CN-122022901-A - Policy parameterized income operation deduction method, device, equipment and medium

Abstract

The application discloses a strategy parameterized income running deduction method, device, equipment and medium, wherein the method comprises the steps of continuously collecting current service running data and current strategy parameter data in a service execution process, calculating a income target base line, generating a predicted income result of a current time point based on the current service running data and the current strategy parameter data, comparing the predicted income result with the income target base line in real time, calculating contribution degree of each income factor to deviation when deviation exists, positioning deviation leading factors, executing directional disturbance on corresponding strategy parameters in the current strategy parameter data according to the deviation leading factors, generating a plurality of groups of strategy combinations after adjustment, and selecting an optimal strategy combination based on the deduction income result corresponding to each strategy combination after adjustment. The initiative, the refinement level and the decision efficiency of the business income management are obviously improved through a dynamic deduction mode with controllable process and intervened strategy.

Inventors

• ZHU YUGUANG

Assignees

• 上海航动科技有限公司

Dates

Publication Date

20260512

Application Date

20260416

Claims (10)

1. 1. A policy-parameterized revenue operation deduction method, comprising: Continuously collecting current service operation data and current strategy parameter data in the service execution process, and dynamically calculating a income target baseline according to historical contemporaneous data; Based on the current service operation data and the current strategy parameter data, decomposing the income into a quantifiable income factor set, and generating a predicted income result of the current time point through an interpretable prediction model; Comparing the predicted income result with a income target baseline in real time, and when deviation exists, calculating the contribution degree of each income factor to the deviation, and positioning a deviation dominant factor; according to the deviation leading factors, performing directional disturbance on corresponding strategy parameters in the current strategy parameter data, generating a plurality of groups of strategy combinations after adjustment, and simulating a business state transition process under each combination; based on the deduction income results corresponding to the strategy combinations after adjustment, selecting an optimal strategy combination, converting the optimal strategy combination into strategy adjustment instructions in a natural language format, and pushing the strategy adjustment instructions to a service execution end to complete closed-loop intervention.
2. 2. The strategically parameterized revenue run deduction method of claim 1, wherein the dynamically calculating revenue target baselines includes: acquiring historical contemporaneous service operation data in a preset time window to extract income change time sequence characteristics; Constructing a baseline model by adopting a sliding window based on the income change time sequence characteristics, and calculating an expected income interval of the current time point, wherein the expected income interval of the current time point is calculated at least according to a baseline value obtained by statistics of historical contemporaneous service operation data in the corresponding sliding window; And dynamically adjusting a baseline weight according to the quantized current business operation rhythm of the current business operation data, and dynamically correcting an expected income interval of the current time point according to the baseline weight to generate a income target baseline of the current time point, wherein the baseline is presented in an interval form.
3. 3. The strategically parameterized revenue operation deduction method of claim 1, wherein said calculating the contribution of each revenue factor to the bias includes: Determining a difference in the predicted revenue result from the revenue target baseline as a total revenue deviation; acquiring the marginal contribution value of each income factor output based on the interpretable predictive model, and calculating the contribution duty ratio of each income factor to the total income deviation based on the marginal contribution value of each income factor; and marking the income factors with the contribution proportion exceeding a preset threshold value as deviation dominant factors.
4. 4. The policy parameterized revenue operation deduction method of claim 1, wherein the simulating business state transition process under each combination comprises: constructing a state transition model, and defining a mapping relation between a current service state vector and a next state vector; Taking the adjusted strategy combinations as transfer driving inputs, and respectively calculating corresponding state transfer paths; and iteratively deducting to a preset time window end point along the state transition path to generate accumulated deduction income results corresponding to each combination.
5. 5. The strategically parameterized revenue operation deduction method of claim 1, wherein the revenue factor collection includes at least a business scale class factor, a price level class factor, a package structure class factor, a channel distribution class factor, a time cadence class factor; the interpretable prediction model is a gradient lifting decision tree model or an integrated learning model and is used for outputting marginal contribution values of all factors to a prediction result.
6. 6. The method for deriving revenue operation by policy parameterization according to claim 5, wherein the method is applied to a communication operation service, and the performing, according to the deviation-dominant factor, a directional perturbation on a corresponding policy parameter in the current policy parameter data comprises: When the deviation dominant factor is a price factor, executing gradient increase and decrease disturbance on a price discount rate in the policy parameter data; When the deviation dominant factor is a structural factor, executing switching disturbance on a main push package identifier in the strategy parameter data; And when the deviation dominant factor is a channel factor, executing reallocation disturbance on the channel throwing proportion in the strategy parameter data.
7. 7. The method for deriving revenue operation by policy parameterization according to claim 1, wherein the selecting an optimal policy combination based on derived revenue results corresponding to each adjusted policy combination, and converting the optimal policy combination into a policy adjustment instruction in a natural language format, comprises: acquiring a strategy parameter adjustment amount corresponding to the optimal strategy combination; Mapping the strategy parameter adjustment amount to a preset instruction template library to generate a structured instruction text; and calling a large language model to carry out semantic color rendering on the structured instruction text, and generating a natural language instruction readable by a service.
8. 8. A strategically parameterized revenue operation deduction device comprising: The data acquisition module is used for continuously acquiring current service operation data and current strategy parameter data in the service execution process and dynamically calculating a income target baseline according to the historical contemporaneous data; the model prediction module is used for decomposing the income into a quantifiable income factor set based on the current service operation data and the current strategy parameter data, and generating a predicted income result of the current time point through an interpretable prediction model; The comparison reasoning module is used for comparing the predicted income result with the income target baseline in real time, and calculating the contribution degree of each income factor to the deviation when the deviation exists, and positioning the deviation dominant factor; The strategy deduction module is used for executing directional disturbance on corresponding strategy parameters in the current strategy parameter data according to the deviation leading factors, generating a plurality of groups of adjusted strategy combinations and simulating a business state transfer process under each combination; And the deduction optimization module is used for selecting an optimal strategy combination based on deduction income results corresponding to each adjusted strategy combination, converting the optimal strategy combination into a strategy adjustment instruction in a natural language format, and pushing the strategy adjustment instruction to a service execution end to complete closed-loop intervention.
9. 9. A computer device comprising a memory and a processor, the memory having stored therein computer readable instructions which, when executed by the processor, cause the processor to perform the steps of the policy parameterized revenue run deduction method of any of claims 1 to 7.
10. 10. A storage medium storing computer readable instructions which, when executed by one or more processors, cause the one or more processors to perform the steps of the policy parameterized revenue run deduction method of any of claims 1 to 7.

Description

Policy parameterized income operation deduction method, device, equipment and medium Technical Field The present application relates to the field of policy deduction in service scenarios, and in particular, to a method, apparatus, device, and medium for policy parameterized revenue operation deduction. Background In recent years, in communication operation services, existing revenue management mainly relies on a billing system to perform post-hoc analysis on service data that has already occurred. The post-statistics mode has obvious limitations that firstly, the acquisition of income results has hysteresis, complete statistics can be generated only after the service execution is finished or account settlement is carried out, the income yaw problem in the operation process is difficult to find in time, secondly, the income forming process lacks interpretability, when the income is not as expected, the income cannot be quantitatively distinguished as the factors such as service scale, price level, package structure, channel putting and the like, and thirdly, the prejudgement capability of strategy adjustment cannot be provided, the effects of different adjustment schemes cannot be evaluated in advance in the service execution process, the decision often depends on experience judgment, especially, the dilemma that the investment is continuously increased but the income is insufficient in the flow type service is frequently generated, and in summary, the current communication operation service has the problems of operation strategy adjustment hysteresis and correctness. Disclosure of Invention The application aims to provide a strategy parameterized income operation deduction method, device, equipment and medium, which at least solve the problem that the strategy cannot be adjusted in real time and efficiently in the current communication operation service. In order to solve the technical problems, the application provides a strategy parameterized income operation deduction method, which comprises the following steps: Continuously collecting current service operation data and current strategy parameter data in the service execution process, and dynamically calculating a income target baseline according to historical contemporaneous data; Based on the current service operation data and the current strategy parameter data, decomposing the income into a quantifiable income factor set, and generating a predicted income result of the current time point through an interpretable prediction model; Comparing the predicted income result with a income target baseline in real time, and when deviation exists, calculating the contribution degree of each income factor to the deviation, and positioning a deviation dominant factor; according to the deviation leading factors, performing directional disturbance on corresponding strategy parameters in the current strategy parameter data, generating a plurality of groups of strategy combinations after adjustment, and simulating a business state transition process under each combination; based on the deduction income results corresponding to the strategy combinations after adjustment, selecting an optimal strategy combination, converting the optimal strategy combination into strategy adjustment instructions in a natural language format, and pushing the strategy adjustment instructions to a service execution end to complete closed-loop intervention. Optionally, the dynamically calculating the revenue target baseline includes: acquiring historical contemporaneous service operation data in a preset time window to extract income change time sequence characteristics; Constructing a baseline model by adopting a sliding window based on the income change time sequence characteristics, and calculating an expected income interval of the current time point, wherein the expected income interval of the current time point is calculated at least according to a baseline value obtained by statistics of historical contemporaneous service operation data in the corresponding sliding window; And dynamically adjusting a baseline weight according to the quantized current business operation rhythm of the current business operation data, and dynamically correcting an expected income interval of the current time point according to the baseline weight to generate a income target baseline of the current time point, wherein the baseline is presented in an interval form. Optionally, the calculating the contribution degree of each income factor to the deviation includes: Determining a difference in the predicted revenue result from the revenue target baseline as a total revenue deviation; acquiring the marginal contribution value of each income factor output based on the interpretable predictive model, and calculating the contribution duty ratio of each income factor to the total income deviation based on the marginal contribution value of each income factor; and marking the income factors with the contribution