A digital e-commerce logistics data comprehensive tracking management method and system

Abstract

The present invention proposes a digital e-commerce logistics data comprehensive tracking management method and system, wherein the method includes: capturing real-time order information from the e-commerce platform, predicting the sales quantity of the goods based on the real-time order information and historical order information, obtaining the standard deviation of the predicted demand, and obtaining the order quantity based on the standard deviation of the predicted demand and the predicted sales quantity of the goods; obtaining historical cargo transportation records, calculating the probability of damage of the goods arriving at the destination based on the historical cargo transportation records, obtaining the transportation combination with the minimum probability of damage of the goods arriving at the destination, and transporting the goods with this transportation combination; obtaining the real-time inventory and the probability of damage of the transportation combination with the minimum probability of damage, and sending an alarm message to the management personnel when the real-time inventory is lower than the preset threshold of the inventory and the probability of damage is higher than the preset threshold of the probability of damage. Through this method and the corresponding system, better inventory and cargo transportation risk management can be achieved.

Description

Technical Field

[0001] The present invention proposes a digital e-commerce logistics data comprehensive tracking management method and system, which relates to the technical field of logistics data management. Background Art

[0002] With the rapid development of the e-commerce industry, the market demand for the efficiency and accuracy of logistics and distribution is growing. Logistics is a crucial part of e-commerce operations and is directly related to consumer satisfaction and the brand image of the company. In this market environment, traditional logistics management methods have been unable to meet the current needs of e-commerce logistics due to slow information update speed, scattered data, and lack of real-time tracking and analysis capabilities. In addition, with the development of technologies such as big data, cloud computing, and the Internet of Things, the e-commerce logistics industry is facing tremendous pressure and opportunities for digital transformation.

[0003] Traditional logistics systems often adopt isolated data management, and data from different departments or links are difficult to share and uniformly analyze, affecting the efficiency and accuracy of decision-making; traditional systems rely more on experience and judgment, lack in-depth data analysis and prediction capabilities, and are difficult to accurately respond to market changes and dynamic adjustments in consumer demand; due to the lack of effective data integration and analysis tools, traditional logistics lacks response speed and flexibility when dealing with order peaks, emergencies, etc., or cannot comprehensively assess the risks encountered during the transportation of goods. Different cargo transportation combinations have a great impact on the probability of damage to the goods when they arrive at the destination, but traditional methods cannot formulate targeted preventive measures to reduce the probability of risks. Summary of the invention

[0004] The present invention provides a digital e-commerce logistics data comprehensive tracking management method and system to solve the above-mentioned problems:

[0005] The present invention proposes a digital e-commerce logistics data comprehensive tracking and management method, the method comprising:

[0006] Capture real-time order information from the e-commerce platform, predict the sales quantity of the product based on the real-time order information and historical order information, obtain the standard deviation of the predicted demand, and obtain the order quantity based on the standard deviation of the predicted demand and the predicted sales quantity of the product;

[0007] Obtain historical cargo transportation records, calculate the probability of damage to the goods upon arrival at the destination based on the historical cargo transportation records, obtain the transportation combination with the lowest probability of damage to the goods upon arrival at the destination, and use this transportation combination to transport the goods;

[0008] The real-time inventory and the damage probability of the transport combination with the minimum damage probability are obtained, and when the real-time inventory is lower than the preset inventory threshold and the damage probability is higher than the preset damage probability threshold, an alarm message is sent to the management personnel.

[0009] Furthermore, real-time order information is captured from the e-commerce platform, the sales quantity of the product is predicted according to the real-time order information and the historical order information, the standard deviation of the predicted demand is obtained, and the order quantity is obtained based on the standard deviation of the predicted demand and the predicted sales quantity of the product, including:

[0010] Preprocessing the real-time order information and the historical order information, wherein the preprocessing includes normalization processing, missing value processing, and time feature extraction;

[0011] After preprocessing, time series data is obtained, the input layer of the RNN receives the time series data, and the LSTM extracts features in the time series data and saves time-dependent features in the time series data;

[0012] The output layer of the RNN converts the output of the LSTM into a feature vector for use by the CNN;

[0013] The input layer of the CNN receives the feature vector from the LSTM, the convolution layer of the CNN identifies the local features to obtain a feature map, and the pooling layer of the CNN reduces the dimension of the feature map;

[0014] After dimensionality reduction, the output layer of the CNN generates the final sales volume forecast;

[0015] Get the actual sales quantity and predicted sales quantity at historical time points, and calculate the standard deviation of predicted demand:

[0016] Where N represents the number of historical time points, X i represents the difference between the actual sales quantity and the predicted sales quantity at time point i,

[0017] The order quantity is obtained through the economic order quantity model based on the standard deviation of the predicted demand and the predicted ideal inventory. Specifically, the economic order quantity model is:

[0018]

[0019] Where D represents the predicted daily sales volume, z represents the standard quantile of the normal distribution corresponding to the service level, represents the standard deviation of the forecast demand, represents the standard deviation of the supply cycle, L represents the average supply cycle, S represents the order cost, H represents the holding cost, that is, the storage cost of each unit of goods for a certain period of time, and d represents the length of the demand cycle.

[0020] Furthermore, historical cargo transportation records are obtained, and the probability of damage to the goods upon arrival at the destination is calculated based on the historical cargo transportation records, and a transportation combination with the lowest probability of damage to the goods upon arrival at the destination is obtained, and the goods are transported using this transportation combination, including:

[0021] Acquire a historical cargo transportation record data set, wherein the historical cargo transportation record data set includes different types of cargo combinations, risk damages, and cargo characteristics, wherein the risk damages are binary values, and the cargo characteristics include cargo fragility, cargo weight, and cargo transportation distance;

[0022] Based on the historical cargo records, the probability of damage to the goods upon arrival at the destination is calculated by a risk probability model. The risk probability model is fitted from the historical cargo transportation record data set and the model coefficients are obtained. Specifically, the risk probability model is:

[0023]

[0024] in, It represents the logarithmic probability of cargo being damaged when transported to the destination. It represents the basic probability of risk occurrence when all other influencing factors are zero. represents the contribution of random factors to risk that cannot be captured by the wind direction probability model, represents the random baseline risk caused by cargo grouping, is the m-order B-spline basis function, represents the spline coefficients, represents the characteristics of the kth commodity;

[0025] Obtain the transport combination with the lowest probability of damage to the goods when arriving at the destination, and use this transport combination to transport the goods.

[0026] Furthermore, fitting the risk probability model from the historical cargo transportation record data set and obtaining the model coefficients include:

[0027] Removing missing values ​​and outliers from the historical cargo transportation record data set, and normalizing the data in the data set after the removal;

[0028] The correlation function of glmm_model was input into the lme4 package of R software, and the correlation function used fragility, weight and distance as fixed effects and cargo grouping as a random effect;

[0029] Then, summary(glmm_model) was executed in the lme4 package of R software to extract detailed information of the risk probability model, including: the basic probability and the random baseline risk caused by the cargo grouping;

[0030] After the model is fitted, the random effects information is extracted from the fitted risk probability model object and the random baseline risk estimate for each cargo group is returned by ranef(glmm_model).

[0031] Furthermore, the real-time inventory and the damage probability of the transport combination with the minimum damage probability are obtained, and when the real-time inventory and the damage probability are respectively lower than corresponding preset thresholds, an alarm message is sent to the management personnel, including:

[0032] Obtain real-time inventory. When the real-time inventory is lower than the preset inventory threshold, an alarm message is sent to the management staff to remind them to place an order according to the order quantity calculated by the economic order quantity model.

[0033] The damage probability of the cargo transportation combination with the smallest damage probability is obtained. When the damage probability is higher than the preset threshold of the damage probability, an alarm message is sent to the alarm personnel to suspend the current combination of cargo transportation, wait for the next batch of cargo transportation, and recombine with the next batch of cargo until the damage probability of the cargo transportation combination is lower than the preset threshold of the damage probability, and then transport by truck.

[0034] The present invention proposes a digital e-commerce logistics data comprehensive tracking and management system, the system comprising:

[0035] An order quantity acquisition module is used to capture real-time order information from an e-commerce platform, predict the sales quantity of a product based on the real-time order information and historical order information, obtain the standard deviation of the predicted demand, and obtain the order quantity based on the standard deviation of the predicted demand and the predicted sales quantity of the product;

[0036] A damage probability calculation module is used to obtain historical cargo transportation records, calculate the damage probability of the goods arriving at the destination based on the historical cargo transportation records, obtain the transportation combination with the lowest damage probability of the goods arriving at the destination, and transport the goods using this transportation combination;

[0037] The alarm module is used to obtain the real-time inventory and the damage probability of the transport combination with the smallest damage probability, and send an alarm message to the management personnel when the real-time inventory is lower than the preset inventory threshold and the damage probability is higher than the preset damage probability threshold.

[0038] Furthermore, the module for obtaining order quantity includes:

[0039] A preprocessing module, used to preprocess the real-time order information and the historical order information, wherein the preprocessing includes normalization processing, missing value processing and time feature extraction;

[0040] A feature extraction module obtains time series data after preprocessing, the input layer of the RNN receives the time series data, and the LSTM extracts features from the time series data and saves the time-dependent features in the time series data;

[0041] A conversion module, used for the output layer of the RNN to convert the output of the LSTM into a feature vector for use by the CNN;

[0042] A dimension reduction module, wherein the input layer of the CNN receives a feature vector from the LSTM, the convolution layer of the CNN identifies local features to obtain a feature map, and the pooling layer of the CNN reduces the dimension of the feature map;

[0043] A prediction module, which is used to generate the final sales volume prediction from the output layer of the CNN after dimensionality reduction;

[0044] The standard deviation calculation module is used to obtain the actual sales quantity and predicted sales quantity at historical time points and calculate the standard deviation of the predicted demand:

[0045] Where N represents the number of historical time points, X i represents the difference between the actual sales quantity and the predicted sales quantity at time point i,

[0046] The module for calculating the economic order quantity is used to obtain the order quantity through the economic order quantity model based on the standard deviation of the predicted demand and the predicted ideal inventory. Specifically, the economic order quantity model is:

[0047]

[0048] Where D represents the predicted daily sales volume, z represents the standard quantile of the normal distribution corresponding to the service level, represents the standard deviation of the forecast demand, represents the standard deviation of the supply cycle, L represents the average supply cycle, S represents the order cost, H represents the holding cost, that is, the storage cost of each unit of goods for a certain period of time, and d represents the length of the demand cycle.

[0049] Furthermore, the damage probability calculation module includes:

[0050] A data set acquisition module is used to acquire a historical cargo transportation record data set, wherein the historical cargo transportation record data set includes different types of cargo combinations, risk damage and cargo characteristics, wherein the risk damage is a binary value, and the cargo characteristics include cargo fragility, cargo weight and cargo transportation distance;

[0051] The risk probability model fitting module is used to calculate the probability of damage of goods arriving at the destination through the risk probability model based on historical cargo records, fit the risk probability model from the historical cargo transportation record data set and obtain the model coefficient. Specifically, the risk probability model is:

[0052]

[0053] in, It represents the logarithmic probability of cargo being damaged when transported to the destination. It represents the basic probability of risk occurrence when all other influencing factors are zero. represents the contribution of random factors to risk that cannot be captured by the wind direction probability model, represents the random baseline risk caused by cargo grouping, is the m-order B-spline basis function, represents the spline coefficients, represents the characteristics of the kth commodity;

[0054] The module for obtaining the transport combination with the minimum probability is used to obtain the transport combination with the minimum probability of damage to the goods when they arrive at the destination, and use this transport combination to transport the goods.

[0055] Furthermore, the risk probability model fitting module includes:

[0056] A data processing module, used to remove missing values ​​and outliers in the historical cargo transportation record data set, and normalize the data in the data set after removal;

[0057] An input function module is used to input the correlation function of glmm_model in the lme4 package of R software, wherein the correlation function takes fragility, weight and distance as fixed effects and cargo group as random effect;

[0058] A running module is used to then execute summary(glmm_model) in the lme4 package of R software to extract detailed information of the risk probability model, including: basic probability and random baseline risk caused by cargo grouping;

[0059] The extraction module is used to extract the random effect information from the fitted risk probability model object after the model is fitted, and return the random baseline risk estimate for each cargo grouping through ranef(glmm_model).

[0060] Furthermore, the alarm module includes:

[0061] The order warning module is used to obtain real-time inventory. When the real-time inventory is lower than the preset inventory threshold, an alarm message is sent to the management personnel to remind them to place an order according to the order quantity calculated by the economic order quantity model;

[0062] The transportation warning module is used to obtain the damage probability of the cargo transportation combination with the lowest damage probability. When the damage probability is higher than the preset threshold of the damage probability, an alarm message is sent to the alarm personnel to suspend the current combination of cargo transportation, wait for the next batch of cargo transportation, and recombine with the next batch of cargo until the damage probability of the cargo transportation combination is lower than the preset threshold of the damage probability, so as to facilitate transportation by truck.

[0063] The beneficial effects of the present invention are as follows: improving forecast accuracy, by comprehensively using real-time and historical data, being able to more accurately forecast the sales demand for goods; optimizing inventory management, by more accurate demand forecasting and order quantity optimization, being able to reduce the problem of over- or under-inventory, reducing holding costs and out-of-stock risks; reducing transportation losses, by analyzing historical transportation data, selecting the transportation plan with the lowest probability of damage, reducing transportation losses and improving customer satisfaction; when inventory levels and transportation risks exceed the safety range, the system can automatically alert management, facilitating timely action to reduce potential losses; providing intelligent decision support based on data analysis to improve the efficiency of inventory and transportation management; by integrating multi-party data and modern statistical models, not only the efficiency of supply chain management is improved, but also the system's ability to respond to uncertainty is enhanced to a certain extent. This approach helps companies achieve better inventory and risk management in the fierce market competition. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 This is a schematic diagram of a digital e-commerce logistics data comprehensive tracking and management method described in the present invention. DETAILED DESCRIPTION

[0065] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0066] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. The embodiments described are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0068] One embodiment of the present invention provides a digital e-commerce logistics data comprehensive tracking and management method, the method comprising:

[0069] Capture real-time order information from the e-commerce platform, predict the sales quantity of the product based on the real-time order information and historical order information, obtain the standard deviation of the predicted demand, and obtain the order quantity based on the standard deviation of the predicted demand and the predicted sales quantity of the product;

[0070] Obtain historical cargo transportation records, calculate the probability of damage to the goods upon arrival at the destination based on the historical cargo transportation records, obtain the transportation combination with the lowest probability of damage to the goods upon arrival at the destination, and use this transportation combination to transport the goods;

[0071] The real-time inventory and the damage probability of the transport combination with the minimum damage probability are obtained, and when the real-time inventory is lower than the preset inventory threshold and the damage probability is higher than the preset damage probability threshold, an alarm message is sent to the management personnel.

[0072] The working principle and effect of the above technical solution are: to capture real-time order information from the e-commerce platform, which can reflect the current consumption and demand trends; to combine the real-time order information with the historical order data to predict the sales volume of the goods;

[0073] Calculate the standard deviation of forecasted demand to quantify forecast uncertainty; use the economic order quantity model to optimize order quantities by taking into account the changes and uncertainties in forecasted demand, which can ensure that inventory matches demand and reduce unnecessary inventory holding and ordering costs; extract data from historical cargo transportation records to analyze and calculate the probability of damage to the goods when they arrive at the destination; based on the analysis of the probability of damage, identify the transportation plan with the lowest probability of damage to the goods when they arrive at the destination, and use that plan to transport the goods. Monitor the current inventory level and the probability of damage to the transportation plan in real time. When the inventory is below the preset threshold and the probability of transportation damage is higher than the set safety value, the system will send an alarm message to the manager to prompt the need for intervention. Improve forecast accuracy. By combining real-time and historical data, it is possible to more accurately predict the sales demand for goods; optimize inventory management. Through more accurate demand forecasts and order quantity optimization, the problem of over- or under-inventory can be reduced, and the holding cost and out-of-stock risk can be reduced;

[0074] Reduce transportation losses. By analyzing historical transportation data, select the transportation plan with the lowest probability of damage, reduce transportation losses and improve customer satisfaction. When inventory levels and transportation risks exceed the safety range, the system can automatically alert management to facilitate timely action to reduce potential losses. Provide intelligent decision support based on data analysis to improve the efficiency of inventory and transportation management. By integrating multi-party data and modern statistical models, it not only improves the efficiency of supply chain management, but also enhances the system's ability to respond to uncertainty to a certain extent. This method helps companies achieve better inventory and risk management in the fierce market competition.

[0075] An embodiment of the present invention captures real-time order information from an e-commerce platform, predicts the sales quantity of a product based on the real-time order information and historical order information, obtains the standard deviation of the predicted demand, and obtains the order quantity based on the standard deviation of the predicted demand and the predicted sales quantity of the product, including:

[0076] Preprocessing the real-time order information and the historical order information, wherein the preprocessing includes normalization processing, missing value processing, and time feature extraction;

[0077] After preprocessing, time series data is obtained, the input layer of the RNN receives the time series data, and the LSTM extracts features in the time series data and saves time-dependent features in the time series data;

[0078] The output layer of the RNN converts the output of the LSTM into a feature vector for use by the CNN;

[0079] The input layer of the CNN receives the feature vector from the LSTM, the convolution layer of the CNN identifies the local features to obtain a feature map, and the pooling layer of the CNN reduces the dimension of the feature map;

[0080] After dimensionality reduction, the output layer of the CNN generates the final sales volume forecast;

[0081] Get the actual sales quantity and predicted sales quantity at historical time points, and calculate the standard deviation of predicted demand:

[0082] Where N represents the number of historical time points, X i represents the difference between the actual sales quantity and the predicted sales quantity at time point i,

[0083] The order quantity is obtained through the economic order quantity model based on the standard deviation of the predicted demand and the predicted ideal inventory. Specifically, the economic order quantity model is:

[0084]

[0085] Where D represents the predicted daily sales volume, z represents the standard quantile of the normal distribution corresponding to the service level, represents the standard deviation of the forecast demand, represents the standard deviation of the supply cycle, L represents the average supply cycle, S represents the order cost, H represents the holding cost, that is, the storage cost of each unit of goods for a certain period of time, and d represents the length of the demand cycle.

[0086] In one embodiment of the present invention, historical cargo transportation records are obtained, the probability of damage to the cargo upon arrival at the destination is calculated based on the historical cargo transportation records, a transportation combination with the lowest probability of damage to the cargo upon arrival at the destination is obtained, and the cargo is transported using the transportation combination, including:

[0087] Acquire a historical cargo transportation record data set, wherein the historical cargo transportation record data set includes different types of cargo combinations, risk damages, and cargo characteristics, wherein the risk damages are binary values, and the cargo characteristics include cargo fragility, cargo weight, and cargo transportation distance;

[0088] Based on the historical cargo records, the probability of damage to the goods upon arrival at the destination is calculated by a risk probability model. The risk probability model is fitted from the historical cargo transportation record data set and the model coefficients are obtained. Specifically, the risk probability model is:

[0089]

[0090] in, It represents the logarithmic probability of cargo being damaged when transported to the destination. It represents the basic probability of risk occurrence when all other influencing factors are zero. represents the contribution of random factors to risk that cannot be captured by the wind direction probability model, represents the random baseline risk caused by cargo grouping, is the m-order B-spline basis function, represents the spline coefficients, represents the characteristics of the kth commodity;

[0091] Obtain the transport combination with the lowest probability of damage to the goods when arriving at the destination, and use this transport combination to transport the goods.

[0092] In one embodiment of the present invention, fitting the risk probability model from a historical cargo transportation record data set and obtaining the model coefficients includes:

[0093] Removing missing values ​​and outliers from the historical cargo transportation record data set, and normalizing the data in the data set after the removal;

[0094] The correlation function of glmm_model was input into the lme4 package of R software, and the correlation function used fragility, weight and distance as fixed effects and cargo group as random effect;

[0095] Then, summary(glmm_model) was executed in the lme4 package of R software to extract detailed information of the risk probability model, including: the basic probability and the random baseline risk caused by the cargo grouping;

[0096] After the model is fitted, the random effects information is extracted from the fitted risk probability model object and the random baseline risk estimate for each cargo group is returned by ranef(glmm_model).

[0097] In one embodiment of the present invention, the real-time inventory and the damage probability of the transport combination with the smallest damage probability are obtained, and when the real-time inventory and the damage probability are respectively lower than corresponding preset thresholds, an alarm message is sent to a management personnel, including:

[0098] Obtain real-time inventory. When the real-time inventory is lower than the preset inventory threshold, an alarm message is sent to the management staff to remind them to place an order according to the order quantity calculated by the economic order quantity model.

[0099] The damage probability of the cargo transportation combination with the smallest damage probability is obtained. When the damage probability is higher than the preset threshold of the damage probability, an alarm message is sent to the alarm personnel to suspend the current combination of cargo transportation, wait for the next batch of cargo transportation, and recombine with the next batch of cargo until the damage probability of the cargo transportation combination is lower than the preset threshold of the damage probability, and then transport by truck.

[0100] One embodiment of the present invention is a digital e-commerce logistics data integrated tracking and management system, the system comprising:

[0101] An order quantity acquisition module is used to capture real-time order information from an e-commerce platform, predict the sales quantity of a product based on the real-time order information and historical order information, obtain the standard deviation of the predicted demand, and obtain the order quantity based on the standard deviation of the predicted demand and the predicted sales quantity of the product;

[0102] A damage probability calculation module is used to obtain historical cargo transportation records, calculate the damage probability of the goods arriving at the destination based on the historical cargo transportation records, obtain the transportation combination with the lowest damage probability of the goods arriving at the destination, and transport the goods using this transportation combination;

[0103] The alarm module is used to obtain the real-time inventory and the damage probability of the transport combination with the smallest damage probability, and send an alarm message to the management personnel when the real-time inventory is lower than the preset inventory threshold and the damage probability is higher than the preset damage probability threshold.

[0104] In one embodiment of the present invention, the module for obtaining order quantity includes:

[0105] A preprocessing module, used to preprocess the real-time order information and the historical order information, wherein the preprocessing includes normalization processing, missing value processing and time feature extraction;

[0106] A feature extraction module obtains time series data after preprocessing, the input layer of the RNN receives the time series data, and the LSTM extracts features from the time series data and saves the time-dependent features in the time series data;

[0107] A conversion module, used for the output layer of the RNN to convert the output of the LSTM into a feature vector for use by the CNN;

[0108] A dimension reduction module, wherein the input layer of the CNN receives a feature vector from the LSTM, the convolution layer of the CNN identifies local features to obtain a feature map, and the pooling layer of the CNN reduces the dimension of the feature map;

[0109] A prediction module, which is used to generate the final sales volume prediction from the output layer of the CNN after dimensionality reduction;

[0110] The standard deviation calculation module is used to obtain the actual sales quantity and predicted sales quantity at historical time points and calculate the standard deviation of the predicted demand:

[0111] Where N represents the number of historical time points, X i represents the difference between the actual sales quantity and the predicted sales quantity at time point i,

[0112] The module for calculating the economic order quantity is used to obtain the order quantity through the economic order quantity model based on the standard deviation of the predicted demand and the predicted ideal inventory. Specifically, the economic order quantity model is:

[0113]

[0114] Where D represents the predicted daily sales volume, z represents the standard quantile of the normal distribution corresponding to the service level, represents the standard deviation of the forecast demand, represents the standard deviation of the supply cycle, L represents the average supply cycle, S represents the order cost, H represents the holding cost, that is, the storage cost of each unit of goods for a certain period of time, and d represents the length of the demand cycle.

[0115] In one embodiment of the present invention, the damage probability calculation module includes:

[0116] A data set acquisition module is used to acquire a historical cargo transportation record data set, wherein the historical cargo transportation record data set includes different types of cargo combinations, risk damage and cargo characteristics, wherein the risk damage is a binary value, and the cargo characteristics include cargo fragility, cargo weight and cargo transportation distance;

[0117] The risk probability model fitting module is used to calculate the probability of damage of goods arriving at the destination through the risk probability model based on historical cargo records, fit the risk probability model from the historical cargo transportation record data set and obtain the model coefficient. Specifically, the risk probability model is:

[0118]

[0119] in, It represents the logarithmic probability of cargo being damaged when transported to the destination. It represents the basic probability of risk occurrence when all other influencing factors are zero. represents the contribution of random factors to risk that cannot be captured by the wind direction probability model, represents the random baseline risk caused by cargo grouping, is the m-order B-spline basis function, represents the spline coefficients, represents the characteristics of the kth commodity;

[0120] The module for obtaining the transport combination with the minimum probability is used to obtain the transport combination with the minimum probability of damage to the goods when they arrive at the destination, and use this transport combination to transport the goods.

[0121] In one embodiment of the present invention, the risk probability model fitting module includes:

[0122] A data processing module, used to remove missing values ​​and outliers in the historical cargo transportation record data set, and normalize the data in the data set after removal;

[0123] An input function module is used to input the correlation function of glmm_model in the lme4 package of R software, wherein the correlation function takes fragility, weight and distance as fixed effects and cargo group as random effect;

[0124] A running module is used to then execute summary(glmm_model) in the lme4 package of R software to extract detailed information of the risk probability model, including: the basic probability and the random baseline risk caused by the cargo grouping;

[0125] The extraction module is used to extract random effect information from the fitted risk probability model object after model fitting, and return the random baseline risk estimate for each cargo grouping through ranef(glmm_model).

[0126] In one embodiment of the present invention, the alarm module comprises:

[0127] The order warning module is used to obtain real-time inventory. When the real-time inventory is lower than the preset inventory threshold, an alarm message is sent to the management personnel to remind them to place an order according to the order quantity calculated by the economic order quantity model;

[0128] The transportation warning module is used to obtain the damage probability of the cargo transportation combination with the lowest damage probability. When the damage probability is higher than the preset threshold of the damage probability, an alarm message is sent to the alarm personnel to suspend the current combination of cargo transportation, wait for the next batch of cargo transportation, and recombine with the next batch of cargo until the damage probability of the cargo transportation combination is lower than the preset threshold of the damage probability, so as to facilitate transportation by truck.

[0129] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A digital e-commerce logistics data comprehensive tracking and management method, characterized in that: The method comprises: Capture real-time order information from the e-commerce platform, predict the sales quantity of the product based on the real-time order information and historical order information, obtain the standard deviation of the predicted demand, and obtain the order quantity based on the standard deviation of the predicted demand and the predicted sales quantity of the product; Obtain historical cargo transportation records, calculate the probability of damage to the goods upon arrival at the destination based on the historical cargo transportation records, obtain a transportation combination with the lowest probability of damage to the goods upon arrival at the destination, and transport the goods using this transportation combination, including; Acquire a historical cargo transportation record data set, where the historical cargo transportation record data set includes different types of cargo combinations, risk damage, and cargo characteristics, where the risk damage is a binary value, and the cargo characteristics include cargo fragility, cargo weight, and cargo transportation distance; Based on the historical cargo records, the probability of damage to the goods upon arrival at the destination is calculated by a risk probability model. The risk probability model is fitted from the historical cargo transportation record data set and the model coefficients are obtained. Specifically, the risk probability model is: in, represents the logarithmic probability of damage to the goods when they are transported to the destination, β0 represents the basic probability of risk occurrence when all other influencing factors are zero, u represents the contribution of random factors to the risk that cannot be captured by the risk probability model, and γ j represents the random baseline risk caused by cargo grouping, N m,p (x k ) is the m-order B-spline basis function, β m represents the spline coefficient, x k represents the characteristics of the kth commodity; Obtain the transport combination with the lowest probability of damage to the goods when they arrive at the destination, and use this transport combination to transport the goods; The real-time inventory and the damage probability of the transport combination with the minimum damage probability are obtained, and when the real-time inventory is lower than the preset inventory threshold and the damage probability is higher than the preset damage probability threshold, an alarm message is sent to the management personnel.

2. According to claim 1, a digital e-commerce logistics data comprehensive tracking and management method is characterized in that: Capture real-time order information from the e-commerce platform, predict the sales quantity of the product based on the real-time order information and historical order information, obtain the standard deviation of the predicted demand, and obtain the order quantity based on the standard deviation of the predicted demand and the predicted sales quantity of the product, including: Preprocessing the real-time order information and the historical order information, wherein the preprocessing includes normalization processing, missing value processing, and time feature extraction; After preprocessing, time series data is obtained, the input layer of the RNN receives the time series data, and the LSTM extracts features in the time series data and saves time-dependent features in the time series data; The output layer of the RNN converts the output of the LSTM into a feature vector for use by the CNN; The input layer of the CNN receives the feature vector from the LSTM, the convolution layer of the CNN identifies the local features to obtain a feature map, and the pooling layer of the CNN reduces the dimension of the feature map; After dimensionality reduction, the output layer of the CNN generates the final sales volume forecast; Get the actual sales quantity and predicted sales quantity at historical time points, and calculate the standard deviation of predicted demand: Where N represents the number of historical time points, X i represents the difference between the actual sales quantity and the predicted sales quantity at time point i, The order quantity is obtained through the economic order quantity model based on the standard deviation of the predicted demand and the predicted ideal inventory. Specifically, the economic order quantity model is: Where D is the predicted daily sales volume, z is the standard quantile of the normal distribution corresponding to the service level, σ is the standard deviation of the predicted demand, and σ L represents the standard deviation of the supply cycle, L represents the average supply cycle, S represents the order cost, H represents the holding cost, that is, the storage cost of each unit of goods for a certain period of time, and d represents the length of the demand cycle.

3. According to claim 1, a digital e-commerce logistics data comprehensive tracking and management method is characterized in that: Fitting the risk probability model from the historical cargo transportation record data set and obtaining the model coefficients include: Removing missing values ​​and outliers from the historical cargo transportation record data set, and normalizing the data in the data set after the removal; The correlation function of glmm_model was input into the lme4 package of R software, and the correlation function used fragility, weight and distance as fixed effects and cargo grouping as a random effect; Then, summary(glmm_model) was executed in the lme4 package of R software to extract detailed information of the risk probability model, including: the basic probability and the random baseline risk caused by the cargo grouping; After the model is fitted, the random effects information is extracted from the fitted risk probability model object and the random baseline risk estimate for each cargo group is returned by ranef(glmm_model).

4. According to claim 1, a digital e-commerce logistics data comprehensive tracking and management method is characterized in that: Obtaining the real-time inventory and the damage probability of the transport combination with the minimum damage probability, and sending an alarm message to the management personnel when the real-time inventory and the damage probability are respectively lower than the corresponding preset thresholds, including: Obtain real-time inventory. When the real-time inventory is lower than the preset inventory threshold, an alarm message is sent to the management staff to remind them to place an order according to the order quantity calculated by the economic order quantity model. The damage probability of the cargo transportation combination with the smallest damage probability is obtained. When the damage probability is higher than the preset threshold of the damage probability, an alarm message is sent to the alarm personnel to suspend the current combination of cargo transportation, wait for the next batch of cargo transportation, and recombine with the next batch of cargo until the damage probability of the cargo transportation combination is lower than the preset threshold of the damage probability, and then transport by truck.

5. A digital e-commerce logistics data integrated tracking and management system, characterized in that: The system comprises: An order quantity acquisition module is used to capture real-time order information from an e-commerce platform, predict the sales quantity of a product based on the real-time order information and historical order information, obtain the standard deviation of the predicted demand, and obtain the order quantity based on the standard deviation of the predicted demand and the predicted sales quantity of the product; The damage probability calculation module is used to obtain historical cargo transportation records, calculate the damage probability of the goods arriving at the destination based on the historical cargo transportation records, obtain the transportation combination with the lowest damage probability of the goods arriving at the destination, and transport the goods using this transportation combination, including: A data set acquisition module is used to acquire a historical cargo transportation record data set, wherein the historical cargo transportation record data set includes different types of cargo combinations, risk damage and cargo characteristics, wherein the risk damage is a binary value, and the cargo characteristics include cargo fragility, cargo weight and cargo transportation distance; The risk probability model fitting module is used to calculate the probability of damage of goods arriving at the destination through the risk probability model based on historical cargo records, fit the risk probability model from the historical cargo transportation record data set and obtain the model coefficient. Specifically, the risk probability model is: in, represents the logarithmic probability of damage to the goods when they are transported to the destination, β0 represents the basic probability of risk occurrence when all other influencing factors are zero, u represents the contribution of random factors to the risk that cannot be captured by the risk probability model, and γ j represents the random baseline risk caused by cargo grouping, N m,p (x k ) is the m-order B-spline basis function, β m represents the spline coefficient, x k represents the characteristics of the kth commodity; A module for obtaining the transport combination with the lowest probability is used to obtain the transport combination with the lowest probability of damage to the goods when they arrive at the destination, and use this transport combination to transport the goods; The alarm module is used to obtain the real-time inventory and the damage probability of the transport combination with the smallest damage probability, and send an alarm message to the management personnel when the real-time inventory is lower than the preset inventory threshold and the damage probability is higher than the preset damage probability threshold.

6. According to claim 5, a digital e-commerce logistics data integrated tracking and management system is characterized in that: The module for obtaining order quantity includes: A preprocessing module, used to preprocess the real-time order information and the historical order information, wherein the preprocessing includes normalization processing, missing value processing and time feature extraction; A feature extraction module obtains time series data after preprocessing, the input layer of the RNN receives the time series data, and the LSTM extracts features from the time series data and saves the time-dependent features in the time series data; A conversion module, used for the output layer of the RNN to convert the output of the LSTM into a feature vector for use by the CNN; A dimension reduction module, wherein the input layer of the CNN receives a feature vector from the LSTM, the convolution layer of the CNN identifies local features to obtain a feature map, and the pooling layer of the CNN reduces the dimension of the feature map; A prediction module, which is used to generate the final sales volume prediction from the output layer of the CNN after dimensionality reduction; The standard deviation calculation module is used to obtain the actual sales quantity and predicted sales quantity at historical time points and calculate the standard deviation of the predicted demand: Where N represents the number of historical time points, X i represents the difference between the actual sales quantity and the predicted sales quantity at time point i, The economic order quantity calculation module is used to obtain the order quantity through the economic order quantity model based on the standard deviation of the predicted demand and the predicted ideal inventory. Specifically, the economic order quantity model is: Where D is the predicted daily sales volume, z is the standard quantile of the normal distribution corresponding to the service level, σ is the standard deviation of the predicted demand, and σ L represents the standard deviation of the supply cycle, L represents the average supply cycle, S represents the order cost, H represents the holding cost, that is, the storage cost of each unit of goods for a certain period of time, and d represents the length of the demand cycle.

7. According to claim 5, a digital e-commerce logistics data integrated tracking and management system is characterized in that: The risk probability model fitting module includes: A data processing module, used to remove missing values ​​and outliers in the historical cargo transportation record data set, and normalize the data in the data set after removal; An input function module is used to input the correlation function of glmm_model in the lme4 package of R software, wherein the correlation function takes fragility, weight and distance as fixed effects and cargo group as random effect; A running module is used to then execute summary(glmm_model) in the lme4 package of R software to extract detailed information of the risk probability model, including: basic probability and random baseline risk caused by cargo grouping; The extraction module is used to extract the random effect information from the fitted risk probability model object after the model is fitted, and return the random baseline risk estimate for each cargo grouping through ranef(glmm_model).

8. According to claim 5, a digital e-commerce logistics data integrated tracking and management system is characterized in that: The alarm module comprises: The order warning module is used to obtain real-time inventory. When the real-time inventory is lower than the preset inventory threshold, an alarm message is sent to the management personnel to remind them to place an order according to the order quantity calculated by the economic order quantity model; The transportation warning module is used to obtain the damage probability of the cargo transportation combination with the lowest damage probability. When the damage probability is higher than the preset threshold of the damage probability, an alarm message is sent to the alarm personnel to suspend the current combination of cargo transportation, wait for the next batch of cargo transportation, and recombine with the next batch of cargo until the damage probability of the cargo transportation combination is lower than the preset threshold of the damage probability, so as to facilitate transportation by truck.

Images