Community operation strategy recommendation method and device, computer equipment and storage medium

Abstract

The invention relates to the technical field of community operation, and discloses a community operation strategy recommendation method and device, computer equipment and a storage medium. The community operation strategy recommendation method comprises the steps of obtaining community attribute data of a to-be-recommended community, and performing prediction processing on the community attribute data through a community strategy prediction model corresponding to the to-be-recommended community to obtain an initial operation strategy; determining an objective function, a decision variable and a constraint condition of a strategy optimization model according to the initial operation strategy; and solving the strategy optimization model based on the target function, the decision variable and the constraint condition, and determining a target operation strategy of the to-be-recommended community according to a solving processing result. According to the invention, real-time dynamic prediction of the community attribute data can be realized based on the community strategy prediction model, and the rationality of the community operation strategy and the strategy making efficiency are improved. Meanwhile, an intelligent recommendation algorithm and an operation planning optimization technology are combined, an optimal community operation strategy putting scheme is customized for the community, and personalized services are achieved.

Description

Technical Field

[0001] The present invention relates to the field of community operation technology, and in particular to a community operation strategy recommendation method, device, computer equipment and storage medium. Background Art

[0002] In urban life, population distribution and living patterns are highly organized, forming a community lifestyle, and community operations have emerged. Community operations are a bridge between enterprises and users. Through community operations, enterprises can better understand user needs and promote business development.

[0003] In the prior art, when formulating community operation strategies, it mainly relies on manual work to formulate strategy plans based on business experience and rules, and select the community scope for strategy deployment. There are many shortcomings in the manual method of formulating community operation strategies: on the one hand, there are many factors that affect community operation strategies, such as community type, user attributes, historical strategy deployment, budget, goals, etc. The manual method of formulating strategies is affected by personal business experience, which is time-consuming and labor-intensive, and is prone to human errors and omissions, resulting in the formulation of unreasonable community operation strategies. On the other hand, when faced with communities that need new deployment strategies, there is no historical deployment strategy as a reference, and it is impossible for humans to formulate appropriate deployment strategies. At the same time, when faced with communities where existing strategies have been deployed, if the effect of the historical deployment strategy is not ideal, it is difficult for humans to make dynamic adjustments in a timely manner according to the actual operating conditions, resulting in unnecessary operational losses. Summary of the invention

[0004] Based on this, it is necessary to provide a community operation strategy recommendation method, device, computer equipment and storage medium to address the above technical issues, so as to solve the problems of low efficiency and unreasonable community operation strategy in manually formulating community operation strategies.

[0005] A community operation strategy recommendation method, comprising: Acquire community attribute data of the community to be recommended, and perform prediction processing on the community attribute data using a community strategy prediction model corresponding to the community to be recommended to obtain an initial operation strategy; Determine the objective function, decision variables and constraints of the strategy optimization model according to the initial operation strategy; The strategy optimization model is solved based on the objective function, decision variables and constraints, and the target operation strategy of the community to be recommended is determined according to the solution result.

[0006] A community operation strategy recommendation device, comprising: A strategy prediction module is used to obtain community attribute data of the community to be recommended, and to perform prediction processing on the community attribute data through a community strategy prediction model corresponding to the community to be recommended to obtain an initial operation strategy; An optimization model determination module, used to determine the objective function, decision variables and constraints of the strategy optimization model according to the initial operation strategy; The target operation strategy determination module is used to solve the strategy optimization model based on the objective function, decision variables and constraints, and determine the target operation strategy of the community to be recommended according to the solution result.

[0007] A computer device includes a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor. When the processor executes the computer-readable instructions, the community operation strategy recommendation method is implemented.

[0008] A computer-readable storage medium stores computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the community operation strategy recommendation method as described above.

[0009] In the above-mentioned community operation strategy recommendation method, device, computer equipment and storage medium, the community operation strategy recommendation method obtains the community attribute data of the community to be recommended, predicts and processes the community attribute data through the community strategy prediction model corresponding to the community to be recommended, and obtains the initial operation strategy; determines the objective function, decision variables and constraints of the strategy optimization model according to the initial operation strategy; solves the strategy optimization model based on the objective function, decision variables and constraints, and determines the target operation strategy of the community to be recommended according to the solution processing result. The community attribute data of the present invention can reflect the various factors affecting the community operation strategy, and the community strategy prediction model transforms the manual decision-making experience into an intelligent recommendation algorithm driven by data and models. Based on the community strategy prediction model, real-time dynamic prediction of community attribute data can be achieved, which improves the rationality of the community operation strategy and the efficiency of strategy formulation. At the same time, the present invention combines the intelligent recommendation algorithm with the operations optimization technology, and the objective function, decision variables and constraints of the strategy optimization model can be dynamically adjusted according to the needs under different application scenarios, and the optimal community operation strategy delivery plan is customized for each community, realizing personalized services. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0011] Figure 1 is a flow chart of a community operation strategy recommendation method in one embodiment of the present invention; Figure 2 is a structural diagram of a community operation strategy recommendation device in one embodiment of the present invention; Figure 3 is a schematic diagram of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

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

[0013] The community operation strategy recommendation method provided in this embodiment can be applied to the application scenario where product merchants or service providers deploy operation strategies in communities, and the operation strategies include content recommendation strategies, advertising deployment strategies, and community activity planning, etc. For example, the pricing strategy adopted by a smart cabinet operator when operating the smart cabinet slots in a community, when deploying smart cabinets in a newly built community or adjusting the prices of smart cabinets that have been deployed in a community, the smart cabinet operation service end will trigger the process of the community operation strategy recommendation method.

[0014] In one embodiment, if Figure 1 As shown, a community operation strategy recommendation method is provided, including the following steps S10-S30. S10: Obtain community attribute data of the community to be recommended, and perform prediction processing on the community attribute data using a community strategy prediction model corresponding to the community to be recommended to obtain an initial operation strategy.

[0015] Understandably, a community to be recommended refers to a community that needs to formulate an operation strategy. A community to be recommended can be a community that has not yet determined an operation strategy, or a community that needs to adopt a new operation strategy. The community attribute data of the community to be recommended is a data set used to describe the various characteristics and conditions of the community to be recommended. Community attribute data includes geographic location, city level, venue type, business data, business district, competitive products, key indicators, and user commonality. These community attribute data may affect the formulation of operation strategies. The community attribute data between different communities are different, and the community operation strategies finally formulated should also be different.

[0016] After obtaining the community attribute data of the community to be recommended, the community attribute data is predicted and processed by the community strategy prediction model corresponding to the community to be recommended, and the initial operation strategy can be obtained. The community strategy prediction model refers to a pre-trained neural network model that uses the community attribute data of the community at the current time to predict the operation strategy of the community at a specific time in the future (such as one week). Since the community attribute data between different communities are different, each community corresponds to a different community strategy prediction model. The initial operation strategy refers to the community operation strategy output by the community strategy prediction model. Since the community strategy prediction model depends on the training samples, the training samples may not meet the expected operation goals. Therefore, the initial operation strategy can be further optimized later.

[0017] In one embodiment, step S10, that is, before the community attribute data is predicted by the community strategy prediction model corresponding to the community to be recommended, includes: S101, determining a target similar community corresponding to the community to be recommended from candidate communities according to the community attribute data; S102, searching the community strategy data table for sample strategy data corresponding to the community to be recommended and the target similar community; S103: training the constructed gradient boosting model according to the sample strategy data to obtain a community strategy prediction model corresponding to the community to be recommended.

[0018] Understandably, before the community attribute data is predicted and processed by the community strategy prediction model corresponding to the community to be recommended, it is necessary to obtain sample data for model training to obtain the community strategy prediction model. In order to obtain a sufficient sample size, not only the sample data of the community to be recommended itself is required, but also the sample data of similar communities to the community to be recommended. Therefore, it is necessary to first determine the target similar community corresponding to the community to be recommended from the candidate communities based on the community attribute data. The candidate community refers to a community that has already implemented an operation strategy and can be used for reference. The target similar community refers to a community similar to the community to be recommended that is determined by comparing the corresponding attributes between the community to be recommended and the candidate community through a similarity measurement method (such as Euclidean distance, cosine similarity, etc.).

[0019] Once the target similarity community is determined, the next step is to find the sample strategy data corresponding to the recommended community and the target similarity community from the community strategy data table. The community strategy data table is a pre-built data table used to record the relationship between the community and historical operation strategy data. The sample strategy data includes the historical operation strategies adopted by the community in a specific historical time period (such as the past month) and the corresponding strategy effect indicator data. The strategy effect indicator data is used to characterize whether the operation strategy has brought improvements after being implemented, including positive and negative values. According to the strategy effect indicator data, positive and negative sample labels can be added to the sample strategy data.

[0020] In one embodiment, model training is performed based on a gradient boosting model, and the constructed gradient boosting model is trained according to sample strategy data to obtain a community strategy prediction model corresponding to the community to be recommended. The gradient boosting model, also known as the gradient boosting decision tree (GBDT), is an iterative decision tree algorithm that uses iterative training of weak classifiers (decision trees) to construct a strong learner, which is suitable for processing data with complex interactions and nonlinear relationships. The gradient boosting model can select a lightweight gradient boosting model (Light GradientBoosting Machine, Light GBM) or an extreme gradient boosting model (eXtreme Gradient Boosting, XGBoost) as needed.

[0021] Before training begins, extract the model features of the sample strategy data based on the community attribute data. The model features include geographic location, city level, venue type, business data, business district, competing products, key indicators, and user commonalities. Then, perform different preprocessing on the extracted model features as needed. For example, check the distribution of positive and negative labels in the sample strategy data to determine whether there is sample imbalance; perform statistical analysis on the model features to determine whether there are outliers, remove abnormal samples and redundant samples; convert model features into one-hot encoding; normalize features; bucketize features; analyze the correlation between features to eliminate redundant features with strong correlation; analyze the combination of features to determine whether the data division ability will be improved, thereby constructing new combination features, etc.

[0022] During the training process, the sample strategy data is divided into a training set and a test set, and the pre-built gradient boosting model is trained using the training set. During the training process, the model will continuously adjust its internal parameters and structure to minimize the loss function, thereby improving the accuracy of future strategy predictions. The loss function of the gradient boosting model is used to quantify the difference between the predicted output of the machine learning algorithm and the actual target value, that is, the loss function is based on the sample processing data as a training sample to evaluate the prediction performance of the neural network model. During the training process, the gradient of the loss function of the back-propagation algorithm relative to the model parameters is used for iterative training. When the loss error is minimized, the iteration is stopped to obtain the community strategy prediction model corresponding to the community to be recommended. The community strategy prediction model can quickly and accurately predict the operation strategy suitable for the community to be recommended, that is, the initial operation strategy, based on real-time community attribute data.

[0023] This embodiment obtains sample strategy data corresponding to the to-be-recommended community and the target similar community based on the target similar community corresponding to the to-be-recommended community, thereby ensuring the richness and diversity of the sample data and helping to improve the model training effect. At the same time, by training the gradient boosting model based on the sample strategy data, a highly personalized community strategy prediction model can be customized for the to-be-recommended community, thereby improving the effectiveness and rationality of the community operation strategy.

[0024] In one embodiment, step S101, that is, determining a target similar community corresponding to the community to be recommended from candidate communities according to the community attribute data, includes: S1011, obtaining candidate community attribute data of each candidate community; S1012, performing similarity calculation processing on the community attribute data and the candidate community attribute data according to community attribute types, to obtain type similarity values ​​of the candidate communities; S1013, performing weighted processing on the type similarity values ​​to obtain weighted similarity values ​​of the candidate communities; S1014: Sort all the candidate communities according to the weighted similarity values, and determine a target similar community corresponding to the to-be-recommended community according to the candidate community list obtained by sorting.

[0025] Understandably, the candidate community attribute data refers to the community attribute data corresponding to the candidate community. The community attribute type refers to the result of classifying the community attribute data according to characteristics. After the community attribute data and the candidate community attribute data are divided according to the community attribute type, they all include data such as geographic location, city level, venue type, business data, business district, competitive products, key indicators and user commonality, but the communities they belong to are different. Among them, the geographic location data is used to characterize the address information of the province, city, and district where the community is located; the city level data is used to characterize the information that the city where the community is located is first-tier, second-tier, third-tier or fourth-tier; the venue type data is used to characterize whether the community is a school district community centered on a school or a business district community centered on a shopping mall; the business data is used to characterize the turnover rate and complaint rate of the community; the business district data is used to characterize the types and number of business districts in the community; the competitive product data is used to characterize the types and number of competitive products in the community; the key indicator data refers to the key data of community users that the operation is concerned about, such as order volume, sales volume, express delivery volume, etc.; the user commonality data is used to characterize the proportion of student users, single users or elderly users in the user population characteristics of the community.

[0026] In one embodiment, when there are multiple candidate communities, the similarity calculation process is performed on the community to be recommended and each candidate community one by one. In the process of calculating the similarity of the community attribute data and the candidate community attribute data, different similarity calculation methods (such as Euclidean distance, cosine similarity, Pearson correlation coefficient, Jaccard similarity coefficient, etc.) are selected according to the community attribute type to obtain the type similarity value of each candidate community. The type similarity value refers to the data used to characterize the similarity between two communities under the same community attribute type, including geographic location similarity value, city level similarity value, venue type similarity value, business data similarity value, business district similarity value, competitor similarity value, key indicator similarity value and user commonality similarity value. For example, when the geographic location data is expressed in the form of coordinates, the Euclidean distance algorithm is used to calculate the geographic location similarity value between the community to be recommended and each candidate community. For example, when the city level data is expressed in the form of a vector, the cosine similarity algorithm is used to calculate the city level similarity value between the community to be recommended and each candidate community.

[0027] In one embodiment, the community attribute type has a preset weighting coefficient. Different community attribute types have different importances and correspond to different weighting coefficients. The sum of the weighting coefficients of all community attribute types is 1. At this time, the type similarity values ​​of the same candidate community are weighted to obtain the weighted similarity value of the candidate community, and then the weighted similarity values ​​of each candidate community are obtained. The weighted similarity value refers to the result of weighted summation of all type similarity values ​​of the same candidate community, and each candidate community corresponds to a weighted similarity value. Specifically, when the corresponding weighted coefficient distribution in the community attribute type is geographical location 0.1, city level 0.1, venue type 0.2, business data 0.2, business district 0.1, competitor 0.1, key indicator 0.1 and user commonality 0.1, the weighted similarity value of a candidate community is geographical location similarity value × 0.1 + city level similarity value × 0.1 + venue type similarity value × 0.2 + business data similarity value × 0.2 + business district similarity value × 0.1 + competitor similarity value × 0.1 + key indicator similarity value × 0.1 + user commonality similarity value × 0.1.

[0028] In one embodiment, all candidate communities are sorted according to the size of the weighted similarity value to obtain a candidate community list. The candidate community list is a list in which the candidate communities are sorted in order from large to small weighted similarity values. Based on the candidate community list, one or more candidate communities with the highest similarity can be selected as the target similarity community corresponding to the community to be recommended. If only one target similarity community is needed, the first one in the list is directly selected; if more than one is needed, the first few in the list can be selected. Specifically, the top K communities (TopK) are selected from the candidate community list, and the selected K communities are determined as the target similarity communities corresponding to the community to be recommended. Among them, the default value of K is set to 5, and it can also be adjusted as needed.

[0029] This embodiment obtains the type similarity value of each candidate community according to the community attribute type, performs weighted processing on the type similarity value to obtain the weighted similarity value, and determines the target similar community based on the sorting result of the weighted similarity value. It can effectively determine the target community that is highly similar to the community to be recommended from the candidate communities, ensure the rationality and comprehensiveness of the similar community, and provide strong support for the subsequent strategy recommendation process.

[0030] In one embodiment, step S102, that is, before searching the community strategy data table for sample strategy data corresponding to the to-be-recommended community and the target similar community, includes: S1021, determining the to-be-recommended community and the candidate community as strategic launch communities, and obtaining historical operation strategy data of the strategic launch communities; S1022, determining a community strategy data group corresponding to each strategy delivery community according to the historical operation strategy data; S1023. Generate the community policy data table according to all the community policy data groups.

[0031] Understandably, before finding the sample strategy data corresponding to the to-be-recommended community and the target-similar community from the community strategy data table, it is necessary to obtain multiple historical operation strategy data from the database and construct the community strategy data table in the strategy and community dimensions. A strategy delivery community refers to a community that has been delivered with a historical operation strategy. The set of to-be-recommended communities and candidate communities includes all communities covered by the operation strategy, and the to-be-recommended communities and candidate communities are determined as strategy delivery communities, and the historical operation strategy data of the strategy delivery communities are obtained. Historical operation strategy data refers to operation strategy data that has been delivered within a specific historical time period (such as 1 month). Each piece of historical operation strategy data includes the historical strategy name, delivery community, delivery time, community geographic location, community city level, community venue type, community business data, business district, competitive products, key indicators, strategy parameters and strategy effect indicators, etc., corresponding to different data table fields. Among them, a strategy delivery community can correspond to multiple historical operation strategies, and a historical operation strategy can also correspond to multiple strategy delivery communities. Taking the strategy delivery community as a unit, according to the mapping relationship between the historical operation strategy data and each data table field, a community strategy data group corresponding to each strategy delivery community can be generated. The community strategy data group is structured data used to record and manage the community and its corresponding historical operation strategy. Finally, the collection of all community strategy data groups is determined as a community strategy data table to facilitate finding all historical operation strategy data corresponding to a known community.

[0032] This embodiment constructs a community strategy data table based on historical operation strategy data according to operation strategy and community dimensions. The community strategy data table can clearly display the historical operation strategy data of each community for subsequent query, comparison and analysis. At the same time, the community strategy data table can add new community strategy data groups when needed, which is easy to use and scalable.

[0033] In one embodiment, in step S102, that is, searching the community strategy data table for sample strategy data corresponding to the community to be recommended and the target similar community, includes: S1024, searching the community strategy data table for first sample strategy data in which the strategy delivery community is the community to be recommended; S1025, searching the community strategy data table for second sample strategy data whose strategy delivery community is the target similarity community; S1026. Determine the sample strategy data according to the first sample strategy data and the second sample strategy data.

[0034] Understandably, in the community strategy data table, according to the data table field of "strategy delivery community", all data corresponding to the community to be recommended are screened out and determined as the first sample strategy data. At the same time, all data corresponding to the target similar community are screened out and determined as the second sample strategy data. The first sample strategy data refers to the historical operation strategy data of the strategy delivery community that includes the community to be recommended, and the second sample strategy data refers to the historical operation strategy data of the strategy delivery community that includes the target similar community. In particular, if the community to be recommended has never been deployed in any operation strategy (such as a newly built community or a community with newly started business), the first sample strategy data is empty. The historical operation strategies in the community strategy data table may have only been deployed in the community to be recommended, or may have only been deployed in the target similar community, or may have been deployed in the community to be recommended and the target similar community at the same time. Therefore, it is necessary to deduplicate the first sample strategy data and the second sample strategy data to obtain the sample strategy data after deduplication.

[0035] This embodiment jointly determines the sample strategy data after searching for the first sample strategy data and the second sample strategy data respectively, thereby avoiding omissions and ensuring the comprehensiveness of the training samples.

[0036] In one embodiment, in step S10, the community attribute data is predicted by a community strategy prediction model corresponding to the community to be recommended to obtain an initial operation strategy, including: S104, performing prediction processing on the community attribute data by using a community strategy prediction model corresponding to the community to be recommended, to obtain a predicted operation strategy and strategy effect indicator information of the predicted operation strategy; S105: Determine the predicted operation strategy with positive strategy effect indicator information as the initial operation strategy.

[0037] Understandably, after the community attribute data corresponding to the community to be recommended is input into the community strategy prediction model, the community strategy prediction model will output multiple predicted operation strategies. Since positive and negative sample labels are added to the sample strategy data according to the strategy effect indicator data when the community strategy prediction model is trained using sample strategy data, the community strategy prediction model will also evaluate each predicted operation strategy, and further output the strategy effect indicator information corresponding to each predicted operation strategy, that is, add positive and negative labels. The predicted operation strategy refers to the prediction result directly output by the community strategy prediction model. The strategy effect indicator information refers to the positive and negative label information corresponding to the predicted operation strategy. In order to ensure the improvement effect of the strategy, the predicted operation strategy with negative strategy effect indicator information is eliminated, and the predicted operation strategy with positive strategy effect indicator information is selected as the initial operation strategy, and the initial operation strategy is subsequently optimized.

[0038] This embodiment selects the predicted operation strategies with positive effects from the predicted operation strategies predicted by the community strategy prediction model and determines them as the initial operation strategies, thereby ensuring the effectiveness of subsequent strategy optimization and reducing the computational complexity of subsequent strategy optimization.

[0039] S20. Determine the objective function, decision variables and constraints of the strategy optimization model according to the initial operation strategy.

[0040] Understandably, since the community strategy prediction model depends on training samples, the training samples may not meet the expected operational goals. Therefore, it is necessary to perform strategy optimization based on the initial operational strategy in the future. The strategy optimization model is a pre-built optimization model for matching operational strategies for the community. In actual business scenarios, the optimal allocation problem can be solved by solving the optimization model. The strategy optimization model can be a model of linear programming, nonlinear programming, dynamic programming, integer programming or system science method, preferably an integer linear programming model. The strategy optimization model includes an objective function, decision variables and constraints. Among them, the objective function refers to an extreme value function to be sought that has a direct relationship with the decision variable, which is used to guide the solution process of the decision variable, such as a benefit function. Decision variables refer to variables to be solved related to constraints and optimization goals involved in the allocation business scenario, such as activity types. Constraints refer to restrictions that decision variables must meet when solving the objective function, such as budget constraints.

[0041] In one embodiment, step S20, i.e., determining the objective function, decision variables and constraints of the strategy optimization model according to the initial operation strategy, includes: S201, obtaining optimization target information of the community to be recommended, and determining the objective function of the strategy optimization model according to the optimization target information; S202: Determine decision variables and constraints of the strategy optimization model according to the strategy name and strategy parameters in the initial operation strategy.

[0042] Understandably, since a community to be recommended may have multiple initial operation strategies, it is necessary to decide what strategy to use and how much the strategy parameters are set to, that is, to determine the objective function, decision variables and constraints of the strategy optimization model. The optimization target information of the community to be recommended refers to the target conditions that need to be achieved by the operation activities of the community to be recommended. The process of determining the objective function of the strategy optimization model according to the optimization target information is the process of converting the optimization target information into the function expression form of the decision variables in the strategy optimization model, and the objective function is the function expression form corresponding to the optimization target information. For example, when the optimization target information is the goal of maximizing the expected return, it indicates that the operation strategy needs to adjust the grid price of the smart cabinet terminal so that the total revenue of the smart cabinet covered by the community to be recommended in a specific time period is maximized. At this time, the objective function can be the max function of revenue minus cost. The strategy name and strategy parameters are necessary elements of each initial operation strategy. The strategy name is a name used to define how the operation strategy is implemented, such as a discount strategy, a voucher strategy, etc. The strategy parameters are parameters used to limit the implementation process of the operation strategy, the discount value of the discount strategy, and the voucher value of the voucher strategy.

[0043] In one embodiment, it is assumed that there are N initial operation strategies, and the strategy parameter is the discount rate and strategy effectiveness indicators , Indicates the sequence number of the initial operation strategy, that is, whether to use the At this time, according to the discount rate and cost conversion formula, the strategy cost corresponding to the discount rate is expressed as (No. The discount rate corresponding to the strategy is ), according to the conversion formula of strategy effect index and income, the strategy income corresponding to the strategy effect index is expressed as (No. The strategy effect indicator corresponding to each strategy is ). The objective function of the strategy optimization model is to maximize profit and is expressed as , which is the maximum value of the policy revenue minus the policy cost. Constraints can be adjusted according to actual business needs. For example, a community can have a maximum number of policies that can be deployed at the same time, budget constraints (such as , represents the lower limit of the cost budget, represents the upper limit of the cost budget), and the range of the minimum convex hull that each community's discount rate and expected strategy effect index need to meet. The minimum convex hull is used to represent the effect improvement range after the operation strategy is implemented in the recommended community. For example, for each strategy name, the minimum convex hull is constructed according to the strategy parameters and the values ​​of the strategy effect index.

[0044] This embodiment determines the objective function of the strategy optimization model according to the optimization target information, and determines the decision variables and constraints according to the strategy name and strategy parameters, ensuring that the strategy optimization process moves in the direction of the optimization target while complying with the constraints in actual operations, reflecting the targeted nature of the strategy optimization.

[0045] S30, solving the strategy optimization model based on the objective function, decision variables and constraints, and determining the target operation strategy of the community to be recommended according to the solution result.

[0046] Understandably, when the policy optimization model is an integer linear programming model, the existing mixed integer linear programming solver can be called to solve the policy optimization model to obtain the decision variable solution that satisfies the constraints and the objective function, and the solution result is expressed as the optimal matching result of the matching operation strategy of the community to be recommended. According to the solution result, the target operation strategy corresponding to the community to be recommended is obtained. The target operation strategy refers to the specific operation strategy and corresponding strategy parameters ultimately adopted by the community to be recommended, such as the strategy name, strategy parameters and budget allocation put in the community to be recommended.

[0047] After obtaining the target operation strategy of the community to be recommended, the target operation strategy is sent to the pre-set operation object for executing the operation strategy plan. For example, the smart cabinet operation service end sends the price discount plan (target operation strategy) of the cabinet slot to the smart cabinet terminal (operation object) within the coverage of the community to be recommended, so that the cabinet terminal updates the slot price according to the price discount plan and completes the slot price adjustment.

[0048] This embodiment obtains the community attribute data of the community to be recommended, and predicts and processes the community attribute data through the community strategy prediction model corresponding to the community to be recommended to obtain the initial operation strategy; determines the objective function, decision variables and constraints of the strategy optimization model according to the initial operation strategy; solves the strategy optimization model based on the objective function, decision variables and constraints, and determines the target operation strategy of the community to be recommended according to the solution processing result. The community attribute data of this embodiment can reflect the various factors that affect the community operation strategy, and the community strategy prediction model transforms the manual decision-making experience into an intelligent recommendation algorithm driven by data and models. Based on the community strategy prediction model, real-time dynamic prediction of community attribute data can be achieved, which improves the rationality of the community operation strategy and the efficiency of strategy formulation. At the same time, this embodiment combines the intelligent recommendation algorithm with the operations optimization technology. The objective function, decision variables and constraints of the strategy optimization model can be dynamically adjusted according to the needs of different application scenarios, and the optimal community operation strategy delivery plan is customized for each community, realizing personalized services.

[0049] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0050] In one embodiment, a community operation strategy recommendation device is provided, and the community operation strategy recommendation device corresponds to the community operation strategy recommendation method in the above embodiment. Figure 2 As shown, the community operation strategy recommendation device includes a strategy prediction module 10, an optimization model determination module 20 and a target operation strategy determination module 30. The functional modules are described in detail as follows: The strategy prediction module 10 is used to obtain community attribute data of the community to be recommended, and to perform prediction processing on the community attribute data through a community strategy prediction model corresponding to the community to be recommended to obtain an initial operation strategy; An optimization model determination module 20, used to determine the objective function, decision variables and constraints of the strategy optimization model according to the initial operation strategy; The target operation strategy determination module 30 is used to solve the strategy optimization model based on the objective function, decision variables and constraints, and determine the target operation strategy of the community to be recommended according to the solution result.

[0051] In one embodiment, the strategy prediction module 10 includes: A target similar community determining unit, configured to determine a target similar community corresponding to the community to be recommended from candidate communities according to the community attribute data; A sample strategy data acquisition unit, used to find sample strategy data corresponding to the to-be-recommended community and the target similar community from the community strategy data table; The community strategy prediction model training unit is used to train the constructed gradient boosting model according to the sample strategy data to obtain a community strategy prediction model corresponding to the community to be recommended.

[0052] In one embodiment, the strategy prediction module 10 further includes: A candidate community attribute data acquisition unit, used to acquire the candidate community attribute data of each candidate community; A similarity calculation processing unit, configured to perform similarity calculation processing on the community attribute data and the candidate community attribute data according to the community attribute type, to obtain a type similarity value of each candidate community; A weighted processing unit, used for performing weighted processing on the type similarity value to obtain a weighted similarity value of each candidate community; The candidate community ranking unit is used to rank all the candidate communities according to the weighted similarity values, and determine a target similar community corresponding to the to-be-recommended community according to the ranked candidate community list.

[0053] In one embodiment, the strategy prediction module 10 further includes: A historical operation strategy data acquisition unit, used to determine the to-be-recommended community and the candidate community as a strategy delivery community, and acquire historical operation strategy data of the strategy delivery community; A community strategy data group generating unit, used to determine the community strategy data group corresponding to each of the strategy delivery communities according to the historical operation strategy data; A community policy data table generating unit is used to generate the community policy data table according to all the community policy data groups.

[0054] In one embodiment, the strategy prediction module 10 further includes: A first sample strategy data searching unit is used to search the community strategy data table for first sample strategy data whose strategy delivery community is the community to be recommended; A second sample strategy data searching unit is used to search the community strategy data table for second sample strategy data whose strategy delivery community is the target similarity community; The sample strategy data determining unit is configured to determine the sample strategy data according to the first sample strategy data and the second sample strategy data.

[0055] In one embodiment, the strategy prediction module 10 further includes: A prediction operation strategy determination unit, configured to perform prediction processing on the community attribute data by using a community strategy prediction model corresponding to the community to be recommended, to obtain a prediction operation strategy and strategy effect indicator information of the prediction operation strategy; The initial operation strategy determining unit is used to determine the predicted operation strategy with positive strategy effect indicator information as the initial operation strategy.

[0056] In one embodiment, the optimization model determination module 20 includes: An objective function determination unit, used to obtain optimization objective information of the community to be recommended, and determine the objective function of the strategy optimization model according to the optimization objective information; The variable and condition determination unit is used to determine the decision variables and constraint conditions of the strategy optimization model according to the strategy name and strategy parameters in the initial operation strategy.

[0057] For the specific definition of the community operation strategy recommendation device, please refer to the definition of the community operation strategy recommendation method above, which will not be repeated here. Each module in the above-mentioned community operation strategy recommendation device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0058] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a readable storage medium and an internal memory. The readable storage medium stores an operating system, computer-readable instructions and a database. The internal memory provides an environment for the operation of the operating system and computer-readable instructions in the readable storage medium. The database of the computer device is used to store data involved in the community operation strategy recommendation method. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer-readable instructions are executed by the processor, a community operation strategy recommendation method is implemented. The readable storage medium provided in this embodiment includes a non-volatile readable storage medium and a volatile readable storage medium.

[0059] In one embodiment, a computer device is provided, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, wherein the processor executes the computer-readable instructions to implement the following steps: Acquire community attribute data of the community to be recommended, and perform prediction processing on the community attribute data using a community strategy prediction model corresponding to the community to be recommended to obtain an initial operation strategy; Determine the objective function, decision variables and constraints of the strategy optimization model according to the initial operation strategy; The strategy optimization model is solved based on the objective function, decision variables and constraints, and the target operation strategy of the community to be recommended is determined according to the solution result.

[0060] In one embodiment, one or more computer-readable storage media storing computer-readable instructions are provided. The readable storage media provided in this embodiment include non-volatile readable storage media and volatile readable storage media. The readable storage media store computer-readable instructions, and when the computer-readable instructions are executed by one or more processors, the following steps are implemented: Acquire community attribute data of the community to be recommended, and perform prediction processing on the community attribute data using a community strategy prediction model corresponding to the community to be recommended to obtain an initial operation strategy; Determine the objective function, decision variables and constraints of the strategy optimization model according to the initial operation strategy; The strategy optimization model is solved based on the objective function, decision variables and constraints, and the target operation strategy of the community to be recommended is determined according to the solution result.

[0061] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions. The computer-readable instructions can be stored in a non-volatile readable storage medium or a volatile readable storage medium. When the computer-readable instructions are executed, they can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0062] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0063] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A community operation strategy recommendation method, characterized in that: include: Acquire community attribute data of the community to be recommended, and perform prediction processing on the community attribute data using a community strategy prediction model corresponding to the community to be recommended to obtain an initial operation strategy; Determine the objective function, decision variables and constraints of the strategy optimization model according to the initial operation strategy; The strategy optimization model is solved based on the objective function, decision variables and constraints, and the target operation strategy of the community to be recommended is determined according to the solution result.

2. The community operation strategy recommendation method according to claim 1, characterized in that: Before the community attribute data is predicted by the community strategy prediction model corresponding to the community to be recommended, the method includes: Determine a target similar community corresponding to the community to be recommended from candidate communities according to the community attribute data; Find out sample strategy data corresponding to the community to be recommended and the target similar community from the community strategy data table; The constructed gradient boosting model is trained according to the sample strategy data to obtain a community strategy prediction model corresponding to the community to be recommended.

3. The community operation strategy recommendation method according to claim 2, characterized in that: The step of determining a target similar community corresponding to the community to be recommended from candidate communities according to the community attribute data includes: Acquiring candidate community attribute data of each candidate community; Performing similarity calculation processing on the community attribute data and the candidate community attribute data according to community attribute types to obtain type similarity values ​​of the candidate communities; Performing weighted processing on the type similarity values ​​to obtain weighted similarity values ​​of the candidate communities; All the candidate communities are sorted according to the weighted similarity values, and a target similar community corresponding to the to-be-recommended community is determined according to the sorted candidate community list.

4. The community operation strategy recommendation method according to claim 2, characterized in that: Before searching the community strategy data table for sample strategy data corresponding to the to-be-recommended community and the target similar community, the method includes: Determine the to-be-recommended communities and candidate communities as strategic launch communities, and obtain historical operation strategy data of the strategic launch communities; Determine, according to the historical operation strategy data, a community strategy data group corresponding to each of the strategy delivery communities; The community policy data table is generated according to all the community policy data groups.

5. The community operation strategy recommendation method according to claim 4, characterized in that: The step of searching the community strategy data table for sample strategy data corresponding to the community to be recommended and the target similar community includes: Find out first sample strategy data in which the strategy delivery community is the community to be recommended from the community strategy data table; Finding second sample strategy data in which the strategy delivery community is the target similar community from the community strategy data table; The sample policy data is determined according to the first sample policy data and the second sample policy data.

6. The community operation strategy recommendation method according to claim 1, characterized in that: The predicting process of the community attribute data by using the community strategy prediction model corresponding to the community to be recommended to obtain the initial operation strategy includes: Performing prediction processing on the community attribute data by using a community strategy prediction model corresponding to the community to be recommended, to obtain a predicted operation strategy and strategy effect indicator information of the predicted operation strategy; The predicted operation strategy for which the strategy effect indicator information is positive is determined as the initial operation strategy.

7. The community operation strategy recommendation method according to claim 1, characterized in that: The objective function, decision variables and constraint conditions of the strategy optimization model are determined according to the initial operation strategy, including: Acquiring optimization target information of the community to be recommended, and determining the objective function of the strategy optimization model according to the optimization target information; The decision variables and constraints of the strategy optimization model are determined according to the strategy name and strategy parameters in the initial operation strategy.

8. A community operation strategy recommendation device, characterized in that: include: A strategy prediction module is used to obtain community attribute data of the community to be recommended, and to perform prediction processing on the community attribute data through a community strategy prediction model corresponding to the community to be recommended to obtain an initial operation strategy; An optimization model determination module, used to determine the objective function, decision variables and constraints of the strategy optimization model according to the initial operation strategy; The target operation strategy determination module is used to solve the strategy optimization model based on the objective function, decision variables and constraints, and determine the target operation strategy of the community to be recommended according to the solution result.

9. A computer device comprising a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, characterized in that: When the processor executes the computer-readable instructions, the community operation strategy recommendation method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing computer-readable instructions, characterized in that: When the computer-readable instructions are executed by one or more processors, the one or more processors execute the community operation strategy recommendation method as described in any one of claims 1 to 7.

Images