A budget allocation optimization method based on big data analysis

Abstract

The present invention discloses a budget allocation optimization method based on big data analysis, which specifically relates to the technical field of budget management. The method collects historical data of budget expenditures from multiple channels, constructs a basic budget allocation data set, extracts budget execution difference characteristics, uses a clustering algorithm to perform group identification, and determines key feature areas; combines external economic indicator data to extract budget sensitivity characteristics; analyzes the budget sensitivity characteristics based on a random forest algorithm to determine the influence weights of the budget sensitivity characteristics; establishes an initial model for budget allocation optimization, and optimizes the combination of budget resource configuration parameters; generates a multi-scenario simulation plan for budget configuration, and evaluates the expected effect of budget execution; selects the optimal budget configuration plan based on the expected effect of budget execution and a preset budget execution efficiency threshold, and outputs a budget optimization allocation strategy, thereby realizing dynamic optimization of the budget allocation process and being suitable for budget management in a complex multi-channel environment.

Description

Technical Field

[0001] The present invention relates to the technical field of budget management, and more specifically, to a budget allocation optimization method based on big data analysis. Background Art

[0002] Existing budget allocation processes generally rely on empirical rules, lacking in-depth analysis of multi-channel historical data and the ability to dynamically respond to changes in the external economic environment. This makes it difficult to promptly identify key influencing factors and adjust budget resources. Traditional budget allocation methods often overlook the nonlinear relationship between budget execution performance and external indicators and lack the ability to intelligently mine and model big data. This results in inefficient and poorly adaptable budget allocation plans, making it difficult to achieve optimal resource allocation. Summary of the Invention

[0003] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a budget allocation optimization method based on big data analysis to solve the problems raised in the above-mentioned background technology.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A budget allocation optimization method based on big data analysis includes the following steps:

[0006] S1: Collect historical data on budget expenditures from multiple channels, build a basic data set for budget allocation, and extract budget execution difference characteristics;

[0007] S2: Group and identify budget execution difference features based on clustering algorithms to determine key feature areas;

[0008] S3: Combine key characteristic regions and external economic indicator data to extract budget sensitivity features;

[0009] S4: Use the random forest algorithm to analyze the budget sensitivity characteristics and determine the impact weight of the budget sensitivity characteristics;

[0010] S5: Establish an initial budget allocation optimization model based on the influence weights of budget sensitivity characteristics, and optimize the parameter combination of budget resource allocation through particle swarm optimization;

[0011] S6: Based on the optimized parameter combination of budget resource allocation, generate multi-scenario simulation plans for budget allocation and evaluate the expected budget execution effects of multiple scenarios;

[0012] S7: Based on the expected budget execution effects in multiple scenarios and the preset budget execution efficiency threshold, the optimal budget configuration plan is selected and the budget optimization allocation strategy is output.

[0013] In a preferred embodiment, S1 is specifically:

[0014] Collect historical budget expenditure data from different budget expenditure channels;

[0015] Arrange historical budget expenditure data in chronological order to construct a basic dataset for budget allocation;

[0016] Calculate the difference between actual expenditure and planned budget expenditure during budget execution based on the budget allocation basic data set;

[0017] Calculate budget execution variance characteristics based on variance data.

[0018] In a preferred embodiment, S2 is specifically:

[0019] Conduct cluster analysis based on budget execution difference characteristics;

[0020] Dividing the budget execution difference characteristics into multiple cluster groups, and determining a cluster center and budget execution difference characteristic data points within each cluster group;

[0021] Calculate the Euclidean distance between the budget execution difference feature data point and the corresponding cluster center within each cluster group;

[0022] Determine the dispersion of budget execution difference feature data points based on Euclidean distance;

[0023] The cluster group with the largest dispersion is defined as the key feature area.

[0024] In a preferred embodiment, S3 is specifically:

[0025] Obtain all budget execution difference feature data points within the key feature area;

[0026] Collect data on external economic indicators related to budget execution;

[0027] Perform data association processing on the budget execution difference feature data points and external economic indicator data according to the corresponding time interval to generate a fused data set;

[0028] Perform correlation analysis on the fused data set, calculate the correlation coefficient between the budget execution difference characteristic data points and the external economic indicator data, and determine the budget sensitivity characteristics.

[0029] In a preferred embodiment, S4 is specifically:

[0030] Establish the input dataset of the random forest algorithm based on the budget sensitivity feature;

[0031] Randomly sample the input data set to generate several training data subsets, and use each training data subset to generate several decision trees;

[0032] Based on the generated decision trees, feature importance analysis is performed on the budget sensitivity features respectively;

[0033] Calculate the number of feature splits and the information gain value of the split nodes for the budget sensitivity feature in each decision tree;

[0034] The comprehensive importance score of the budget sensitivity feature is determined based on the number of feature splits and the information gain value of the split node;

[0035] Based on the comprehensive importance score of budget sensitivity features, the impact weight of budget sensitivity features is determined.

[0036] In a preferred embodiment, S5 is specifically:

[0037] Construct an initial model for budget allocation optimization based on the impact weights of budget sensitivity characteristics;

[0038] Based on the budget allocation, the initial model is optimized and the search space of the particle swarm algorithm is established;

[0039] Continuously iterate and update in the search space through particle swarm algorithm;

[0040] The particle swarm algorithm is terminated when the maximum number of iterations is reached or the stability threshold condition of the overall historical optimal position of the particle swarm is met.

[0041] The budget resource configuration parameter combination corresponding to the overall historical optimal position of the particle swarm is output as the optimized budget resource configuration parameter combination.

[0042] In a preferred embodiment, S6 is specifically:

[0043] Construct a set of multi-scenario simulation solutions for budget configuration based on the optimized combination of budget resource configuration parameters;

[0044] Build a simulation computing environment for a set of multi-scenario simulation solutions configured with a budget;

[0045] Based on the simulation computing environment, the expected budget execution effect corresponding to each budget expenditure channel under each multi-scenario simulation plan is calculated respectively;

[0046] Each multi-scenario simulation scheme in the multi-scenario simulation scheme set is evaluated according to the expected effect of budget execution, and the evaluation index corresponding to each multi-scenario simulation scheme is output.

[0047] In a preferred embodiment, S7 is specifically:

[0048] Compare the evaluation indicators of each multi-scenario simulation plan with the preset budget execution efficiency threshold to determine whether each multi-scenario simulation plan meets the budget execution efficiency threshold;

[0049] Select the multi-scenario simulation scheme with the highest evaluation index from the multi-scenario simulation schemes that reach the budget execution efficiency threshold, and define it as the optimal budget configuration scheme;

[0050] Output the budget resource configuration parameter combination contained in the optimal budget configuration plan as the budget optimization allocation strategy.

[0051] Technical effects and advantages of the budget allocation optimization method based on big data analysis of the present invention:

[0052] By collecting historical data on budget expenditures from multiple channels and constructing a basic data set for budget allocation, we can accurately extract the characteristics of budget execution differences, thereby improving the comprehensiveness and basic accuracy of data analysis; by identifying key feature areas through clustering algorithms, we can effectively locate areas with large fluctuations in budget execution efficiency; by combining external economic indicators to extract budget sensitivity characteristics, we can enhance the ability to respond to changes in the external environment; by using a random forest algorithm to analyze the impact weights of budget sensitivity characteristics, we can improve the accuracy of feature evaluation; by establishing an initial budget allocation optimization model based on the impact weights of budget sensitivity characteristics and using a particle swarm algorithm to optimize budget resource allocation, we can improve the global optimality of the budget plan; through multi-scenario simulation and simulation evaluation, we can give the budget plan dynamic adaptability; by comparing with the budget execution efficiency threshold to select the optimal plan, we can achieve optimal budget allocation decisions and improve the flexibility of budget allocation and resource utilization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a schematic diagram of a budget allocation optimization method based on big data analysis in the present invention. DETAILED DESCRIPTION

[0054] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0055] Example

[0056] Figure 1 The present invention provides a budget allocation optimization method based on big data analysis, which includes the following steps:

[0057] S1: Collect historical data on budget expenditures from multiple channels, build a basic data set for budget allocation, and extract budget execution difference characteristics;

[0058] S2: Group and identify budget execution difference features based on clustering algorithms to determine key feature areas;

[0059] S3: Combine key characteristic regions and external economic indicator data to extract budget sensitivity features;

[0060] S4: Use the random forest algorithm to analyze the budget sensitivity characteristics and determine the impact weight of the budget sensitivity characteristics;

[0061] S5: Establish an initial budget allocation optimization model based on the influence weights of budget sensitivity characteristics, and optimize the parameter combination of budget resource allocation through particle swarm optimization;

[0062] S6: Based on the optimized parameter combination of budget resource allocation, generate multi-scenario simulation plans for budget allocation and evaluate the expected budget execution effects of multiple scenarios;

[0063] S7: Based on the expected budget execution effects in multiple scenarios and the preset budget execution efficiency threshold, the optimal budget configuration plan is selected and the budget optimization allocation strategy is output.

[0064] S1: Collect historical data on budget expenditures from multiple channels, build a basic dataset for budget allocation, and extract budget execution difference characteristics, including:

[0065] Collect historical budget expenditure data from different budget expenditure channels;

[0066] Different budget expenditure channels include, but are not limited to, search engine advertising, social media advertising, mobile in-app advertising, and online video advertising. The advertising platform collects historical budget expenditure data from these channels. This historical budget expenditure data includes the budget amount, budget consumption period, and budget expenditure type. The budget amount represents the total amount actually invested in different advertising channels during a specific period. The budget consumption period records the start and end dates of the actual budget, for example, the start and end dates of each month. The budget expenditure type records the advertising format or advertising resource used for the budget expenditure, such as brand exposure ads and performance-based ads in video advertising channels. The advertising platform automatically accesses the APIs exposed by each channel to obtain historical budget expenditure data from the backend database or advertising management system of each advertising channel. The APIs exposed by each channel are exposed as web interfaces and return data in a standard format. The advertising platform system parses and stores this data in a standardized format according to its own unified data specifications to ensure the timeliness, accuracy, and effectiveness of data acquisition.

[0067] Arrange historical budget expenditure data in chronological order to construct a basic dataset for budget allocation;

[0068] To ensure the continuity and completeness of the budget allocation basic dataset, the advertising platform sorts the collected historical budget expenditure data from different budget expenditure channels according to the start and end dates of the budget expenditures. Specifically, the advertising platform arranges the budget expenditure data for each channel in descending order of budget consumption time periods to construct a budget allocation basic dataset with a time series structure. This sorted budget allocation basic dataset ensures the temporal nature of the data and the integrity and accuracy of the data for each channel.

[0069] Calculate the difference between actual expenditure and planned budget expenditure during budget execution based on the budget allocation basic data set;

[0070] Actual expenditure refers to the actual expenditure incurred by each channel during each time period, as recorded by the advertising platform. Planned budget expenditure refers to the pre-planned budget expenditure for the corresponding period. The advertising platform calculates actual expenditure based on the budget amounts recorded in historical budget expenditure data. Planned budget expenditure is derived from the advertiser's pre-set budget plan data. This process is repeated for each channel during each budget period, generating a data set of differences. This data set reflects the degree of deviation between actual budget execution and the advertiser's planned expectations.

[0071] Calculate budget execution variance characteristics based on variance data;

[0072] Budget execution variance characteristics include the mean, variance, and coefficient of variation of the variance data. The advertising platform performs statistical calculations on the variance data set for each channel's corresponding period: the mean is calculated by dividing the sum of all period variance data within the channel by the number of periods; the variance is calculated by taking the sum of the squares of the difference between each period's variance data and the mean, divided by the number of periods; and the coefficient of variation is calculated by dividing the standard deviation of the variance data set by the mean, where the standard deviation is the square root of the variance. This method is used to calculate budget execution variance characteristics for each channel, such as search engine advertising channels and social media advertising channels.

[0073] S2: Group and identify budget execution difference features based on a clustering algorithm to determine key feature areas, including:

[0074] Conduct cluster analysis based on budget execution difference characteristics;

[0075] During cluster analysis, the advertising platform uses budget execution variance characteristics as the basic unit of clustering, classifying them based on the degree of numerical similarity. Using the K-means clustering algorithm, the platform iterates multiple times to group budget execution variance characteristics with similar values ​​into the same category, gradually aggregating and grouping budget channels with similar budget execution results.

[0076] Dividing the budget execution difference characteristics into multiple cluster groups, and determining a cluster center and budget execution difference characteristic data points within each cluster group;

[0077] The clustering algorithm first performs a preliminary division of the budget execution difference characteristics based on their numerical values, dividing them into several distinct cluster groups. Within each cluster group, the budget execution difference characteristic data points have high numerical similarity. Taking the search engine advertising channel as an example, after division, one cluster group may contain budget execution difference characteristic data points with a low mean difference and small variance; another cluster group may contain budget execution difference characteristic data points with a high mean difference and large variance. After division, the cluster center is determined. The cluster center is the arithmetic mean of the numerical values ​​of all budget execution difference characteristic data points within each cluster group, thus forming a cluster center. Each budget execution difference characteristic data point belongs to the corresponding cluster group and is distributed around the corresponding cluster center.

[0078] Calculate the Euclidean distance between the budget execution difference feature data point and the corresponding cluster center within each cluster group;

[0079] The Euclidean distance calculation process is as follows: For each budget execution difference feature data point within each cluster group, the difference between the budget execution difference feature data point and the cluster center of the cluster group is calculated. The difference is then squared and summed, and the square root of the resulting sum is taken to obtain the Euclidean distance between the budget execution difference feature data point and the corresponding cluster center. This calculation process is repeated for each budget execution difference feature data point within each cluster group to obtain a complete Euclidean distance data set.

[0080] Determine the dispersion of budget execution difference feature data points based on Euclidean distance;

[0081] The dispersion of budget execution difference feature data points is measured by the mean of the Euclidean distances within the cluster groups. This is calculated by adding the Euclidean distance values ​​of all budget execution difference feature data points within a cluster group and dividing the sum by the total number of budget execution difference feature data points within the cluster group. The result of this calculation reflects the average dispersion of budget execution difference feature data points relative to the cluster center. A larger value indicates a greater dispersion of budget execution difference feature data points within the cluster group. By calculating the dispersion of budget execution difference feature data points across all cluster groups, a numerical value reflecting the dispersion within all cluster groups is obtained, facilitating comparison and analysis of differences in budget execution difference features between cluster groups.

[0082] The cluster group with the largest dispersion is defined as the key feature area;

[0083] By calculating the dispersion of the budget execution difference feature data points for all cluster groups, the cluster group with the greatest dispersion is determined. This cluster group with the greatest dispersion is defined as the key feature region. The key feature region includes all budget execution difference feature data points and their corresponding cluster centers within the cluster group with the greatest dispersion.

[0084] S3: Combine key feature regions and external economic indicator data to extract budget sensitivity features, including:

[0085] Obtain all budget execution difference feature data points within the key feature area;

[0086] Key feature areas are the most dispersed clusters in cluster analysis. Data points within these clusters include budget execution differences across budget spending channels, including search engine advertising, social media advertising, mobile in-app advertising, and online video advertising, within the corresponding period.

[0087] Collect data on external economic indicators related to budget execution;

[0088] External economic indicator data refers to economic indicators collected by advertising platforms from publicly available external data sources that may be correlated with budget execution results. This includes, but is not limited to, macroeconomic growth rate data, inflation rate data, and other indicators that reflect changing trends in the economic environment. The collection cycle for external economic indicator data aligns with the corresponding cycle for budget execution variance data to ensure consistency and alignment across the data timeline.

[0089] Perform data association processing on the budget execution difference feature data points and external economic indicator data according to the corresponding time interval to generate a fused data set;

[0090] The data association process involves establishing a clear correspondence between budget execution variance data points and external economic indicators based on time periods. For example, an advertising platform might associate a budget execution variance data point for a search engine advertising channel for a particular month with the corresponding macroeconomic growth rate data for that month. Each data point in the fused data set contains both budget execution variance data and external economic indicator data for the same period.

[0091] Conduct correlation analysis on the fused data set, calculate the correlation coefficient between the budget execution difference characteristic data points and the external economic indicator data, and determine the budget sensitivity characteristics;

[0092] Correlation analysis determines the strength of the correlation between budget execution difference characteristic data points and external economic indicator data by fusing the data set. The correlation coefficient is calculated by calculating the covariance between the numerical values ​​of the budget execution difference characteristic data points and the numerical values ​​of the external economic indicator data, then calculating the standard deviation of the numerical values ​​of the budget execution difference characteristic data points and the numerical standard deviation of the external economic indicator data, and dividing the calculated covariance by the product of the two standard deviation values ​​to obtain the correlation coefficient between the budget execution difference characteristic data points and the external economic indicator data. For example, the advertising platform calculates the covariance between the average value of the budget execution difference characteristic of the search engine advertising channel and the consumer confidence index data, and calculates the standard deviation of the two data separately, and then performs the above correlation coefficient calculation process. The above calculation process is repeated for all budget execution difference characteristic data points and external economic indicator data in the fused data set to obtain a complete set of correlation coefficients.

[0093] Set screening criteria to determine budget sensitivity characteristics. The screening criteria are: select data items whose standardized correlation coefficients are in the top 20 percent of the entire set of standardized correlation coefficients, and define the corresponding data items as budget sensitivity characteristics. For example, if the standardized correlation coefficient value calculated between the average value of the budget execution difference characteristic of a certain advertising channel, such as a search engine advertising channel, and the industry prosperity index data is within the top 20 percent of the entire set of standardized correlation coefficients, then the industry prosperity index data is one of the budget sensitivity characteristics; if the standardized correlation coefficient corresponding to the consumer confidence index data is within the top 20 percent, then the consumer confidence index data is also a budget sensitivity characteristic.

[0094] S4: Use the random forest algorithm to analyze budget sensitivity features and determine the impact weights of budget sensitivity features, including:

[0095] Establish the input dataset of the random forest algorithm based on the budget sensitivity feature;

[0096] The advertising platform obtains budget sensitivity features identified through screening. These features include data items whose standardized correlation coefficients between budget expenditure channels (e.g., search engine advertising, social media advertising, mobile in-app advertising, and online video advertising) and external economic indicators exceed the screening criteria. These budget sensitivity features and the corresponding budget execution variance features constitute the input dataset for the random forest algorithm. The budget sensitivity features for each budget channel and each budget cycle are fully collected and organized to form the input dataset for the random forest algorithm.

[0097] Randomly sample the input data set to generate several training data subsets, and use each training data subset to generate several decision trees;

[0098] The random forest algorithm performs random sampling on the input dataset. Random sampling extracts data points from the input dataset with replacement. The number of data points extracted each time is the same as the number in the input dataset. By repeating the random sampling operation multiple times, multiple different training data subsets are obtained to ensure the difference and diversity between the datasets. Each training data subset is used to train and generate several decision trees. The decision trees are generated by partitioning the feature data in the training data subset using a decision tree algorithm to gradually partition the data into different feature dimensions, thus forming a decision tree. This training process is repeated until multiple decision trees are generated for each training data subset, forming a random forest model.

[0099] Based on the generated decision trees, feature importance analysis is performed on the budget sensitivity features respectively;

[0100] Feature importance analysis involves calculating the number of times a budget-sensitive feature is used when splitting a node in a decision tree and the contribution of that node split to the classification result. The contribution of a node split is measured by the information gain value.

[0101] Calculate the number of feature splits and the information gain value of the split nodes for the budget sensitivity feature in each decision tree;

[0102] The specific number of times the budget sensitivity feature is used to split the decision tree node is calculated for each decision tree in the random forest model. The number of node splits refers to the total number of times a budget sensitivity feature is explicitly used to perform data differentiation operations in a decision tree model. The information gain value is calculated by calculating the information entropy of the decision tree node set before the split, then calculating the information entropy of the child nodes after the node split, and calculating the difference in information entropy, which is the information gain value of the budget sensitivity feature when the node splits. The information gain value is calculated for all nodes in each decision tree to ensure that the importance of each budget sensitivity feature is evaluated. For example, if the industry prosperity index data is used for node splitting multiple times in the decision tree, and the information gain generated at each node split is high, then the industry prosperity index data has a high feature importance in the decision tree.

[0103] The comprehensive importance score of the budget sensitivity feature is determined based on the number of feature splits and the information gain value of the split node;

[0104] The calculation method of the comprehensive importance score is as follows: normalize the number of node splits of each budget-sensitive feature in all decision trees, that is, divide the number of feature splits by the total number of feature splits; normalize the information gain value of each budget-sensitive feature in all decision trees, that is, divide the information gain value of the feature by the total number of information gain values ​​of all features; perform weighted summation on the normalized number of node splits and the normalized information gain value, and then add them together to obtain the comprehensive importance score of the budget-sensitive feature.

[0105] Determine the impact weight of budget sensitivity features based on their comprehensive importance scores;

[0106] The comprehensive importance scores of budget sensitivity features are normalized and the calculated values ​​are used as the impact weights of budget sensitivity features.

[0107] S5: Establish an initial budget allocation optimization model based on the impact weights of budget sensitivity characteristics, and optimize the parameter combination of budget resource allocation through particle swarm optimization, including:

[0108] Construct an initial model for budget allocation optimization based on the impact weights of budget sensitivity characteristics;

[0109] The initial budget allocation optimization model includes a two-dimensional combination of budget resource allocation parameters: the first dimension is the funding ratio parameter for each budget expenditure channel, and the second dimension is the total budget allocation parameter. Budget expenditure channels include search engine advertising, social media advertising, mobile in-app advertising, and online video advertising. The funding ratio parameter represents the proportion of the corresponding funding input to the total budget allocation. The total budget allocation parameter represents the total budget funds available to advertisers for allocation across channels during a specific budget execution cycle. When constructing the initial budget allocation optimization model, the initial values ​​of the funding ratio parameters for each channel are determined based on the influence weights of budget sensitivity characteristics. For example, a higher weight for the industry prosperity index will result in a corresponding increase in the funding ratio parameter for the search engine channel. This results in a specific combination of budget resource allocation parameters, which serves as the initial budget allocation optimization model.

[0110] Based on the budget allocation, the initial model is optimized and the search space of the particle swarm algorithm is established;

[0111] The search space is defined as a multidimensional space composed of combinations of budget resource allocation parameters. Each dimension represents a possible value for the budget allocation ratio parameter and the total budget allocation parameter. Each particle in the search space represents a combination of budget resource allocation parameters, that is, a potential budget allocation solution. For example, the potential solution corresponding to a particle position is a 20% allocation ratio for search engine channels, a 30% allocation ratio for social media channels, a 30% allocation ratio for mobile in-app channels, a 20% allocation ratio for online video channels, and a specific value for the total budget allocation parameter. The advertising platform records and tracks the specific position of each particle in the search space to distinguish between different potential budget allocation solutions.

[0112] Continuously iterate and update in the search space through particle swarm algorithm;

[0113] Determine the parameter settings for the Particle Swarm Optimizer. First, set the particle size, which is the total number of particles in the Particle Swarm Optimizer's search space. This number is determined through multiple experiments before the algorithm's initial run to ensure the algorithm's effectiveness and stability. The initial position is the initial value of the budget resource configuration parameter combination corresponding to each particle in the Particle Swarm Optimizer. The initial position is based on the initial budget allocation optimization model and is adjusted to ensure that the initial positions of the particles are evenly distributed. The initial velocity is the rate of change of each particle's initial movement state in the search space. The initial velocity value is set by randomly selecting a value within a small range.

[0114] The inertia weight controls the degree to which particles in the particle swarm algorithm are affected by their own inertia during motion. The individual learning factor indicates the degree to which a particle is influenced by its own historical best position, while the group learning factor indicates the degree to which a particle is influenced by the historical best position of the entire swarm. To determine the values ​​for the inertia weight, individual learning factor, and group learning factor, multiple sets of experiments and data testing were conducted in advance to select values ​​that maximize the algorithm's convergence and speed. The inertia weight ranges from zero to one, while the individual and group learning factors range from one to three to ensure that the particle swarm algorithm maintains a balanced search process.

[0115] Specifically, the iterative update process of the particle swarm algorithm involves continuous updates in two dimensions: position and velocity. During each iteration of the algorithm, the updated velocity and new position of the particle are calculated. The velocity update is calculated by multiplying the particle's inertia weight by its current velocity, adding the product of the individual learning factor and the difference between the particle's own historical optimal position, and then adding the product of the group learning factor and the difference between the particle's overall historical optimal position, to form the particle's new velocity. The position update is calculated by summing the particle's original position and its new velocity to determine its new position. The historical optimal position of each particle is recorded, and the overall historical optimal position of the entire particle swarm is tracked and recorded in real time for the next iterative calculation.

[0116] The particle swarm algorithm is terminated when the maximum number of iterations is reached or the stability threshold condition of the overall historical optimal position of the particle swarm is met.

[0117] Set termination conditions for the particle swarm algorithm, including a maximum number of iterations and a stability threshold for the overall historical optimal position. The maximum number of iterations is the total number of times the algorithm performs iterative calculations. The stability threshold for the overall historical optimal position is set to terminate when the change in the particle swarm's overall historical optimal position over multiple consecutive iterations is less than a preset change threshold. After each iteration, determine whether the algorithm meets one of the termination conditions. If so, the algorithm stops iterating immediately.

[0118] Output the budget resource configuration parameter combination corresponding to the overall historical optimal position of the particle swarm as the optimized budget resource configuration parameter combination;

[0119] When the particle swarm algorithm iteration terminates, the budget resource configuration parameter combination corresponding to the overall historical optimal position of the particle swarm is output as the optimized budget resource configuration parameter combination.

[0120] S6: Based on the optimized parameter combination of budget resource configuration, generate multi-scenario simulation plans for budget configuration and evaluate the expected budget execution effects of multiple scenarios, including:

[0121] Construct a set of multi-scenario simulation solutions for budget configuration based on the optimized combination of budget resource configuration parameters;

[0122] The budget resource configuration parameter combination includes the funding allocation ratio parameters and the total budget allocation parameters for each budget expenditure channel. The budget expenditure channels include search engine advertising channels, social media advertising channels, mobile application advertising channels, and online video advertising channels. Based on the budget resource configuration parameter combination, the advertising platform sets multiple scenarios. The scenarios are generated by adjusting different combinations of funding allocation ratio parameters and total budget allocation parameters. For example, the first simulation scenario allocates a higher proportion of funds to the search engine advertising channel and a lower proportion of funds to the social media advertising channel; the second simulation scenario increases the funding allocation ratio of the online video advertising channel and reduces the funding ratios of other channels accordingly; and the third simulation scenario distributes the funding ratios equally among all channels. Through the above methods, the advertising platform constructs multiple scenarios to form a multi-scenario simulation plan set for budget configuration, which is used for simulation calculation and analysis.

[0123] Build a simulation computing environment for a set of multi-scenario simulation solutions configured with a budget;

[0124] The advertising platform constructs a simulation computing environment for a collection of multi-scenario simulation scenarios to evaluate the expected budget execution effects of each scenario. The simulation computing environment includes the time dimension of budget expenditure, the budget expenditure channel dimension, and the budget sensitivity characteristic dimension. The time dimension of budget expenditure represents the cyclical characteristics of budget input and expenditure, such as a month or a quarter; the budget expenditure channel dimension includes search engine advertising channels, social media advertising channels, mobile in-app advertising channels, and online video advertising channels; and the budget sensitivity characteristic dimension represents external economic indicators that affect budget execution effects. The advertising platform determines the parameters of the simulation computing environment, such as the capital utilization efficiency indicators of the budget expenditure channels and the expected impact of budget sensitivity characteristics on budget execution effects, which constitute the input data of the simulation environment.

[0125] Based on the simulation computing environment, the expected budget execution effect corresponding to each budget expenditure channel under each multi-scenario simulation plan is calculated respectively;

[0126] In a simulation environment, each scenario in the set of multi-scenario budget configuration simulation scenarios is simulated to determine the expected budget execution effect for each budget expenditure channel. The expected budget execution effect is expressed as the expected deviation between the actual budget expenditure and the expected budget execution target. The calculation process is as follows: Based on the funding allocation ratio parameters and the total budget allocation parameters for the budget expenditure channel, the expected value of the actual expenditure funds for each channel is calculated; based on the impact weight of the budget sensitivity characteristics, the impact of external economic indicators on the budget execution effect is calculated. Specifically, the funding allocation ratio parameter is multiplied by the economic indicator impact weight, and then combined with the total budget allocation parameter to determine the expected budget execution expenditure effect for each channel. For example, in a scenario, if the search engine advertising channel has a higher funding allocation ratio and the industry prosperity index data has a greater influence weight, the corresponding budget execution expected effect for the search engine advertising channel will be better. The above calculation process is repeated for each scenario to obtain a complete data set of expected budget execution effects.

[0127] Evaluate each multi-scenario simulation scheme in the multi-scenario simulation scheme set according to the expected effect of budget execution, and output the evaluation index corresponding to each multi-scenario simulation scheme;

[0128] Based on the data set of expected budget execution results, each simulation scenario in the multi-scenario simulation scenario set is evaluated. Evaluation indicators include budget efficiency, fund accuracy, and budget resource comprehensive utilization. The budget efficiency indicator represents the expected advertising effect of budget expenditures; the fund accuracy indicator represents the difference between the expected actual budget expenditures and the expected target budget expenditures; and the budget resource comprehensive utilization effectiveness indicator represents the degree of alignment between the overall allocation of budget resources and external economic indicators. Evaluation indicators are calculated as follows: the budget efficiency indicator is the average of the expected budget execution results across all budget channels; the fund accuracy indicator is the absolute value of the difference between the expected actual budget expenditures and the target budget expenditures; and the budget resource comprehensive utilization effectiveness indicator is the weighted average of the degree of alignment between the expected budget execution results and external economic indicators, with the weights determined by the influence weights of budget sensitivity characteristics. For example, if a simulation scenario has a high budget efficiency indicator, a low fund accuracy indicator, and a high budget resource comprehensive utilization effectiveness indicator, this scenario is marked as the preferred option. The advertising platform evaluates all multi-scenario simulation scenarios separately, recording and outputting the corresponding evaluation indicators for each scenario.

[0129] S7: Based on the expected budget execution results in multiple scenarios and the preset budget execution efficiency threshold, the optimal budget configuration plan is selected and the budget optimization allocation strategy is output, including:

[0130] Compare the evaluation indicators of each multi-scenario simulation plan with the preset budget execution efficiency threshold to determine whether each multi-scenario simulation plan meets the budget execution efficiency threshold;

[0131] The evaluation indicator set includes a budget utilization efficiency indicator, a fund utilization accuracy indicator, and a budget resource comprehensive utilization effectiveness indicator. A budget execution efficiency threshold is pre-set. The budget execution efficiency threshold represents the minimum acceptable ratio of advertising effectiveness achieved by budget expenditures to the amount of funds invested. The budget execution efficiency threshold is determined by referencing historical data and numerical standards set by business objectives. For example, the budget execution efficiency threshold can be expressed as follows: the budget utilization efficiency indicator must meet a specific numerical standard, the fund utilization accuracy indicator's numerical difference must not exceed a certain limit, and the budget resource comprehensive utilization effectiveness indicator must exceed a specific standard value. The budget utilization efficiency indicator, fund utilization accuracy indicator, and budget resource comprehensive utilization effectiveness indicator are individually compared with the specific standards in the corresponding budget execution efficiency threshold. The difference between each indicator of each simulation scenario and the budget execution efficiency threshold is recorded to determine whether the scenario meets the budget execution efficiency threshold requirements. Through this comparison process, all multi-scenario budget configuration simulation scenarios that meet the budget execution efficiency threshold are screened out.

[0132] Select the multi-scenario simulation scheme with the highest evaluation index from the multi-scenario simulation schemes that reach the budget execution efficiency threshold, and define it as the optimal budget configuration scheme;

[0133] Each simulation plan that meets the budget execution efficiency threshold is ranked based on a comprehensive evaluation index consisting of the budget efficiency index, the fund accuracy index, and the budget resource comprehensive utilization effect index. The comprehensive evaluation index is calculated by normalizing the values ​​of the budget efficiency index, the fund accuracy index, and the budget resource comprehensive utilization effect index, then assigning weights and adding them together. The budget efficiency index is weighted the most, the fund accuracy index the second most, and the budget resource comprehensive utilization effect index the least. These weights are determined through analysis of the advertising platform's internal historical data and market experience. The ranking process then identifies the multi-scenario simulation plan with the highest comprehensive evaluation index among all simulation plans that meet the budget execution efficiency threshold. The multi-scenario simulation plan with the highest comprehensive evaluation index is defined as the optimal budget allocation plan.

[0134] Output the budget resource configuration parameter combination contained in the optimal budget configuration plan as the budget optimization allocation strategy;

[0135] The optimal budget allocation plan includes a specific combination of budget resource allocation parameters, including the funding ratio and total budget allocation parameters for each budget expenditure channel. These channels include search engine advertising, social media advertising, mobile in-app advertising, and online video advertising. This combination of budget resource allocation parameters indicates the funding ratio and total budget allocation allocated to each channel in the optimal budget allocation plan.

[0136] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters and thresholds in the formulas are set by technicians in this field according to actual conditions.

[0137] The above embodiments can be implemented in whole or in part via software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions or computer programs. When loaded or executed on a computer, the processes or functions described in the embodiments of this application are fully or partially performed. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired means (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be magnetic media (e.g., floppy disks, hard disks, tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0138] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0139] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and modules described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0140] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0141] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, and may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0142] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0143] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0144] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0145] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A budget allocation optimization method based on big data analysis, characterized in that: The steps include: S1: Collect historical data on budget expenditures from multiple channels, build a basic data set for budget allocation, and extract budget execution difference characteristics; S2: Group and identify budget execution difference features based on clustering algorithms to determine key feature areas; S3: Combine key feature areas and external economic indicator data to extract budget sensitivity features; Obtain all budget execution difference feature data points within the key feature area; Collect data on external economic indicators related to budget execution; Perform data association processing on the budget execution difference feature data points and external economic indicator data according to the corresponding time interval to generate a fused data set; Conduct correlation analysis on the fused data set, calculate the correlation coefficient between the budget execution difference characteristic data points and the external economic indicator data, and determine the budget sensitivity characteristics; S4: Use the random forest algorithm to analyze the budget sensitivity characteristics and determine the impact weight of the budget sensitivity characteristics; S5: Establish an initial budget allocation optimization model based on the influence weights of budget sensitivity characteristics, and optimize the parameter combination of budget resource allocation through particle swarm optimization; S6: Based on the optimized parameter combination of budget resource allocation, generate multi-scenario simulation plans for budget allocation and evaluate the expected budget execution effects of multiple scenarios; S7: Based on the expected budget execution effects in multiple scenarios and the preset budget execution efficiency threshold, the optimal budget configuration plan is selected and the budget optimization allocation strategy is output.

2. The budget allocation optimization method based on big data analysis according to claim 1, characterized in that: S1, specifically: Collect historical budget expenditure data from different budget expenditure channels; Arrange historical budget expenditure data in chronological order to construct a basic dataset for budget allocation; Calculate the difference between actual expenditure and planned budget expenditure during budget execution based on the budget allocation basic data set; Calculate budget execution variance characteristics based on variance data.

3. The budget allocation optimization method based on big data analysis according to claim 2, characterized in that: S2, specifically: Conduct cluster analysis based on budget execution difference characteristics; Dividing the budget execution difference characteristics into multiple cluster groups, and determining a cluster center and budget execution difference characteristic data points within each cluster group; Calculate the Euclidean distance between the budget execution difference feature data point and the corresponding cluster center within each cluster group; Determine the dispersion of budget execution difference feature data points based on Euclidean distance; The cluster group with the largest dispersion is defined as the key feature area.

4. The budget allocation optimization method based on big data analysis according to claim 3 is characterized in that: S4, specifically: Establish the input dataset of the random forest algorithm based on the budget sensitivity feature; Randomly sample the input data set to generate several training data subsets, and use each training data subset to generate several decision trees; Based on the generated decision trees, feature importance analysis is performed on the budget sensitivity features respectively; Calculate the number of feature splits and the information gain value of the split nodes for the budget sensitivity feature in each decision tree; The comprehensive importance score of the budget sensitivity feature is determined based on the number of feature splits and the information gain value of the split node; Based on the comprehensive importance score of budget sensitivity features, the impact weight of budget sensitivity features is determined.

5. The budget allocation optimization method based on big data analysis according to claim 4 is characterized in that: S5, specifically: Construct an initial model for budget allocation optimization based on the impact weights of budget sensitivity characteristics; Based on the budget allocation, the initial model is optimized and the search space of the particle swarm algorithm is established; Continuously iterate and update in the search space through particle swarm algorithm; The particle swarm algorithm is terminated when the maximum number of iterations is reached or the stability threshold condition of the overall historical optimal position of the particle swarm is met. The budget resource configuration parameter combination corresponding to the overall historical optimal position of the particle swarm is output as the optimized budget resource configuration parameter combination.

6. The budget allocation optimization method based on big data analysis according to claim 5, characterized in that: S6, specifically: Construct a set of multi-scenario simulation solutions for budget configuration based on the optimized combination of budget resource configuration parameters; Build a simulation computing environment for a set of multi-scenario simulation solutions configured with a budget; Based on the simulation computing environment, the expected budget execution effect corresponding to each budget expenditure channel under each multi-scenario simulation plan is calculated respectively; Each multi-scenario simulation scheme in the multi-scenario simulation scheme set is evaluated according to the expected effect of budget execution, and the evaluation index corresponding to each multi-scenario simulation scheme is output.

7. The budget allocation optimization method based on big data analysis according to claim 6, characterized in that: S7, specifically: Compare the evaluation indicators of each multi-scenario simulation plan with the preset budget execution efficiency threshold to determine whether each multi-scenario simulation plan meets the budget execution efficiency threshold; Select the multi-scenario simulation scheme with the highest evaluation index from the multi-scenario simulation schemes that reach the budget execution efficiency threshold, and define it as the optimal budget configuration scheme; Output the budget resource configuration parameter combination contained in the optimal budget configuration plan as the budget optimization allocation strategy.

Images