Supply risk level confirmation method and device, equipment and storage medium

Abstract

The invention belongs to the technical field of artificial intelligence, and discloses a supply risk level confirmation method and device, equipment and a storage medium, and the method comprises the steps: obtaining the multi-dimensional customer information of a customer, and constructing a customer portrait of the customer according to the multi-dimensional customer information; according to the customer portrait, grouping the customers by adopting a first clustering algorithm to obtain a first customer group, and performing secondary grouping on the first customer group by adopting a second clustering algorithm to obtain a second customer group; predicting a first service demand quantity of each group in the second customer groups based on the time sequence, and combining the first service demand quantity with a second service demand quantity corresponding to the region growth rate to obtain a demand prediction result of each region; and obtaining service resource data of service supply in each region in real time, and calculating a supply risk level of each region according to the demand prediction result of each region and the service resource data of each region. The efficiency and accuracy of supply risk level confirmation are improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and in particular to a method, apparatus, device, and storage medium for confirming a supply risk level. Background Art

[0002] Traditionally, customer needs and home-based elderly care operators are manually matched to determine the direction of operations. However, when faced with a large and complex business scale, the traditional approach lacks effective technical means for demand forecasting and supply risk monitoring. Manual matching makes it difficult to accurately grasp the dynamic changes in customer needs and the service supply situation in various places. In other words, with existing technologies, the degree of match between service supply and customer needs is poor, resulting in the inability of service supply to meet customer needs. Summary of the Invention

[0003] The purpose of the embodiments of the present application is to provide a method, device, equipment and storage medium for confirming the supply risk level, the main purpose of which is to improve the efficiency and accuracy of confirming the supply risk level.

[0004] In order to solve the above technical problems, the present invention provides a method for confirming the supply risk level, which adopts the following technical solutions:

[0005] Acquire multi-dimensional customer information of the customer, and construct a customer profile of the customer based on the multi-dimensional customer information;

[0006] Based on the customer portraits, the customers are grouped using a first clustering algorithm to obtain a first customer group, and the first customer group is grouped again using a second clustering algorithm to obtain a second customer group;

[0007] Predicting the first service demand of each group in the second customer group based on the time series, combining the first service demand with the second service demand corresponding to the regional growth rate to obtain a demand forecast result for each region;

[0008] The service resource data of the service supply in each region is obtained in real time, and the supply risk level of each region is calculated based on the demand forecast results of each region and the service resource data of each region.

[0009] Secondly, in order to solve the above technical problems, the embodiment of the present application further provides a device for confirming the supply risk level, which adopts the following technical solution:

[0010] A portrait building module is used to obtain multi-dimensional customer information of a customer and build a customer portrait of the customer based on the multi-dimensional customer information;

[0011] A secondary clustering module, configured to group the customers using a first clustering algorithm based on the customer portraits to obtain a first customer group, and to perform secondary clustering on the first customer group using a second clustering algorithm to obtain a second customer group;

[0012] a demand forecasting module configured to forecast the first service demand of each group in the second customer group based on a time series, and to obtain a demand forecast result for each region by combining the first service demand with the second service demand corresponding to the regional growth rate;

[0013] The risk level calculation module is used to obtain the service resource data of the service supply in each region in real time, and calculate the supply risk level of each region based on the demand forecast results of each region and the service resource data of each region.

[0014] On the third aspect, in order to solve the above-mentioned technical problems, an embodiment of the present application also provides a computer device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the supply risk level confirmation method as described above.

[0015] Fourthly, in order to solve the above technical problems, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the supply risk level confirmation method as described above.

[0016] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0017] By obtaining multi-dimensional customer information and building customer profiles based on this information, demand forecast results are calculated based on the user profiles, thereby improving the accuracy of subsequent demand forecasts and thus improving the efficiency and accuracy of risk level confirmation.

[0018] By mining deep features through dual clustering, we can quickly identify customer groups with obvious commonalities, and then calculate demand forecast results based on these customer groups, thereby improving the accuracy of demand forecast results and, in turn, the efficiency and accuracy of confirming supply risk levels.

[0019] By weighting and summing the first service demand volume predicted by time series and the second service demand volume predicted by regional growth rate, we can obtain the demand forecast results for each region. This can prevent the demand forecast of a single factor from being affected by interference factors, which may lead to deviations in the demand forecast results. It can more comprehensively and accurately predict the demand of each region, thereby improving the accuracy of the demand forecast results.

[0020] By acquiring service resource data for each region in real time and comparing it with demand forecast results, we can accurately identify the supply risk level in each region. After calculating the supply risk level for each region, enterprises can rationally allocate service resources based on actual conditions, shifting resources from areas with lower supply risks to areas with higher risks, thereby achieving optimal resource allocation and improving customer loyalty. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the solutions in this application, a brief introduction will be given below to the drawings required for use in the description of the embodiments of this application. Obviously, the drawings described below are some embodiments of this application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0022] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0023] Figure 2 A flow chart of an embodiment of a method for confirming a supply risk level according to the present application;

[0024] Figure 3 is a structural diagram of an embodiment of a device for confirming a supply risk level according to the present application;

[0025] Figure 4 It is a structural diagram of an embodiment of a device according to the present application. DETAILED DESCRIPTION

[0026] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as commonly understood by those skilled in the art to which this application belongs. The terms used in the specification of the application are for the purpose of describing specific embodiments only and are not intended to limit this application. The terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of this application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0027] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0028] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0029] like Figure 1 As shown, system architecture 100 may include a terminal device 101, a network 102, and a server 103. Terminal device 101 may be a laptop computer 1011, a tablet computer 1012, or a mobile phone 1013. Network 102 is a medium for providing a communication link between terminal device 101 and server 103. Network 102 may include various connection types, such as wired or wireless communication links or fiber optic cables.

[0030] The user can use the terminal device 101 to interact with the server 103 via the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0031] The terminal device 101 can be various electronic devices with a display screen and supporting web browsing. In addition to the laptop computer 1011, tablet computer 1012 or mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV) player, a laptop computer and a desktop computer, etc.

[0032] The server 103 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal device 101 .

[0033] It should be noted that the method for confirming the supply risk level provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the device for confirming the supply risk level is generally set in the server / terminal device.

[0034] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is merely illustrative. Any number of terminal devices, networks and servers may be provided as required.

[0035] Continue to refer Figure 2, shows a flow chart of an embodiment of the method for confirming the supply risk level according to the present application. According to different needs, the order of the steps in the flow chart can be changed, and some steps can be omitted. The method for confirming the supply risk level provided in the embodiment of the present application can be applied to any scenario that requires confirmation of the supply risk level, and the method for confirming the supply risk level can be applied to products in these scenarios. The method for confirming the supply risk level includes the following steps:

[0036] Step S201: Acquire multi-dimensional customer information of a customer, and construct a customer profile of the customer based on the multi-dimensional customer information.

[0037] In this embodiment, the electronic device (eg Figure 1 The server / terminal device shown in the figure) can obtain the multi-dimensional customer information of the customer through a wired connection or a wireless connection. It should be noted that the above-mentioned wireless connection method may include but is not limited to 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future.

[0038] In this embodiment, multi-dimensional customer information of the customer is obtained from the customer information database. The multi-dimensional customer information refers to the first contact between the staff and the customer when the customer first enters the platform, wherein the first contact includes a Ferris wheel assessment of the customer and a Ferris wheel assessment of the family members; the Ferris wheel assessment of the customer includes obtaining the customer's physical health dimension, mental health dimension, living habit dimension, and social environment dimension; the Ferris wheel assessment of the family members includes obtaining the family support capacity dimension and the demand for elderly care services dimension.

[0039] In this embodiment, the above-mentioned physical health dimensions include basic health status (such as physiological data and medical history, etc.), physical function status (such as the client's mobility, etc.), nutritional status (such as malnutrition, normal nutrition or overnutrition); the above-mentioned mental health dimensions include emotional state (such as whether negative emotions such as anxiety, depression, irritability, loneliness, etc. often occur, as well as the stability and fluctuation of emotions), cognitive function (such as assessing whether there is Alzheimer's disease, etc.), and psychological needs (such as whether the elderly need emotional support); the above-mentioned living habits dimension includes daily routines, interests and hobbies; the above-mentioned social environment dimension includes participation in social activities; the above-mentioned family support capacity dimension includes economic support capacity, time support capacity and spiritual support capacity; the above-mentioned demand for elderly care services dimension includes the expected content of elderly care services, the degree of acceptance of service fees, etc.

[0040] In this embodiment, after collecting the multidimensional customer information of the customer, the staff stores the multidimensional customer information in a customer information database, obtains the multidimensional customer information of the customer from the customer information database, and performs preprocessing operations on the multidimensional customer information. The preprocessing operations include data cleaning and data standardization to obtain multidimensional customer information after the preprocessing operations; the data cleaning refers to cleaning the collected multidimensional customer information, removing duplicate data, and correcting and completing the data (such as checking whether the format of the ID card number and mobile phone number is correct. If not, delete them directly and then complete the deleted data); the data standardization refers to standardizing data of different formats and units in the multidimensional customer information; based on the multidimensional customer information after the preprocessing operation, a customer profile of the customer is constructed, and the customer profile includes social attributes, behavioral preferences, health status and customer labels.

[0041] In this embodiment, by obtaining multi-dimensional customer information of the customer and constructing a customer portrait based on the multi-dimensional customer information, the demand forecast results are calculated based on the user portrait, thereby improving the accuracy of subsequent demand forecasts and thus improving the efficiency and accuracy of risk level confirmation.

[0042] In one embodiment, constructing a customer profile of the customer based on the multi-dimensional customer information includes:

[0043] Extracting features from the multidimensional customer information using a feature extraction algorithm to obtain multidimensional customer features;

[0044] Extracting social attribute features from the multi-dimensional customer features, and constructing a social attribute profile based on the social attribute features;

[0045] extracting behavioral features from the multi-dimensional customer features, and constructing a behavioral preference profile based on the behavioral features;

[0046] extracting health status characteristics from the multi-dimensional customer characteristics, and constructing a health status profile based on the health status characteristics;

[0047] According to the social attribute portrait, the behavioral preference portrait, and the health status portrait, a corresponding label is assigned using a preset labeling rule to obtain a customer label;

[0048] The social attribute portrait, the behavioral preference portrait, the health status portrait and the customer label are aggregated to obtain a customer portrait of the customer.

[0049] In this embodiment, the feature extraction algorithms include but are not limited to Bag of Words, Word Embedding, ng) and feature coding, etc., identify keywords of unstructured data in multidimensional customer information through a feature extraction algorithm, and perform feature extraction based on the identified keywords to obtain a first customer feature, perform feature coding on the structured data to obtain a second customer feature, combine the first customer feature and the second customer feature to obtain a multidimensional customer feature, and classify the multidimensional information features into features of multiple dimensions, including social attribute features, behavioral features, and health status features; according to the features of each dimension, respectively construct corresponding portraits through a pre-trained portrait construction model to obtain a social attribute portrait, a behavioral preference portrait, and a health status portrait, wherein the above-mentioned preset portrait construction model refers to the model type corresponding to the features of each dimension, for example, the social attribute portrait can be constructed by a statistical binning model, the behavioral preference portrait can be constructed by cluster analysis, and the health status portrait can be constructed by a neural network; use a preset labeling rule to assign a label to each portrait respectively to obtain multiple customer labels, and summarize the social attribute portrait, behavioral preference portrait, health status portrait, and customer label to obtain a customer portrait of the customer, wherein the preset labeling rule refers to assigning corresponding labels based on the characteristics of the portrait, for example, if the user's health status portrait is healthy, then a health label is assigned.

[0050] In this embodiment, by extracting features from the collected customer information, multi-dimensional customer features are obtained, and a customer portrait is constructed based on the multi-dimensional customer features. This customer portrait can provide a basis for subsequent calculations of supply and demand risk relationships, thereby improving the efficiency and accuracy of confirming the supply risk level.

[0051] Step S202: Based on the customer portrait, a first clustering algorithm is used to group the customers to obtain a first customer group, and a second clustering algorithm is used to perform a secondary clustering on the first customer group to obtain a second customer group.

[0052] In this embodiment, the matching degree between the service supply and the customer's requirements is predicted through the target risk model, and the target risk model includes a large exposure risk model and a small exposure risk model, wherein the large exposure risk model is used to predict the demand on the demand side, and the small exposure risk model is used to calculate the supply on the supply side. The future demand is predicted based on the large exposure risk model, and the future supply is calculated based on the small exposure risk model. Warning information is issued based on the demand and supply, thereby improving the matching degree between the service supply and customer requirements and reducing supply risks.

[0053] In this embodiment, the customer profile is input into the large exposure risk model of the target risk model, and the customer profile is processed using the large exposure risk model to group the customers using a first clustering algorithm to obtain a first customer group, wherein the first clustering algorithm refers to the K-Means clustering algorithm, and the first customer group includes six groups: healthy, young chronic diseases, old chronic diseases, semi-self-care, disabled, and dementia; these six groups are subjected to agglomerative hierarchical clustering through a second clustering algorithm to obtain a second customer group, wherein the second customer group refers to each group obtained by clustering each category in the first customer group (such as clustering from the healthy customer group to obtain: active sports + high consumption, inactive sports + high consumption, inactive sports + low consumption).

[0054] In this embodiment, by mining deep features through double clustering, customer groups with obvious commonalities can be quickly identified, and then demand forecast results are calculated based on the customer groups, thereby improving the accuracy of demand forecast results and further improving the efficiency and accuracy of confirming the supply risk level.

[0055] In one embodiment, grouping the customers using a first clustering algorithm based on the customer portrait to obtain a first customer group includes:

[0056] pre-defining the types and clustering rules of the first customer group;

[0057] Using a preset screening method to screen out clustering features related to the type of the first customer group from the customer portrait;

[0058] The customers are grouped according to the clustering rule and the clustering characteristics to obtain a first customer group.

[0059] In this embodiment, the types of the first customer group (healthy, low-age chronic diseases, high-age chronic diseases, semi-self-care, disability, and dementia) are pre-defined, and the clustering rules are defined, wherein health refers to having no chronic diseases and being able to take care of oneself; low-age chronic diseases refer to those whose age is less than or equal to the preset age threshold, and who suffer from at least one chronic disease (such as diabetes, etc.) and are completely self-care; high-age chronic diseases refer to those whose age is greater than the preset age threshold, and who suffer from at least one chronic disease and are completely self-care; semi-self-care refers to those who need assistance in some daily life and rely on mild nursing; disability refers to the complete loss of self-care ability; dementia refers to diagnosed cognitive impairment; and the chi-square test is used to analyze the difference between the two groups. The method evaluates the correlation between the customer portrait and the type of the first customer group, and extracts features with correlation greater than a preset correlation threshold from the customer portrait based on the correlation to obtain clustering features; clusters the customers according to the clustering rules and clustering features to obtain the first customer group, specifically excluding customers whose age is younger than the minimum age threshold (such as directly excluding customers who are younger than 18 years old and do not participate in subsequent clustering operations) to obtain screened customers, and calculates the mean of each cluster on the clustering features for the screened customers, and judges which type of the first customer group the customer belongs to based on the calculated value to obtain the first customer group.

[0060] In this embodiment, by pre-defining the type of the first customer group and the first rule, and combining the customer portrait for grouping, customers can be accurately divided into groups with similar characteristics, thereby improving the accuracy of the first customer group clustering; a preset screening method is used to filter out clustering features related to the type of the first customer group from the customer portrait, which can remove irrelevant or redundant information and improve the accuracy and efficiency of clustering; thereby improving the accuracy of subsequent supply and demand relationship predictions.

[0061] In one embodiment, the second clustering algorithm is used to perform secondary clustering on the first customer group to obtain a second customer group, including:

[0062] Using elbow rule and silhouette coefficient cross validation to determine the number of clusters for each category in the first customer group;

[0063] Select the core features of each cluster in each category according to the feature importance analysis method to obtain the target cluster features;

[0064] Each category in the first customer group is grouped according to the target cluster characteristics to obtain the second customer group.

[0065] In this embodiment, in order to prevent misjudgment due to data noise or uneven distribution, the elbow rule and silhouette coefficient cross-validation are used to determine the number of clusters for each category in the first customer group. Specifically, for each category in the first customer group, the K value range (such as K=2-K=10) is traversed, the SSE under each K value is calculated, and a variation curve of SSE with K value is plotted. The K value at the inflection point is selected according to the variation curve to obtain the target K value, and the adjacent K values ​​of the target K value are obtained (such as K=3, then K=2 and K=4 are obtained). The silhouette coefficient calculation formula is used to calculate the silhouette coefficients of the target K value and the adjacent K values ​​respectively. The final K value (such as K=2) is selected from the target K value and the adjacent K values ​​according to the silhouette coefficient, and the final K value is used as the number of clusters for each category in the first customer group; the corresponding K value is added according to the number of clusters of each category. The number of cluster labels is obtained, and the cluster labels are used as the target variables to train a random forest classifier. According to the trained random forest classifier, the importance of the clustering features of each category in the first customer group to the clustering is extracted, and the clustering features are ranked according to the importance to obtain the sorted clustering features. According to the number of clusters obtained above, the top N clustering features are selected (for example, if the number of clusters K = 2, the top 2 clustering features are selected) to obtain the target cluster features (for example, the target cluster features of the health category in the first customer group are: exercise frequency and consumption ability). According to the target cluster features, each category in the first customer group is clustered according to the number of clusters to obtain the second customer group. For example, customer 1: initial group = health, secondary clustering label = high consumption + active exercise, customer 2: initial group = elderly chronic disease, secondary clustering label = multiple complications + high dependence on nursing care.

[0066] In this embodiment, the optimal number of clusters for each category is determined through cross-validation using the elbow rule and silhouette coefficient, which avoids the problem of inaccurate clustering caused by subjectively setting the number of clusters. The most appropriate number of clusters can be automatically determined based on the category of the first customer group, thereby improving the accuracy of clustering of the second customer group. The target cluster features are selected using the feature importance analysis method, which can improve the accuracy and efficiency of clustering. The elbow rule and silhouette coefficient are used in combination to evaluate the clustering effect through the two indicators of SSE (sum of squared errors) and silhouette coefficient, thereby ensuring the stability and reliability of the clustering quality. This in turn improves the efficiency and accuracy of the subsequent confirmation of the supply risk level.

[0067] Step S203: predicting the first service demand of each group in the second customer group based on the time series, combining the first service demand with the second service demand corresponding to the regional growth rate to obtain a demand forecast result for each region.

[0068] In this embodiment, historical order data of all customers in the second customer group is obtained from the internal database of the enterprise. The historical order data includes a customer group identifier (i.e., which customer group in the second customer group the customer belongs to), historical order time, ordered service products, customer region label, etc. The historical order data is preprocessed, and the data preprocessing includes missing value processing, removal of outliers and duplicate data. For example, missing order records can be filled by interpolation, and outliers can be identified and corrected by statistical methods to obtain the historical order data after data preprocessing. The historical order time and service products are extracted from the historical order data after data preprocessing. , divide the customers according to the region they belong to, and based on the divided regions, predict the first service demand of each group in each region at a certain time point in the future, where the first service demand includes the service demand of each service product; screen out customers who have not placed orders from the second customer group according to the customer group identifier in the historical data, divide the customers who have not placed orders into regions, and based on the customers in each region, predict the orders of customers who have not placed orders in the same region according to the pre-trained order prediction model to obtain the first order quantity, combine the first order quantity with the historical growth rate data to obtain the second service demand quantity, and combine the first service demand quantity with the second service demand quantity to obtain the demand forecast result for each region.

[0069] In this embodiment, the demand forecast result for each region is obtained by weightedly summing the first service demand predicted by the time series and the second service demand predicted by the regional growth rate. This can avoid the demand forecast of a single factor being affected by interference factors, which may lead to deviations in the demand forecast results. It can predict the demand of each region more comprehensively and accurately, thereby improving the accuracy of the demand forecast results.

[0070] In one embodiment, predicting the first service demand of each group in the second customer group based on the time series includes:

[0071] Acquire historical order data from the second customer group, and classify the historical order data according to regional labels in the historical order data to obtain at least one regional order data set;

[0072] Based on each of the regions, extracting the order time series of each group in the second customer group from the order data set corresponding to the region;

[0073] Performing stationary analysis and seasonal analysis on the order time series to obtain a time series analysis result;

[0074] A corresponding demand analysis model is called according to the time series analysis result, and a prediction is performed according to the demand analysis model and the historical order data to obtain the first service demand.

[0075] In this embodiment, the historical order data of the second customer group is extracted from the internal database of the enterprise through a pre-written SQL query statement. As mentioned above, the historical order data includes the customer group identifier (i.e., which customer group in the second customer group belongs to), the historical order time, the ordered service product, the customer's region label, etc. The obtained historical order data is classified according to the customer's region label to obtain at least one regional order data set. The above-mentioned customer region label can be divided according to administrative regions, such as provinces, cities, counties (districts), etc., and can also be customized according to the business needs of the enterprise, such as the eastern region, central region, northern region, etc. region, etc.; based on each region, extract the historical order time of the customer and the corresponding service products from the order data set corresponding to the region, sort the historical order data according to the historical order time, and extract the order time series of each group in the second customer group; perform stationarity analysis and seasonality analysis on the order time series to obtain time series analysis results; specifically, the stationarity analysis includes using a data analysis tool to draw a time series graph of the order time series of each customer group, judging whether it is stationary according to the time series graph, or using a statistical test method to perform stationarity analysis on the time series, and the statistical test method includes but is not limited to ADF (Augmented Derivatives) Dickey-Fuller) test, the null hypothesis of the ADF test is that the order time series is non-stationary; the alternative hypothesis is that the order time series is stationary. By calculating the p-value of the test statistic, if the p-value is less than the significance level (such as 0.05), the null hypothesis is rejected and the order time series is judged to be stationary; otherwise, the order time series is considered to be non-stationary; seasonal analysis includes using a preset drawing method (such as using Python's Statsmodels library) to draw the ACF graph and PACF graph of the order time series of each customer group, where the ACF graph can show the autocorrelation coefficient of the order time series at different lag orders, and the PACF graph can show the autocorrelation coefficient after removing the influence of the intermediate lag order. By observing the position and intensity of the peaks in the ACF and PACF graphs, it is determined whether the time series has seasonality.For example, if there is a significant peak at a specific seasonal lag order (such as a 3-month lag) in the ACF diagram, it indicates that the order time series may have quarterly seasonality; the time series analysis results of the order time series are obtained according to the above analysis method; a suitable demand analysis model is selected according to the time series analysis results. Specifically, when the time series analysis results are stable and non-seasonal, the first demand analysis model is selected to predict the historical order data to obtain the first service demand; when the time series analysis results are stable and seasonal, the second demand analysis model is selected to predict the historical order data; the above-mentioned first demand analysis model refers to a pre-trained ARIMA (Auto Regressive Integrated Moving Average) model, which predicts the historical order data through autoregressive (AR) terms, difference (I) terms and moving average (MA) terms to obtain the first service demand; the above-mentioned second demand analysis model refers to a pre-trained SARIMA (Seasonal ARIMA) model, which The ARIMA) model adds seasonal autoregression (SAR), seasonal difference (SI) and seasonal moving average (SMA) terms to the ARIMA model, and uses the second demand analysis model to predict historical order data to obtain the first service demand.

[0076] In this embodiment, the historical order data is classified by the regional labels, and the order time series of the historical order data is analyzed to call the corresponding model for processing, which can improve the accuracy of the first service demand and thereby improve the efficiency and accuracy of the confirmation of the supply risk level.

[0077] In one embodiment, combining the first service demand with the second service demand corresponding to the regional growth rate to obtain a demand forecast result for each region includes:

[0078] Acquiring historical order data from the second customer group, and filtering out customers who have not placed an order from the second customer group based on the historical order data;

[0079] classifying the non-ordering customers according to their regional labels to obtain at least one non-ordering customer data set;

[0080] Based on each of the regions, performing order prediction on the dataset of customers who have not placed an order using an order prediction model to obtain a first order quantity;

[0081] Obtaining historical growth rate data for a preset time period in each of the regions, and calculating a second order quantity based on the historical growth rate data for the preset time period;

[0082] Performing a weighted summation of the first order quantity and the second order quantity according to a preset weight to obtain the second service demand quantity;

[0083] The first service demand amount and the second service demand amount are summed to obtain a demand forecast result for each of the regions.

[0084] In this embodiment, historical order data of a second customer group is extracted from an internal database of the enterprise using a pre-written SQL query statement. Non-ordering customers are screened out from the second customer group based on the historical order data. Non-ordering customers are classified based on the extracted regional labels of the non-ordering customers to obtain a dataset of non-ordering customers in at least one region. Based on each region, order prediction is performed on the dataset of non-ordering customers using a preset trained order prediction model, wherein the order prediction model refers to a neural network model. Prediction features are obtained by extracting feature data of the dataset of non-ordering customers (such as the number of views, customer attributes, necessary demand features, etc.). The prediction features are processed by the order prediction model to obtain a prediction probability. , multiply the predicted probability by the total number of customers in the non-ordering customer data set to obtain the first order quantity; extract the historical growth rate data of each region in a preset time period (such as the past three years) from the enterprise database, calculate the average growth rate of each region based on the historical growth rate data, predict the second order quantity in the future based on the average growth rate, set the weights of the first order quantity and the second order quantity according to business needs or historical experience, and perform weighted summation of the first order quantity and the second order quantity in each region to obtain the second service demand quantity, integrate the first service demand quantity (the demand quantity obtained based on the time series prediction) and the second service demand quantity (the demand quantity obtained based on the non-ordering customer prediction and the growth rate prediction) to obtain the final demand forecast result.

[0085] In this embodiment, the second order quantity corresponding to the growth rate of each region and the first order quantity of the non-ordering customer data set are predicted, and the first order quantity and the second order quantity are weightedly summed to obtain the second service demand quantity. This can more comprehensively consider the impact of different factors, improve the accuracy of the second service demand quantity prediction, and thus improve the accuracy of subsequent supply and demand relationship predictions; and the first service demand quantity and the second service demand quantity are combined to obtain the final demand forecast result, which can combine data from multiple dimensions, improve the accuracy of the demand forecast result, and thus improve the efficiency and accuracy of the subsequent confirmation of the supply risk level.

[0086] Step S204: Acquire service resource data of service supply in each region in real time, and calculate the supply risk level of each region based on the demand forecast result of each region and the service resource data of each region.

[0087] In this embodiment, the supplier system is accessed through the API interface, and service resource data of service supply in each region is obtained from the supplier system in real time. The service resource data refers to information such as cities, counties, resource quantities, dates, and time periods where the supplier can provide services. The demand forecast result obtained in step S203 is integrated with the service resource data for each region to obtain target matching data. The target matching data is calculated using the small exposure risk model in the target risk model in step S202 to obtain the supply risk level of each region.

[0088] In this embodiment, by acquiring the service resource data of each region in real time and comparing and analyzing it with the demand forecast results, the supply risk level of each region can be accurately identified. After calculating the supply risk level of each region, the enterprise can reasonably allocate service resources according to actual conditions, and tilt resources from areas with lower supply risks to areas with higher risks, thereby achieving optimal resource allocation and improving customer loyalty.

[0089] In one embodiment, calculating the supply risk level of each region based on the demand forecast result of each region and the service resource data of each region includes:

[0090] Within a preset calculation period, calculating the difference between the demand forecast result and the service resource data to obtain a supply-demand gap value;

[0091] Obtaining the supply risk level by comparing the supply-demand gap value with a pre-set risk level threshold;

[0092] Using a visualization tool, color-code the supply risk level on a preset map;

[0093] The demand forecast result and the service resource data are updated regularly to obtain updated demand forecast result and updated service resource data, and the supply risk level is updated on the preset map according to the updated demand forecast result and the updated service resource data.

[0094] In this embodiment, the above-mentioned demand forecast results and service resource data are all within a preset calculation period. For example, if the calculation period is within one month, the demand forecast result refers to the demand forecast result for the next month, and the service resource data refers to the service resource data available for the next month. The demand forecast data for each region within the preset calculation period is obtained from the large exposure risk model, and the service resource data of the service supply in each region is obtained in real time from the supplier system. A calculation program is written or a data analysis tool (such as Excel) is used to calculate the service resource data. l. Python's Pandas library) calculates the supply-demand gap value according to the formula "supply-demand gap value = demand forecast result - service resource data". Pre-set thresholds for different supply risk levels based on business experience. For example, the supply risk level can be divided into three levels: low risk, medium risk, and high risk. Comparison programs or data analysis tools are written to compare the supply-demand gap value of each region with the pre-set supply risk level thresholds to obtain comparison results. The comparison results are reviewed and the accuracy of the threshold comparison and risk level determination is ensured through manual sampling of the comparison process in some areas. Ultimately, the supply-demand gap value is determined to which supply-demand risk level it belongs. Pre-set map data is obtained to ensure that the map includes all areas to be assessed and that the boundaries and identification of the areas are clear and accurate. The supply risk level of each region is associated with the corresponding area on the map. A mapping relationship between supply risk level and geographic information is established. Different color labeling rules are set according to the supply risk level. For example, green is used to represent low-risk areas, yellow is used to represent medium-risk areas, and red is used to represent high-risk areas. Visualization tools are used to render the map associated with the supply and demand risk levels, and the risk level of each area is displayed on the map according to the pre-set color labeling rules. Based on the frequency of business changes and the importance of data updates, set a reasonable update cycle (such as once a day). When each update cycle arrives, collect updated demand forecast results from the large exposure risk model and re-acquire updated service resource data from the supplier system to obtain updated demand forecast results and updated service resource data. Based on the updated demand forecast results and updated service resource data, recalculate the supply and demand gap value and determine the supply risk level. Update the risk level color label for each area on the preset map to ensure that the visual display reflects the latest risk status.

[0095] In this embodiment, the difference between the demand forecast results and the service resource data within a preset period is calculated to assign a corresponding supply and demand risk level, and visualized on a preset map. This not only improves the accuracy of the supply and demand risk level marking, but also can intuitively visualize the areas of supply and demand risk levels, thereby improving the accuracy of supply and demand relationship predictions; and by regularly updating the demand forecast results and service resource data, it ensures that the supply risk level reflects the latest market dynamics, and thus the operating strategy can be dynamically adjusted according to the latest risk situation.

[0096] It should be emphasized that in order to further ensure the privacy and security of the multi-dimensional customer information of the above-mentioned customers, the multi-dimensional customer information of the above-mentioned customers can also be stored in a node of a blockchain.

[0097] The blockchain referred to in this application refers to a new application model for computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Blockchain is essentially a decentralized database, a series of data blocks generated using cryptographic methods. Each data block contains information about a batch of network transactions, which is used to verify the validity of the information (to prevent counterfeiting) and generate the next block. Blockchain can include the blockchain underlying platform, the platform product service layer, and the application service layer.

[0098] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to achieve optimal results.

[0099] Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0100] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware via computer-readable instructions. The computer-readable instructions can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes in the above-described method embodiments. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0101] It should be understood that although the steps in the flowcharts of the accompanying drawings are shown in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some of the steps in the flowcharts of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0102] Further references Figure 3 , as a response to the above Figure 2 In order to realize the method shown in FIG, the present application provides an embodiment of a device for confirming the supply risk level. Figure 2 Corresponding to the method embodiment shown, the apparatus can be specifically applied to various computer devices.

[0103] like Figure 3 As shown, the supply risk level confirmation device 300 of this embodiment includes: a portrait construction module 301, a secondary clustering module 302, a demand forecasting module 303, and a risk level calculation module 304. Among them:

[0104] The portrait construction module 301 is used to obtain multi-dimensional customer information of a customer and construct a customer portrait of the customer based on the multi-dimensional customer information.

[0105] In one embodiment, the portrait building module includes:

[0106] A feature extraction submodule, configured to extract features from the multidimensional customer information using a feature extraction algorithm to obtain multidimensional customer features;

[0107] A first portrait construction submodule is configured to extract social attribute features from the multi-dimensional customer features and construct a social attribute portrait based on the social attribute features;

[0108] A second profile building submodule is configured to extract behavioral features from the multi-dimensional customer features and build a behavioral preference profile based on the behavioral features;

[0109] A third portrait construction submodule is configured to extract health status characteristics from the multi-dimensional customer characteristics and construct a health status portrait based on the health status characteristics;

[0110] A label assignment submodule is used to assign corresponding labels according to the social attribute portrait, the behavioral preference portrait, and the health status portrait using preset labeling rules to obtain a customer label;

[0111] The customer portrait construction submodule is used to aggregate the social attribute portrait, the behavioral preference portrait, the health status portrait and the customer label to obtain the customer portrait of the customer.

[0112] The secondary clustering module 302 is configured to cluster the customers according to the customer portraits using a first clustering algorithm to obtain a first customer group, and to perform secondary clustering on the first customer group using a second clustering algorithm to obtain a second customer group.

[0113] In one embodiment, the secondary clustering module includes:

[0114] A definition submodule, used to predefine the types and clustering rules of the first customer group;

[0115] A clustering feature screening submodule, configured to screen out clustering features related to the type of the first customer group from the customer portrait using a preset screening method;

[0116] The first clustering submodule is configured to cluster the customers according to the clustering rules and the clustering characteristics to obtain a first customer group.

[0117] In one embodiment, the secondary clustering module includes:

[0118] a cluster number determination submodule, configured to determine the number of clusters for each category in the first customer group by cross-validation using the elbow rule and the silhouette coefficient;

[0119] The cluster feature selection submodule is used to select the core features of each cluster in each category according to the feature importance analysis method to obtain the target cluster features;

[0120] The second clustering submodule is configured to cluster each category in the first customer group according to the target cluster characteristics to obtain the second customer group.

[0121] The demand forecasting module 303 is used to forecast the first service demand of each group in the second customer group based on time series, and combine the first service demand with the second service demand corresponding to the regional growth rate to obtain a demand forecast result for each region.

[0122] In one embodiment, the demand forecasting module includes:

[0123] a first region classification submodule, configured to obtain historical order data from the second customer group, classify the historical order data according to region labels in the historical order data, and obtain at least one regional order data set;

[0124] A time series extraction submodule is used to extract, based on each of the regions, the order time series of each group in the second customer group from the order data set corresponding to the region;

[0125] The time series analysis submodule is used to perform stationarity analysis and seasonality analysis on the order time series to obtain the time series analysis results;

[0126] The first service demand prediction submodule is used to call the corresponding demand analysis model according to the time series analysis result, and perform prediction based on the demand analysis model and the historical order data to obtain the first service demand.

[0127] In one embodiment, the demand forecasting module includes:

[0128] a non-ordering customer screening submodule, configured to obtain historical order data from the second customer group and screen out non-ordering customers from the second customer group based on the historical order data;

[0129] A second region classification submodule is configured to classify the non-ordering customers according to their region labels to obtain at least one non-ordering customer data set;

[0130] A first order quantity prediction submodule is configured to perform order prediction on the data set of customers who have not placed an order based on each of the regions using an order prediction model to obtain a first order quantity;

[0131] A second order quantity prediction submodule is configured to obtain historical growth rate data for a preset time period in each of the regions, and calculate a second order quantity based on the historical growth rate data for the preset time period;

[0132] A second service demand prediction submodule is configured to perform a weighted summation of the first order quantity and the second order quantity according to a preset weight to obtain the second service demand;

[0133] The demand forecast result calculation submodule is used to sum the first service demand amount and the second service demand amount to obtain the demand forecast result for each of the regions.

[0134] The risk level calculation module 304 is used to obtain the service resource data of the service supply in each region in real time, and calculate the supply risk level of each region based on the demand forecast result of each region and the service resource data of each region.

[0135] In one embodiment, the risk level calculation module includes:

[0136] A supply-demand gap value calculation submodule, configured to calculate the difference between the demand forecast result and the service resource data within a preset calculation period to obtain a supply-demand gap value;

[0137] A supply risk level confirmation submodule is used to compare the supply-demand gap value with a preset risk level threshold to obtain the supply risk level;

[0138] A map marking submodule is used to mark colors according to the supply risk level on a preset map using a visualization tool;

[0139] The supply risk level update submodule is used to regularly update the demand forecast results and the service resource data, obtain updated demand forecast results and updated service resource data, and update the supply risk level on the preset map based on the updated demand forecast results and the updated service resource data.

[0140] In this embodiment, by obtaining multi-dimensional customer information of customers and constructing customer profiles based on the multi-dimensional customer information, the demand forecast results are calculated based on the user profiles, thereby improving the accuracy of subsequent demand forecasts and thus improving the efficiency and accuracy of risk level confirmation.

[0141] By mining deep features through dual clustering, we can quickly identify customer groups with obvious commonalities, and then calculate demand forecast results based on these customer groups, thereby improving the accuracy of demand forecast results and, in turn, the efficiency and accuracy of confirming supply risk levels.

[0142] By weighting and summing the first service demand volume predicted by time series and the second service demand volume predicted by regional growth rate, we can obtain the demand forecast results for each region. This can prevent the demand forecast of a single factor from being affected by interference factors, which may lead to deviations in the demand forecast results. It can more comprehensively and accurately predict the demand of each region, thereby improving the accuracy of the demand forecast results.

[0143] By acquiring service resource data for each region in real time and comparing it with demand forecast results, we can accurately identify the supply risk level in each region. After calculating the supply risk level for each region, enterprises can rationally allocate service resources based on actual conditions, shifting resources from areas with lower supply risks to areas with higher risks, thereby achieving optimal resource allocation and improving customer loyalty.

[0144] In order to solve the above technical problems, the embodiment of the present application also provides a device (computer device). Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0145] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 4 with a memory 41, a processor 42, and a network interface 43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0146] The computer device may be a desktop computer, notebook computer, PDA, cloud server, etc. The computer device may interact with the user via a keyboard, mouse, remote control, touchpad, or voice control device.

[0147] The memory 41 includes at least one type of readable storage medium, including flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic storage, a magnetic disk, an optical disk, etc. In some embodiments, the memory 41 may be an internal storage unit of the computer device 4, such as the hard disk or memory of the computer device 4. In other embodiments, the memory 41 may also be an external storage device of the computer device 4, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash memory card, etc. Of course, the memory 41 may also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions for providing a risk level confirmation method. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.

[0148] In some embodiments, the processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to execute computer-readable instructions or process data stored in the memory 41, such as computer-readable instructions for executing the method for determining the supply risk level.

[0149] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0150] During the implementation of the electronic device of the present application, by obtaining multi-dimensional customer information of the customer and constructing a customer profile based on the multi-dimensional customer information, the demand forecast result is calculated based on the user profile, thereby improving the accuracy of subsequent demand forecasts and thus providing efficiency and accuracy in confirming the risk level;

[0151] By mining deep features through dual clustering, we can quickly identify customer groups with obvious commonalities, and then calculate demand forecast results based on these customer groups, thereby improving the accuracy of demand forecast results and, in turn, the efficiency and accuracy of confirming supply risk levels.

[0152] By weighting and summing the first service demand volume predicted by time series and the second service demand volume predicted by regional growth rate, we can obtain the demand forecast results for each region. This can prevent the demand forecast of a single factor from being affected by interference factors, which may lead to deviations in the demand forecast results. It can more comprehensively and accurately predict the demand of each region, thereby improving the accuracy of the demand forecast results.

[0153] By acquiring service resource data for each region in real time and comparing it with demand forecast results, we can accurately identify the supply risk level in each region. After calculating the supply risk level for each region, enterprises can rationally allocate service resources based on actual conditions, shifting resources from areas with lower supply risks to areas with higher risks, thereby achieving optimal resource allocation and improving customer loyalty.

[0154] The present application also provides another embodiment, namely, providing a storage medium (computer-readable storage medium), wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the supply risk level confirmation method as described above.

[0155] During implementation, the computer-readable storage medium of the present application obtains multi-dimensional customer information of a customer, constructs a customer profile based on the multi-dimensional customer information, and calculates demand forecast results based on the user profile, thereby improving the accuracy of subsequent demand forecasts and thus providing efficiency and accuracy in confirming risk levels.

[0156] By mining deep features through dual clustering, we can quickly identify customer groups with obvious commonalities, and then calculate demand forecast results based on these customer groups, thereby improving the accuracy of demand forecast results and, in turn, the efficiency and accuracy of confirming supply risk levels.

[0157] By weighting and summing the first service demand volume predicted by time series and the second service demand volume predicted by regional growth rate, we can obtain the demand forecast results for each region. This can prevent the demand forecast of a single factor from being affected by interference factors, which may lead to deviations in the demand forecast results. It can more comprehensively and accurately predict the demand of each region, thereby improving the accuracy of the demand forecast results.

[0158] By acquiring service resource data for each region in real time and comparing it with demand forecast results, we can accurately identify the supply risk level in each region. After calculating the supply risk level for each region, enterprises can rationally allocate service resources based on actual conditions, shifting resources from areas with lower supply risks to areas with higher risks, thereby achieving optimal resource allocation and improving customer loyalty.

[0159] The non-Company software tools or components appearing in the embodiments of this application are merely examples and do not represent actual use.

[0160] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0161] Obviously, the embodiments described above are only some of the embodiments of the present application, rather than all of the embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application has been described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions described in the aforementioned specific embodiments, or to make equivalent replacements for some of the technical features therein. Any equivalent structure made using the contents of the present application specification and the accompanying drawings, directly or indirectly used in other related technical fields, is also within the scope of patent protection of the present application.

Claims

1. A method for confirming supply risk level, characterized in that: The method comprises: Acquire multi-dimensional customer information of the customer, and construct a customer profile of the customer based on the multi-dimensional customer information; Based on the customer portraits, the customers are grouped using a first clustering algorithm to obtain a first customer group, and the first customer group is grouped again using a second clustering algorithm to obtain a second customer group; Predicting the first service demand of each group in the second customer group based on the time series, combining the first service demand with the second service demand corresponding to the regional growth rate to obtain a demand forecast result for each region; The service resource data of the service supply in each region is obtained in real time, and the supply risk level of each region is calculated based on the demand forecast results of each region and the service resource data of each region.

2. The method for confirming the supply risk level according to claim 1, wherein: The step of constructing a customer profile of the customer based on the multi-dimensional customer information includes: Extracting features from the multidimensional customer information using a feature extraction algorithm to obtain multidimensional customer features; Extracting social attribute features from the multi-dimensional customer features, and constructing a social attribute profile based on the social attribute features; extracting behavioral features from the multi-dimensional customer features, and constructing a behavioral preference profile based on the behavioral features; extracting health status characteristics from the multi-dimensional customer characteristics, and constructing a health status profile based on the health status characteristics; According to the social attribute portrait, the behavioral preference portrait, and the health status portrait, a corresponding label is assigned using a preset labeling rule to obtain a customer label; The social attribute portrait, the behavioral preference portrait, the health status portrait and the customer label are aggregated to obtain a customer portrait of the customer.

3. The method for confirming the supply risk level according to claim 1, wherein: The step of grouping the customers using a first clustering algorithm based on the customer portrait to obtain a first customer group includes: pre-defining the types and clustering rules of the first customer group; Using a preset screening method to screen out clustering features related to the type of the first customer group from the customer portrait; The customers are grouped according to the clustering rule and the clustering characteristics to obtain a first customer group.

4. The method for confirming the supply risk level according to claim 1, wherein: The second clustering algorithm is used to perform secondary clustering on the first customer group to obtain a second customer group, including: Using elbow rule and silhouette coefficient cross validation to determine the number of clusters for each category in the first customer group; Select the core features of each cluster in each category according to the feature importance analysis method to obtain the target cluster features; Each category in the first customer group is grouped according to the target cluster characteristics to obtain the second customer group.

5. The method for confirming the supply risk level according to claim 1, wherein: The predicting the first service demand of each group in the second customer group based on the time series includes: Acquire historical order data from the second customer group, and classify the historical order data according to regional labels in the historical order data to obtain at least one regional order data set; Based on each of the regions, extracting the order time series of each group in the second customer group from the order data set corresponding to the region; Performing stationary analysis and seasonal analysis on the order time series to obtain a time series analysis result; A corresponding demand analysis model is called according to the time series analysis result, and a prediction is performed based on the demand analysis model and the historical order data to obtain the first service demand.

6. The method for confirming the supply risk level according to claim 1, wherein: The combining of the first service demand with the second service demand corresponding to the regional growth rate to obtain a demand forecast result for each region includes: Acquiring historical order data from the second customer group, and filtering out customers who have not placed an order from the second customer group based on the historical order data; classifying the non-ordering customers according to their regional labels to obtain at least one non-ordering customer data set; Based on each of the regions, performing order prediction on the dataset of customers who have not placed an order using an order prediction model to obtain a first order quantity; Obtaining historical growth rate data for a preset time period in each of the regions, and calculating a second order quantity based on the historical growth rate data for the preset time period; Performing a weighted summation of the first order quantity and the second order quantity according to a preset weight to obtain the second service demand quantity; The first service demand amount and the second service demand amount are summed to obtain a demand forecast result for each of the regions.

7. The method for confirming the supply risk level according to claim 1, wherein: Calculating the supply risk level of each region based on the demand forecast result of each region and the service resource data of each region includes: Within a preset calculation period, calculating the difference between the demand forecast result and the service resource data to obtain a supply-demand gap value; Obtaining the supply risk level by comparing the supply-demand gap value with a pre-set risk level threshold; Using a visualization tool, color-code the supply risk level on a preset map; The demand forecast result and the service resource data are updated regularly to obtain updated demand forecast result and updated service resource data, and the supply risk level is updated on the preset map according to the updated demand forecast result and the updated service resource data.

8. A device for confirming supply risk level, characterized in that: The device comprises: A portrait building module is used to obtain multi-dimensional customer information of a customer and build a customer portrait of the customer based on the multi-dimensional customer information; A secondary clustering module, configured to group the customers using a first clustering algorithm based on the customer portraits to obtain a first customer group, and to perform secondary clustering on the first customer group using a second clustering algorithm to obtain a second customer group; a demand forecasting module configured to forecast the first service demand of each group in the second customer group based on a time series, and to obtain a demand forecast result for each region by combining the first service demand with the second service demand corresponding to the regional growth rate; The risk level calculation module is used to obtain the service resource data of the service supply in each region in real time, and calculate the supply risk level of each region based on the demand forecast results of each region and the service resource data of each region.

9. A computer device, characterized in that: The computer device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method for confirming the supply risk level according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for confirming the supply risk level according to any one of claims 1 to 7 is implemented.

Images