Logistics industry customer identification method and device

Abstract

The invention relates to the technical field of data analysis and processing based on IAO and LSTM algorithms, and discloses a logistics industry customer identification method. Comprises: obtaining customer data; establishing a customer long-term business volume prediction model through an LSTM (Long Short-Term Memory) network; generating an IAO algorithm based on the AO algorithm; using an IAO algorithm to optimize the customer long-term business volume prediction model, and generating an IAO-LSTM-based customer long-term business volume prediction model; a prediction result is generated through an IAO-LSTM-based customer long-term business volume prediction model, information integration is carried out, a comprehensive score is generated through an analytic hierarchy process, and waist customer information is generated; sending the waist customer information to the collaboration end, receiving feedback information, adjusting an analytic hierarchy algorithm and the weight of the customer long-term business volume prediction model based on the IAO-LSTM, and optimizing the waist customer information; the method has the advantages of improving the data screening efficiency and improving the data screening accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis and processing based on IAO and LSTM algorithms, and in particular to a method and device for identifying customers in the logistics industry. Background Art

[0002] Mid-tier customers (i.e., customers with an average daily shipment volume of 50-1,000 orders) have achieved a good balance between profit and stability, and are an important source of profit growth for express delivery companies. With the vigorous development of e-commerce and the rise of small and medium-sized enterprises, the mid-tier customer group continues to expand, which contains huge market potential. Express delivery companies can further expand their market share and enhance their market competitiveness. Providing mid-tier customers with better quality and more customized services can help express delivery companies improve their service system and improve the overall service level, but the existing technology lacks a method for customer identification.

[0003] Therefore, a method and device for identifying customers in the logistics industry are provided to solve the above problems. Summary of the invention

[0004] The main purpose of the present invention is to solve the problem in the prior art that there is a lack of a method for accurately identifying and classifying customers.

[0005] A first aspect of the present invention provides a method for identifying customers in the logistics industry, the method comprising: Acquiring customer data, the customer data including shipment volume data information, customer type data information, goods characteristics data information, service demand data information and financial data information; Establish a customer long-term business volume prediction model through LSTM long short-term memory network; Based on the AO algorithm, the hypercube strategy, adaptive spiral strategy and Gaussian mutation strategy are introduced to generate the IAO algorithm; the customer long-term business volume prediction model is optimized using the IAO algorithm to generate a customer long-term business volume prediction model based on IAO-LSTM; Generate forecast results through the IAO-LSTM-based customer long-term business volume forecast model, integrate the forecast results, shipment data information, customer type data information, cargo characteristics data information, service demand data information and financial data information, generate comprehensive scores through the hierarchical analysis algorithm, and generate mid-level customer information; The mid-level customer information is sent to the collaborative end, feedback information from the collaborative end is received, the weights of the hierarchical analysis algorithm and the IAO-LSTM-based customer long-term business volume prediction model are adjusted according to the feedback information, the mid-level customer information is optimized, and the optimized mid-level customer information is sent to the mobile terminal.

[0006] Further, for the acquisition of customer data, the customer data includes shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information, including: Through the order data system, obtain shipment volume data information, customer type data information, and goods characteristic data information; the shipment volume data information includes daily average shipment volume information, monthly average shipment volume information, and annual average shipment volume information, the customer type data information includes industry attribute information, enterprise scale information, and business model information, and the goods characteristic data information includes goods type information, goods weight information, and goods volume information; Obtain service demand data information through the questionnaire system, and the service demand information includes distribution timeliness information, distribution range information, and value-added service demand information; Obtain financial data information through a third-party data platform, and the financial data information includes enterprise revenue information and the proportion of logistics costs.

[0007] Further, the establishment of the customer long-term business volume prediction model through the LSTM long short-term memory network includes: Based on the TensorFlow or PyTorch deep learning framework, construct an LSTM long short-term memory network model; Determine the number of input layer nodes of the LSTM long short-term memory network model, and use the customer data after feature engineering processing as input features; Design the number of layers of the hidden layer of the LSTM long short-term memory network model and the number of neurons in each layer of the hidden layer; Set the output layer of the LSTM long short-term memory network model as the customer's future long-term business volume information, and complete the establishment of the customer long-term business volume prediction model.

[0008] Further, introducing the hypercube strategy, the adaptive spiral strategy, and the Gaussian mutation strategy based on the AO algorithm to generate the IAO algorithm; using the IAO algorithm to optimize the customer long-term business volume prediction model, and generating the customer long-term business volume prediction model based on IAO-LSTM includes: Based on the AO algorithm, introduce the hypercube strategy, divide the search space into multiple hypercube sub-regions, and randomly generate initial individuals within each sub-region according to a certain probability distribution; Based on the AO algorithm, introduce the self-spiral strategy. When the algorithm is in the global search stage, use a larger search step size and a wider search direction. When the algorithm approaches the potential optimal solution region, gradually reduce the search step size, and perform local search through a quasi-spiral search path; Based on the AO algorithm, introduce the Gaussian mutation strategy to generate the IAO algorithm; Optimize the long-term business volume prediction model of customers using the IAO algorithm to generate a long-term business volume prediction model of customers based on IAO-LSTM.

[0009] Furthermore, generate a prediction result through the long-term business volume prediction model of customers based on IAO-LSTM, integrate the prediction result, shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information, and generate a comprehensive score through the analytic hierarchy process algorithm. The generated waist customer information includes: Generate a prediction result through the long-term business volume prediction model of customers based on IAO-LSTM; Perform data preprocessing on the prediction result, shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information; Convert the processed shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information into timeliness index values, key financial index values, unit logistics cost values, and profit contribution information; Establish an analytic hierarchy model and use the analytic hierarchy model to calculate the quantitative comprehensive score of each customer; Sort the customers from high to low according to the scores, and select the customers within the preset ranking as waist customers.

[0010] Furthermore, the establishment of the analytic hierarchy model includes: Divide the customer evaluation problem into an objective layer, a criterion layer, and a scheme layer; the objective layer is to screen waist users; the criterion layer includes several evaluation dimensions, and several evaluation dimensions include: long-term business volume, financial status, service timeliness requirements, logistics cost, and profit contribution; the scheme layer is each customer individual; Construct a judgment matrix, use the 1-9 scale method proposed by Saaty to compare and score the importance between all pairs of evaluation dimensions to form a judgment matrix; use the consistency index and random consistency ratio to conduct a consistency test on the judgment matrix. When the random consistency ratio is less than 0.1, it is considered that the judgment matrix has consistency, otherwise, readjust the element values of the judgment matrix until the consistency test is passed.

[0011] Furthermore, sending the waist customer information to the collaboration terminal, receiving the feedback information from the collaboration terminal, adjusting the weights of the analytic hierarchy process algorithm and the long-term business volume prediction model of customers based on IAO-LSTM according to the feedback information, optimizing the waist customer information, and sending the optimized waist customer information to the mobile terminal includes: Send the waist customer information to the management collaboration terminal and the marketing collaboration terminal; Obtain the feedback information from the management collaboration terminal and the feedback information from the marketing collaboration terminal; Adjust the weights of the analytic hierarchy process algorithm and the customer long-term business volume prediction model based on IAO-LSTM according to the feedback information from the management collaboration terminal and the marketing collaboration terminal; optimize the information of mid-tier customers; Send the optimized mid-tier customer information to the mobile terminal.

[0012] The second aspect of the present invention provides a customer identification device for the logistics industry, and the customer identification device for the logistics industry includes: A customer data acquisition module, configured to acquire customer data, where the customer data includes shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information; A prediction model establishment module, configured to establish a customer long-term business volume prediction model through an LSTM long short-term memory network; A prediction model adjustment module, configured to generate an IAO algorithm by introducing a hypercube strategy, an adaptive spiral strategy, and a Gaussian mutation strategy based on the AO algorithm; optimize the customer long-term business volume prediction model using the IAO algorithm to generate a customer long-term business volume prediction model based on IAO-LSTM; A mid-tier customer information generation module, configured to generate a prediction result through the customer long-term business volume prediction model based on IAO-LSTM, integrate the prediction result, shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information, generate a comprehensive score through the analytic hierarchy process algorithm, and generate mid-tier customer information; A mid-tier customer information optimization module, configured to send the mid-tier customer information to the collaboration terminal, receive the feedback information from the collaboration terminal, adjust the weights of the analytic hierarchy process algorithm and the customer long-term business volume prediction model based on IAO-LSTM according to the feedback information, optimize the mid-tier customer information, and send the optimized mid-tier customer information to the mobile terminal.

[0013] The third aspect of the present invention provides an electronic device, and the electronic device includes a memory and at least one processor, and instructions and data are stored in the memory; The at least one processor calls the instructions and data in the memory so that the electronic device executes each step of the above-mentioned customer identification method for the logistics industry.

[0014] The fourth aspect of the present invention provides a readable storage medium, and instructions and data are stored on the readable storage medium, and when the instructions are executed by a processor, each step of the customer identification method for the logistics industry as described above is implemented.

[0015] The present invention obtains basic customer information, order data information, financial data information, service requirements, etc., and performs feature engineering data processing on customer information, including: calculating the average daily shipment volume, average monthly shipment volume, and average annual shipment volume; converting categorical features into numerical features; establishing an LSTM long-term customer business volume prediction model through an LSTM long short-term memory network, and identifying waist customers with stable growth potential, having the advantages of accurate identification and high identification rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flowchart of the customer identification method in the logistics industry provided by the present invention; Figure 2 It is a schematic structural diagram of the customer identification device in the logistics industry provided by the present invention; Figure 3 It is a schematic structural diagram of the electronic device provided by the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] The embodiment of the present invention provides a customer identification method in the logistics industry, including obtaining customer data, where the customer data includes shipment volume data information, customer type data information, goods characteristic data information, service requirement data information, and financial data information; establishing a long-term customer business volume prediction model through an LSTM long short-term memory network; generating an IAO algorithm based on the AO algorithm by introducing a hypercube strategy, an adaptive spiral strategy, and a Gaussian mutation strategy; optimizing the long-term customer business volume prediction model by using the IAO algorithm to generate a long-term customer business volume prediction model based on IAO-LSTM; generating a prediction result through the long-term customer business volume prediction model based on IAO-LSTM, integrating the prediction result, shipment volume data information, customer type data information, goods characteristic data information, service requirement data information, and financial data information, generating a comprehensive score through an analytic hierarchy process algorithm to generate waist customer information; sending the waist customer information to a collaborative end, receiving feedback information from the collaborative end, adjusting the weights of the analytic hierarchy process algorithm and the long-term customer business volume prediction model based on IAO-LSTM according to the feedback information, optimizing the waist customer information, and sending the optimized waist customer information to a mobile terminal. The main purpose of the present invention is to solve the problem in the prior art that there is a lack of a method for predicting and ranking the long-term business volume of customers.

[0018] In the description, claims and the above-mentioned drawings of the present invention, terms such as "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the term "comprising" or "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily limit to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0019] For ease of understanding, the specific process of the embodiments of the present invention will be described below. Please refer to Figure 1 , the first embodiment of the customer identification method for the logistics industry provided by the present invention includes: S1. Obtain customer data, where the customer data includes shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information; Specifically, it includes: obtaining shipment volume data information, customer type data information, and goods characteristic data information through an order data system; the shipment volume data information includes daily average shipment volume information, monthly average shipment volume information, and annual average shipment volume information, the customer type data information includes industry attribute information, enterprise scale information, and business model information, and the goods characteristic data information includes goods type information, goods weight information, and goods volume information; Obtain service demand data information through a questionnaire system, where the service demand information includes distribution timeliness information, distribution range information, and value-added service demand information; Obtain financial data information through a third-party data platform, where the financial data information includes enterprise revenue information and logistics cost ratio information.

[0020] S2. Establish a customer long-term business volume prediction model through an LSTM long short-term memory network; Specifically, it includes: Based on the TensorFlow or PyTorch deep learning framework, construct an LSTM long short-term memory network model; Determine the number of input layer nodes of the LSTM long short-term memory network model, and use the customer data after feature engineering processing as input features; Design the number of layers of the hidden layer of the LSTM long short-term memory network model and the number of neurons in each layer of the hidden layer; Set the output layer of the LSTM long short-term memory network model as the customer's future long-term business volume information, and complete the establishment of the customer long-term business volume prediction model.

[0021] S3. Introduce the hypercube strategy, adaptive spiral strategy, and Gaussian mutation strategy based on the AO algorithm to generate the IAO algorithm; use the IAO algorithm to optimize the long-term customer business volume prediction model to generate a long-term customer business volume prediction model based on IAO-LSTM; Specifically, it includes: introducing the hypercube strategy based on the AO algorithm, dividing the search space into multiple hypercube sub-regions, and randomly generating initial individuals within each sub-region according to a certain probability distribution; Introduce the self-spiral strategy based on the AO algorithm. When the algorithm is in the global search stage, use a larger search step and a broader search direction. When the algorithm approaches the potential optimal solution region, gradually reduce the search step and perform local search through a quasi-spiral search path; Introduce the Gaussian mutation strategy based on the AO algorithm to generate the IAO algorithm; Use the IAO algorithm to optimize the long-term customer business volume prediction model to generate a long-term customer business volume prediction model based on IAO-LSTM.

[0022] S4. Generate a prediction result through the long-term customer business volume prediction model based on IAO-LSTM, integrate the prediction result, shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information, and generate a comprehensive score through the analytic hierarchy process algorithm to generate waist customer information; Specifically, it includes: generating a prediction result through the long-term customer business volume prediction model based on IAO-LSTM; Perform data preprocessing on the prediction result, shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information; Convert the processed shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information into timeliness index values, key financial index values, unit logistics cost values, and profit contribution information; Establish a hierarchical structure model, and use the hierarchical structure model to calculate the quantitative comprehensive score of each customer; the specific process of establishing the hierarchical structure model is: Divide the customer evaluation problem into an objective layer, a criterion layer, and a scheme layer; the objective layer is to screen waist users; the criterion layer includes several evaluation dimensions, and several evaluation dimensions include: long-term business volume, financial status, service timeliness requirements, logistics cost, and profit contribution; the scheme layer is each customer individual; Construct a judgment matrix. By using the 1-9 scale method proposed by Saaty, compare and score the importance between every two evaluation dimensions to form a judgment matrix. Use the consistency index and random consistency ratio to conduct a consistency test on the judgment matrix. When the random consistency ratio is less than 0.1, it is considered that the judgment matrix has consistency; otherwise, readjust the element values of the judgment matrix until the consistency test is passed. Rank the customers from high to low according to the scores, and select the customers within the preset ranking as the waist customers.

[0023] S5. Send the waist customer information to the collaboration terminal, receive the feedback information from the collaboration terminal, adjust the weights of the analytic hierarchy process algorithm and the long-term customer business volume prediction model based on IAO-LSTM according to the feedback information, optimize the waist customer information, and send the optimized waist customer information to the mobile terminal; Among them, it specifically includes: sending the waist customer information to the management collaboration terminal and the marketing collaboration terminal; Obtain the feedback information from the management collaboration terminal and the feedback information from the marketing collaboration terminal; Adjust the weights of the analytic hierarchy process algorithm and the long-term customer business volume prediction model based on IAO-LSTM according to the feedback information from the management collaboration terminal and the feedback information from the marketing collaboration terminal; optimize the waist customer information; Send the optimized waist customer information to the mobile terminal.

[0024] Please refer to Figure 1 , the second embodiment of the customer identification method for the logistics industry provided by the present invention includes: S1. Obtain customer data, where the customer data includes shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information; Specifically include: through the order data system, obtain the shipment volume data information, customer type data information, and goods characteristic data information; the shipment volume data information includes daily average shipment volume information, monthly average shipment volume information, and annual average shipment volume information, the customer type data information includes industry attribute information, enterprise scale information, and business model information, and the goods characteristic data information includes goods type information, goods weight information, and goods volume information; Obtain the service demand data information through the questionnaire system, where the service demand information includes distribution timeliness information, distribution range information, and value-added service demand information; Obtain the financial data information through a third-party data platform, where the financial data information includes enterprise revenue information and logistics cost ratio information.

[0025] In the actual operation process, data collection generally includes platform data collection, collection of customer basic information, financial condition research, analysis of shipping behavior and courier selection, and collation of service requirements and historical cooperation records, etc.; Platform data collection: Establish data docking interfaces with major e-commerce platforms, logistics trading platforms, etc. Use data scraping tools (such as the Scrapy framework in Python) to obtain the sender statistics data of customers on these platforms according to predefined rules and frequencies. Include detailed information such as the time distribution of sending (statistical by day, week, month, quarter, year), geographical distribution of sending destinations, classification statistics of types of items sent (such as documents, clothing, electronic products, etc.), and the proportion of the number of sent items in different weight segments and volume segments, ensuring the accuracy and integrity of the data. At the same time, follow the data usage specifications and privacy policies of each platform and take good data security protection measures.

[0026] Collection of basic customer information: Extract the basic information of customers from the enterprise's customer relationship management system (CRM), such as customer name, contact information, registered address, nature of the enterprise (individual, small and medium-sized enterprise, large enterprise, etc.), industry category (manufacturing, trading, service, etc.), and relevant qualification information such as business licenses, to provide basic dimension data for subsequent customer classification and analysis. Through data cleaning and standardization processing, ensure the consistency and usability of the information, for example, unify the format of contact information and standardize the filling method of addresses.

[0027] Investigation of financial status: Collaborate with the financial department of the enterprise to obtain the financial status data of customers. For enterprise customers, collect data from key financial statements such as balance sheets, income statements, and cash flow statements, and analyze indicators such as operating income, net profit, asset scale, solvency (such as asset-liability ratio, current ratio, etc.), profitability (such as gross profit margin, net profit margin, etc.), and cash flow status (such as net cash flow from operating activities, cash flow ratio, etc.); for individual customers, evaluate their personal income level, credit score, debt situation, etc. by cooperating with third-party credit assessment agencies or collecting their consumer credit records, bank statements, etc., in order to understand the payment ability and potential business risks of customers. During the data acquisition process, strictly abide by the confidentiality regulations of financial data and the requirements of laws and regulations.

[0028] Analysis of Sender Behavior and Courier Selection: Record the detailed information of customers' daily shipments in the logistics management system, including shipment frequency (average number of shipments per month, number of shipments during peak periods, etc.), shipment time pattern (whether there are fixed shipment time periods, seasonal shipment fluctuations, etc.), urgency of shipments (proportion of express shipments, average time limit requirements for ordinary shipments, etc.), and customers' preference and switching frequency among different courier companies. Combine the service characteristics of each courier company in the market (such as price, delivery time, service scope, value-added services, etc.) to analyze the driving factors behind customers' selection behavior, provide a basis for formulating targeted marketing strategies and service optimization plans, and discover customers' potential needs and market trends through in-depth mining and correlation analysis of logistics data. For example, by analyzing the changes in customers' shipment destinations, layout logistics network resources in advance.

[0029] Sorting of Service Requirements and Historical Cooperation Records: Collect customers' specific requirements for courier services through methods such as questionnaires, customer service feedback records, and direct communication with customers, such as packaging standards for goods, temperature control requirements during transportation, insurance needs for goods, expected degree of door-to-door delivery, timeliness and accuracy requirements for information tracking, etc.; at the same time, sort out the historical cooperation between customers and the enterprise, including cooperation duration, past shipment volume trends, service disputes and solutions that have occurred, historical customer satisfaction evaluations, etc., in order to comprehensively understand the cooperation relationship between customers and the enterprise, discover the advantages and deficiencies in the cooperation process, provide a reference for subsequent customer relationship maintenance and business expansion, establish a customer service requirements and historical cooperation database for convenient data storage, query, and analysis, and use data visualization tools to intuitively display customers' service requirements and historical cooperation situations for the management and relevant departments to quickly understand customer situations.

[0030] S2. Establish a long-term customer business volume prediction model through the LSTM long short-term memory network; Specifically, it includes: Based on the TensorFlow or PyTorch deep learning framework, construct an LSTM long short-term memory network model; Determine the number of input layer nodes of the LSTM long short-term memory network model, and use the customer data after feature engineering processing as input features; Design the number of layers of the hidden layer of the LSTM long short-term memory network model and the number of neurons in each layer of the hidden layer; Set the output layer of the LSTM long short-term memory network model as the long-term future business volume information of customers to complete the establishment of the long-term customer business volume prediction model.

[0031] In the actual operation process, an LSTM (Long Short-Term Memory) long short-term memory network model can be constructed based on deep learning frameworks such as TensorFlow or PyTorch. Determine the number of nodes in the input layer, and use the numerical data after feature engineering processing of the customer's sender statistics data, basic information, financial status, etc. as input features. For example, input the normalized sender time series data, and at the same time, perform one-hot encoding on the categorical data and convert it into a numerical vector before inputting it into the network. Design a suitable hidden layer structure, including the number of hidden layers and the number of neurons in each layer. Generally, the best hidden layer configuration is determined through multiple experiments and tuning to balance the learning ability and computational complexity of the model. The output layer is set to predict the customer's future long-term business volume. According to business needs, the business volume values for the next month, quarter, half-year, or year can be selected for prediction, and a linear activation function is used to output continuous business volume prediction values. During the model construction process, reasonably set the hyperparameters of the model, such as the learning rate, number of iterations, batch size, etc., and initialize the weights and thresholds of the LSTM network. Usually, the method of random initialization or initialization based on a specific distribution (such as Xavier initialization) is adopted to ensure that the model has a good parameter distribution at the initial stage of training, which is beneficial to the subsequent learning process.

[0032] S3. Introduce the hypercube strategy, adaptive spiral strategy, and Gaussian mutation strategy based on the AO algorithm to generate the IAO algorithm; use the IAO algorithm to optimize the customer long-term business volume prediction model to generate a customer long-term business volume prediction model based on IAO-LSTM; Specifically, it includes: introducing the hypercube strategy based on the AO algorithm, dividing the search space into multiple hypercube sub-regions, and randomly generating initial individuals according to a certain probability distribution within each sub-region; for example, for the search space where the weights and thresholds of the LSTM model are located, set a suitable hypercube division method according to its value range and experience, and generate initial weight and threshold vectors with a certain degree of randomness but conforming to the overall distribution law within each sub-region, providing a richer starting point for the subsequent optimization process. In this way, the possibility of the algorithm finding the global optimal solution in the complex search space is increased, and the risk of falling into the local optimum is reduced.

[0033] Based on the AO algorithm, a self-spiral strategy is introduced. When the algorithm is in the global search stage, a larger search step size and a broader search direction are adopted. When the algorithm approaches the potential optimal solution region, the search step size is gradually reduced, and local search is carried out through a spiral-like search path. Through this adaptive strategy, the IAO algorithm can flexibly switch between global and local search modes according to the actual situation during the search process, improving the search efficiency and convergence speed of the algorithm. During the implementation process, mathematical functions are used to describe the spiral search path, and the search parameters are dynamically adjusted according to the feedback information of the fitness function to ensure that the algorithm can continuously evolve towards the optimal solution.

[0034] Based on the AO algorithm, a Gaussian mutation strategy is introduced. During the iterative process of the algorithm, certain dimensions of individuals are subjected to Gaussian mutation operations with a certain probability. Gaussian mutation means generating a random perturbation vector according to the Gaussian distribution and adding it to the corresponding dimensions of the current individual, thereby changing the position of the individual, enabling the algorithm to potentially jump out of the current local optimal solution region and continue to explore other potential better solutions. By reasonably setting the probability and mutation intensity parameters of Gaussian mutation, the diversity of the population can be effectively increased without destroying the overall search stability of the algorithm, avoiding the algorithm from prematurely converging to the local optimal solution. In practical applications, according to the complexity of the problem and the characteristics of the search space, the Gaussian mutation parameters are repeatedly tested and adjusted to find the best parameter configuration to improve the performance of the algorithm in complex optimization problems.

[0035] After generating the process of the IAO algorithm by introducing the hypercube strategy, self-adaptive spiral strategy and Gaussian mutation strategy based on the above AO algorithm, the IAO algorithm is generated. The IAO algorithm is used to optimize the customer long-term business volume prediction model, and a customer long-term business volume prediction model based on IAO-LSTM is generated.

[0036] In the actual operation process, specifically, the improved Tianying optimization algorithm (IAO) is used to optimize the weights and thresholds of the LSTM customer long-term business volume prediction model. The prediction errors of the LSTM model (such as mean square error, mean absolute error, etc.) are used as the fitness function of the IAO algorithm. By continuously iterating the IAO algorithm, the weights and thresholds of the LSTM model are updated, so that the prediction performance of the LSTM model is gradually improved. In each iteration, the IAO algorithm selects excellent individuals according to the fitness values of the individuals in the current population for operations such as crossover and mutation to generate new individuals. Then, the weights and thresholds corresponding to the new individuals are updated into the LSTM model, and the prediction error of the model is recalculated and fed back to the IAO algorithm as the basis for the next iteration. Through multiple iterations of optimization, the LSTM model can better fit the customer's business volume data, improve the accuracy and stability of the prediction. At the same time, visualization tools (such as TensorBoard) are used to monitor the training process of the model, including the change trend of the loss function, the update of model parameters, etc., so as to timely discover problems and adjust the optimization strategy.

[0037] S4. Generate a prediction result through the customer long-term business volume prediction model based on IAO-LSTM, integrate the prediction result, shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information, and generate a comprehensive score through the analytic hierarchy process algorithm to generate the information of waist customers; Specifically, it includes: generating a prediction result through the customer long-term business volume prediction model based on IAO-LSTM; Perform data preprocessing on the prediction result, shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information; From the processed shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information, convert the timeliness index value, key financial index value, unit logistics cost value, and profit contribution information; integrate the long-term business volume prediction results of customers obtained from the LSTM prediction model with data such as the customer service timeliness requirement, customer financial status, customer logistics quotation, and logistics cost. For the customer service timeliness requirement, convert the customer's expectation of the express delivery time into specific timeliness index values, such as the average expected delivery days, the maximum acceptable delay duration, etc.; for the customer financial status data, further refine and standardize the key financial indicators to make them comparable with other data; for the logistics quotation and cost data, calculate the unit logistics cost and profit contribution of each customer, and at the same time consider the price elasticity and cost structure differences of different logistics service products. Through preprocessing steps such as data cleaning, missing value processing, and feature normalization, ensure the reliability of the integrated data quality, provide accurate input data for the subsequent Analytic Hierarchy Process (AHP) calculation, establish the work process and specifications for data integration and preprocessing, ensure the consistency and integrity of the data, and at the same time perform data backup and version management for retrospective and verification when needed; Establish a hierarchical structure model and calculate the quantitative comprehensive score of each customer using the hierarchical structure model; the specific process of establishing the hierarchical structure model is as follows: Divide the customer evaluation problem into an objective layer, a criterion layer, and a scheme layer; the objective layer is to screen out waist users; the criterion layer includes several evaluation dimensions, and the several evaluation dimensions include: long-term business volume, financial status, service timeliness requirement, logistics cost, and profit contribution; the scheme layer is each customer individual; Construct a judgment matrix, and through the 1-9 scale method proposed by Saaty, compare and score the importance between every two of all evaluation dimensions to form a judgment matrix; use the consistency index and the random consistency ratio to conduct a consistency test on the judgment matrix. When the random consistency ratio is less than 0.1, it is considered that the judgment matrix has consistency, otherwise readjust the element values of the judgment matrix until the consistency test is passed; Sort customers according to their scores from high to low, and select the customers within the preset ranking as the waist customers. Specifically, it includes sorting customers according to the calculated comprehensive customer scores from high to low, and selecting the customers in the middle position with relatively high comprehensive scores as high-quality waist customers. When determining the selection range of high-quality waist customers, combine the enterprise's market positioning, business development strategy, and resource allocation capabilities, and comprehensively consider the scale and structure of the customer group. For example, customers with comprehensive score rankings in the top 30%-50% can be selected as the high-quality waist customer group, and at the same time, set a certain minimum score threshold to ensure that the selected customers have certain advantages and potential in each evaluation dimension. For the selected high-quality waist customers, establish special customer files and management lists, and record the customer's various information and scoring situations in detail to provide strong support for subsequent precision marketing and customer relationship management. Through data analysis and market research, continuously optimize the screening criteria and methods of high-quality waist customers to meet the needs of market changes and enterprise development.

[0038] S5. Send the waist customer information to the collaboration end, receive the feedback information from the collaboration end, adjust the weights of the analytic hierarchy process algorithm and the long-term customer business volume prediction model based on IAO-LSTM according to the feedback information, optimize the waist customer information, and send the optimized waist customer information to the mobile terminal; Among them, specifically including: sending the waist customer information to the management collaboration end and the marketing collaboration end; Obtain the feedback information from the management collaboration end and the feedback information from the marketing collaboration end; Adjust the weights of the analytic hierarchy process algorithm and the long-term customer business volume prediction model based on IAO-LSTM according to the feedback information from the management collaboration end and the feedback information from the marketing collaboration end; optimize the waist customer information; Send the optimized waist customer information to the mobile terminal. Visualize the optimized customer data results and display them on the mobile terminal so that relevant enterprise personnel can conveniently view and analyze customer information anytime and anywhere. Utilize mobile application development technologies (such as ReactNative, Flutter, etc.) to design and develop a dedicated mobile application for customer data management. On the mobile terminal interface, display the detailed information of high-quality waist customers in various forms such as charts (such as bar charts, line charts, pie charts, etc.) and reports (such as customer lists, scoring ranking tables, business volume trend analysis tables, etc.), including the basic information of customers, business volume prediction trends, comprehensive scoring situations, score details of each evaluation dimension, etc. At the same time, provide functions such as data filtering, searching, and sorting to facilitate users to quickly locate and view the information of specific customers according to different needs. During the data display process, pay attention to the user experience and the friendliness of the interface design to ensure that the data is displayed clearly, intuitively, and easily understandable, while ensuring the security and privacy of the data. Through the push notification function of the mobile application, timely push information such as the update of customer data and the business change warning of important customers to relevant personnel, improve the enterprise's response speed and management efficiency for customer information, provide strong support for the enterprise's marketing, customer service, operation management, etc., and promote the enterprise's digital transformation and intelligent development.

[0039] The above describes the customer identification method in the logistics industry in the embodiments of the present invention. Next, the customer identification device in the embodiments of the present invention will be described. Please refer to Figure 2 The customer identification device in the embodiments of the present invention includes, for the above embodiments: A customer data acquisition module 201, configured to acquire customer data, where the customer data includes shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information; A prediction model establishment module 202, configured to establish a long-term customer business volume prediction model through an LSTM long short-term memory network; A prediction model adjustment module 203, configured to generate an IAO algorithm based on the AO algorithm by introducing a hypercube strategy, an adaptive spiral strategy, and a Gaussian mutation strategy; use the IAO algorithm to optimize the long-term customer business volume prediction model to generate a long-term customer business volume prediction model based on IAO-LSTM; A waist customer information generation module 204, configured to generate a prediction result through the long-term customer business volume prediction model based on IAO-LSTM, integrate the prediction result, shipment volume data information, customer type data information, goods characteristic data information, service demand data information, and financial data information, generate a comprehensive score through an analytic hierarchy process algorithm, and generate waist customer information; The waist customer information optimization module 205 is used to send the waist customer information to the collaboration end, receive the feedback information from the collaboration end, adjust the weights of the hierarchical analysis algorithm and the customer long-term business volume prediction model based on IAO-LSTM according to the feedback information, optimize the waist customer information, and send the optimized waist customer information to the mobile terminal.

[0040] above Figure 2 The logistics industry customer identification device in the embodiment of the present invention is described in detail from the perspective of modular functional entities. Next, the electronic device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0041] Figure 3 FIG. is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. The electronic device 700 may vary greatly due to configuration or performance differences, and may include one or more processors 710 (for example, one or more processors) and a memory 720, and one or more storage media 730 for storing application programs 733 or data 732 (for example, one or more storage devices, including RAM, FLASH, etc.). Among them, the memory 720 and the storage media 730 may be transient storage or persistent storage. The program stored in the storage media 730 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the electronic device 700. Further, the processor 710 may be configured to communicate with the storage media 730 and execute a series of instruction operations in the storage media 730 on the electronic device 700.

[0042] The electronic device 700 may further include one or more power supplies 740, one or more input / output interfaces 750, and / or one or more operating systems 731, such as FreeRTOS, Android, etc. Those skilled in the art can understand that Figure 3 The shown structure of the electronic device does not constitute a limitation on the electronic device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0043] 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, and when the instructions are run on a computer, the computer is caused to execute the steps of the logistics industry customer identification method.

[0044] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the above-described system or device and unit can refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again.

[0045] When 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, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a mobile device, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0046] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for identifying customers in the logistics industry, characterized in that: The logistics industry customer identification method comprises: Acquiring customer data, the customer data including shipment volume data information, customer type data information, goods characteristics data information, service demand data information and financial data information; Establish a customer long-term business volume prediction model through LSTM long short-term memory network; Based on the AO algorithm, the hypercube strategy, adaptive spiral strategy and Gaussian mutation strategy are introduced to generate the IAO algorithm; the customer long-term business volume prediction model is optimized using the IAO algorithm to generate a customer long-term business volume prediction model based on IAO-LSTM; Generate forecast results through the IAO-LSTM-based customer long-term business volume forecast model, integrate the forecast results, shipment data information, customer type data information, cargo characteristics data information, service demand data information and financial data information, generate comprehensive scores through the hierarchical analysis algorithm, and generate mid-level customer information; The mid-level customer information is sent to the collaborative end, feedback information from the collaborative end is received, the weights of the hierarchical analysis algorithm and the IAO-LSTM-based customer long-term business volume prediction model are adjusted according to the feedback information, the mid-level customer information is optimized, and the optimized mid-level customer information is sent to the mobile terminal.

2. The method for identifying customers in the logistics industry according to claim 1, characterized in that: The customer data obtained includes shipment data, customer type data, cargo characteristics data, service demand data and financial data, including: Obtain shipment data information, customer type data information, and cargo characteristic data information through the order data system; the shipment data information includes daily average shipment information, monthly average shipment information, and annual average shipment information; the customer type data information includes industry attribute information, enterprise scale information, and business model information; and the cargo characteristic data information includes cargo type information, cargo weight information, and cargo volume information; Obtain service demand data information through a questionnaire system, wherein the service demand information includes delivery time information, delivery range information and value-added service demand information; Financial data information is obtained through a third-party data platform, wherein the financial data information includes corporate revenue information and logistics cost ratio information.

3. The method for identifying customers in the logistics industry according to claim 1, characterized in that: The establishment of a customer long-term business volume prediction model through an LSTM long short-term memory network includes: Build LSTM long short-term memory network model based on TensorFlow or PyTorch deep learning framework; Determine the number of input layer nodes of the LSTM long short-term memory network model, and use customer data as input features after feature engineering processing; Design the number of hidden layers and the number of neurons in each hidden layer of the LSTM long short-term memory network model; The output layer of the LSTM long short-term memory network model is set as the customer's future long-term business volume information to complete the establishment of the customer's long-term business volume prediction model.

4. The method for identifying customers in the logistics industry according to claim 1, characterized in that: The method of introducing the hypercube strategy, the adaptive spiral strategy and the Gaussian mutation strategy based on the AO algorithm to generate the IAO algorithm; optimizing the customer long-term business volume prediction model by using the IAO algorithm to generate the customer long-term business volume prediction model based on the IAO-LSTM includes: Based on the AO algorithm, the hypercube strategy is introduced to divide the search space into multiple hypercube sub-regions, and the initial individuals are randomly generated in each sub-region according to a certain probability distribution; Based on the introduction of the self-spiral strategy of the AO algorithm, when the algorithm is in the global search stage, a larger search step and a wider search direction are adopted. When the algorithm approaches the potential optimal solution area, the search step is gradually reduced, and a local search is performed through a spiral-like search path. Based on the AO algorithm, the Gaussian mutation strategy is introduced to generate the IAO algorithm; The IAO algorithm is used to optimize the customer long-term business volume prediction model and generate a customer long-term business volume prediction model based on IAO-LSTM.

5. The method for identifying customers in the logistics industry according to claim 1, characterized in that: The prediction results are generated by the customer long-term business volume prediction model based on IAO-LSTM, and the prediction results, shipment data information, customer type data information, cargo characteristics data information, service demand data information and financial data information are integrated to generate a comprehensive score through a hierarchical analysis algorithm. The generated mid-level customer information includes: Generate forecast results through the customer long-term business volume forecast model based on IAO-LSTM; Preprocess the forecast results, shipment data, customer type data, cargo characteristics data, service demand data and financial data; Convert the processed shipment volume data, customer type data, cargo characteristics data, service demand data and financial data into time efficiency index values, key financial index values, unit logistics cost values ​​and profit contribution information; Establish a hierarchical model and use it to calculate the quantitative comprehensive score of each customer; Sort the customers by scores from high to low, and select those within the preset ranking as mid-level customers.

6. The method for identifying customers in the logistics industry according to claim 5, characterized in that: The said establishing a hierarchical structure model comprises: The customer evaluation problem is divided into the target layer, the criterion layer and the solution layer; the target layer is to screen mid-level users; the criterion layer includes several evaluation dimensions, including: long-term business volume, financial status, service timeliness requirements, logistics costs and profit contribution; the solution layer is for each individual customer; A judgment matrix was constructed. The importance of all evaluation dimensions was compared and scored pairwise using the 1-9 scaling method proposed by Saaty to form a judgment matrix. The judgment matrix was tested for consistency using consistency indicators and random consistency ratios. When the random consistency ratio was less than 0.1, the judgment matrix was considered consistent. Otherwise, the element values ​​of the judgment matrix were readjusted until the consistency test passed.

7. The method for identifying customers in the logistics industry according to claim 1, characterized in that: The sending of the mid-level customer information to the collaborative end, receiving feedback information from the collaborative end, adjusting the weights of the hierarchical analysis algorithm and the customer long-term business volume prediction model based on IAO-LSTM according to the feedback information, optimizing the mid-level customer information, and sending the optimized mid-level customer information to the mobile terminal includes: Send mid-level customer information to the management collaboration end and the marketing collaboration end; Obtain feedback information from the management collaboration end and the marketing collaboration end; Adjust the weights of the hierarchical analysis algorithm and the customer long-term business volume prediction model based on IAO-LSTM according to the feedback information from the management collaboration end and the marketing collaboration end; optimize the information of mid-level customers; The optimized waist customer information is sent to the mobile terminal.

8. A customer identification device for the logistics industry, characterized in that: include: A customer data acquisition module is used to acquire customer data, wherein the customer data includes shipment data information, customer type data information, goods characteristic data information, service demand data information and financial data information; The prediction model building module is used to build a customer long-term business volume prediction model through the LSTM long short-term memory network; The prediction model adjustment module is used to generate the IAO algorithm by introducing the hypercube strategy, adaptive spiral strategy and Gaussian mutation strategy based on the AO algorithm; the customer long-term business volume prediction model is optimized by using the IAO algorithm to generate the customer long-term business volume prediction model based on IAO-LSTM; The mid-level customer information generation module is used to generate prediction results through the customer long-term business volume prediction model based on IAO-LSTM, integrate the prediction results, shipment data information, customer type data information, cargo characteristics data information, service demand data information and financial data information, generate a comprehensive score through the hierarchical analysis algorithm, and generate mid-level customer information; The mid-level customer information optimization module is used to send the mid-level customer information to the collaborative end, receive feedback information from the collaborative end, adjust the weights of the hierarchical analysis algorithm and the IAO-LSTM-based customer long-term business volume prediction model according to the feedback information, optimize the mid-level customer information, and send the optimized mid-level customer information to the mobile terminal.

9. An electronic device, comprising a memory and at least one processor, wherein the memory stores instructions and data; The at least one processor calls the instructions and data in the memory so that the electronic device executes each step of the logistics industry customer identification method as described in any one of claims 1-7.

10. A readable storage medium having instructions and data stored thereon, characterized in that: When the instructions are executed by the processor, the various steps of the method for identifying customers in the logistics industry as claimed in any one of claims 1 to 7 are implemented.

Images