Full-product data optimization method of logistics system, electronic equipment and storage medium

Abstract

The invention relates to the technical field of data processing, in particular to a full-product data optimization method of a logistics system, electronic equipment and a storage medium, and the method comprises the steps: receiving a product logistics request information set sent by each product service platform in a preset time period, analyzing each piece of product logistics request information, and sending the analyzed product logistics request information to a server; obtaining a plurality of product logistics features corresponding to each piece of product logistics request information, inputting the product logistics feature set of the product and the optimized target feature weight corresponding to each product logistics feature into a trained given classification model, and distributing a corresponding target logistics instruction for the product; data optimization is performed on the feature weight of the logistics feature of each product, so that the logistics instruction allocated to each product is more reasonable, and the logistics sending efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a full-product data optimization method, electronic equipment and storage medium for a logistics system. Background Art

[0002] At present, users' expectations for logistics services are no longer limited to the safe delivery of goods, but they hope to receive goods faster. However, due to the large number of product service platforms, such as a large number of shopping apps and shopping mini-programs, the total number of orders will be large. When users place orders on different product service platforms, due to the different properties of different product service platforms, for example, the order is placed early but the delivery time to the logistics system is late due to the large sending interval. When the logistics system integrates the received order information, it will lead to unreasonable logistics distribution and low logistics efficiency, affecting user satisfaction. Summary of the invention

[0003] In response to the above technical problems, the present invention provides a full-product data optimization method, electronic device and storage medium for a logistics system. By optimizing the feature weights of each product's logistics features, the logistics instructions assigned to each product are more reasonable, thereby improving the logistics delivery efficiency.

[0004] According to a first aspect of the present invention, a method for optimizing full product data of a logistics system is provided, comprising the following steps: S100, receiving a product logistics request information set sent by each product service platform within a preset time period , where A i is the i-th product logistics request information, i=1, 2, …, m, and m is the number of product logistics requests sent by any product service platform.

[0005] S200, for A i Analyze and obtain A i Corresponding product logistics feature set ,in, A i The corresponding j-th product logistics characteristics, j = 1, 2, ..., n, n is A i The number of corresponding product logistics characteristics.

[0006] S300, A i The corresponding product logistics feature set B i and each The corresponding optimized target feature weights are input into the trained given classification model, and the output result of the given classification model is B i The corresponding products are assigned corresponding target logistics instructions to achieve A according to the target logistics instructions. iThe logistics information of the corresponding products is updated.

[0007] According to a second aspect of the present invention, a non-transitory computer-readable storage medium is provided, in which at least one instruction or at least one program is stored, and the at least one instruction or the at least one program is loaded and executed by a processor to implement the above-mentioned full-product data optimization method of the logistics system.

[0008] According to a third aspect of the present invention, there is provided an electronic device comprising a processor and the above-mentioned non-transitory computer-readable storage medium.

[0009] The present invention has at least the following beneficial effects: The present invention provides a method for optimizing the data of all products in a logistics system. The method receives a set of product logistics request information sent by each product service platform within a preset time period, parses each product logistics request information, obtains a number of product logistics features corresponding to each product logistics request information, inputs the product logistics feature set of the product and the optimized target feature weight corresponding to each product logistics feature into a trained given classification model, allocates corresponding target logistics instructions to the product, and optimizes the feature weight of each product logistics feature to make the logistics instructions allocated to each product more reasonable, thereby improving the logistics sending efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 A flowchart of a method for optimizing all product data of a logistics system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0012] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0013] The embodiment of the present invention provides a method for optimizing all product data of a logistics system. Figure 1 As shown, the method comprises the following steps: S100, receiving a product logistics request information set sent by each product service platform within a preset time period , where A i is the i-th product logistics request information, i=1, 2, ..., m, and m is the number of product logistics requests sent by any product service platform. In a specific implementation, different product service platforms send different numbers of product logistics request information within a preset time period, which is related to the actual number of messages sent within the preset time period.

[0014] Specifically, the preset time period is any time period divided according to the preset time period. For example, the duration of the preset time period may be 10 seconds.

[0015] Furthermore, the product service platform refers to a platform that provides users with products to choose from, such as a shopping app or a shopping mini-program.

[0016] S200, for A i Analyze and obtain A i Corresponding product logistics feature set ,in, A i The corresponding j-th product logistics characteristics, j = 1, 2, ..., n, n is A i The number of corresponding product logistics characteristics.

[0017] Furthermore, the product logistics feature is any one of the order time, delivery time limit, shipping address and delivery address corresponding to the product. In a specific implementation, other product logistics features may also be added, such as whether the shipping is free.

[0018] As mentioned above, when obtaining the logistics distribution situation corresponding to each order product, it is necessary to take into account several product logistics characteristics corresponding to the order, and each product logistics characteristic has a different degree of influence on the logistics distribution situation. Therefore, it is necessary to first parse and obtain several logistics characteristics, and then realize the subsequent logistics instruction generation based on the obtained logistics characteristics.

[0019] S300, A i The corresponding product logistics feature set B i and each The corresponding optimized target feature weights are input into the trained given classification model, and the output result of the given classification model is B i The corresponding products are assigned corresponding target logistics instructions to achieve A according to the target logistics instructions. i The logistics information of the corresponding products is updated; it can be understood that: the target logistics instruction refers to the instruction to assign any product to any logistics vehicle, and the corresponding characteristics of the logistics vehicle include the departure address, departure time, and vehicle size.

[0020] Furthermore, the given classification model is a trained target random forest model.

[0021] Specifically, the given classification model is trained through the following steps: S10, obtaining a historical product logistics feature set corresponding to several historical products and a preset logistics instruction corresponding to each historical product, and constructing a sample data set. The preset logistics instruction corresponding to the historical product is a logistics instruction pre-set according to the actual situation, so that the historical product has a higher logistics delivery efficiency.

[0022] S20, for any decision tree in the preset random forest model, a number of samples are randomly extracted from the sample data set with replacement as a sub-training set of the decision tree.

[0023] S30, inputting several historical product logistics feature sets in each sub-training set into the corresponding decision tree in the preset random forest model to obtain the logistics instruction output result of each decision tree.

[0024] S40, adjusting the parameters of the decision tree according to the preset logistics instructions in each sub-training set and the logistics instruction output results of each decision tree, to obtain a trained given classification model. Among them, the parameter adjustment process of the random forest model during the training process is a prior art, or those skilled in the art adjust the parameters of the model according to actual needs, and the specific process is not repeated here.

[0025] Preferably, the following steps are performed to obtain The corresponding target feature weights are: S310, obtain Corresponding to several target impact indicators; the target impact indicators are pre-acquired impact The corresponding target feature weight indicator.

[0026] Further, the following steps are performed to obtain Corresponding target impact indicators: S311, obtain Corresponding to several preset impact indicators.

[0027] Specifically, the preset impact indicator is a preset impact The corresponding target feature weight indicator.

[0028] Furthermore, the preset influencing indicators include but are not limited to indicators such as the API interface stability of the product service platform, network delay duration, number of orders placed within a preset time period, decompression time of the decompressed package, and address resolution accuracy.

[0029] S312, preprocessing and normalizing the collected historical data corresponding to each preset impact indicator to obtain a number of target data corresponding to each preset impact indicator.

[0030] Specifically, preprocessing the historical data corresponding to each preset influencing indicator includes cleaning and one-hot encoding operations on the historical data.

[0031] S313, according to the target data corresponding to each preset influencing indicator, calculate the Pearson correlation coefficient between each preset influencing indicator and the corresponding logistics instruction data. The logistics instruction data may be a vector combination obtained by one-hot encoding a plurality of logistics information in the logistics instruction. For example, the logistics information may be the departure time of the train, the size of the vehicle, etc. Those skilled in the art know the calculation formula and calculation process of the Pearson correlation coefficient, which will not be described in detail here.

[0032] S314, each preset influence indicator whose corresponding Pearson correlation coefficient is greater than a preset Pearson correlation coefficient threshold is selected and used as a target influence indicator. Those skilled in the art can set the preset Pearson correlation coefficient threshold according to actual needs, for example, 0.6.

[0033] As mentioned above, by calculating the Pearson correlation coefficient between each preset influencing indicator and the corresponding logistics instruction data, the preset influencing indicators with high correlation with the logistics instruction data and greater influence on the logistics instruction data can be screened out, and by inputting the data corresponding to the screened preset influencing indicators into the classification model, the allocation results of the logistics instructions are made more accurate and reasonable.

[0034] S320, dimensionless processing is performed on the data corresponding to each target impact indicator, and the indicator weight corresponding to each target impact indicator is obtained based on the processed data; it can be understood that: before the dimensionless processing, different types of data are uniformly converted into the same type of data, for example, they are all converted into extremely large data or extremely small data. The data conversion process is a prior art and will not be described in detail.

[0035] Specifically, step S320 includes the following steps: S321, dimensionless processing is performed on the data corresponding to each target impact indicator to obtain Corresponding dimensionless indicator dataset ,in, for The corresponding dimensionless data of the target impact indicator, , f is The number of corresponding target impact indicators.

[0036] Furthermore, the initialization method or the averaging method can be used to dimensionlessly process the data corresponding to each target impact indicator to eliminate the differences in dimensions and orders of magnitude of different indicators. Technical personnel in this field are aware of the specific process of using the initialization method and the averaging method for dimensionless processing and will not be described in detail.

[0037] S322, calculate each and The difference between the corresponding preset reference values ​​can be understood as: The corresponding preset reference value is The corresponding The preset optimal value of each target impact indicator.

[0038] S323, according to the maximum difference and the minimum difference obtained from the plurality of differences, and each and The difference between the corresponding preset reference values ​​is calculated to calculate the impact index of each target and The correlation coefficient of the corresponding target feature weight; it can be understood that: the several differences refer to the differences corresponding to all dimensionless data, that is, there are a total of f×n differences.

[0039] Specifically, the target impact indicators and The correlation coefficient of the corresponding target feature weight meets the following conditions: ,in, express The corresponding Target impact indicators and The corresponding correlation coefficient of the target feature weight, d is the minimum difference, D is the maximum difference, for and The difference between the corresponding preset reference values, ρ is the preset resolution coefficient, where ρ∈[0,1]. For example, the value of ρ is 0.5.

[0040] S324, according to each target impact indicator and The correlation coefficient of the corresponding target feature weight is used to obtain the indicator weight corresponding to each target impact indicator.

[0041] Specifically, the indicator weights corresponding to the target impact indicators meet the following conditions: ,in, express The corresponding The indicator weight of each target impact indicator, express The corresponding Target impact indicators and The corresponding correlation coefficient of the target feature weight, A i The corresponding g-th product logistics characteristic, the value of g ranges from 1 to n.

[0042] In the above, considering that different product service platforms have different attributes, their corresponding feature weights are also different. The attributes include several target influence indicators. Different target influence indicators have different degrees of influence on the feature weights. The correlation coefficient between each target influence indicator and the corresponding target feature weight is calculated through the above process, and the indicator weight corresponding to the target influence indicator is obtained according to the correlation coefficient, which makes the setting of the indicator weight more accurate and reliable.

[0043] S330, calculating a weighted sum based on the values ​​of the plurality of target impact indicators and the indicator weight of each target impact indicator to obtain The corresponding target impact factor.

[0044] S340, the target impact factor and The product of the corresponding initial feature weights is The corresponding target feature weight can be understood as: The corresponding initial feature weight is a pre-set weight. In a specific implementation, several target feature weights may also be normalized to between 0 and 1.

[0045] As mentioned above, since the feature weight of each product logistics feature is affected by multiple factors, by analyzing multiple influencing factors and determining the degree of influence of each influencing factor on the product logistics feature, the feature weight of the product logistics feature is made more accurate and reliable, thereby realizing the optimization of product data. Through the optimized data, the logistics instructions assigned by the classification model to each product are more reasonable, thereby improving the logistics delivery efficiency.

[0046] In summary, an embodiment of the present invention provides a method for optimizing the data of all products in a logistics system, which receives a set of product logistics request information sent by each product service platform within a preset time period, parses each product logistics request information, obtains a number of product logistics features corresponding to each product logistics request information, inputs the product logistics feature set of the product and the optimized target feature weight corresponding to each product logistics feature into a trained given classification model, and assigns corresponding target logistics instructions to the product. By performing data optimization on the feature weight of each product logistics feature, the logistics instructions assigned to each product are made more reasonable, thereby improving the logistics delivery efficiency.

[0047] An embodiment of the present invention also provides a non-transitory computer-readable storage medium, which can be set in an electronic device to store at least one instruction or at least one program related to implementing a method in a method embodiment. The at least one instruction or the at least one program is loaded and executed by the processor to implement the method provided in the above embodiment.

[0048] An embodiment of the present invention further provides an electronic device, comprising a processor and the aforementioned non-transitory computer-readable storage medium.

[0049] Although some specific embodiments of the present invention have been described in detail by way of example, it will be appreciated by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It will also be appreciated by those skilled in the art that various modifications may be made to the embodiments without departing from the scope and spirit of the present invention. The scope of the present invention is defined by the appended claims.

Claims

1. A method for optimizing all product data of a logistics system, characterized in that: The method comprises the following steps: S100, receiving a product logistics request information set sent by each product service platform within a preset time period , where A i is the i-th product logistics request information, i=1, 2, ..., m, m is the number of product logistics requests sent by any product service platform; S200, for A i Analyze and obtain A i Corresponding product logistics feature set ,in, A i The corresponding j-th product logistics characteristics, j = 1, 2, ..., n, n is A i the number of corresponding product logistics characteristics; S300, A i The corresponding product logistics feature set B i and each The corresponding optimized target feature weights are input into the trained given classification model, and the output result of the given classification model is B i The corresponding products are assigned corresponding target logistics instructions to achieve A according to the target logistics instructions. i The logistics information of the corresponding products is updated.

2. The method for optimizing the full product data of a logistics system according to claim 1, characterized in that: The given classification model is trained through the following steps: S10, obtaining a historical product logistics feature set corresponding to a number of historical products and a preset logistics instruction corresponding to each historical product, and constructing a sample data set; S20, for any decision tree in the preset random forest model, randomly extracting a number of samples from the sample data set with replacement as a sub-training set for the decision tree; S30, inputting a number of historical product logistics feature sets in each sub-training set into a corresponding decision tree in a preset random forest model to obtain a logistics instruction output result of each decision tree; S40, adjusting the parameters of the decision tree according to the preset logistics instructions in each sub-training set and the logistics instruction output results of each decision tree to obtain a trained given classification model.

3. The method for optimizing the full product data of a logistics system according to claim 1, characterized in that: Obtain it by following the steps below The corresponding target feature weights are: S310, obtain Corresponding to several target impact indicators; the target impact indicators are pre-acquired impact The corresponding target feature weight indicator; S320, performing dimensionless processing on the data corresponding to each target impact indicator, and obtaining the indicator weight corresponding to each target impact indicator according to the processed data; S330, calculating a weighted sum based on the values ​​of the plurality of target impact indicators and the indicator weight of each target impact indicator to obtain The corresponding target impact factor; S340, the target impact factor and The product of the corresponding initial feature weights is The corresponding target feature weight.

4. The method for optimizing the full product data of a logistics system according to claim 3, characterized in that: Obtained through the following steps Corresponding target impact indicators: S311, obtain Corresponding several preset impact indicators; S312, preprocessing and normalizing the collected historical data corresponding to each preset impact indicator to obtain a number of target data corresponding to each preset impact indicator; S313, calculating the Pearson correlation coefficient between each preset influencing indicator and the corresponding logistics instruction data according to a number of target data corresponding to each preset influencing indicator; S314, each preset influence indicator whose corresponding Pearson correlation coefficient is greater than a preset Pearson correlation coefficient threshold is selected and used as a target influence indicator.

5. The method for optimizing the full product data of a logistics system according to claim 3, characterized in that: Step S320 includes the following steps: S321, dimensionless processing is performed on the data corresponding to each target impact indicator to obtain Corresponding dimensionless indicator dataset ,in, for The corresponding dimensionless data of the target impact indicator, , f is The number of corresponding target impact indicators; S322, calculate each and The difference between the corresponding preset reference values; S323, according to the maximum difference and the minimum difference obtained from the plurality of differences, and each and The difference between the corresponding preset reference values ​​is calculated to calculate the impact index of each target and The correlation coefficient of the corresponding target feature weight; Among them, the target impact index and The correlation coefficient of the corresponding target feature weight meets the following conditions: ,in, express The corresponding Target impact indicators and The corresponding correlation coefficient of the target feature weight, d is the minimum difference, D is the maximum difference, for and The difference between the corresponding preset reference values, ρ is the preset resolution coefficient; S324, according to each target impact indicator and The correlation coefficient of the corresponding target feature weight is used to obtain the indicator weight corresponding to each target influencing indicator; Among them, the indicator weights corresponding to the target impact indicators meet the following conditions: ,in, express The corresponding The indicator weight of each target impact indicator, express The corresponding Target impact indicators and The corresponding correlation coefficient of the target feature weight, A i The corresponding g-th product logistics characteristic, the value of g ranges from 1 to n.

6. The method for optimizing the full product data of a logistics system according to claim 4, characterized in that: The preset influencing indicators include the API interface stability of the product service platform, network delay duration, number of orders placed within a preset time period, decompression time of the decompressed package, and address resolution accuracy.

7. A non-transitory computer-readable storage medium, wherein at least one instruction or at least one program is stored in the storage medium, characterized in that: The at least one instruction or the at least one program segment is loaded and executed by the processor to implement the full product data optimization method for the logistics system as described in any one of claims 1-6.

8. An electronic device, characterized in that: The invention comprises a processor and the non-transitory computer-readable storage medium as claimed in claim 7.

Images