Transaction strategy recommendation method and device, computer equipment and storage medium

Abstract

The invention relates to a transaction strategy recommendation method and device, computer equipment and a storage medium. The method comprises the steps of determining a standardized score of a user evaluation index based on enterprise operation capability, upstream cooperation capability and downstream reach capability of a target transaction enterprise where a target user is located; according to the expert score, determining a weighted average score of the user evaluation index; determining a comprehensive index weight of the user evaluation index according to the weighted average score; according to the comprehensive index weight and the standardized score, determining a first-dimension evaluation score of the target user in the enterprise operation capability dimension, a second-dimension evaluation score of the target user in the upstream cooperative capability dimension and a third-dimension evaluation score of the target user in the downstream reach capability dimension; and determining a target label of the target user, and recommending a transaction strategy to the target user based on the target label. According to the scheme, different label users can be accurately distinguished, and the pertinence and effectiveness of strategy recommendation are improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, computer device, and storage medium for recommending trading strategies. Background Technology

[0002] Cigarette retailers, as key connecting nodes between buyers and sellers, directly impact product market performance and brand growth through their operational capabilities and brand development abilities. Current methods for classifying retailers, often based on historical sales volume or user scale, suffer from drawbacks such as limited evaluation dimensions, subjective stratification standards, homogenized operational strategies, severe data silos, and lagging dynamic updates. While some research attempts to introduce more user evaluation indicators, these often suffer from arbitrarily selected indicators, reliance on experience for weighting, a lack of systematic theoretical models and objective methods for determining weights, and low accuracy in user classification results. Therefore, improving the efficiency and accuracy of retail user classification to recommend appropriate transaction strategies is a problem that needs to be addressed. Summary of the Invention

[0003] Therefore, it is necessary to provide a method, device, computer equipment, and storage medium that can improve the classification efficiency and accuracy of retail users, thereby recommending reasonable transaction strategies to retail users, in order to address the above-mentioned technical problems.

[0004] Firstly, this application provides a method for recommending trading strategies, the method comprising:

[0005] Obtain user evaluation metrics for the target users and determine the standardized scores for the user evaluation metrics;

[0006] Obtain the expert scores for the user evaluation indicators, and determine the weighted average score of the user evaluation indicators based on the expert scores.

[0007] The comprehensive index weight of the user evaluation index is determined based on the weighted average score.

[0008] Based on the comprehensive indicator weights and the standardized scores, the target user's first-dimensional evaluation score in the enterprise operation capability dimension, the second-dimensional evaluation score in the upstream collaboration capability dimension, and the third-dimensional evaluation score in the downstream reach capability dimension are determined.

[0009] Based on the evaluation scores of the first dimension, the second dimension, the third dimension, and the evaluation score ranking threshold, the target user's target tag is determined, and a transaction strategy recommendation is made for the target user based on the target tag.

[0010] In one embodiment, determining the standardized score of the user evaluation metric includes:

[0011] The user evaluation indicators are divided into numerical indicators and categorical indicators.

[0012] The numerical indicators are standardized using the Z-score standardization algorithm to determine their scores.

[0013] The categorized indicators are standardized using one-hot encoding technology to determine their scores. The scores of the numerical indicators and the categorized indicators are then used as the standardized scores of the user evaluation indicators.

[0014] In one embodiment, determining the comprehensive index weight of the user evaluation index based on the weighted average score includes:

[0015] The first indicator weight of the user evaluation index is determined by using the range normalization method based on the weighted average score.

[0016] The weight of the second indicator of the user evaluation index is determined by a data-driven algorithm.

[0017] The product of the weight of the first indicator and the weight of the second indicator is used as the comprehensive indicator weight of the user evaluation indicator.

[0018] In one embodiment, the second indicator weight of the user evaluation indicator is determined using a data-driven algorithm, including:

[0019] The proportion of the standardized scores corresponding to the user evaluation indicators is determined by using data-driven algorithms.

[0020] The information entropy of the standardized score is determined based on the sample proportion of the aforementioned indicator.

[0021] The difference coefficient of the standardized score is determined based on the information entropy;

[0022] The difference coefficients are normalized to determine the weight of the second indicator of the user evaluation index.

[0023] In one embodiment, determining the target user's target tag based on the first dimension evaluation score, the second dimension evaluation score, the third dimension evaluation score, and an evaluation score ranking threshold includes:

[0024] The user tags of the target user are determined based on the evaluation scores of the first dimension, the second dimension, the third dimension, and the evaluation score sorting threshold.

[0025] The initial profile coefficient of the target user is determined based on the user tags using the profile coefficient method.

[0026] Based on the initial profile coefficient, invalid tags are determined from the user tags, and the target users corresponding to the invalid tags are designated as users to be verified.

[0027] Based on user coding, obtain the tag classification results of the operation and maintenance personnel for the user to be verified, and determine the tag accuracy rate according to the tag classification results and the user tags corresponding to the user to be verified.

[0028] If the accuracy of the tag is less than a preset accuracy threshold, then the user tags of the target user are adjusted using a grid search method to determine the target tags of the target user.

[0029] In one embodiment, if the label accuracy is less than a preset accuracy threshold, a grid search method is used to adjust the user labels of the target user to determine the target labels of the target user, including:

[0030] Determine the threshold search range and threshold adjustment step size, and determine the candidate threshold based on the threshold search range and threshold adjustment step size;

[0031] Using a grid search method, candidate tags for the target user are determined based on the candidate threshold.

[0032] Based on the user tags, determine the updated profile coefficients of the target user and the verification accuracy of the candidate tags;

[0033] The optimal score threshold is determined from the candidate thresholds based on the updated contour coefficients and the verification accuracy.

[0034] Based on the optimal score threshold, the first dimension evaluation score, the second dimension evaluation score, and the third dimension evaluation score, the user tags of the target user are adjusted to determine the target tags of the target user.

[0035] In one embodiment, the above-mentioned trading strategy recommendation method further includes:

[0036] Using the user code of the target user as the core association key, a data association algorithm is used to clean the heterogeneous data in the heterogeneous data source based on the core association key to determine the valid user data; the heterogeneous data source includes a transaction management database, a terminal information database, and a public account user information database.

[0037] Based on the valid user data, a user view of the target user is generated.

[0038] Secondly, this application also provides a trading strategy recommendation device, the device comprising:

[0039] The standardized score determination module is used to obtain the user evaluation index of the target user and determine the standardized score of the user evaluation index;

[0040] The weighted average score determination module is used to obtain the expert scores of the user evaluation index and determine the weighted average score of the user evaluation index based on the expert scores.

[0041] The comprehensive indicator weight determination module is used to determine the comprehensive indicator weight of the user evaluation indicator based on the weighted average score.

[0042] The evaluation score determination module is used to determine the target user's first-dimensional evaluation score in the enterprise operation capability dimension, the second-dimensional evaluation score in the upstream collaboration capability dimension, and the third-dimensional evaluation score in the downstream reach capability dimension based on the comprehensive indicator weights and the standardized scores.

[0043] The trading strategy recommendation module is used to determine the target user's target tag based on the first dimension evaluation score, the second dimension evaluation score, the third dimension evaluation score, and the evaluation score ranking threshold, and to recommend trading strategies to the target user based on the target tag.

[0044] Thirdly, this application also provides a computer device, the computer device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to perform the following steps:

[0045] Obtain user evaluation metrics for the target users and determine the standardized scores for the user evaluation metrics;

[0046] Obtain the expert scores for the user evaluation indicators, and determine the weighted average score of the user evaluation indicators based on the expert scores.

[0047] The comprehensive index weight of the user evaluation index is determined based on the weighted average score.

[0048] Based on the comprehensive indicator weights and the standardized scores, the target user's first-dimensional evaluation score in the enterprise operation capability dimension, the second-dimensional evaluation score in the upstream collaboration capability dimension, and the third-dimensional evaluation score in the downstream reach capability dimension are determined.

[0049] Based on the evaluation scores of the first dimension, the second dimension, the third dimension, and the evaluation score ranking threshold, the target user's target tag is determined, and a transaction strategy recommendation is made for the target user based on the target tag.

[0050] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0051] Obtain user evaluation metrics for the target users and determine the standardized scores for the user evaluation metrics;

[0052] Obtain the expert scores for the user evaluation indicators, and determine the weighted average score of the user evaluation indicators based on the expert scores.

[0053] The comprehensive index weight of the user evaluation index is determined based on the weighted average score.

[0054] Based on the comprehensive indicator weights and the standardized scores, the target user's first-dimensional evaluation score in the enterprise operation capability dimension, the second-dimensional evaluation score in the upstream collaboration capability dimension, and the third-dimensional evaluation score in the downstream reach capability dimension are determined.

[0055] Based on the evaluation scores of the first dimension, the second dimension, the third dimension, and the evaluation score ranking threshold, the target user's target tag is determined, and a transaction strategy recommendation is made for the target user based on the target tag.

[0056] The aforementioned transaction strategy recommendation method, apparatus, computer equipment, and storage medium acquire user evaluation indicators for target users and determine standardized scores for these indicators; acquire expert scores for the user evaluation indicators and determine a weighted average score based on these scores; determine a comprehensive indicator weight for the user evaluation indicators based on the weighted average score; determine the target user's first-dimensional evaluation score (enterprise operational capability), second-dimensional evaluation score (upstream collaboration capability), and third-dimensional evaluation score (downstream reach capability) based on the comprehensive indicator weight and the standardized scores; determine the target user's target tag based on the first-dimensional evaluation score, second-dimensional evaluation score, third-dimensional evaluation score, and evaluation score ranking threshold; and recommend transaction strategies to the target user based on the target tag. This addresses the problems of current retail user classification methods, which commonly rely on historical sales volume or user scale, resulting in single evaluation dimensions, subjective stratification standards, homogenized operational strategies, severe data silos, and delayed dynamic updates. The above solution first determines and standardizes user evaluation indicators based on the three capabilities of the target user's trading company: operation, upstream collaboration, and downstream reach. Then, it combines expert ratings to obtain a weighted average score for each indicator, thus determining the overall indicator weight. Next, it calculates the user's evaluation score across the three capability dimensions using the overall weight and standardized scores. Finally, it determines the user's target tags based on these dimensional scores and recommends suitable trading strategies accordingly. By determining the overall weight through indicator standardization and expert rating weighting, the objectivity of the evaluation indicators is ensured while incorporating professional experience, making the evaluation scores across the three dimensions more aligned with actual business needs. This approach accurately distinguishes users with different tags, avoids homogenization of trading strategies, and improves the targeting and effectiveness of strategy recommendations. Attached Figure Description

[0057] Figure 1 This is a diagram illustrating the application environment of a transaction strategy recommendation method in one embodiment.

[0058] Figure 2 This is a flowchart illustrating a trading strategy recommendation method in one embodiment;

[0059] Figure 3 This is a flowchart illustrating a method for determining standardized scores for user evaluation metrics in one embodiment.

[0060] Figure 4 This is a flowchart illustrating a method for determining the weights of comprehensive indicators in one embodiment;

[0061] Figure 5 This is a flowchart illustrating a method for determining the target tag of a target user in one embodiment;

[0062] Figure 6This is a structural block diagram of a transaction strategy recommendation device in one embodiment;

[0063] Figure 7 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0065] The trading strategy recommendation method provided in this application can be applied to, for example, Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or placed on a cloud or other network server. Server 104 obtains user evaluation metrics for the target user and determines the standardized score of the user evaluation metrics; obtains expert scores for the user evaluation metrics and determines the weighted average score of the user evaluation metrics based on the expert scores; determines the comprehensive index weight of the user evaluation metrics based on the weighted average score; determines the target user's first-dimensional evaluation score in the enterprise operation capability dimension, the second-dimensional evaluation score in the upstream collaboration capability dimension, and the third-dimensional evaluation score in the downstream reach capability dimension based on the comprehensive index weight and the standardized score; determines the target user's target tag based on the first-dimensional evaluation score, the second-dimensional evaluation score, the third-dimensional evaluation score, and the evaluation score ranking threshold; and recommends transaction strategies for the target user based on the target tag, sending the corresponding item transaction strategy to terminal 102 via the communication network. The terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, and smart in-vehicle systems. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted devices. The server 104 can be implemented using a standalone server or a server cluster consisting of multiple servers.

[0066] In one embodiment, such as Figure 2 As shown, a trading strategy recommendation method is provided. This embodiment illustrates the method applied to a terminal, but it is understood that the method can also be applied to a server, or to a system including both a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0067] S210. Obtain user evaluation metrics for the target users and determine the standardized scores for the user evaluation metrics.

[0068] Among them, target users are cigarette retailers who allow information such as the operational capabilities, upstream collaboration capabilities, and downstream reach capabilities of the target trading company they belong to to be accessed. Operational capabilities, upstream collaboration capabilities, and downstream reach capabilities are the three core dimensions for characterizing cigarette retail users. Operational capabilities focus on the stability, standardization, and hardware support of the user's internal operations, independent of external collaboration, and are capabilities that the customer can control independently. Upstream collaboration capabilities focus on the target user's performance as the company's trading terminal in areas such as sourcing, transaction execution, and transaction receipt collection. Downstream reach capabilities focus on the target user's execution capabilities in areas such as maintaining trading customers and recommending traded items.

[0069] Specifically, user evaluation metrics for target users are determined based on the target transaction company's operational capabilities, upstream collaboration capabilities, and downstream reach capabilities.

[0070] For example, user evaluation indicators for enterprise operational capabilities include the target user's location, operating area, years of operation, business district type, store image, completeness of terminal equipment configuration, compliance rating, and business qualification level of the target trading enterprise. User evaluation indicators for upstream collaboration capabilities include the target user's average monthly number of transaction vouchers, total transaction amount, transaction profit, number of transaction vouchers per box, percentage of key transaction items, timely transaction voucher collection rate, and transaction item turnover rate of the target trading enterprise. User evaluation indicators for downstream reach capabilities include the number of key transaction items displayed for the target user's target trading enterprise, the reservation transaction rate for newly released transaction items, the participation rate of both parties in the transaction, the number of members developed, member activity, party satisfaction, efficiency of transaction item recommendation, quality of market information feedback, and electronic settlement rate. These user evaluation indicators for target users can be stored in an evaluation indicator database.

[0071] For example, such as Figure 3 As shown, the standardized scores for user evaluation metrics are determined, including:

[0072] S2101. Divide user evaluation indicators into numerical indicators and categorical indicators.

[0073] Numerical indicators are those that can be represented by specific numerical values, and the values ​​have mathematical meaning. They can reflect the quantification degree of a customer's characteristics in a certain dimension. Categorical indicators are used to describe a customer's category attribute in a certain dimension. Their values ​​are discrete category labels, and they do not have mathematical meaning. They reflect the qualitative differences in customer characteristics.

[0074] S2102. The numerical indicators are standardized using the Z-score standardization algorithm to determine their scores.

[0075] Among them, the Z-score standardization algorithm can eliminate the differences in dimensions and numerical ranges between different indicators, allowing indicators of different dimensions to be compared and calculated on the same order of magnitude.

[0076] Specifically, the numerical indicators are standardized using the Z-score standardization algorithm, which converts the numerical indicators to a uniform scale to eliminate dimensional differences and determines the standardized score of the numerical indicators.

[0077] S2103. The categorized indicators are standardized using one-hot encoding technology to determine the indicator scores of the categorized indicators. The indicator scores of the numerical indicators and the indicator scores of the categorized indicators are used as the standardized scores of the user evaluation indicators.

[0078] Specifically, one-hot encoding technology is used to convert categorical indicators into machine-recognizable binary vectors, thereby quantifying qualitative indicators and determining standardized scores for categorical indicators. The scores of numerical indicators and categorical indicators are then used as the standardized scores for the user evaluation indicators.

[0079] Understandably, the Z-score standardization algorithm eliminates the dimensional differences between different user evaluation metrics. By calculating the standard deviation of each metric value from the mean of all user evaluation metrics, it transforms each numerical metric into a standardized score with a mean of 0 and a standard deviation of 1. The standardized scores of these numerical metrics are no longer affected by their original dimensions and numerical ranges, allowing them to participate in subsequent experience-driven and data-driven fusion weight calculations on the same order of magnitude. This ensures that the weighted summation of customer value scores across various dimensions for determining user tags is objective and reasonable, avoiding situations where excessively large or small original values ​​of some metrics dominate the user tag determination. One-hot encoding, for categorical metrics lacking mathematical computational meaning, generates a binary vector for each category, eliminating the disorder and incalculability of category labels and outputting quantifiable standardized scores. This provides a consistent data foundation for subsequent metric weight fusion and user tag determination.

[0080] S220. Obtain expert scores for user evaluation metrics, and determine the weighted average score of user evaluation metrics based on the expert scores.

[0081] Specifically, the expert scores for user evaluation metrics, ranging from 0 to 5 points, are obtained from the evaluation metric database. Based on the weight of each expert score, a weighted average score for each user evaluation metric is determined.

[0082] S230. Determine the comprehensive index weight of user evaluation indicators based on the weighted average score.

[0083] For example, the comprehensive index weight of user evaluation indicators can be determined by using the range normalization method based on the weighted average score.

[0084] Preferred, such as Figure 4 As shown, the comprehensive index weights of user evaluation indicators are determined based on the weighted average score, including:

[0085] S2301. Using the range normalization method, determine the weight of the first indicator of the user evaluation index based on the weighted average score.

[0086] Specifically, the weight of the first indicator of the user evaluation index is determined based on formula (1) by using the range normalization method.

[0087] W_j=(S_j-S_min) / (S_max-S_min) (1).

[0088] Where W_j is the pre-set weight of the j-th user evaluation indicator; S_j is the weighted average score of the j-th user evaluation indicator; S_min is the minimum value among the weighted average scores of user evaluation indicators; and S_max is the maximum value among the weighted average scores of user evaluation indicators.

[0089] For example, if the average monthly transaction voucher count S_j=4.3, S_min=1.2, and S_max=4.8, then the final weight of the average monthly transaction voucher count W_j=(4.3-1.2) / (4.8-1.2)=3.1 / 3.6≈0.86; if another user evaluation indicator, the company's operating years, S_j=2.5, then W_j=(2.5-1.2) / 3.6≈0.36, which means that the weight of the average monthly transaction voucher count is higher than the weight of the company's operating years. When determining the user tags of target users, the standardized score corresponding to the average monthly transaction voucher count has a greater influence.

[0090] S2302. Determine the weight of the second indicator of the user evaluation index through a data-driven algorithm.

[0091] Among them, the data-driven algorithm can be the entropy weight method, the standard deviation coefficient method, the coefficient of variation method, or the CRITIC (Criteria Importance Through Intercriteria Correlation) evaluation method.

[0092] For example, the weight of the second indicator of the user evaluation metric is determined through a data-driven algorithm, including:

[0093] Using a data-driven algorithm, the sample proportion of the standardized scores corresponding to user evaluation indicators is determined; the information entropy of the standardized scores is determined based on the sample proportion; the difference coefficient of the standardized scores is determined based on the information entropy; and the difference coefficient is normalized to determine the weight of the second indicator of the user evaluation indicators.

[0094] Here, the indicator sample proportion refers to the ratio of the value of a user evaluation indicator in different samples to the total number of values ​​in all samples. A sample refers to the target users, and a value refers to the standardized value of a specific evaluation indicator for a single retail customer. The data-driven algorithm is the entropy weight method.

[0095] Specifically, the proportion of standardized scores corresponding to user evaluation indicators is determined by data-driven algorithms, and the information entropy of standardized scores is determined based on the proportion of standardized scores according to formula (2).

[0096] (2).

[0097] in, Let k be the information entropy corresponding to the i-th user evaluation metric; k is a constant. Let m be the sample percentage of the i-th user evaluation metric in the q-th sample, and m be the total number of user evaluation metrics.

[0098] The difference coefficient of the standardized score is determined by formula (3) based on the information entropy.

[0099] (3).

[0100] in, Let be the difference coefficient of the standardized score corresponding to the i-th user evaluation indicator.

[0101] Based on formula (4), the difference coefficient is normalized to determine the weight of the second indicator of the user evaluation index.

[0102] (4).

[0103] in, The second indicator weight for user evaluation metrics.

[0104] The above scheme calculates the second weight of user evaluation indicators using the entropy weight method, a data-driven algorithm. Based on the objective data characteristics of standardized scores, it quantifies the ability of indicators to distinguish customer value, avoids the subjective bias of experience-driven weights, and provides a scientific and reliable objective basis for the subsequent integration of experience weights to form a comprehensive weight.

[0105] S2303. The product of the weight of the first indicator and the weight of the second indicator shall be used as the comprehensive indicator weight of the user evaluation indicator.

[0106] For example, the comprehensive index weight of the user evaluation index is determined by formula (5).

[0107] (5).

[0108] in, The comprehensive weighting of user evaluation metrics; The first indicator weight.

[0109] The above scheme determines the comprehensive index weight of user evaluation by multiplying the first index weight obtained by the range normalization method with the second index weight obtained by the data-driven algorithm. This comprehensive index weight takes into account both the empirical rationality and data objectivity of the evaluation process, thereby improving the reliability of the comprehensive index weight.

[0110] It should be noted that, based on the above embodiments, the system automatically performs a model evaluation every quarter. When the market environment changes significantly, the weights are recalculated, and the weights of the first indicator, the second indicator, and the comprehensive indicator are redefined.

[0111] S240. Based on the comprehensive indicator weights and standardized scores, determine the target user's first-dimensional evaluation score in the enterprise operation capability dimension, the second-dimensional evaluation score in the upstream collaboration capability dimension, and the third-dimensional evaluation score in the downstream reach capability dimension.

[0112] For example, the formula for calculating the first dimension evaluation score is shown in formula (6):

[0113] F_i=Σ_{j1∈Corresponding indicators of enterprise operational capabilities}S_ij1×w i ×100 (6).

[0114] F_i represents the first dimension evaluation score; S_ij1 represents the standardized score of the user evaluation indicator corresponding to the enterprise's operational capabilities.

[0115] The formula for calculating the score of the second dimension evaluation is shown in formula (7):

[0116] O_i=Σ_{j2∈Upstream Collaboration Capability Corresponding Index}S_ij2×w i ×100 (7).

[0117] Where O_i is the evaluation score of the second dimension; S_ij2 is the standardized score of the user evaluation index corresponding to the upstream collaboration capability.

[0118] The formula for calculating the third dimension evaluation score is shown in formula (8):

[0119] M_i=Σ_{j3∈Downstream reach capability corresponding index}S_ij3×wi ×100 (8).

[0120] Where M_i is the evaluation score of the third dimension; S_ij3 is the standardized score of the user evaluation index corresponding to the downstream reach capability.

[0121] S250. Based on the evaluation scores of the first dimension, the second dimension, the third dimension, and the evaluation score ranking threshold, determine the target user's target tag, and recommend transaction strategies to the target user based on the target tag.

[0122] The target user's target tags can include Category 1, Category 2, Category 3, and Category 4 users. The ranking threshold for the evaluation score can be s%, where the value of s can be preset according to actual needs, and 5 ≤ s ≤ 50.

[0123] For example, taking cities as units, the evaluation scores of the first dimension, second dimension, and third dimension are sorted in descending order to determine the first dimension score queue, the second dimension score queue, and the third dimension score queue. The top s% and s% of the first dimension evaluation scores in the first dimension score queue are designated as Grade A scores for the first dimension, and the remaining s% are designated as Grade B scores for the first dimension. Similarly, the top s% and s% of the second dimension evaluation scores in the second dimension score queue are designated as Grade A scores for the second dimension, and the remaining s% are designated as Grade B scores for the second dimension. Likewise, the top s% and s% of the third dimension evaluation scores in the third dimension score queue are designated as Grade A scores for the third dimension, and the remaining s% are designated as Grade B scores for the third dimension. If a target user's evaluation scores in all three dimensions are A-level, then the target user is classified as a Category 1 user. If the target user has two A-level scores and one B-level score in these three dimensions, then the target user is classified as a Category 2 user. If the target user has one A-level score and two B-level scores in these three dimensions, then the target user is classified as a Category 3 user. If the target user's evaluation scores in all three dimensions are B-level, then the target user is classified as a Category 4 user. Based on the target tags, a target transaction strategy is determined from the candidate transaction strategies stored in a pre-built recommendation strategy database. The user tag and target transaction strategy are then pushed to the user service system via a RESTful API interface to recommend transaction strategies to the target user. The recommendation strategy database stores candidate tags and candidate transaction strategies, as well as the mapping relationship between candidate tags and candidate transaction strategies.

[0124] The aforementioned transaction strategy recommendation method involves: obtaining user evaluation metrics for target users and determining standardized scores for these metrics; obtaining expert scores for these metrics and determining a weighted average score; determining a comprehensive indicator weight for these metrics based on the weighted average score; determining the target user's evaluation score in three dimensions based on the comprehensive indicator weight and the standardized scores: a first-dimensional evaluation score in enterprise operational capability, a second-dimensional evaluation score in upstream collaboration capability, and a third-dimensional evaluation score in downstream reach capability; determining the target user's target tag based on the first-dimensional evaluation score, the second-dimensional evaluation score, the third-dimensional evaluation score, and a ranking threshold; and recommending transaction strategies to the target user based on these target tags. This method addresses the problems of current retail user classification methods, which commonly rely on historical sales volume or user scale, resulting in single evaluation dimensions, subjective stratification standards, homogenized operational strategies, severe data silos, and delayed dynamic updates. The above solution first determines and standardizes user evaluation indicators based on the three capabilities of the target user's trading company: operation, upstream collaboration, and downstream reach. Then, it combines expert ratings to obtain a weighted average score for each indicator, thus determining the overall indicator weight. Next, it calculates the user's evaluation score across the three capability dimensions using the overall weight and standardized scores. Finally, it determines the user's target tags based on these dimensional scores and recommends suitable trading strategies accordingly. By determining the overall weight through indicator standardization and expert rating weighting, the objectivity of the evaluation indicators is ensured while incorporating professional experience, making the evaluation scores across the three dimensions more aligned with actual business needs. This approach accurately distinguishes users with different tags, avoids homogenization of trading strategies, and improves the targeting and effectiveness of strategy recommendations.

[0125] For example, such as Figure 5 As shown, based on the evaluation scores of the first dimension, the second dimension, the third dimension, and the evaluation score ranking threshold, the target tags for the target users are determined, including:

[0126] S310. Determine the user tags of the target user based on the evaluation scores of the first dimension, the second dimension, the third dimension, and the evaluation score sorting threshold.

[0127] S320. Using the profile coefficient method, determine the initial profile coefficient of the target user based on the user tags.

[0128] For example, the formula for determining the initial contour coefficient is shown in formula (9):

[0129] s(u)=(b(u)-a(u)) / max(a(u),b(u))(9).

[0130] Where a(u) is the average distance from the u-th target user to other target users with different user labels, reflecting the degree of aggregation within the category; b(u) is the minimum average distance from the u-th target user to other target users with different user labels, reflecting the degree of separation between categories.

[0131] S330. Based on the initial profile coefficients, determine invalid tags from the user tags and use the target users corresponding to the invalid tags as users to be verified.

[0132] Specifically, the user labels corresponding to the top 10 target users with the smallest s(i) among all user labels are selected as invalid labels.

[0133] S340. Based on user coding, obtain the label classification results of the users to be verified by the operation and maintenance personnel, and determine the label accuracy rate according to the label classification results and the user labels corresponding to the users to be verified.

[0134] The tag classification results of the users to be verified by the operations and maintenance personnel can be the user tags of the users to be verified determined by the operations and maintenance personnel during offline verification. The user code is a unique identifier for each retail user. Different business systems will record different dimensions of data for the same user. Using the user code as the association key, the data of the same user in different systems can be accurately matched and bound together, avoiding data confusion.

[0135] Specifically, obtain the tag classification results of the users to be verified by the operation and maintenance personnel, determine the consistency between the tag classification results of the users to be verified and the user tags corresponding to the users to be verified, and determine the tag accuracy based on the consistency between the tag classification results of the users to be verified and the user tags corresponding to the users to be verified.

[0136] S350. If the label accuracy is less than the preset accuracy threshold, the grid search method is used to adjust the user labels of the target user and determine the target label of the target user.

[0137] The accuracy threshold can be set according to actual needs, for example, it can be 95%.

[0138] Specifically, if the tag accuracy is less than the preset accuracy threshold of 95%, a grid search method is used to adjust the user tags for the target users and determine the target tags corresponding to the target users. For example, the grid search method traverses all threshold combinations within the parameter space, re-runs the customer segmentation process for each threshold combination, and calculates the average profile coefficient of the corresponding segmentation results. The threshold combination with the highest average profile coefficient is determined as the optimal threshold, automatically replacing the original threshold rules for determining user tags, and regenerating the optimized user tag determination results.

[0139] For example, if the tag accuracy is less than a preset accuracy threshold, a grid search method is used to adjust the user tags of the target user to determine the target tags of the target user, including:

[0140] The threshold search range and threshold adjustment step size are determined, and candidate thresholds are determined based on the threshold search range and threshold adjustment step size. A grid search method is used to determine candidate labels for the target user based on the candidate thresholds. The updated profile coefficients for the target user are determined based on the user labels, and the verification accuracy of the candidate labels is determined. The optimal score threshold is determined from the candidate thresholds based on the updated profile coefficients and the verification accuracy. Based on the optimal score threshold, the first dimension evaluation score, the second dimension evaluation score, and the third dimension evaluation score, the user labels for the target user are adjusted to determine the target label for the target user.

[0141] Specifically, when the accuracy of the target user's user tag corresponding to the tag is lower than a preset threshold, the search range and adjustment step size of the threshold are first defined to generate a series of candidate thresholds. These candidate thresholds are then substituted into the user tag determination process described above, combined with the evaluation scores of the first, second, and third dimensions, to obtain the corresponding candidate tags. The updated profile coefficients corresponding to the candidate tags are calculated, and based on the target user's user tags determined by the operations personnel, the accuracy of the candidate tags is verified. The optimal score threshold is selected from the candidate thresholds using the updated profile coefficients and the accuracy of the candidate tags as screening criteria. Based on the optimal threshold and the evaluation scores of the three dimensions, the original user tags are adjusted to determine the final target tags. For example, if the threshold range is set to [20, 40] and the step size is set to 5, then the candidate threshold values ​​to be traversed are 20, 25, 30, 35, and 40.

[0142] The above scheme generates candidate labels by traversing candidate thresholds within a preset range using a grid search method, and then adjusts user labels by selecting the optimal thresholds based on updated profile coefficients and verification accuracy. This effectively corrects the deviation of the initial labels, improves the matching degree between the labels and the actual value characteristics of the users, and ensures the accuracy and rationality of subsequent trading strategy recommendations.

[0143] For example, based on the above embodiments, the trading strategy recommendation method further includes:

[0144] Using the target user's user code as the core association key, a data association algorithm is used to clean the heterogeneous data in the heterogeneous data source based on the core association key to determine the valid user data. The heterogeneous data source includes a transaction management database, a terminal information database, and a public account user information database. Based on the valid user data, a user view of the target user is generated.

[0145] Heterogeneous data refers to data with different formats, structures, and / or uses across different systems. A user view is a structured data set that reflects the core characteristics of a user, formed by integrating valid user data from heterogeneous data sources, with the user's user code as the core association key.

[0146] Specifically, using the target user's user code as the unique core association key, and with the help of data association algorithms, data is cleaned from three heterogeneous data sources: the transaction management database, the terminal information database, and the public account user information database. Redundant, erroneous, and invalid data are filtered out to determine the valid user data. Based on this valid user data, a user view that can comprehensively reflect the target user's transaction behavior, terminal attributes, rights and interests is generated.

[0147] The above scheme can filter out redundant and invalid information, forming a structured data set that comprehensively reflects the core characteristics of users, providing unified, accurate and reliable data support for subsequent user evaluation and strategy recommendation.

[0148] For example, based on the above embodiments, the trading strategy recommendation method includes:

[0149] Obtain user evaluation metrics from target users. These metrics are categorized into numerical and categorical metrics. Numerical metrics are standardized using the Z-score standardization algorithm to convert them to a uniform scale, eliminating dimensional differences and determining their standardized scores. Categorical metrics are converted into machine-readable binary vectors using one-hot encoding, quantifying qualitative metrics and determining their standardized scores. The scores of both the numerical and categorical metrics are used as the standardized scores for the user evaluation metrics. Expert scores for user evaluation metrics from the evaluation metric database are obtained, ranging from 0 to 5. A weighted average score for each expert score is determined.

[0150] The first indicator weight of the user evaluation metric is determined using the range normalization method and the weighted average score. A data-driven algorithm is then used to determine the sample proportion of the standardized score corresponding to the user evaluation metric; the information entropy of the standardized score is determined based on the sample proportion; the difference coefficient of the standardized score is determined based on the information entropy; and the difference coefficient is normalized to determine the second indicator weight of the user evaluation metric. The product of the first and second indicator weights is used as the comprehensive indicator weight of the user evaluation metric. Based on the comprehensive indicator weight and the standardized score, the target user's evaluation score is determined in three dimensions: first dimension for enterprise operational capabilities, second dimension for upstream collaboration capabilities, and third dimension for downstream reach capabilities.

[0151] Using cities as units, the evaluation scores for the first, second, and third dimensions are sorted in descending order to determine the first, second, and third dimension score queues. The top s% and s% of the first dimension score in the first dimension score queue are designated as Grade A scores, and the remaining s% are designated as Grade B scores. Similarly, the top s% and s% of the second dimension score in the second dimension score queue are designated as Grade A scores, and the remaining s% are designated as Grade B scores. Likewise, the top s% and s% of the third dimension score queue are designated as Grade A scores, and the remaining s% are designated as Grade B scores. If a target user's evaluation scores in all three dimensions are A, then the user is classified as a Category 1 user. If a target user has two A scores and one B score in these three dimensions, then the user is classified as a Category 2 user. If a target user has one A score and two B scores in these three dimensions, then the user is classified as a Category 3 user. If a target user has all three dimensions are B, then the user is classified as a Category 4 user. Using the silhouette coefficient method, initial silhouette coefficients are determined for the target users based on their user tags. Based on these initial silhouette coefficients, invalid tags are identified from the user tags, and the target users corresponding to these invalid tags are designated as users to be verified. The tag classification results of the users to be verified are obtained from the operations and maintenance personnel. The consistency between the tag classification results and the corresponding user tags of the users to be verified is determined, and the tag accuracy is determined based on this consistency.

[0152] When the accuracy of the target user's user tag corresponding to the tag is lower than a preset threshold, the search range and adjustment step size of the threshold are first defined, and a series of candidate thresholds are generated. These candidate thresholds are then substituted into the user tag determination process described above, combined with the first-dimensional evaluation score, the second-dimensional evaluation score, and the third-dimensional evaluation score to obtain the corresponding candidate tags. The updated profile coefficient corresponding to the candidate tags is calculated, and based on the target user's user tags determined by the operations personnel, the accuracy of the candidate tags is verified. The optimal score threshold is selected from the candidate thresholds using the updated profile coefficient and the accuracy of the candidate tags as screening criteria. Based on the optimal threshold and the evaluation scores of the three dimensions, the original user tags are adjusted to determine the final target tags. Based on the target tags, the target transaction strategy is determined from the candidate transaction strategies stored in the pre-built recommendation strategy database. The user tags and target transaction strategies are pushed to the user service system via a RESTful API interface to achieve transaction strategy recommendation for the target user. The recommendation strategy database stores candidate tags and candidate transaction strategies, as well as the mapping relationship between candidate tags and candidate transaction strategies.

[0153] Using the target user's user code as the core association key, a data association algorithm is used to clean the heterogeneous data in the heterogeneous data source based on the core association key to determine the valid user data. The heterogeneous data source includes a transaction management database, a terminal information database, and a public account user information database. Based on the valid user data, a user view of the target user is generated.

[0154] Based on the above embodiments, in addition to the core functions of user stratification evaluation and strategy matching, the trading strategy recommendation system also incorporates a scheduled task configured based on CRON expressions. For example, the task can be configured to run every six months or every year. This scheduled task will automatically trigger a full evaluation of the current stratification model. The evaluation includes calculating the overall average profile coefficient to measure the clustering rationality of the stratification results, statistically analyzing the rationality of the customer proportion of each stratum to determine whether the stratification meets business expectations, and comparing the stability of historical stratification results to detect whether the model has drifted. Through this series of automated evaluation actions, the system can promptly identify the adaptation deviation between the stratification model and business needs, ensuring that the model aligns with business development needs in the long term, and providing stable support for the continuous output of accurate user tags and trading strategies.

[0155] Meanwhile, the trading strategy recommendation system is designed with a dynamic weight adjustment trigger mechanism. This mechanism connects to market environment monitoring data sources such as changes in the regional cigarette trading structure, brand strategy adjustments, and user trading behavior trends to capture significant market change signals in real time, such as a sudden increase or decrease in the sales share of a certain brand or a significant change in the consumption capacity of a regional business district. Once such signal changes are detected, the trading strategy recommendation system automatically triggers the indicator weight recalculation process. It calls the weight allocation and calculation module, re-collects expert scores, and generates new weights adapted to the new market environment using the range normalization method. The new weights then replace the original weights, and the customer segmentation process is re-run. This achieves dynamic iteration of indicator weights, ensuring that the user tag determination method can adapt to dynamic market changes in a timely manner, avoiding the lag in segmentation results caused by changes in the market environment, and improving the timeliness and rationality of trading strategy recommendations.

[0156] The aforementioned transaction strategy recommendation method involves: obtaining user evaluation metrics for target users and determining standardized scores for these metrics; obtaining expert scores for these metrics and determining a weighted average score; determining a comprehensive indicator weight for these metrics based on the weighted average score; determining the target user's evaluation score in three dimensions based on the comprehensive indicator weight and the standardized scores: a first-dimensional evaluation score in enterprise operational capability, a second-dimensional evaluation score in upstream collaboration capability, and a third-dimensional evaluation score in downstream reach capability; determining the target user's target tag based on the first-dimensional evaluation score, the second-dimensional evaluation score, the third-dimensional evaluation score, and a ranking threshold; and recommending transaction strategies to the target user based on these target tags. This method addresses the problems of current retail user classification methods, which commonly rely on historical sales volume or user scale, resulting in single evaluation dimensions, subjective stratification standards, homogenized operational strategies, severe data silos, and delayed dynamic updates. The above solution first determines and standardizes user evaluation indicators based on the three capabilities of the target user's trading company: operation, upstream collaboration, and downstream reach. Then, it combines expert ratings to obtain a weighted average score for each indicator, thus determining the overall indicator weight. Next, it calculates the user's evaluation score across the three capability dimensions using the overall weight and standardized scores. Finally, it determines the user's target tags based on these dimensional scores and recommends suitable trading strategies accordingly. By determining the overall weight through indicator standardization and expert rating weighting, the objectivity of the evaluation indicators is ensured while incorporating professional experience, making the evaluation scores across the three dimensions more aligned with actual business needs. This approach accurately distinguishes users with different tags, avoids homogenization of trading strategies, and improves the targeting and effectiveness of strategy recommendations.

[0157] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated 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 steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0158] Based on the same inventive concept, this application also provides a trading strategy recommendation device for implementing the trading strategy recommendation method described above. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations in one or more trading strategy recommendation device embodiments provided below can be found in the limitations of the trading strategy recommendation method described above, and will not be repeated here.

[0159] In one embodiment, such as Figure 6 As shown, a trading strategy recommendation device is provided, including: a standardized score determination module 601, a weighted average score determination module 602, a comprehensive indicator weight determination module 603, an evaluation score determination module 604, and a trading strategy recommendation module 605, wherein:

[0160] The standardized score determination module 601 is used to obtain the user evaluation index of the target user and determine the standardized score of the user evaluation index;

[0161] The weighted average score determination module 602 is used to obtain the expert scores of the user evaluation index and determine the weighted average score of the user evaluation index based on the expert scores.

[0162] The comprehensive indicator weight determination module 603 is used to determine the comprehensive indicator weight of the user evaluation indicator based on the weighted average score.

[0163] The evaluation score determination module 604 is used to determine the target user's first-dimensional evaluation score in the enterprise operation capability dimension, the second-dimensional evaluation score in the upstream collaboration capability dimension, and the third-dimensional evaluation score in the downstream reach capability dimension based on the comprehensive indicator weight and the standardized score.

[0164] The transaction strategy recommendation module 605 is used to determine the target user's target tag based on the first dimension evaluation score, the second dimension evaluation score, the third dimension evaluation score and the evaluation score ranking threshold, and to recommend transaction strategies to the target user based on the target tag.

[0165] For example, the standardized score determination module 601 is specifically used for:

[0166] The user evaluation indicators are divided into numerical indicators and categorical indicators.

[0167] The numerical indicators are standardized using the Z-score standardization algorithm to determine their scores.

[0168] The categorized indicators are standardized using one-hot encoding technology to determine their scores. The scores of the numerical indicators and the categorized indicators are then used as the standardized scores of the user evaluation indicators.

[0169] For example, the comprehensive indicator weight determination module 603 is specifically used for:

[0170] The first indicator weight of the user evaluation index is determined by using the range normalization method based on the weighted average score.

[0171] The weight of the second indicator of the user evaluation index is determined by a data-driven algorithm.

[0172] The product of the weight of the first indicator and the weight of the second indicator is used as the comprehensive indicator weight of the user evaluation indicator.

[0173] For example, the comprehensive index weight determination module 603 is also specifically used for:

[0174] The proportion of the standardized scores corresponding to the user evaluation indicators is determined by using data-driven algorithms.

[0175] The information entropy of the standardized score is determined based on the sample proportion of the aforementioned indicator.

[0176] The difference coefficient of the standardized score is determined based on the information entropy;

[0177] The difference coefficients are normalized to determine the weight of the second indicator of the user evaluation index.

[0178] For example, the trading strategy recommendation module 605 is specifically used for:

[0179] The user tags of the target user are determined based on the evaluation scores of the first dimension, the second dimension, the third dimension, and the evaluation score sorting threshold.

[0180] The initial profile coefficient of the target user is determined based on the user tags using the profile coefficient method.

[0181] Based on the initial profile coefficient, invalid tags are determined from the user tags, and the target users corresponding to the invalid tags are designated as users to be verified.

[0182] Based on user coding, obtain the tag classification results of the operation and maintenance personnel for the user to be verified, and determine the tag accuracy rate according to the tag classification results and the user tags corresponding to the user to be verified.

[0183] If the accuracy of the tag is less than a preset accuracy threshold, then the user tags of the target user are adjusted using a grid search method to determine the target tags of the target user.

[0184] For example, the trading strategy recommendation module 605 is also specifically used for:

[0185] Determine the threshold search range and threshold adjustment step size, and determine the candidate threshold based on the threshold search range and threshold adjustment step size;

[0186] Using a grid search method, candidate tags for the target user are determined based on the candidate threshold.

[0187] Based on the user tags, determine the updated profile coefficients of the target user and the verification accuracy of the candidate tags;

[0188] The optimal score threshold is determined from the candidate thresholds based on the updated contour coefficients and the verification accuracy.

[0189] Based on the optimal score threshold, the first dimension evaluation score, the second dimension evaluation score, and the third dimension evaluation score, the user tags of the target user are adjusted to determine the target tags of the target user.

[0190] For example, the above-mentioned trading strategy recommendation device further includes:

[0191] The user view generation module is used to use the user code of the target user as the core association key, and through a data association algorithm, to perform data cleaning on heterogeneous data in heterogeneous data sources based on the core association key to determine valid user data; the heterogeneous data sources include a transaction management database, a terminal information database, and a public account user information database; and to generate the user view of the target user based on the valid user data.

[0192] The modules in the aforementioned trading strategy recommendation device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0193] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 7 As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a trading strategy recommendation method. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or a key vector, trackball, or touchpad set on the computer device casing, or an external key vector disk, touchpad, or mouse, etc.

[0194] Those skilled in the art will understand that Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0195] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0196] Step 1: Obtain the user evaluation metrics of the target users and determine the standardized scores of the user evaluation metrics;

[0197] Step 2: Obtain the expert scores for the user evaluation indicators, and determine the weighted average score for the user evaluation indicators based on the expert scores.

[0198] Step 3: Determine the comprehensive index weight of the user evaluation index based on the weighted average score;

[0199] Step 4: Based on the comprehensive indicator weights and the standardized scores, determine the target user's first-dimensional evaluation score in the enterprise operation capability dimension, the second-dimensional evaluation score in the upstream collaboration capability dimension, and the third-dimensional evaluation score in the downstream reach capability dimension.

[0200] Step 5: Determine the target user's target tag based on the first dimension evaluation score, the second dimension evaluation score, the third dimension evaluation score, and the evaluation score sorting threshold, and recommend transaction strategies to the target user based on the target tag.

[0201] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0202] Step 1: Obtain the user evaluation metrics of the target users and determine the standardized scores of the user evaluation metrics;

[0203] Step 2: Obtain the expert scores for the user evaluation indicators, and determine the weighted average score for the user evaluation indicators based on the expert scores.

[0204] Step 3: Determine the comprehensive index weight of the user evaluation index based on the weighted average score;

[0205] Step 4: Based on the comprehensive indicator weights and the standardized scores, determine the target user's first-dimensional evaluation score in the enterprise operation capability dimension, the second-dimensional evaluation score in the upstream collaboration capability dimension, and the third-dimensional evaluation score in the downstream reach capability dimension.

[0206] Step 5: Determine the target user's target tag based on the first dimension evaluation score, the second dimension evaluation score, the third dimension evaluation score, and the evaluation score sorting threshold, and recommend transaction strategies to the target user based on the target tag.

[0207] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps:

[0208] Step 1: Obtain the user evaluation metrics of the target users and determine the standardized scores of the user evaluation metrics;

[0209] Step 2: Obtain the expert scores for the user evaluation indicators, and determine the weighted average score for the user evaluation indicators based on the expert scores.

[0210] Step 3: Determine the comprehensive index weight of the user evaluation index based on the weighted average score;

[0211] Step 4: Based on the comprehensive indicator weights and the standardized scores, determine the target user's first-dimensional evaluation score in the enterprise operation capability dimension, the second-dimensional evaluation score in the upstream collaboration capability dimension, and the third-dimensional evaluation score in the downstream reach capability dimension.

[0212] Step 5: Determine the target user's target tag based on the first dimension evaluation score, the second dimension evaluation score, the third dimension evaluation score, and the evaluation score sorting threshold, and recommend transaction strategies to the target user based on the target tag.

[0213] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0214] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0215] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0216] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A trading strategy recommendation method, characterized in that, include: Obtain user evaluation metrics for the target users and determine the standardized scores for the user evaluation metrics; Obtain the expert scores for the user evaluation indicators, and determine the weighted average score of the user evaluation indicators based on the expert scores. The comprehensive index weight of the user evaluation index is determined based on the weighted average score. Based on the comprehensive indicator weights and the standardized scores, the target user's first-dimensional evaluation score in the enterprise operation capability dimension, the second-dimensional evaluation score in the upstream collaboration capability dimension, and the third-dimensional evaluation score in the downstream reach capability dimension are determined. Based on the evaluation scores of the first dimension, the second dimension, the third dimension, and the evaluation score ranking threshold, the target user's target tag is determined, and a transaction strategy recommendation is made for the target user based on the target tag.

2. The method according to claim 1, characterized in that, Determining the standardized score of the user evaluation indicator includes: The user evaluation indicators are divided into numerical indicators and categorical indicators. The numerical indicators are standardized using the Z-score standardization algorithm to determine their scores. The categorized indicators are standardized using one-hot encoding technology to determine their scores. The scores of the numerical indicators and the categorized indicators are then used as the standardized scores of the user evaluation indicators.

3. The method according to claim 2, characterized in that, Based on the weighted average score, the comprehensive index weight of the user evaluation index is determined, including: The first indicator weight of the user evaluation index is determined by using the range normalization method based on the weighted average score. The weight of the second indicator of the user evaluation index is determined by a data-driven algorithm. The product of the weight of the first indicator and the weight of the second indicator is used as the comprehensive indicator weight of the user evaluation indicator.

4. The method according to claim 3, characterized in that, The second indicator weight of the user evaluation metric is determined using a data-driven algorithm, including: The proportion of the standardized scores corresponding to the user evaluation indicators is determined by using data-driven algorithms. The information entropy of the standardized score is determined based on the sample proportion of the aforementioned indicator. The difference coefficient of the standardized score is determined based on the information entropy; The difference coefficients are normalized to determine the weight of the second indicator of the user evaluation index.

5. The method according to claim 1, characterized in that, Based on the evaluation scores of the first dimension, the second dimension, the third dimension, and the evaluation score ranking threshold, the target tags for the target user are determined, including: The user tags of the target user are determined based on the evaluation scores of the first dimension, the second dimension, the third dimension, and the evaluation score sorting threshold. The initial profile coefficient of the target user is determined based on the user tags using the profile coefficient method. Based on the initial profile coefficient, invalid tags are determined from the user tags, and the target users corresponding to the invalid tags are designated as users to be verified. Based on user coding, obtain the tag classification results of the operation and maintenance personnel for the user to be verified, and determine the tag accuracy rate according to the tag classification results and the user tags corresponding to the user to be verified. If the accuracy of the tag is less than a preset accuracy threshold, then the user tags of the target user are adjusted using a grid search method to determine the target tags of the target user.

6. The method according to claim 5, characterized in that, If the accuracy of the tags is less than a preset accuracy threshold, a grid search method is used to adjust the user tags of the target user to determine the target tags of the target user, including: Determine the threshold search range and threshold adjustment step size, and determine the candidate threshold based on the threshold search range and threshold adjustment step size; Using a grid search method, candidate tags for the target user are determined based on the candidate threshold. Based on the user tags, determine the updated profile coefficients of the target user and the verification accuracy of the candidate tags; The optimal score threshold is determined from the candidate thresholds based on the updated contour coefficients and the verification accuracy. Based on the optimal score threshold, the first dimension evaluation score, the second dimension evaluation score, and the third dimension evaluation score, the user tags of the target user are adjusted to determine the target tags of the target user.

7. The method according to claim 1, characterized in that, Also includes: Using the user code of the target user as the core association key, a data association algorithm is used to clean the heterogeneous data in the heterogeneous data source based on the core association key to determine the valid user data; the heterogeneous data source includes a transaction management database, a terminal information database, and a public account user information database. Based on the valid user data, a user view of the target user is generated.

8. A trading strategy recommendation device, characterized in that, The trading strategy recommendation device includes: The standardized score determination module is used to obtain the user evaluation index of the target user and determine the standardized score of the user evaluation index; The weighted average score determination module is used to obtain the expert scores of the user evaluation index and determine the weighted average score of the user evaluation index based on the expert scores. The comprehensive indicator weight determination module is used to determine the comprehensive indicator weight of the user evaluation indicator based on the weighted average score. The evaluation score determination module is used to determine the target user's first-dimensional evaluation score in the enterprise operation capability dimension, the second-dimensional evaluation score in the upstream collaboration capability dimension, and the third-dimensional evaluation score in the downstream reach capability dimension based on the comprehensive indicator weights and the standardized scores. The trading strategy recommendation module is used to determine the target user's target tag based on the first dimension evaluation score, the second dimension evaluation score, the third dimension evaluation score, and the evaluation score ranking threshold, and to recommend trading strategies to the target user based on the target tag.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images