each branch in the prepared destination branch information is calculated by weighted average based on the characteristic variable information and weight information. The best destination branch information is generated based on the branch with the highest score, and the best destination branch information is sent to the client. The changing factors are comprehensively considered to improve flexibility, accurately predict the best destination branch, improve the accuracy of package delivery, prevent branch overload or delivery delays, reduce dependence on human resources, reduce labor costs, increase the degree of automation, improve the efficiency of package sorting and matching, respond to orders in real time, and improve customer satisfaction.

[0105] See also Figure 6 Another embodiment of the device for intelligently matching destination network point information in the embodiment of the present invention includes:

[0106] The collection and extraction module 501 is used to collect historical order data and extract key feature sets from the historical order data, where the key feature sets include delivery outlets, delivery time and customer preferences;

[0107] A receiving module 502 is constructed to construct a network point matching model according to the key feature set and receive real-time order data sent by the client;

[0108] An input acquisition module 503 is used to input the real-time order data into the network point matching model, output a plurality of preliminary destination network point information, and acquire characteristic variable information in real time, wherein the characteristic variable information includes the traffic conditions, weather conditions, and load conditions of each network point;

[0109] A query calculation module 504 is used to query the weight information corresponding to the traffic conditions, weather conditions and load conditions, and calculate the score of each network point in the prepared destination network point information by weighted average according to the characteristic variable information and the weight information;

[0110] A generating and sending module 505 is used to generate the best destination network point information based on the network point with the highest score, and send the best destination network point information to the client;

[0111] In this embodiment, the collection and extraction module 501 includes: a collection and parsing unit 5011, which is used to collect historical order data, parse the historical order data, and obtain historical order parsed data; a removal unit 5012, which is used to remove duplicate data in the historical order parsed data, and obtain historical order deduplication data; a correction unit 5013, which is used to correct erroneous data in the historical order deduplication data using a regular expression, and obtain historical order correction data; a filling unit 5014, which is used to fill missing values ​​in the historical order correction data, and obtain historical order filled data; a conversion unit 5015, which is used to convert the format of the historical order filled data based on a preset target standard format, and obtain historical order standard data; a first extraction unit 5016, which is used to extract a key feature set in the historical order standard data, and the key feature set includes delivery outlets, delivery time and customer preferences.

[0112] In this embodiment, the receiving module 502 is constructed to include: a division unit 5021, which is used to divide the key feature set according to a preset split ratio to obtain a training set, a verification set and a test set; a training selection unit 5022, which is used to select a basic model, use a random forest learning algorithm and train the basic model through the training set to obtain a network matching model; a tuning and evaluation unit 5023, which is used to tune and evaluate the network matching model through the verification set and the test set respectively; a receiving unit 5024, which is used to receive real-time order data sent by the client.

[0113] In this embodiment, the input acquisition module 503 includes: an input unit 5031, which is used to input real-time order data into the network matching model, and output a plurality of prepared destination network information; a second extraction unit 5032, which is used to extract order destination information from the real-time order data; an identification unit 5033, which is used to identify the delivery range of the order destination information; an acquisition unit 5034, which is used to acquire characteristic variable information in real time according to the delivery range, and the characteristic variable information includes the traffic conditions, weather conditions and load conditions of each network.

[0114] In this embodiment, the query calculation module 504 includes: a calling unit 5041, used to call a preset weight configuration list; a query unit 5042, used to query the weight information corresponding to the traffic conditions, weather conditions and load conditions from the weight configuration list; a calculation unit 5043, used to calculate the score of each network point in the prepared destination network point information by weighted average based on the characteristic variable information and the weight information.

[0115] In this embodiment, the generation and sending module 505 includes: a sorting unit 5051, which is used to sort the outlets in the prepared destination outlet information according to the score to obtain the sorting information; a determination and identification unit 5052, which is used to determine the best outlet based on the sorting information, and identify the outlet name, outlet address, outlet number and outlet contact information corresponding to the best outlet; a generation unit 5053, which is used to generate the best destination outlet information according to the outlet name, outlet address, outlet number and outlet contact information; a sending unit 5054, which is used to send the best destination outlet information to the client.

[0116] In this embodiment, it also includes: creating a monitoring module 506, which is used to create a monitor and use the monitor to monitor the actual delivery point information; obtaining a module 507, which is used to obtain the current actual delivery point information when the monitor monitors the actual delivery point information; judging module 508, which is used to judge whether the actual delivery point information is consistent with the best destination point information; generating an upload module 509, which is used to generate feedback information if not, and upload the feedback information to the blockchain.

[0117] above Figure 5 and Figure 6 The destination network point information intelligent matching device in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The destination network point information intelligent matching device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0118] Figure 7 6 is a schematic diagram of the structure of a destination network point information intelligent matching device provided by an embodiment of the present invention. The destination network point information intelligent matching device 600 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 610 (for example, one or more processors) and a memory 620, and one or more storage media 630 (for example, one or more mass storage devices) storing application programs 633 or data 632. Among them, the memory 620 and the storage medium 630 can be temporary storage or permanent storage. The program stored in the storage medium 630 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the destination network point information intelligent matching device 600. Furthermore, the processor 610 can be configured to communicate with the storage medium 630, and execute a series of instruction operations in the storage medium 630 on the destination network point information intelligent matching device 600 to implement the steps of the destination network point information intelligent matching method provided in the above-mentioned method embodiments.

[0119] The destination network point information intelligent matching device 600 may also include one or more power supplies 640, one or more wired or wireless network interfaces 650, one or more input and output interfaces 660, and / or one or more operating systems 631, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will appreciate that Figure 7 The structure of the destination network point information intelligent matching device shown does not constitute a limitation on the destination network point information-based intelligent matching device, 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 the method for intelligently matching destination network information.

[0121] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system, device, or unit 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] Finally, it should be noted that the above description is only a preferred example of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for intelligently matching destination network point information, characterized in that: include: Collecting historical order data and extracting key feature sets from the historical order data, wherein the key feature sets include delivery outlets, delivery time, and customer preferences; Building a network point matching model according to the key feature set, and receiving real-time order data sent by the client; Input the real-time order data into the network point matching model, output a plurality of preliminary destination network point information, and obtain characteristic variable information in real time, wherein the characteristic variable information includes the traffic conditions, weather conditions and load conditions of each network point; querying the weight information corresponding to the traffic condition, the weather condition and the load condition respectively, and calculating the score of each network point in the prepared destination network point information by weighted average according to the characteristic variable information and the weight information; The best destination network point information is generated based on the network point with the highest score, and the best destination network point information is sent to the client.

2. The method for intelligently matching destination network point information according to claim 1, characterized in that: The collecting of historical order data and extracting key feature sets from the historical order data, wherein the key feature sets include delivery outlets, delivery time and customer preferences, include: Collecting historical order data, parsing the historical order data, and obtaining historical order parsed data; Removing duplicate data from the historical order parsed data to obtain deduplicated historical order data; Using a regular expression to correct erroneous data in the historical order deduplication data to obtain historical order corrected data; Filling missing values ​​in the historical order correction data to obtain historical order filling data; Based on a preset target standard format, the historical order filling data is formatted to obtain historical order standard data; A key feature set is extracted from the historical order standard data, wherein the key feature set includes a delivery point, a delivery time, and a customer preference.

3. The method for intelligently matching destination network point information according to claim 1, characterized in that: The step of building a network point matching model according to the key feature set and receiving real-time order data sent by a client includes: Dividing the key feature set according to a preset split ratio to obtain a training set, a validation set, and a test set; Select a basic model, and train the basic model using a random forest learning algorithm and the training set to obtain a network point matching model; Using the validation set and the test set to respectively tune and evaluate the dot matching model; Receive real-time order data sent by the client.

4. The method for intelligently matching destination network point information according to claim 1, characterized in that: The real-time order data is input into the network point matching model, and multiple preliminary destination network point information is outputted to obtain feature variable information in real time. The feature variable information includes the traffic conditions, weather conditions and load conditions of each network point, including: Inputting the real-time order data into the network point matching model, and outputting information of multiple prepared destination network points; Extracting order destination information from the real-time order data; Identify the delivery range of the order destination information; Characteristic variable information is acquired in real time according to the delivery range, and the characteristic variable information includes traffic conditions, weather conditions and load conditions of each network point.

5. The method for intelligently matching destination network point information according to claim 1, characterized in that: The querying of the weight information corresponding to the traffic condition, the weather condition, and the load condition, and calculating the score of each network point in the prepared destination network point information by weighted average according to the characteristic variable information and the weight information, includes: Call the preset weight configuration list; Query the weight information corresponding to the traffic condition, the weather condition and the load condition respectively from the weight configuration list; The score of each network point in the preparation destination network point information is calculated by weighted average according to the characteristic variable information and the weight information.

6. The method for intelligently matching destination network point information according to claim 1, characterized in that: The generating the best destination network point information based on the network point with the highest score, and sending the best destination network point information to the client, comprises: Sort each network point in the preliminary destination network point information according to the score to obtain sorting information; Determine the best outlet based on the ranking information, and identify the outlet name, outlet address, outlet number and outlet contact information corresponding to the best outlet; Generate optimal destination network point information according to the network point name, the network point address, the network point number and the network point contact information; The optimal destination network point information is sent to the client.

7. The method for intelligently matching destination network point information according to claim 1, characterized in that: After generating the best destination network point information based on the network point with the highest score and sending the best destination network point information to the client, the method further includes: Create a listener, and use the listener to monitor the actual delivery point information; When the listener listens to the actual delivery network point information, the current actual delivery network point information is obtained; Determining whether the actual delivery point information is consistent with the optimal destination point information; If not, feedback information is generated and uploaded to the blockchain.

8. A device for intelligently matching destination network point information, characterized in that: include: A collection and extraction module, used to collect historical order data and extract key feature sets from the historical order data, wherein the key feature sets include delivery outlets, delivery time and customer preferences; Constructing a receiving module, used to construct an outlet matching model according to the key feature set, and receiving real-time order data sent by the client; An input acquisition module, used to input the real-time order data into the network point matching model, output a plurality of preliminary destination network point information, and acquire characteristic variable information in real time, wherein the characteristic variable information includes the traffic conditions, weather conditions and load conditions of each network point; A query calculation module, used to query the weight information corresponding to the traffic condition, the weather condition and the load condition, and calculate the score of each network point in the prepared destination network point information by weighted average according to the characteristic variable information and the weight information; A generating and sending module is used to generate the best destination network point information based on the network point with the highest score, and send the best destination network point information to the client.

9. A device for intelligently matching target network point information, characterized in that: The target network point information intelligent matching device comprises: a memory and at least one processor, wherein the memory stores instructions; At least one of the processors calls the instructions in the memory to enable the destination network point information intelligent matching device to perform each step of the destination network point information intelligent matching method as described in any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the steps of the method for intelligently matching destination network point information as described in any one of claims 1 to 7 are implemented.

Images