Business data prediction method, device, electronic device and computer-readable storage medium
Through the two-stage prediction method, the ARIMA model parameters are adjusted using historical data to adjust the prediction deviation problem of large changes in the outside world of the ARIMA model, and the prediction accuracy is improved.
Patent Information
- Application Number
- CN202010276366.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-04-09
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2040-07-27
AI Technical Summary
The ARIMA model has large deviations in the prediction data when there are major changes in the outside world, and the changes in other related random variables cannot be effectively considered.
By obtaining historical contemporary data for the specified period, determining the initial model parameters and performing two-stage predictions: first use the initial model parameter prediction to obtain the first prediction result and the residual value, and then adjust the model parameters according to the residual value to make the second prediction, Improve prediction accuracy.
Improve the accuracy of data prediction, especially in scenarios where external changes are large, the model can be better adjusted to reduce prediction bias.
Smart Images

Figure CN111523084B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data processing technology, and in particular to a business data prediction method, device, electronic device and computer-readable storage medium. Background Art
[0002] ARIMA model (Autoregressive Integrated Moving Average model), also known as integrated moving average autoregressive model (moving can also be called sliding), is one of the time series forecasting and analysis methods. In ARIMA (p, d, q), AR is "autoregressive", p is the number of autoregressive terms; MA is "sliding average", q is the number of sliding average terms, and d is the number of differences (order) made to make it a stationary series. The characteristic of the model is that it does not directly consider the changes of other related random variables.
[0003] As one of the conventional time series prediction models, the ARIMA model is widely used in prediction scenarios, such as stock prices, housing prices, sales, visit volume, temperature, etc. However, the characteristic of the model itself is that it does not directly consider the changes of other related random variables, so the disadvantage of the model is that when encountering large changes in the external environment, the prediction often has a large deviation. Summary of the invention
[0004] The present application provides a business data prediction method, device, electronic device and computer-readable storage medium to solve the problem that the ARIMA model in the prior art has large prediction deviations when encountering large changes in the external environment.
[0005] In order to solve the above problems, the present application provides a business data prediction method, including:
[0006] When it is necessary to predict business data within a specified period in the future, obtain historical business data for the same period of the past corresponding to the specified period;
[0007] Determining initial model parameters corresponding to the prediction model based on the historical business data;
[0008] Calling the prediction model to predict a first prediction result and a prediction residual value corresponding to the specified time period according to the initial model parameters and the historical business data;
[0009] Determining target model parameters corresponding to the prediction model according to the prediction residual value;
[0010] Calling the prediction model to process the prediction residual value according to the target model parameters and the historical business data to obtain a second prediction result corresponding to the specified time period;
[0011] Determine prediction data corresponding to the specified time period based on the first prediction result and the second prediction result.
[0012] Optionally, the business data is enterprise user load, and when the business data in a future specified period needs to be predicted, obtaining historical business data of the same period corresponding to the specified period includes:
[0013] When it is necessary to predict the enterprise user load within a specified future period, obtain the historical enterprise user load for the same period of the specified period;
[0014] Determining the initial model parameters corresponding to the prediction model according to the historical business data includes:
[0015] Determining initial model parameters corresponding to the prediction model according to the historical enterprise user load;
[0016] The calling of the prediction model to predict the first prediction result and the prediction residual value corresponding to the specified time period according to the initial model parameters and the historical business data includes:
[0017] The prediction model is called to predict a first prediction result and a prediction residual value corresponding to the specified time period according to the initial model parameters and the historical enterprise user load;
[0018] The calling of the prediction model to process the prediction residual value according to the target model parameter and the historical business data to obtain a second prediction result corresponding to the specified time period includes:
[0019] The prediction model is called to process the prediction residual value according to the target model parameter and the historical enterprise user load to obtain a second prediction result corresponding to the specified time period;
[0020] The determining, according to the first prediction result and the second prediction result, prediction data corresponding to the specified time period includes:
[0021] The predicted enterprise user load corresponding to the specified time period is determined according to the first prediction result and the second prediction result.
[0022] Optionally, the acquiring of historical business data corresponding to the specified period includes:
[0023] Obtaining initial historical business data for the same period of history;
[0024] The initial historical business data is screened to obtain the historical business data.
[0025] Optionally, the screening the initial historical business data to obtain the historical business data includes:
[0026] Calculate the average value and standard deviation of the initial historical business data;
[0027] The historical business data is obtained by eliminating business data whose difference from the average value is greater than a preset multiple of the standard deviation from the initial historical business data.
[0028] Optionally, determining initial model parameters corresponding to the prediction model according to the historical business data includes:
[0029] Performing differential processing on the historical business data to obtain a first model parameter corresponding to the prediction model;
[0030] The historical business data is subjected to order determination processing to obtain second model parameters and third model parameters corresponding to the prediction model.
[0031] Optionally, performing differential processing on the historical business data to obtain a first model parameter corresponding to the prediction model includes:
[0032] Determining data stability of the historical business data;
[0033] When the historical business data is stationary data, determining the first model parameter to be 0;
[0034] When the historical business data is non-stationary data, performing n-order difference processing on the historical business data; wherein n is a positive integer greater than or equal to 1;
[0035] When the historical business data after the n-th order difference processing is stationary data, the n is used as the first model parameter.
[0036] Optionally, the historical period includes a first historical period and a second historical period, the first historical period is earlier than the second historical period, and the calling of the prediction model to predict the first prediction result and the prediction residual value corresponding to the specified time period according to the initial model parameters and the historical business data includes:
[0037] The prediction model is called to predict the business data of the second historical period according to the initial model parameters and the first historical business data of the first historical period, so as to obtain predicted business data;
[0038] The prediction residual value is calculated based on the predicted business data and the second historical business data of the second historical period;
[0039] The prediction model is called to predict the business data within the specified time period according to the initial model parameters, the first historical business data and the second historical business data to obtain the first prediction result.
[0040] Optionally, determining the prediction data corresponding to the specified time period according to the first prediction result and the second prediction result includes:
[0041] Obtain a sum of the first prediction result and the second prediction result, and use the sum as prediction data for the specified time period.
[0042] In order to solve the above problems, the present application provides a service data prediction device, comprising:
[0043] A historical data acquisition module is used to acquire historical business data corresponding to the specified period of time when it is necessary to predict business data within a specified period of time in the future;
[0044] A model parameter determination module, used to determine the initial model parameters corresponding to the prediction model according to the historical business data;
[0045] A first result prediction module, used for calling the prediction model to predict the first prediction result and prediction residual value corresponding to the specified time period according to the initial model parameters and the historical business data;
[0046] A target parameter determination module, used to determine the target model parameters corresponding to the prediction model according to the prediction residual value;
[0047] A second result prediction module, used for calling the prediction model to process the prediction residual value according to the target model parameters and the historical business data to obtain a second prediction result corresponding to the specified time period;
[0048] A prediction data determination module is used to determine the prediction data corresponding to the specified time period based on the first prediction result and the second prediction result.
[0049] Optionally, the business data is enterprise user load, and the historical data acquisition module includes:
[0050] A historical load acquisition unit, used to acquire the historical enterprise user loads of the same period of history corresponding to the specified time period when it is necessary to predict the enterprise user loads in the future specified time period;
[0051] The model parameter determination module comprises:
[0052] An initial parameter determination unit, used to determine initial model parameters corresponding to the prediction model according to the historical enterprise user load;
[0053] The first result prediction module comprises:
[0054] A first result prediction unit, configured to call the prediction model to predict a first prediction result and a prediction residual value corresponding to the specified time period according to the initial model parameters and the historical enterprise user load;
[0055] The second result prediction module includes:
[0056] A second result acquisition unit is used to call the prediction model to process the prediction residual value according to the target model parameters and the historical enterprise user load to obtain a second prediction result corresponding to the specified time period;
[0057] The prediction data determination module comprises:
[0058] The predicted load determination unit is used to determine the predicted enterprise user load corresponding to the specified time period based on the first prediction result and the second prediction result.
[0059] Optionally, the historical data acquisition module includes:
[0060] An initial business data acquisition unit, used to acquire the initial historical business data of the same historical period;
[0061] The historical business data screening unit is used to screen the initial historical business data to obtain the historical business data.
[0062] Optionally, the historical business data screening unit includes:
[0063] A standard deviation calculation subunit, used to calculate the average value and standard deviation corresponding to the initial historical business data;
[0064] The historical data acquisition subunit is used to eliminate the business data whose difference from the average value is greater than a preset multiple of the standard deviation in the initial historical business data to obtain the historical business data.
[0065] Optionally, the model parameter determination module includes:
[0066] A first model parameter acquisition unit, configured to perform differential processing on the historical business data to obtain a first model parameter corresponding to the prediction model;
[0067] The second model parameter acquisition unit is used to perform order determination processing on the historical business data to obtain second model parameters and third model parameters corresponding to the prediction model.
[0068] Optionally, the first model parameter acquisition unit includes:
[0069] A data stability determination subunit, used to determine the data stability of the historical business data;
[0070] A model parameter determination subunit, configured to determine that the first model parameter is 0 when the historical business data is stationary data;
[0071] A data difference processing subunit, used for performing n-order difference processing on the historical business data when the historical business data is non-stationary data; wherein n is a positive integer greater than or equal to 1;
[0072] The first parameter acquisition subunit is used to use n as the first model parameter when the historical business data after the n-th order difference processing is stationary data.
[0073] Optionally, the historical period includes a first historical period and a second historical period, the first historical period is earlier than the second historical period, and the first result prediction module includes:
[0074] A predicted business data acquisition unit, configured to call the prediction model to predict the business data of the second historical period according to the initial model parameters and the first historical business data of the first historical period, so as to obtain predicted business data;
[0075] A prediction residual value calculation unit, used for calculating the prediction residual value according to the predicted business data and the second historical business data of the second historical period;
[0076] The first prediction result acquisition unit is used to call the prediction model to predict the business data within the specified time period according to the initial model parameters, the first historical business data and the second historical business data to obtain the first prediction result.
[0077] Optionally, the prediction data determination module includes:
[0078] The prediction data acquisition unit is used to acquire the sum of the first prediction result and the second prediction result, and use the sum as the prediction data for the specified time period.
[0079] In order to solve the above problems, the present application provides an electronic device, including:
[0080] A processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the above-mentioned business data prediction methods is implemented.
[0081] In order to solve the above problems, the present application provides a computer-readable storage medium. When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device can execute any of the business data prediction methods described above.
[0082] Compared with the prior art, this application has the following advantages:
[0083] The embodiment of the present application provides a solution for business data prediction, by obtaining historical business data of the same period corresponding to the specified time period when it is necessary to predict business data within a specified time period in the future, determining the initial model parameters corresponding to the prediction model based on the historical business data, calling the prediction model to predict the first prediction result and prediction residual value corresponding to the specified time period based on the initial model parameters and the historical business data, determining the target model parameters corresponding to the prediction model based on the prediction residual value, calling the prediction model to process the prediction residual value based on the target model parameters and the historical business data, obtaining the second prediction result corresponding to the specified time period, and determining the prediction data corresponding to the specified time period based on the first prediction result and the second prediction result. The embodiment of the present application can improve the accuracy of data prediction by adopting a two-stage prediction method. BRIEF DESCRIPTION OF THE DRAWINGS
[0084] Figure 1 A flowchart of a method for predicting business data provided by an embodiment of the present application is shown;
[0085] Figure 2 A flowchart showing another method for predicting business data provided by an embodiment of the present application is shown;
[0086] Figure 2a A timing diagram for testing data stability provided by an embodiment of the present application is shown;
[0087] Figure 2b A schematic diagram of an autocorrelation function provided by an embodiment of the present application is shown;
[0088] Figure 2c A schematic diagram of a partial autocorrelation function provided in an embodiment of the present application is shown;
[0089] Figure 2d A schematic diagram showing a predicted value and a true value provided by an embodiment of the present application is shown;
[0090] Figure 3 A schematic diagram of the structure of a business data prediction device provided in an embodiment of the present application is shown;
[0091] Figure 4 A schematic diagram of the structure of another business data prediction device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION
[0092] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0093] Reference Figure 1 , shows a flowchart of a method for predicting business data provided by an embodiment of the present application, and the method for predicting business data may specifically include the following steps:
[0094] Step 101: when it is necessary to predict business data within a specified period in the future, obtain historical business data of the same period in history corresponding to the specified period.
[0095] The embodiments of the present application can be applied to scenarios where the ARIMA model is used to predict business data within a certain period in the future.
[0096] The business data may be stock prices, housing prices, sales, visits, GDP (Gross Domestic Product), temperature and other data. Specifically, it may be determined according to business needs, and this embodiment does not impose any restrictions on this.
[0097] In ARIMA (p, d, q), AR stands for "autoregression", p is the number of autoregression terms; MA stands for "moving average", q is the number of moving average terms, and d is the number of differences (order) made to make it a stationary series. The ARIMA model is described as follows:
[0098] 1. AR autoregressive model
[0099] The model needs to determine the order p, which means how many periods of historical values are used to predict the current value. The model formula is:
[0100]
[0101] In the above formula (1), y t is the current value, u is a constant, p is the order, ε t is the error value.
[0102] 2. Moving Average MA
[0103] The moving average model focuses on the accumulation of error terms in the autoregressive model. The formula for the q-order autoregressive process is:
[0104]
[0105] In the above formula (2), y t is the current value, u is a constant, q is the order, θ i is the error coefficient, ε t It's an error.
[0106] 3. Autoregressive Moving Average Model ARMA
[0107] Combining the autoregressive model AR and the moving average model MA model, we can get the autoregressive moving average model ARMA (p, q), the calculation formula is as follows:
[0108]
[0109] 4. The order of difference d
[0110] The order of difference when processing data stationarity, that is, the non-stationary data becomes stationary after several orders of difference.
[0111] 5. ARIMA (Autoregressive Moving Average)
[0112] By combining the autoregressive model, the moving average model and the difference method, we can get the differenced autoregressive moving average model ARIMA (p, d, q), where d is the order of the data that needs to be differentiated.
[0113] It can be understood that the above is only a scheme listed for understanding the difference autoregressive moving average model.
[0114] The designated time period refers to a time period in the future during which business data forecasting is required. For example, if you want to forecast the stock price for the whole day tomorrow, then the whole day tomorrow will be the designated time period. If you want to forecast the electricity consumption between 9:00 and 12:00 tomorrow, then 9:00 to 12:00 tomorrow will be the designated time period, and so on.
[0115] Historical business data refers to historical data that belongs to the same period as the specified time period, that is, business data of the same historical period. For example, if the electricity consumption between 9:00 and 12:00 tomorrow is required, 9:00 to 12:00 tomorrow will be the specified period, and the historical period can be 9:00 to 12:00 yesterday, 9:00 to 12:00 the day before yesterday, and so on.
[0116] When it is necessary to predict business data within a specified period in the future, historical business data corresponding to the specified period can be obtained. Specifically, based on a pre-saved historical business database, business data that is in the same historical period as the specified period can be filtered out from the database as historical business data.
[0117] Of course, in this embodiment, other methods may be used to obtain historical business data of the same period of history corresponding to the specified time period, and this embodiment does not limit this.
[0118] After acquiring the historical business data for the same period of the history corresponding to the designated time period, step 102 is executed.
[0119] Step 102: Determine initial model parameters corresponding to the prediction model based on the historical business data.
[0120] The prediction model refers to a model used to predict future business data. In this embodiment, the prediction model is preferably an ARIMA model. In subsequent examples, this embodiment is described using the ARIMA model.
[0121] The initial model parameters refer to the parameters required to run the forecasting model. For example, when running the ARIMA model, the necessary parameters are d, p, and q (i.e., the parameters described in the above step 101). In this embodiment, the second-order forecasting method is adopted, and the initial model parameters are the model parameters required when the forecasting model is running during the first-order forecasting.
[0122] After obtaining the historical business data corresponding to the specified period, the initial model parameters corresponding to the ARIMA model can be determined according to the historical business data. Only after the model parameters are obtained can the ARIMA model be operated normally. The process of determining the initial model parameters will be described in detail in the following embodiments, and this embodiment will not be repeated here.
[0123] After the initial model parameters corresponding to the prediction model are determined according to the historical business data, step 103 is executed.
[0124] Step 103: Call the prediction model to predict the first prediction result and prediction residual value corresponding to the specified time period according to the initial model parameters and the historical business data.
[0125] This embodiment adopts a two-order prediction method, and the first prediction result is the result obtained by the first-order prediction. The prediction residual value refers to the residual value obtained when predicting the prediction result of the specified time period.
[0126] After determining the initial model parameters corresponding to the prediction model, the ARIMA model can be called to predict the business data within a specified time period based on the initial model parameters and historical business data to obtain a first prediction result and a prediction residual value. The specific prediction process can be described in detail in the following embodiments, which will not be repeated in this embodiment.
[0127] After calling the prediction model to predict the first prediction result and the prediction residual value corresponding to the specified time period according to the initial model parameters and the historical business data, step 104 is executed.
[0128] Step 104: Determine the target model parameters corresponding to the prediction model according to the prediction residual value.
[0129] The target model parameters refer to the model parameters of the running prediction model obtained by retraining the prediction model based on the initial prediction residual values obtained through the first-order prediction.
[0130] When the ARIMA model is called to predict business data for a specified period, an initial prediction residual value is generated, and the model parameters of the prediction model, i.e., the target model parameters, can be re-determined by the initial prediction residual value. This process is similar to the process of determining the initial model parameters, and this embodiment will not be repeated here.
[0131] After calling the prediction model to determine the target model parameters of the prediction model according to the initial prediction residual value, step 105 is then executed.
[0132] Step 105: Call the prediction model to process the prediction residual value according to the target model parameters and the historical business data to obtain a second prediction result corresponding to the specified time period.
[0133] The second prediction result refers to the prediction result obtained by calling the ARIMA model to perform a second-order prediction on the prediction residual value generated in the first-order prediction process.
[0134] When the ARIMA model is called to predict business data for a specified period, a prediction residual value is generated. Then, ARIMA can be called to perform a second-order prediction on the prediction residual value based on the re-acquired target model parameters and historical business data, and then the second prediction result corresponding to the specified period can be obtained. This process will be described in detail in the following embodiments, and this embodiment will not be repeated here.
[0135] When calling the prediction model, the prediction residual value is processed according to the target model parameters and the historical business data to obtain a second prediction result corresponding to the specified time period, and then step 106 is executed.
[0136] Step 106: Determine the prediction data corresponding to the specified time period according to the first prediction result and the second prediction result.
[0137] Forecast data refers to the forecasted business data within a specified future period obtained by using the ARIMA model.
[0138] After performing a second-order prediction for the specified time period and obtaining the corresponding first prediction results and second prediction results, the prediction data corresponding to the specified time period can be determined by combining the first prediction results and the second prediction results, that is, the first prediction result is corrected using the second prediction result to obtain the final predicted business data within the specified time period.
[0139] This embodiment uses a secondary prediction method in the process of using the ARIMA model for prediction, which can greatly improve the accuracy of data prediction.
[0140] The business data prediction method provided by the embodiment of the present application obtains historical business data of the same period of history corresponding to the specified time period when it is necessary to predict business data within a specified time period in the future, determines the initial model parameters corresponding to the prediction model based on the historical business data, calls the prediction model to predict the first prediction result and prediction residual value corresponding to the specified time period based on the initial model parameters and the historical business data, determines the target model parameters corresponding to the prediction model based on the prediction residual value, calls the prediction model to process the prediction residual value based on the target model parameters and the historical business data, obtains the second prediction result corresponding to the specified time period, and determines the prediction data corresponding to the specified time period based on the first prediction result and the second prediction result. The embodiment of the present application can improve the accuracy of data prediction by adopting a two-stage prediction method.
[0141] Reference Figure 2 , shows a flowchart of another business data prediction method provided by an embodiment of the present application, and the business data prediction method may specifically include the following steps:
[0142] Step 201: Obtain the initial historical business data for the same historical period.
[0143] The embodiments of the present application can be applied to scenarios where the ARIMA model is used to predict business data within a certain period in the future.
[0144] The business data may be stock prices, housing prices, sales, visits, GDP (Gross Domestic Product), temperature and other data. Specifically, it may be determined according to business needs, and this embodiment does not impose any restrictions on this.
[0145] In ARIMA (p, d, q), AR stands for "autoregression", p is the number of autoregression terms; MA stands for "moving average", q is the number of moving average terms, and d is the number of differences (order) made to make it a stationary series. The ARIMA model is described as follows:
[0146] 1. AR autoregressive model
[0147] The model needs to determine the order p, which means how many periods of historical values are used to predict the current value. The model formula is:
[0148]
[0149] In the above formula (1), y t is the current value, u is a constant, p is the order, ε t is the error value.
[0150] 2. Moving Average MA
[0151] The moving average model focuses on the accumulation of error terms in the autoregressive model. The formula for the q-order autoregressive process is:
[0152]
[0153] In the above formula (2), y t is the current value, u is a constant, q is the order, θ i is the error coefficient, ε t It's an error.
[0154] 3. Autoregressive Moving Average Model ARMA
[0155] Combining the autoregressive model AR and the moving average model MA model, we can get the autoregressive moving average model ARMA (p, q), the calculation formula is as follows:
[0156]
[0157] 4. The order of difference d
[0158] The order of difference when processing data stationarity, that is, the non-stationary data becomes stationary after several orders of difference.
[0159] 5. ARIMA (Autoregressive Moving Average)
[0160] By combining the autoregressive model, the moving average model and the difference method, we can get the differenced autoregressive moving average model ARIMA (p, d, q), where d is the order of the data that needs to be differentiated.
[0161] It can be understood that the above is only a scheme listed for understanding the difference autoregressive moving average model.
[0162] The core function of short-term load forecasting is to predict the future short-term (15 minutes, day) power load based on the historical power load of enterprise-level users, and provide a reference for the power sales business to make transaction declarations and calculate the deviation price difference in power spot transactions. This embodiment uses power load as business data for forecasting.
[0163] The designated time period refers to a time period in the future during which business data forecasting is required. For example, if you want to forecast the stock price for the whole day tomorrow, then the whole day tomorrow will be the designated time period. If you want to forecast the electricity consumption between 9:00 and 12:00 tomorrow, then 9:00 to 12:00 tomorrow will be the designated time period, and so on.
[0164] The historical period refers to the same period in the past that corresponds to the specified time period. For example, if the electricity consumption between 9:00 and 12:00 tomorrow is required, 9:00 to 12:00 tomorrow will be the specified time period. The historical period can be 9:00 to 12:00 yesterday, 9:00 to 12:00 the day before yesterday, and so on.
[0165] Initial historical business data refers to the business data obtained during the same period of history that has not yet been filtered.
[0166] When it is necessary to predict business data within a specified period in the future, the initial historical business data of the same historical period corresponding to the specified period can be obtained. Based on the pre-saved historical business database, the business data in the same historical period as the specified period can be filtered out from the database as the initial historical business data.
[0167] Of course, in this embodiment, other methods may be used to obtain the initial historical business data of the same historical period corresponding to the specified time period, and this embodiment does not limit this.
[0168] After the initial historical business data is acquired, step 202 is executed.
[0169] Step 202: Screen the initial historical business data to obtain the historical business data.
[0170] After the initial historical business data is obtained, the initial historical business data may be screened to obtain the historical business data. The specific screening process may be described in detail in conjunction with the following specific implementation.
[0171] In a specific implementation of this embodiment, the above step 202 may include:
[0172] Sub-step S1: Calculate the mean value and standard deviation corresponding to the initial historical business data.
[0173] In this embodiment, the average value refers to the average value of the initial historical business data. For example, when the initial historical business data is 10 pieces of data, the 10 pieces of data can be added together and then divided by 10 to obtain the average value.
[0174] Standard deviation, also known as mean square error, refers to the standard deviation of initial historical business data. It is most commonly used as a measure of statistical dispersion in probability statistics. The standard deviation is defined as the square root of the arithmetic mean of the squared deviations of the standard values of each unit in the population from their mean.
[0175] After the initial historical business data is acquired, the average value and standard deviation corresponding to the initial historical business data may be calculated. The specific calculation process is not described in detail in this embodiment.
[0176] After the average value and standard deviation corresponding to the initial historical business data are calculated, sub-step S2 is executed.
[0177] Sub-step S2: Eliminate the business data whose difference from the average value is greater than a preset multiple of the standard deviation in the initial historical business data to obtain the historical business data.
[0178] The preset multiple refers to the multiple pre-set by the business personnel for filtering the initial historical business data. In this embodiment, the preset multiple is preferably 3 times.
[0179] After calculating the mean value and standard deviation corresponding to the initial historical business data, the difference between the initial historical business data and the mean value can be calculated. Then, the initial historical business data with a difference greater than a preset multiple is eliminated, and the remaining ones are the historical business data that need to be predicted later.
[0180] After acquiring the historical business data, step 203 and step 204 are executed.
[0181] Step 203: Perform differential processing on the historical business data to obtain first model parameters corresponding to the prediction model.
[0182] The first model parameter refers to the parameter required for ARIMA operation. In this embodiment, the first model parameter is the parameter d shown in the above step 201. That is, the first model parameter is the difference order.
[0183] In this embodiment, after acquiring the historical business data, the historical business data may be differentially processed to obtain the first model parameter. Specifically, it may be described in detail in combination with the following specific implementation manner.
[0184] In a specific implementation of this embodiment, the above step 203 may include:
[0185] Sub-step M1: Determine the data stationarity of the historical business data.
[0186] In this embodiment, data stability refers to whether the historical business data is stable.
[0187] Before determining the first model parameter of the ARIMA model, first determine the data stationarity of the historical business data. Specifically, the following two methods can be used:
[0188] 1. Time series diagram: If the sequence always fluctuates randomly around a constant value, the fluctuation range is bounded, and there is no obvious trend or periodicity, it can be considered a stationary sequence. Figure 2a As shown, the left picture is non-stationary and the right picture is a stationary sequence.
[0189] 2. ADF unit root test can accurately determine stationarity. Call the python statsmodels module tsa.stattools.adfuller function. If the P value corresponding to the test statistic is less than 0.05, it can be confirmed as a stationary series.
[0190] Of course, in a specific implementation, other methods may be used to detect the stability of historical business data. Specifically, it may be determined according to business requirements, and this embodiment does not limit this.
[0191] After determining the data stationarity of the historical business data, execute sub-step M2 or M3.
[0192] Sub-step M2: When the historical business data is stationary data, determine that the first model parameter is 0.
[0193] When it is determined that the historical business data is stationary data, there is no need to perform n-order difference processing on the historical business data, and 0 can be used as the first model parameter.
[0194] Sub-step M3: When the historical business data is non-stationary data, perform n-order difference processing on the historical business data; wherein n is a positive integer greater than or equal to 1.
[0195] When it is determined that the historical business data is non-stationary data, the historical business data can be processed by n-order difference. Specifically, when the historical business data is non-stationary data, the historical business data can be processed by first-order difference, second-order difference, third-order difference or first-order and second-order difference methods after taking the logarithm, and then the stationarity can be tested until a stationary sequence is obtained.
[0196] Sub-step M4: When the historical business data after the n-order difference processing is stationary data, use the n as the first model parameter.
[0197] When n-order difference processing is performed on the historical business data, and the historical business data after the n-order difference processing is stationary satellite data, n is used as the first model parameter. For example, when the historical business data is processed by second-order difference processing and the historical business data is obtained as stationary data, 2 is used as the first model parameter.
[0198] Step 204: performing order determination processing on the historical business data to obtain second model parameters and third model parameters corresponding to the prediction model.
[0199] The second model parameter and the third model parameter are p and q shown in the above step 201, that is, the order of the optimal ARIMA model. In this embodiment, when the second model parameter is p, the third model parameter is q; and when the second model parameter is q, the third model parameter is q. Specifically, it can be determined according to business needs, and this embodiment does not limit this.
[0200] After the historical service data is acquired, the historical service data may be subjected to order determination processing to obtain the second model parameter and the third model parameter. Specifically, the second model parameter and the third model parameter may be acquired by the following two methods.
[0201] Method 1: Determine the values of p and q based on the autocorrelation function and partial autocorrelation function. The autocorrelation function ACF describes the linear correlation between the time series observations and their past observations. The partial autocorrelation function PACF describes the linear correlation between the time series observations and their expected past observations under the condition of given intermediate observations. As shown in the ACF graph ( Figure 2b As shown in Figure 2c, the autocorrelation coefficient is truncated after the first order, and the partial autocorrelation coefficient is truncated after the zero order, so p = 0, q = 1.
[0202] Method 2: Information criterion for order determination. Currently, the following criteria are commonly used to select models: (where L is the likelihood function, k is the number of parameters, and n is the number of observations).
[0203] AIC = -2ln(L) + 2k: Akaike information criterion
[0204] BIC = -2ln(L) + ln(n)*k: Bayesian information criterion
[0205] HQ=-2ln(L)+ln(ln(n))*k:hannan-quinn criterion
[0206] The AIC criterion is commonly used, and overfitting should be avoided as much as possible. Therefore, the model that is given priority should be the one with the smallest AIC value. You can also try each criterion in turn and select the best one.
[0207] Call the python statsmodels module tsa.arma_order_select_ic function, set the order range of p and q and the model selection criteria, and get the orders p and q of the optimal ARMA model.
[0208] After determining the parameters d, p, and q, the ARIMA model is used for prediction. If d is not equal to 0, the difference or logarithmic operation is used in the stationary processing, and the predicted load value can be obtained by inverse operation of the prediction result. Figure 2d As shown, the results of the predicted value and the true value are shown, where the upper curve when time is 30 is the curve corresponding to the true value, and the lower curve when time is 30 is the curve corresponding to the predicted value.
[0209] After the second model parameter and the third model parameter are determined, step 205 is performed.
[0210] Step 205: Call the prediction model to predict the business data of the second historical period according to the initial model parameters and the first historical business data of the first historical period to obtain predicted business data.
[0211] In this embodiment, the historical period may include a first historical period and a second historical period. The time of the first historical period must be earlier than that of the second historical period. For example, if the specified time period is 9:00-12:00 tomorrow morning, the second historical period may be 9:00-12:00 yesterday morning, and the third historical period may be 9:00-12:00 the day before yesterday morning.
[0212] It can be understood that the above examples are merely examples listed for a better understanding of the technical solutions of the embodiments of the present application, and are not intended to be the sole limitation on the embodiments of the present application.
[0213] The first historical business data refers to business data obtained from the first historical period.
[0214] The second historical business data refers to business data obtained from the second historical period.
[0215] The predicted business data refers to the business data obtained by predicting the business data in the second historical period, that is, the predicted business data in the second historical period.
[0216] After obtaining the initial model parameters required for running the ARIMA model, the ARIMA model may be called to predict the business data of the second historical period according to the initial model parameters and the first historical business data of the first historical period to obtain predicted business data.
[0217] Step 206: Calculate the prediction residual value based on the predicted business data and the second historical business data of the second historical period.
[0218] After the predicted business data of the second historical period is obtained, the predicted business data and the second historical business data of the second historical period can be used to obtain the predicted residual value. Specifically, the predicted residual value can be calculated based on the predicted business data and the real second historical business data. For example, taking the prediction of 15-minute electricity consumption in the next day as an example, y1, y2, ..., y n are true values, p1,p2,...,p n is the predicted value, then the residual sequence is r1,r2,...,r n , the corresponding calculation formula is: i =y i -p i , where i=1, 2, ..., n, and n is a positive integer.
[0219] After the prediction residual value is calculated based on the predicted business data and the second historical business data of the second historical period, step 207 is executed.
[0220] Step 207: Call the prediction model to predict the business data within the specified time period according to the initial model parameters, the first historical business data and the second historical business data to obtain the first prediction result.
[0221] This embodiment adopts a two-order prediction method, and the first prediction result is the result obtained by the first-order prediction. The prediction residual value refers to the residual value obtained when predicting the prediction result of the specified time period.
[0222] After obtaining the initial model parameters (i.e., the first model parameters, the second model parameters, and the third model parameters mentioned in the above steps), the ARIMA model can be called to predict the business data within a specified time period based on the model parameters and the first historical business data and the second historical business data to obtain a first prediction result.
[0223] After the first prediction result is obtained, step 208 is executed.
[0224] Step 208: Determine the target model parameters corresponding to the prediction model according to the prediction residual value.
[0225] The target model parameters refer to the model parameters of the running prediction model obtained by retraining the prediction model based on the initial prediction residual values obtained through the first-order prediction.
[0226] When the ARIMA model is called to predict business data for a specified period, an initial prediction residual value is generated, and the model parameters of the prediction model, i.e., the target model parameters, can be re-determined by the initial prediction residual value. This process is similar to the process of determining the initial model parameters, and this embodiment will not be repeated here.
[0227] After calling the prediction model to determine the target model parameters of the prediction model according to the initial prediction residual value, step 209 is then executed.
[0228] Step 209: Call the prediction model to process the prediction residual value according to the target model parameters and the historical business data to obtain a second prediction result corresponding to the specified time period.
[0229] The second prediction result refers to the prediction result obtained by calling the ARIMA model to perform a second-order prediction on the prediction residual value generated in the first-order prediction process.
[0230] When the ARIMA model is called to predict business data for a specified period, a prediction residual value is generated. Then, ARIMA can be called to perform a second-order prediction on the prediction residual value according to the re-determined target model parameters and historical business data, and then the second prediction result corresponding to the specified period can be obtained. Continuing with the example in step 206 above, the obtained residual sequence r1, r2, ..., r n Make a prediction to get the second prediction result The specific prediction method can be: if d is not equal to 0, a difference or logarithmic operation is used in the stationarity processing, and the predicted load value can be obtained by inversely calculating the prediction result.
[0231] In this embodiment, the second prediction result is the correction parameter of the first prediction result, that is, the first prediction result is corrected by the second prediction result, so as to obtain the required predicted business data.
[0232] After the second prediction result is obtained, step 210 is executed.
[0233] Step 210: Obtain a sum of the first prediction result and the second prediction result, and use the sum as prediction data for the specified time period.
[0234] After obtaining the first prediction result and the second prediction result, the first prediction result and the second prediction result may be added together, thereby obtaining prediction data for a specified period of time.
[0235] In this embodiment, by adopting a two-stage prediction method, the beneficial effect of improving the accuracy of the prediction data can be achieved.
[0236] The business data prediction method provided by the embodiment of the present application obtains historical business data of the same period of history corresponding to the specified time period when it is necessary to predict business data within a specified time period in the future, determines the initial model parameters corresponding to the prediction model based on the historical business data, calls the prediction model to predict the first prediction result and prediction residual value corresponding to the specified time period based on the initial model parameters and the historical business data, determines the target model parameters corresponding to the prediction model based on the prediction residual value, calls the prediction model to process the prediction residual value based on the target model parameters and the historical business data, obtains the second prediction result corresponding to the specified time period, and determines the prediction data corresponding to the specified time period based on the first prediction result and the second prediction result. The embodiment of the present application can improve the accuracy of data prediction by adopting a two-stage prediction method.
[0237] Reference Figure 3 , shows a schematic diagram of the structure of a business data prediction device provided in an embodiment of the present application, and the business data prediction device may specifically include the following modules:
[0238] The historical data acquisition module 310 is used to acquire historical business data of the same period of the past corresponding to the specified period when it is necessary to predict business data within a specified period in the future;
[0239] A model parameter determination module 320, used to determine initial model parameters corresponding to the prediction model based on the historical business data;
[0240] A first result prediction module 330 is used to call the prediction model to predict the first prediction result and prediction residual value corresponding to the specified time period according to the initial model parameters and the historical business data;
[0241] A target parameter determination module 340 is used to determine the target model parameters corresponding to the prediction model according to the prediction residual value;
[0242] A second prediction result module 350 is used to call the target prediction model to process the prediction residual value according to the model parameters and the historical business data to obtain a second prediction result corresponding to the specified time period;
[0243] The prediction data determination module 360 is used to determine the prediction data corresponding to the specified time period according to the first prediction result and the second prediction result.
[0244] The business data prediction device provided by the embodiment of the present application obtains historical business data of the same period of history corresponding to the specified time period when it is necessary to predict business data within a specified time period in the future, determines the initial model parameters corresponding to the prediction model based on the historical business data, calls the prediction model to predict the first prediction result and prediction residual value corresponding to the specified time period based on the initial model parameters and the historical business data, determines the target model parameters corresponding to the prediction model based on the prediction residual value, calls the prediction model to process the prediction residual value based on the target model parameters and the historical business data, obtains the second prediction result corresponding to the specified time period, and determines the prediction data corresponding to the specified time period based on the first prediction result and the second prediction result. The embodiment of the present application can improve the accuracy of data prediction by adopting a two-stage prediction method.
[0245] Reference Figure 4 , shows a schematic diagram of the structure of another service data prediction device provided in an embodiment of the present application, and the service data prediction device may specifically include the following modules:
[0246] The historical data acquisition module 410 is used to acquire historical business data of the same period of the past corresponding to the specified period when it is necessary to predict business data within a specified period in the future;
[0247] A model parameter determination module 420, configured to determine initial model parameters corresponding to a prediction model according to the historical business data;
[0248] A first result prediction module 430 is used to call the prediction model to predict the first prediction result and prediction residual value corresponding to the specified time period according to the initial model parameters and the historical business data;
[0249] A target parameter determination model 440, used to determine the target model parameters corresponding to the prediction model according to the prediction residual value;
[0250] A second result prediction module 450 is used to call the prediction model to process the prediction residual value according to the target model parameters and the historical business data to obtain a second prediction result corresponding to the specified time period;
[0251] The prediction data determination module 460 is used to determine the prediction data corresponding to the specified time period according to the first prediction result and the second prediction result.
[0252] Optionally, the business data is enterprise user load, and the historical data acquisition module 410 includes:
[0253] A historical load acquisition unit, used to acquire the historical enterprise user loads of the same period of history corresponding to the specified time period when it is necessary to predict the enterprise user loads in the future specified time period;
[0254] The model parameter determination module 420 includes:
[0255] An initial parameter determination unit, used to determine initial model parameters corresponding to the prediction model according to the historical enterprise user load;
[0256] The first result prediction module 430 includes:
[0257] A first result prediction unit, configured to call the prediction model to predict a first prediction result and a prediction residual value corresponding to the specified time period according to the initial model parameters and the historical enterprise user load;
[0258] The second result prediction module 450 includes:
[0259] A second result acquisition unit is used to call the prediction model to process the prediction residual value according to the target model parameters and the historical enterprise user load to obtain a second prediction result corresponding to the specified time period;
[0260] The prediction data determination module 460 includes:
[0261] The predicted load determination unit is used to determine the predicted enterprise user load corresponding to the specified time period based on the first prediction result and the second prediction result.
[0262] Optionally, the historical data acquisition module 410 includes:
[0263] The initial business data acquisition unit 411 is used to acquire the initial historical business data of the same period of history;
[0264] The historical business data screening unit 412 is used to screen the initial historical business data to obtain the historical business data.
[0265] Optionally, the historical business data screening unit 412 includes:
[0266] A standard deviation calculation subunit, used to calculate the average value and standard deviation corresponding to the initial historical business data;
[0267] The historical data acquisition subunit is used to eliminate the business data whose difference from the average value is greater than a preset multiple of the standard deviation in the initial historical business data to obtain the historical business data.
[0268] Optionally, the model parameter determination module 420 includes:
[0269] A first model parameter acquisition unit 421 is used to perform differential processing on the historical business data to obtain a first model parameter corresponding to the prediction model;
[0270] The second model parameter acquisition unit 422 is used to perform order determination processing on the historical business data to obtain second model parameters and third model parameters corresponding to the prediction model.
[0271] Optionally, the first model parameter acquisition unit 421 includes:
[0272] A data stability determination subunit, used to determine the data stability of the historical business data;
[0273] A model parameter determination subunit, configured to determine that the first model parameter is 0 when the historical business data is stationary data;
[0274] A data difference processing subunit, used for performing n-order difference processing on the historical business data when the historical business data is non-stationary data; wherein n is a positive integer greater than or equal to 1;
[0275] The first parameter acquisition subunit is used to use n as the first model parameter when the historical business data after the n-th order difference processing is stationary data.
[0276] Optionally, the historical period includes a first historical period and a second historical period, the first historical period is earlier than the second historical period, and the first result prediction module 430 includes:
[0277] The predicted business data acquisition unit 431 is used to call the prediction model to predict the business data of the second historical period according to the initial model parameters and the first historical business data of the first historical period, so as to obtain predicted business data;
[0278] A prediction residual value calculation unit 432, configured to calculate the prediction residual value according to the predicted business data and the second historical business data of the second historical period;
[0279] The first prediction result acquisition unit 433 is used to call the prediction model to predict the business data within the specified time period according to the initial model parameters, the first historical business data and the second historical business data to obtain the first prediction result.
[0280] Optionally, the prediction data determination module 460 includes:
[0281] The prediction data acquisition unit 461 is used to acquire the sum of the first prediction result and the second prediction result, and use the sum as the prediction data for the specified time period.
[0282] The business data prediction device provided by the embodiment of the present application obtains historical business data of the same period of history corresponding to the specified time period when it is necessary to predict business data within a specified time period in the future, determines the initial model parameters corresponding to the prediction model based on the historical business data, calls the prediction model to predict the first prediction result and prediction residual value corresponding to the specified time period based on the initial model parameters and the historical business data, determines the target model parameters corresponding to the prediction model based on the prediction residual value, calls the prediction model to process the prediction residual value based on the target model parameters and the historical business data, obtains the second prediction result corresponding to the specified time period, and determines the prediction data corresponding to the specified time period based on the first prediction result and the second prediction result. The embodiment of the present application can improve the accuracy of data prediction by adopting a two-stage prediction method.
[0283] For the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the order of the actions described, because according to the present application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.
[0284] In addition, an embodiment of the present application also provides an electronic device, including: a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the business data prediction methods described above is implemented.
[0285] An embodiment of the present application also provides a computer-readable storage medium. When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute any of the business data prediction methods described above.
[0286] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.
[0287] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0288] The above is a detailed introduction to a business data prediction method, a business data prediction device, an electronic device and a computer-readable storage medium provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for general technical personnel in this field, according to the idea of the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.
Claims
1. A business data prediction method, characterized in that: include: When it is necessary to predict business data within a specified period in the future, obtain historical business data for the same period of the past corresponding to the specified period; Determining initial model parameters corresponding to the prediction model based on the historical business data; Calling the prediction model to predict a first prediction result and a prediction residual value corresponding to the specified time period according to the initial model parameters and the historical business data; Determine the target model parameters corresponding to the prediction model according to the prediction residual value, wherein the target model parameters are the model parameters of the prediction model obtained by retraining the prediction model with the prediction residual value obtained by the first-order prediction; Calling the prediction model to process the prediction residual value according to the target model parameters and the historical business data to obtain a second prediction result corresponding to the specified time period; Determining prediction data corresponding to the specified time period according to the first prediction result and the second prediction result; Determining the initial model parameters corresponding to the prediction model according to the historical business data includes: Performing differential processing on the historical business data to obtain a first model parameter corresponding to the prediction model; Performing order determination processing on the historical business data to obtain second model parameters and third model parameters corresponding to the prediction model; The performing differential processing on the historical business data to obtain a first model parameter corresponding to the prediction model includes: Determining data stability of the historical business data; When the historical business data is stationary data, determining the first model parameter to be 0; When the historical business data is non-stationary data, performing n-order difference processing on the historical business data; wherein n is a positive integer greater than or equal to 1; When the historical business data after the n-th order difference processing is stationary data, the n is used as the first model parameter.
2. The method according to claim 1, characterized in that The business data is the enterprise user load. When the business data in a future specified period needs to be predicted, the historical business data corresponding to the specified period is obtained, including: When it is necessary to predict the enterprise user load within a specified future period, obtain the historical enterprise user load for the same period of the specified period; Determining the initial model parameters corresponding to the prediction model according to the historical business data includes: Determining initial model parameters corresponding to the prediction model according to the historical enterprise user load; The calling of the prediction model to predict the first prediction result and the prediction residual value corresponding to the specified time period according to the initial model parameters and the historical business data includes: The prediction model is called to predict a first prediction result and a prediction residual value corresponding to the specified time period according to the initial model parameters and the historical enterprise user load; The calling of the prediction model to process the prediction residual value according to the target model parameter and the historical business data to obtain a second prediction result corresponding to the specified time period includes: The prediction model is called to process the prediction residual value according to the target model parameter and the historical enterprise user load to obtain a second prediction result corresponding to the specified time period; The determining, according to the first prediction result and the second prediction result, prediction data corresponding to the specified time period includes: The predicted enterprise user load corresponding to the specified time period is determined according to the first prediction result and the second prediction result.
3. The method according to claim 1 or 2, characterized in that: The acquiring of historical business data corresponding to the specified period includes: Obtaining initial historical business data for the same period of history; The initial historical business data is screened to obtain the historical business data.
4. The method according to claim 3, characterized in that The screening of the initial historical business data to obtain the historical business data includes: Calculate the average value and standard deviation of the initial historical business data; The historical business data is obtained by eliminating business data whose difference from the average value is greater than a preset multiple of the standard deviation from the initial historical business data.
5. The method according to claim 1 or 2, characterized in that: The historical period includes a first historical period and a second historical period, the first historical period is earlier than the second historical period, and the calling of the prediction model to predict the first prediction result and the prediction residual value corresponding to the specified time period according to the initial model parameters and the historical business data includes: The prediction model is called to predict the business data of the second historical period according to the initial model parameters and the first historical business data of the first historical period, so as to obtain predicted business data; The prediction residual value is calculated based on the predicted business data and the second historical business data of the second historical period; The prediction model is called to predict the business data within the specified time period according to the initial model parameters, the first historical business data and the second historical business data to obtain the first prediction result.
6. The method according to claim 1, characterized in that The determining, according to the first prediction result and the second prediction result, prediction data corresponding to the specified time period includes: Obtain a sum of the first prediction result and the second prediction result, and use the sum as prediction data for the specified time period.
7. A business data prediction device, characterized in that: include: A historical data acquisition module is used to acquire historical business data corresponding to the specified period of time when it is necessary to predict business data within a specified period of time in the future; A model parameter determination module, used to determine the initial model parameters corresponding to the prediction model according to the historical business data; A first result prediction module, used for calling the prediction model to predict the first prediction result and prediction residual value corresponding to the specified time period according to the initial model parameters and the historical business data; A target parameter determination module, used to determine the target model parameters corresponding to the prediction model according to the prediction residual value, wherein the target model parameters are model parameters of the prediction model obtained by retraining the prediction model with the prediction residual value obtained by first-order prediction; A second result prediction module, used for calling the prediction model to process the prediction residual value according to the target model parameters and the historical business data to obtain a second prediction result corresponding to the specified time period; A prediction data determination module, configured to determine the prediction data corresponding to the specified time period according to the first prediction result and the second prediction result; The model parameter determination module comprises: A first model parameter acquisition unit, configured to perform differential processing on the historical business data to obtain a first model parameter corresponding to the prediction model; A second model parameter acquisition unit, used for performing order determination processing on the historical business data to obtain second model parameters and third model parameters corresponding to the prediction model; Optionally, the first model parameter acquisition unit includes: A data stability determination subunit, used to determine the data stability of the historical business data; A model parameter determination subunit, configured to determine that the first model parameter is 0 when the historical business data is stationary data; A data difference processing subunit, used for performing n-order difference processing on the historical business data when the historical business data is non-stationary data; wherein n is a positive integer greater than or equal to 1; The first parameter acquisition subunit is used to use n as the first model parameter when the historical business data after the n-th order difference processing is stationary data.
8. An electronic device, characterized in that: include: A processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor implements the business data prediction method according to any one of claims 1 to 6 when executing the program.
9. A computer-readable storage medium, characterized in that: When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the business data prediction method described in any one of claims 1 to 6.
