Busbar load prediction method and system based on Bayes-XGBoost, and storage medium

The Bayes-XGBoost method optimizes the correlation analysis of XGBoost model and meteorological data, and solves the problems of long training time and impact of meteorological data in bus load prediction, achieving efficient and accurate load prediction.

CN119066548APending Publication Date: 2024-12-03POWER DISPATCHING CONTROL CENT OF GUANGDONG POWER GRID CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411050391.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-01
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing bus load prediction methods have a long training time and are inefficient under large sample sizes, and the meteorological data has a great impact on model accuracy, resulting in limited prediction efficiency and accuracy.

Method used

The bus load prediction method based on Bayes-XGBoost is adopted, and the XGBoost model is optimized through Bayesian optimization algorithm, combined with the correlation analysis of meteorological data, the matching weather station is determined, the data calculation complexity is reduced, and the prediction efficiency and accuracy are improved.

Benefits of technology

It improves the accuracy and efficiency of bus load prediction, shortens training time, enhances the prediction performance of the model, and reduces the complexity of data processing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119066548A_ABST
    Figure CN119066548A_ABST
Patent Text Reader

Abstract

The invention discloses a Bayes-XGBoost-based bus load prediction method and system and a storage medium. The method comprises the steps of obtaining historical load data of a to-be-predicted bus; performing correlation analysis on the historical load data corresponding to the to-be-predicted bus and meteorological data, and determining a meteorological station matched with the to-be-predicted bus; an XGBoost model is optimized through a Bayesian optimization algorithm, a bus load prediction model of the to-be-predicted bus is obtained, and the XGBoost model is obtained through training according to the historical load data of the to-be-predicted bus, meteorological data corresponding to the meteorological station and time characteristics corresponding to the meteorological data; and inputting to-be-predicted time into the bus load prediction model, and predicting the load of the to-be-predicted bus at the to-be-predicted time. The method can improve the prediction precision and efficiency of the bus load.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of power system load analysis, and in particular to a bus load prediction method, system and storage medium based on Bayes-XGBoost. Background Art

[0002] Bus load forecasting is a basic task in power system management. It involves predicting the total terminal load of a relatively small power supply area provided by the main transformer of a substation. It plays an important role in the safe and economical operation of modern power systems. Its prediction results can reflect the changes in real-time load in advance, ensure efficient power generation, effectively avoid the operational risks of real-time scheduling, and minimize the adverse effects of load fluctuations on the power system. It plays an extremely important role in the fields of power grid security early warning and intelligent scheduling technology.

[0003] In bus load forecasting, meteorological factors are important factors in achieving accurate modeling, which will directly affect the accuracy of bus load forecasting. However, due to the complexity of the geographical environment and the differences in the matching degree between meteorological stations and substations, meteorological data from different meteorological stations will have different effects on the training and prediction accuracy of the model. At present, support vector regression (SVR) and least squares support vector machine (LSSVM) are mostly used for bus load forecasting. Since SVR and LSSVM have excellent learning performance and promotion capabilities, they can provide more accurate prediction results. However, when the sample size is large, the training process of SVR and LSSVM takes a long time, the complexity is too high and the efficiency is low. Therefore, this application proposes a bus load forecasting method, system and storage medium based on Bayes-XGBoost, which improves the prediction efficiency while reducing the impact of meteorological data on the prediction accuracy.

[0004] Application Contents

[0005] The present application provides a bus load prediction method, system and storage medium based on Bayes-XGBoost, which can improve the prediction accuracy and efficiency of bus load.

[0006] In a first aspect, the present application provides a bus load forecasting method based on Bayes-XGBoost, comprising:

[0007] Obtain historical load data of the bus to be predicted;

[0008] Performing correlation analysis on the historical load data and meteorological data corresponding to the bus to be predicted, and determining a meteorological station matching the bus to be predicted;

[0009] Optimizing the XGBoost model by using a Bayesian optimization algorithm to obtain a bus load prediction model for the bus to be predicted, wherein the XGBoost model is obtained by training based on the historical load data of the bus to be predicted, the meteorological data corresponding to the meteorological station, and the time characteristics corresponding to the meteorological data;

[0010] The time to be predicted is input into the bus load prediction model, and the load of the bus to be predicted at the time to be predicted is predicted.

[0011] The embodiment of the present application performs a correlation analysis on the historical load data and meteorological data corresponding to the bus to be predicted, and then determines the meteorological station that matches the bus to be predicted, thereby reducing the complexity of subsequent data calculation, improving data processing efficiency, and providing more accurate meteorological characteristics for subsequent model training; utilizing the XGBoost model can improve the load prediction function of the overall model, and utilizing the Bayesian optimization algorithm to search for the optimal parameters of the XGBoost model can quickly search for all optimal parameter points of the XGBoost model, thereby improving the load prediction performance and prediction efficiency of the overall model.

[0012] Furthermore, the historical data of the bus to be predicted is obtained, specifically:

[0013] Acquire actual load data of the bus to be predicted in a preset time period;

[0014] Based on the actual load data, a maximum load value, a minimum load value, and a load change rate are determined.

[0015] In this way, by obtaining a variety of load data as the training set of the XGBoost model, the performance of the XGBoost model prediction can be improved.

[0016] Further, correlation analysis is performed on the historical load data and meteorological data corresponding to the bus to be predicted to determine a meteorological station matching the bus to be predicted, specifically:

[0017] Obtaining the distances between the bus to be predicted and several meteorological stations and sorting them, screening the distances of the meteorological stations according to a preset rule, and preliminarily matching the meteorological station combination corresponding to the bus to be predicted;

[0018] Acquiring meteorological data corresponding to the historical load data from the meteorological station combination;

[0019] The Pearson correlation coefficient method and the Spearman correlation coefficient method are used to analyze the correlation between the historical load data corresponding to the bus to be predicted and the meteorological data corresponding to the meteorological station combination, and obtain the meteorological index with the greatest correlation with the historical load data; wherein the meteorological data includes single meteorological data, multi-day accumulated meteorological data and comprehensive meteorological index;

[0020] Based on the meteorological index, a meteorological station matching the bus to be predicted is determined.

[0021] In this way, by performing correlation analysis on the historical load data corresponding to the bus to be predicted and the meteorological data, and then determining the meteorological station that matches the bus to be predicted, the complexity of later data calculation can be reduced, the data processing efficiency can be improved, and more accurate meteorological features can be provided for subsequent model training. Furthermore, the formula of the XGBoost model is specifically as follows:

[0022]

[0023] Where ξ is the objective function of the XGBoost model, is the loss function of the XGBoost model, ∑ k φ(f k ) is the regularization term of the XGBoost model, T is the number of decision trees, α is the penalty coefficient of T, β is the regularization penalty coefficient, and ω is the column vector of leaf nodes.

[0024] In this way, by considering the deviation between the predicted value and the true target value and constraining the complexity of the model through the loss function, overfitting can be avoided and the load forecasting performance of the XGBoost model can be improved.

[0025] Furthermore, the XGBoost model is optimized by the Bayesian optimization algorithm to obtain the bus load prediction model of the bus to be predicted, which is specifically:

[0026] Presetting the fitness function of the Bayesian optimization algorithm and initializing the parameters and ranges of the fitness function;

[0027] Based on the fitness function and the parameters, the fitness value is iteratively calculated until a termination condition is met, and a fitness value representing a hyperparameter is output; wherein, in each iteration, the fitness value is evaluated by a probabilistic proxy model, and the parameters of the Bayesian optimization algorithm and the probabilistic proxy model are updated based on the evaluation result;

[0028] The hyperparameters are input into the XGBoost model for training to obtain a bus load prediction model for the bus to be predicted.

[0029] In this way, by searching for the optimal parameters of the XGBoost model through the Bayesian optimization algorithm, the next optimal parameter point can be searched, which can improve the prediction accuracy and efficiency of the bus load.

[0030] Furthermore, the inputting of the time to be predicted into the bus load prediction model and predicting the load of the bus to be predicted at the time to be predicted is specifically as follows:

[0031] Confirming the time to be predicted, and acquiring actual meteorological data and forecast meteorological data corresponding to the bus to be predicted based on the time to be predicted;

[0032] The actual meteorological data and the forecast meteorological data corresponding to the time to be predicted and the bus to be predicted are input into the bus load prediction model to predict the load of the bus to be predicted at the time to be predicted.

[0033] In this way, by inputting the time to be predicted, actual meteorological data and forecast meteorological data into the bus load prediction model, the load of the bus to be predicted at the time to be predicted can be accurately predicted.

[0034] In a second aspect, the present application provides a bus load forecasting system based on Bayes-XGBoost, including: an acquisition module, an analysis module, an optimization module and a prediction module;

[0035] The acquisition module is used to acquire the historical load data of the bus to be predicted;

[0036] The analysis module is used to perform correlation analysis on the historical load data and meteorological data corresponding to the bus to be predicted, and determine the meteorological station that matches the bus to be predicted;

[0037] The optimization module is used to optimize the XGBoost model through a Bayesian optimization algorithm to obtain a bus load prediction model for the bus to be predicted, wherein the XGBoost model is obtained by training based on the historical load data of the bus to be predicted, the meteorological data corresponding to the meteorological station, and the time characteristics corresponding to the meteorological data;

[0038] The prediction module is used to input a time to be predicted, and predict the load of the bus to be predicted at the time to be predicted based on the bus load prediction model.

[0039] The embodiment of the present application performs a correlation analysis on the historical load data and meteorological data corresponding to the bus to be predicted, and then determines the meteorological station that matches the bus to be predicted, thereby reducing the complexity of subsequent data calculation, improving data processing efficiency, and providing more accurate meteorological characteristics for subsequent model training; utilizing the XGBoost model can improve the load prediction function of the overall model, and utilizing the Bayesian optimization algorithm to search for the optimal parameters of the XGBoost model can quickly search for all optimal parameter points of the XGBoost model, thereby improving the load prediction performance and prediction efficiency of the overall model.

[0040] Further, the acquisition module includes: a first acquisition unit and a calculation unit;

[0041] The first acquisition unit is used for actual load data of the bus to be predicted in a preset time period;

[0042] The calculation unit is used for the maximum load value, the minimum load value and the load change rate in the actual load data.

[0043] Further, the analysis module includes: a second acquisition unit, a third acquisition unit, an analysis unit and a matching unit;

[0044] The second acquisition unit is used to acquire the distances between the bus to be predicted and a plurality of meteorological stations and sort them, screen the distances of the meteorological stations according to a preset rule, and preliminarily match the meteorological station combination corresponding to the bus to be predicted;

[0045] The third acquisition unit is used to acquire meteorological data corresponding to the historical load data from the meteorological station combination;

[0046] The analysis unit is used to analyze the correlation between the historical load data corresponding to the bus to be predicted and the meteorological data corresponding to the meteorological station combination by using the Pearson correlation coefficient method and the Spearman correlation coefficient method, so as to obtain the meteorological index with the greatest correlation with the historical load data; wherein the meteorological data includes single meteorological data, multi-day accumulated meteorological data and comprehensive meteorological index;

[0047] The matching unit is used to determine a meteorological station that matches the bus to be predicted based on the meteorological index.

[0048] Furthermore, the formula of the XGBoost model in the optimization module is specifically:

[0049]

[0050] Where ξ is the objective function of the XGBoost model, is the loss function of the XGBoost model, ∑ k φ(f k ) is the regularization term of the XGBoost model, T is the number of decision trees, α is the penalty coefficient of T, β is the regularization penalty coefficient, and ω is the column vector of leaf nodes.

[0051] Further, the optimization module includes: an initialization unit, an update unit and a training unit;

[0052] The initialization unit is used to preset the fitness function of the Bayesian optimization algorithm and initialize the parameters and ranges of the fitness function;

[0053] The updating unit is used to iteratively calculate the fitness value based on the fitness function and the parameters until a termination condition is met, and output the fitness value representing the hyperparameter; wherein, in each iteration, the fitness value is evaluated by a probabilistic proxy model, and the parameters of the Bayesian optimization algorithm and the probabilistic proxy model are updated based on the evaluation results;

[0054] The training unit is used to input the hyperparameters into the XGBoost model for training to obtain a bus load prediction model for the bus to be predicted.

[0055] Further, the prediction module includes: a determination unit and a prediction unit;

[0056] The determining unit is used to confirm the time to be predicted, and obtain the actual meteorological data and the forecast meteorological data corresponding to the bus to be predicted based on the time to be predicted;

[0057] The prediction unit is used to input the actual meteorological data and the forecast meteorological data corresponding to the time to be predicted and the bus to be predicted into the bus load prediction model to predict the load of the bus to be predicted at the time to be predicted.

[0058] In a third aspect, the present application provides a computer-readable storage medium, characterized in that the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the bus load prediction method based on Bayes-XGBoost described in the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a flow chart of an embodiment of a bus load prediction method based on Bayes-XGBoost provided in this application;

[0060] Figure 2is a flow chart for determining the correlation between the historical load data and the meteorological data provided by the present application;

[0061] Figure 3 is a flow chart of the optimized XGBoost model of the Bayesian optimization algorithm provided in this application;

[0062] Figure 4 It is a structural schematic diagram of an embodiment of a bus load prediction system based on Bayes-XGBoost provided in this application. DETAILED DESCRIPTION

[0063] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0064] It should be understood that the step numbers used in this article are only for the convenience of description and are not intended to limit the order in which the steps are executed.

[0065] It should be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an" and "the" are intended to include plural forms unless the context clearly indicates otherwise.

[0066] The terms “include” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0067] The term "and / or" means and includes any and all possible combinations of one or more of the associated listed items.

[0068] Bus load forecasting is a basic task in power system management, which involves predicting the total regional load provided by the main transformer of the substation, and is crucial to the safe and economic operation of the power system. The forecast results can reflect load changes in advance, ensure efficient power generation and avoid scheduling risks, and play a significant role in power grid security warning and intelligent scheduling. Meteorological factors are very important for accurate modeling and directly affect the accuracy of prediction, but data from different meteorological stations have different effects on model training and prediction accuracy. Currently, support vector regression (SVR) and least squares support vector machine (LSSVM) are mostly used for prediction. Although they have strong learning performance and generalization capabilities, they have long training time and low efficiency under large sample sizes. Therefore, this application proposes a bus load forecasting method, system and storage medium based on Bayes-XGBoost to reduce the impact of meteorological data on prediction accuracy and improve prediction efficiency.

[0069] Embodiment 1

[0070] Please refer to Figure 1 , Figure 1 A schematic flow chart of an embodiment of a bus load prediction method based on Bayes-XGBoost provided in an embodiment of the present application, comprising steps S1 to S4;

[0071] Step S1, obtaining historical load data of the bus to be predicted;

[0072] Specifically, actual load data of the bus to be predicted in a preset time period is obtained; based on the actual load data, a maximum load value, a minimum load value and a load change rate are determined.

[0073] It can be understood that the actual load data of the bus to be predicted in the preset time period may be the load at the same moment of the previous day, the actual load at two points before and after the load at the same moment of the previous day, the load at the same moment two days before the prediction point, the load at one point before and after the load at the same moment two days before the prediction point, and the load at the same moment three days before; and calculate the maximum load of the actual load data one, two, and three days before, the minimum load of the prediction point, and the load change rate at the same moment one, two, and three days before the prediction point.

[0074] It should be noted that the prediction point is the substation corresponding to the bus to be predicted, and the historical data can be collected using existing technology, which will not be repeated in this application.

[0075] In this way, by obtaining a variety of load data as the training set of the XGBoost model, the performance of the XGBoost model prediction can be improved.

[0076] Step S2, performing correlation analysis on the historical load data and meteorological data corresponding to the bus to be predicted, and determining a meteorological station matching the bus to be predicted;

[0077] First, the distances between the bus to be predicted and several meteorological stations are obtained and sorted, and the distances of the meteorological stations are screened according to preset rules to preliminarily match the meteorological station combination corresponding to the bus to be predicted; it can be understood that the distances between the bus to be predicted and several meteorological stations are first calculated, sorted from small to large, and meteorological stations are extracted according to preset rules, wherein the preset rules may be the top three meteorological stations or the top five meteorological stations, and the present application does not limit this. If the extracted meteorological stations are located in the same direction, only the meteorological station closest to them is retained. Exemplarily, if three meteorological stations are located in different directions, all three meteorological stations are retained. If two of the three meteorological stations are located in the same direction, only the closest meteorological station in the same direction is retained, and finally only two meteorological stations are left in total.

[0078] Secondly, meteorological data corresponding to the historical load data are obtained from the meteorological station combination; it can be understood that the meteorological data is the meteorological data corresponding to the historical load data obtained from the meteorological station combination, and the meteorological data can be single meteorological data, multi-day accumulated meteorological data and comprehensive meteorological index, wherein the single meteorological data can include 7 factors including maximum temperature, minimum temperature, average temperature, temperature, relative humidity, rainfall and wind speed, and the multi-day accumulated meteorological data and comprehensive meteorological index can be the cumulative effect data or comprehensive data among temperature, relative humidity, rainfall and wind speed.

[0079] It should be noted that meteorological data can be collected using existing technologies, which will not be elaborated in this application.

[0080] Again, the Pearson correlation coefficient method and the Spearman correlation coefficient method are used to analyze the correlation between the historical load data corresponding to the bus to be predicted and the meteorological data corresponding to the meteorological station combination, and the meteorological index with the greatest correlation with the historical load data is obtained; wherein the meteorological data includes single meteorological data, multi-day accumulated meteorological data and comprehensive meteorological index; it can be understood that the linear correlation and nonlinear correlation of the three dimensions of single meteorological data, multi-day accumulated meteorological data and comprehensive meteorological index with the historical load data corresponding to the bus to be predicted are calculated respectively, wherein the linear correlation can be calculated by the Pearson correlation coefficient method, and the nonlinear correlation can be calculated by the Spearman correlation coefficient method, and the meteorological index with the greatest correlation with the historical load data is obtained. The flow chart for determining the correlation between the historical load data and the meteorological data is shown in the following figure. Figure 2 shown.

[0081] The calculation formula of Pearson correlation coefficient is:

[0082]

[0083] In the formula, ρ XY is the Pearson correlation value, X, Y are the meteorological data and the historical load data respectively; cov(X,Y) is the covariance of the meteorological data and the historical load data; σX, σY are the standard deviations of the meteorological data and the historical load data respectively.

[0084] The calculation formula of Spearman correlation coefficient is:

[0085]

[0086] Where ρ is the Spearman correlation value, D i is the rank difference corresponding to the variables of the meteorological data and the historical load data, and n is the sample size.

[0087] D i It should be noted that the correlation can only be confirmed when the correlation values ​​of Pearson and Spearman are greater than the preset values, where the preset values ​​can be set according to needs.

[0088] Finally, based on the meteorological indicators, the meteorological station that matches the bus to be predicted is determined; it can be understood that if there is only one meteorological indicator with the largest correlation value, the meteorological station that matches the bus to be predicted can be directly determined based on the meteorological indicator; if there are several meteorological indicators with the same correlation value, then by confirming the size of the correlation value in the meteorological station matched by the meteorological indicator, the meteorological station with the largest correlation value is finally selected as the meteorological station that matches the bus to be predicted.

[0089] In this way, by performing a correlation analysis on the historical load data corresponding to the bus to be predicted and the meteorological data, and then determining the meteorological station that matches the bus to be predicted, the complexity of subsequent data calculation can be reduced, the data processing efficiency can be improved, and more accurate meteorological characteristics can be provided for subsequent model training.

[0090] Step S3, optimizing the XGBoost model by using a Bayesian optimization algorithm to obtain a bus load prediction model for the bus to be predicted, wherein the XGBoost model is obtained by training based on the historical load data of the bus to be predicted, the meteorological data corresponding to the meteorological station, and the time characteristics corresponding to the meteorological data;

[0091] First, the historical load data of the bus to be predicted, the meteorological data corresponding to the meteorological station and the time features corresponding to the meteorological data are input into the XGBoost model, wherein the time features corresponding to the meteorological data include holiday marks, marks of days to holidays, marks of working days and marks of time points;

[0092] Among them, the calculation formula of the XGBoost model is:

[0093]

[0094] Where: is the predicted value of the XGBoost model; v i is the i-th sample; ω(v i ) is about sample v i v i function, T is the number of decision trees; f T is an independent function of the function space; F is the function space composed of decision trees.

[0095] It should be noted that XGBoost is an integrated model that improves on the GBDT algorithm and uses a gradient boosted tree as an internal decision tree learner. XGBoost iteratively builds a series of weak learners (usually decision trees) into a strong classifier to enhance the algorithm's learning ability for input features. By optimizing the loss function, XGBoost can improve the predictive performance of the overall model.

[0096] When a given sample set has n samples and m features, the data set of the XGBoost model is:

[0097]

[0098] In the formula, A is the i ,q i The sample set of q i is the historical load data of the ith category, p i is the meteorological data of the ith category, and R is a real number set.

[0099] Among them, the function space calculation formula of the decision tree in the XGBoost model is:

[0100]

[0101] In the formula, q(x i ) means that the sample x i Assigned to the i-th leaf node, Indicates that the value is q(x i )’s leaf node weights.

[0102] The objective function of the XGBoost model can be divided into two parts: the loss function and the regularization term. The loss function reflects the deviation between the predicted value and the true target value, while the regularization term is used to constrain the complexity of the tree model to avoid overfitting. The formula of the objective function of the XGBoost model is as follows:

[0103]

[0104] Where ξ is the objective function of the XGBoost model, is the loss function of the XGBoost model, ∑ k φ(f k ) is the regularization term of the XGBoost model, T is the number of decision trees, α is the penalty coefficient of T, β is the regularization penalty coefficient, and ω is the column vector of leaf nodes.

[0105] In this way, by considering the deviation between the predicted value and the true target value and constraining the complexity of the model through the loss function, overfitting can be avoided and the load forecasting performance of the XGBoost model can be improved.

[0106] Secondly, the XGBoost model is optimized by the Bayesian optimization algorithm to obtain the bus load prediction model of the bus to be predicted, which is specifically:

[0107] The fitness function of the Bayesian optimization algorithm is preset, and the parameters and range of the fitness function are initialized; wherein, the fitness function is fun(x) = -mse, and it should be noted that the mean square error (MSE) uses 10-fold cross validation, and the average MSE of 10 times is taken as the fitness function; and the maximum number of iterations, the minimum leaf node, the maximum number of features, and the depth of the tree are initialized by n_estimators, min_samples_split, max_features, and max_depth, respectively. Exemplarily, when the maximum number of iterations is 100, the minimum leaf node is 3, the maximum number of features is 0.12, and the depth of the tree is 15, the parameters to be optimized are defined as: n_estimators = (10, 250), min_samples_split = (2, 25), max_features = (0.1, 0.999), max_depth = (5, 15).

[0108] It should be noted that the parameters and ranges of the fitness function are determined according to the sample set of the Bayesian optimization algorithm, and the sample set of the Bayesian optimization algorithm is the sample set A of the XGBoost model.

[0109] Based on the fitness function and the parameters, the fitness value is iteratively calculated until the termination condition is met, and the fitness value representing the hyperparameter is output; wherein, in each iteration, the fitness value is evaluated by the probabilistic proxy model, and the parameters of the Bayesian optimization algorithm and the probabilistic proxy model are updated based on the evaluation result; it can be understood that the fitness function is calculated by the Bayesian optimization algorithm, and the average value of the root mean square error of the cross-validation is used as the fitness value output, and the fitness value is also the initial hyperparameter combination, and the initial hyperparameter combination is evaluated by the probabilistic proxy model, and new hyperparameter combinations are continuously generated, and the probabilistic proxy model is updated based on the hyperparameter combination, and the initial hyperparameter combination is continuously evaluated by the probabilistic proxy model until the termination condition is met, at which time the fitness value representing the hyperparameter can be output, and the flowchart of the Bayesian optimization algorithm for optimizing the XGBoost model is shown as follows Figure 3 shown.

[0110] It should be noted that when judging whether the termination condition is met, it is necessary to pre-set the error that meets the hyperparameter accuracy. In each iteration, the fitness value of this iteration needs to be calculated. If the fitness value of the current iteration is less than the target accuracy error, the termination condition is met. If the termination condition is not met, it enters the next round of iteration.

[0111] The hyperparameters are input into the XGBoost model for training to obtain the bus load prediction model of the bus to be predicted; it should be noted that in the XGBoost model, the hyperparameters are crucial to the prediction performance of the XGBoost model. The Bayesian optimization method is a method of driving optimization decisions by establishing a proxy model. By establishing a model that can be updated and queried, the optimization decision is driven. The Bayesian optimization algorithm mainly includes a probabilistic proxy model and an acquisition function; wherein, the probabilistic proxy model includes a priori probability proxy model p(f) and an observation model that describes the mechanism of generating observation data, namely, the likelihood distribution p(D|f); wherein, updating the probabilistic proxy model means obtaining a posterior probability distribution p(f|Di) containing more data information. The acquisition function is constructed based on the Gaussian process of the posterior probability distribution p(f|Di), and the next most "potential" evaluation point can be selected by maximizing the acquisition function. At the same time, an effective acquisition function can ensure that the selected evaluation point sequence minimizes the total loss regret; wherein, the calculation formula for the loss regret is:

[0112] r n =|y*-y n |;

[0113] In the formula, r n is the loss, y* represents the current optimal solution, y n Indicates the current actual value.

[0114] In this way, by searching for the optimal parameters of the XGBoost model through the Bayesian optimization algorithm, the next optimal parameter point can be searched, which can improve the prediction accuracy and efficiency of the bus load.

[0115] Step S4: input the time to be predicted into the bus load prediction model, and predict the load of the bus to be predicted at the time to be predicted.

[0116] Specifically, confirm the time to be predicted, and obtain the actual meteorological data and the forecast meteorological data corresponding to the bus to be predicted based on the time to be predicted; input the actual meteorological data and the forecast meteorological data corresponding to the time to be predicted and the bus to be predicted into the bus load prediction model to predict the load of the bus to be predicted at the time to be predicted.

[0117] It should be noted that a bus load prediction model will be established for each bus to be predicted so as to be able to predict the load of each bus to be predicted.

[0118] In this way, by inputting the time to be predicted, actual meteorological data and forecast meteorological data into the bus load prediction model, the load of the bus to be predicted at the time to be predicted can be accurately predicted.

[0119] The embodiment of the present application performs a correlation analysis on the historical load data and meteorological data corresponding to the bus to be predicted, and then determines the meteorological station that matches the bus to be predicted, thereby reducing the complexity of subsequent data calculation, improving data processing efficiency, and providing more accurate meteorological characteristics for subsequent model training; utilizing the XGBoost model can improve the load prediction function of the overall model, and utilizing the Bayesian optimization algorithm to search for the optimal parameters of the XGBoost model can quickly search for all optimal parameter points of the XGBoost model, thereby improving the load prediction performance and prediction efficiency of the overall model.

[0120] Embodiment 2

[0121] Please refer to Figure 3 , Figure 3 A schematic diagram of the structure of an embodiment of a bus load prediction system based on Bayes-XGBoost provided in an embodiment of the present application, comprising: an acquisition module 01, an analysis module 02, an optimization module 03 and a prediction module 04;

[0122] The acquisition module 01 is used to acquire the historical load data of the bus to be predicted;

[0123] Specifically, the acquisition module 01 includes: a first acquisition unit and a calculation unit; the first acquisition unit is used for the actual load data of the bus to be predicted in a preset time period; the calculation unit is used for the maximum load value, the minimum load value and the load change rate in the actual load data.

[0124] It can be understood that the actual load data of the bus to be predicted in the preset time period may be the load at the same moment of the previous day, the actual load at two points before and after the load at the same moment of the previous day, the load at the same moment two days before the prediction point, the load at one point before and after the load at the same moment two days before the prediction point, and the load at the same moment three days before; and calculate the maximum load of the actual load data one, two, and three days before, the minimum load of the prediction point, and the load change rate at the same moment one, two, and three days before the prediction point.

[0125] It should be noted that the prediction point is the substation corresponding to the bus to be predicted, and the historical data can be collected using existing technology, which will not be repeated in this application.

[0126] In this way, by obtaining a variety of load data as the training set of the XGBoost model, the performance of the XGBoost model prediction can be improved.

[0127] The analysis module 02 is used to perform correlation analysis on the historical load data and meteorological data corresponding to the bus to be predicted, and determine the meteorological station that matches the bus to be predicted;

[0128] Specifically, the analysis module 02 includes: a second acquisition unit, a third acquisition unit, an analysis unit and a matching unit;

[0129] The second acquisition unit is used to obtain the distance between the bus to be predicted and several meteorological stations and sort them, and screen the distances of the meteorological stations according to preset rules to preliminarily match the meteorological station combination corresponding to the bus to be predicted; it can be understood that the distance between the bus to be predicted and several meteorological stations is first calculated, sorted from small to large, and the meteorological stations are extracted according to the preset rules, wherein the preset rules can be the top three meteorological stations or the top five meteorological stations, and the present application does not limit this. If the extracted meteorological stations are located in the same direction, only the meteorological station closest to them is retained. Exemplarily, if three meteorological stations are located in different directions, all three meteorological stations are retained. If two of the three meteorological stations are located in the same direction, only the closest meteorological station in the same direction is retained, and finally only two meteorological stations are left in total.

[0130] The third acquisition unit is used to acquire meteorological data corresponding to the historical load data from the meteorological station combination; it can be understood that the meteorological data is the meteorological data corresponding to the historical load data acquired from the meteorological station combination, and the meteorological data may be single meteorological data, multi-day accumulated meteorological data and comprehensive meteorological index, wherein the single meteorological data may include seven factors: maximum temperature, minimum temperature, average temperature, temperature, relative humidity, rainfall and wind speed, and the multi-day accumulated meteorological data and comprehensive meteorological index may be cumulative effect data or comprehensive data between temperature, relative humidity, rainfall and wind speed.

[0131] It should be noted that meteorological data can be collected using existing technologies, which will not be elaborated in this application.

[0132] The analysis unit is used to analyze the correlation between the historical load data corresponding to the bus to be predicted and the meteorological data corresponding to the meteorological station combination by using the Pearson correlation coefficient method and the Spearman correlation coefficient method, so as to obtain the meteorological index with the greatest correlation with the historical load data; wherein the meteorological data includes single meteorological data, multi-day accumulated meteorological data and comprehensive meteorological index; it can be understood that the linear correlation and nonlinear correlation of the three dimensions of single meteorological data, multi-day accumulated meteorological data and comprehensive meteorological index with the historical load data corresponding to the bus to be predicted are calculated respectively, wherein the linear correlation can be calculated by using the Pearson correlation coefficient method, and the nonlinear correlation can be calculated by using the Spearman correlation coefficient method, so as to obtain the meteorological index with the greatest correlation with the historical load data. The flowchart for determining the correlation between the historical load data and the meteorological data is shown in the following figure. Figure 2 shown.

[0133] The calculation formula of Pearson correlation coefficient is:

[0134]

[0135] In the formula, ρ XY is the Pearson correlation value, X, Y are the meteorological data and the historical load data respectively; cov(X,Y) is the covariance of the meteorological data and the historical load data; σX, σY are the standard deviations of the meteorological data and the historical load data respectively.

[0136] The calculation formula of Spearman correlation coefficient is:

[0137]

[0138] Where ρ is the Spearman correlation value, D i is the rank difference corresponding to the variables of the meteorological data and the historical load data, and n is the sample size.

[0139] D i It should be noted that the correlation can only be confirmed when the correlation values ​​of Pearson and Spearman are greater than the preset values, where the preset values ​​can be set according to needs.

[0140] The matching unit is used to determine the meteorological station that matches the bus to be predicted based on the meteorological indicators; it can be understood that if there is only one meteorological indicator with the largest correlation value, the meteorological station that matches the bus to be predicted can be directly determined based on the meteorological indicator; if there are several meteorological indicators with the same correlation value, the correlation value of the meteorological station matched by the meteorological indicator is confirmed, and finally the meteorological station with the largest correlation value is selected as the meteorological station that matches the bus to be predicted.

[0141] In this way, by performing a correlation analysis on the historical load data corresponding to the bus to be predicted and the meteorological data, and then determining the meteorological station that matches the bus to be predicted, the complexity of subsequent data calculation can be reduced, the data processing efficiency can be improved, and more accurate meteorological characteristics can be provided for subsequent model training.

[0142] The optimization module 03 is used to optimize the XGBoost model through the Bayesian optimization algorithm to obtain the bus load prediction model of the bus to be predicted, wherein the XGBoost model is obtained by training based on the historical load data of the bus to be predicted, the meteorological data corresponding to the meteorological station, and the time characteristics corresponding to the meteorological data;

[0143] First, the historical load data of the bus to be predicted, the meteorological data corresponding to the meteorological station and the time features corresponding to the meteorological data are input into the XGBoost model, wherein the time features corresponding to the meteorological data include holiday marks, marks of days to holidays, marks of working days and marks of time points;

[0144] Among them, the calculation formula of the XGBoost model is:

[0145]

[0146] Where: is the predicted value of the XGBoost model; v i is the i-th sample; ω(v i ) is about sample v i v i function, T is the number of decision trees; f T is an independent function of the function space; F is the function space composed of decision trees.

[0147] It should be noted that XGBoost is an integrated model that improves on the GBDT algorithm and uses a gradient boosted tree as an internal decision tree learner. XGBoost iteratively builds a series of weak learners (usually decision trees) into a strong classifier to enhance the algorithm's learning ability for input features. By optimizing the loss function, XGBoost can improve the predictive performance of the overall model.

[0148] When a given sample set has n samples and m features, the data set of the XGBoost model is:

[0149]

[0150] In the formula, A is the i ,q i The sample set of q i is the historical load data of the ith category, p i is the meteorological data of the ith category, and R is a real number set.

[0151] Among them, the function space calculation formula of the decision tree in the XGBoost model is:

[0152]

[0153] In the formula, q(x i ) means that the sample x i Assigned to the i-th leaf node, Indicates that the value is q(x i )’s leaf node weights.

[0154] The objective function of the XGBoost model can be divided into two parts: the loss function and the regularization term. The loss function reflects the deviation between the predicted value and the true target value, while the regularization term is used to constrain the complexity of the tree model to avoid overfitting. The formula of the objective function of the XGBoost model is as follows:

[0155]

[0156] Where ξ is the objective function of the XGBoost model, is the loss function of the XGBoost model, ∑ k φ(f k ) is the regularization term of the XGBoost model, T is the number of decision trees, α is the penalty coefficient of T, β is the regularization penalty coefficient, and ω is the column vector of leaf nodes.

[0157] In this way, by considering the deviation between the predicted value and the true target value and constraining the complexity of the model through the loss function, overfitting can be avoided and the load forecasting performance of the XGBoost model can be improved.

[0158] Specifically, the optimization module 03 includes: an initialization unit, an update unit and a training unit;

[0159] The initialization unit is used to preset the fitness function of the Bayesian optimization algorithm and initialize the parameters and range of the fitness function; wherein the fitness function is fun(x) = -mse, and it should be noted that the mean square error (MSE) uses 10-fold cross validation, and the average MSE of 10 times is taken as the fitness function; and the maximum number of iterations, the minimum leaf node, the maximum number of features and the depth of the tree are initialized by n_estimators, min_samples_split, max_features, and max_depth respectively. Exemplarily, when the maximum number of iterations is 100, the minimum leaf node is 3, the maximum number of features is 0.12 and the depth of the tree is 15, the parameters to be optimized are defined as: n_estimators = (10, 250), min_samples_split = (2, 25), max_features = (0.1, 0.999), max_depth = (5, 15).

[0160] It should be noted that the parameters and ranges of the fitness function are determined according to the sample set of the Bayesian optimization algorithm, and the sample set of the Bayesian optimization algorithm is the sample set A of the XGBoost model.

[0161] The updating unit is used to iteratively calculate the fitness value based on the fitness function and the parameters until the termination condition is met, and output the fitness value representing the hyperparameter; wherein, in each iteration, the fitness value is evaluated by the probabilistic proxy model, and the parameters of the Bayesian optimization algorithm and the probabilistic proxy model are updated based on the evaluation result; it can be understood that the fitness function is calculated by the Bayesian optimization algorithm, and the average value of the root mean square error of the cross-validation is used as the fitness value output, and the fitness value is also the initial hyperparameter combination. The initial hyperparameter combination is evaluated by the probabilistic proxy model, and new hyperparameter combinations are continuously generated. The probabilistic proxy model is updated based on the hyperparameter combination, and the initial hyperparameter combination is continuously evaluated by the probabilistic proxy model until the termination condition is met, at which time the fitness value representing the hyperparameter can be output. The flowchart of the Bayesian optimization algorithm for optimizing the XGBoost model is shown as follows. Figure 3 shown.

[0162] It should be noted that when judging whether the termination condition is met, it is necessary to pre-set the error that meets the hyperparameter accuracy. In each iteration, the fitness value of this iteration needs to be calculated. If the fitness value of the current iteration is less than the target accuracy error, the termination condition is met. If the termination condition is not met, it enters the next round of iteration.

[0163] The training unit is used to input the hyperparameters into the XGBoost model for training to obtain the bus load prediction model of the bus to be predicted; it should be noted that in the XGBoost model, the hyperparameters are crucial to the prediction performance of the XGBoost model. The Bayesian optimization method is a method of driving optimization decisions by establishing a proxy model, and driving optimization decisions by establishing a model that can be updated and queried. The Bayesian optimization algorithm mainly includes a probabilistic proxy model and an acquisition function; wherein the probabilistic proxy model includes a priori probabilistic proxy model p(f) and an observation model that describes the mechanism of generating observation data, namely, the likelihood distribution p(D|f); wherein updating the probabilistic proxy model means obtaining a posterior probability distribution p(f|Di) containing more data information. The acquisition function is constructed based on the Gaussian process of the posterior probability distribution p(f|Di), and the next most "potential" evaluation point can be selected by maximizing the acquisition function. At the same time, an effective acquisition function can ensure that the selected evaluation point sequence minimizes the total loss regret; wherein, the calculation formula for the loss regret is:

[0164] r n =|y*-y n |;

[0165] In the formula, r n is the loss, y* represents the current optimal solution, y n Indicates the current actual value.

[0166] In this way, by searching for the optimal parameters of the XGBoost model through the Bayesian optimization algorithm, the next optimal parameter point can be searched, which can improve the prediction accuracy and efficiency of the bus load.

[0167] The prediction module 04 is used to input a time to be predicted, and predict the load of the bus to be predicted at the time to be predicted based on the bus load prediction model.

[0168] Specifically, the prediction module 04 includes: a determination unit and a prediction unit; the determination unit is used to confirm the time to be predicted, and obtain the actual meteorological data and forecast meteorological data corresponding to the bus to be predicted based on the time to be predicted; the prediction unit is used to input the actual meteorological data and forecast meteorological data corresponding to the bus to be predicted at the time to be predicted into the bus load prediction model, and predict the load of the bus to be predicted at the time to be predicted.

[0169] It should be noted that a bus load prediction model will be established for each bus to be predicted so as to be able to predict the load of each bus to be predicted.

[0170] In this way, by inputting the time to be predicted, actual meteorological data and forecast meteorological data into the bus load prediction model, the load of the bus to be predicted at the time to be predicted can be accurately predicted.

[0171] The embodiment of the present application performs a correlation analysis on the historical load data and meteorological data corresponding to the bus to be predicted, and then determines the meteorological station that matches the bus to be predicted, thereby reducing the complexity of subsequent data calculation, improving data processing efficiency, and providing more accurate meteorological characteristics for subsequent model training; utilizing the XGBoost model can improve the load prediction function of the overall model, and utilizing the Bayesian optimization algorithm to search for the optimal parameters of the XGBoost model can quickly search for all optimal parameter points of the XGBoost model, thereby improving the load prediction performance and prediction efficiency of the overall model.

[0172] Embodiment 3

[0173] An embodiment of the present application also provides a computer-readable storage medium, which stores processor-executable instructions. When the processor-executable instructions are executed by the processor, they are used to implement the above-mentioned bus load prediction method based on Bayes-XGBoost.

[0174] The contents of the above-mentioned bus load prediction method based on Bayes-XGBoost are all applicable to the present storage medium embodiment. The functions specifically implemented by the present storage medium embodiment are the same as those of the above-mentioned bus load prediction method based on Bayes-XGBoost embodiment, and the beneficial effects achieved are also the same as those achieved by the above-mentioned bus load prediction method based on Bayes-XGBoost embodiment. Therefore, this embodiment can also improve the load prediction performance and prediction efficiency of the overall model.

[0175] Those skilled in the art can understand that all or part of the processes in the above embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes in the above embodiments. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0176] The specific embodiments described above further describe the purpose, technical solutions and beneficial effects of the present application in detail. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the scope of protection of the present application. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A bus load forecasting method based on Bayes-XGBoost, characterized in that: include: Obtain historical load data of the bus to be predicted; Performing correlation analysis on the historical load data and meteorological data corresponding to the bus to be predicted, and determining a meteorological station matching the bus to be predicted; Optimizing the XGBoost model by using a Bayesian optimization algorithm to obtain a bus load prediction model for the bus to be predicted, wherein the XGBoost model is obtained by training based on the historical load data of the bus to be predicted, the meteorological data corresponding to the meteorological station, and the time characteristics corresponding to the meteorological data; The time to be predicted is input into the bus load prediction model, and the load of the bus to be predicted at the time to be predicted is predicted.

2. The bus load prediction method based on Bayes-XGBoost according to claim 1 is characterized in that: Get the historical data of the bus to be predicted, specifically: Acquire actual load data of the bus to be predicted in a preset time period; Based on the actual load data, a maximum load value, a minimum load value, and a load change rate are determined.

3. The bus load prediction method based on Bayes-XGBoost according to claim 1 is characterized in that: The performing correlation analysis on the historical load data and meteorological data corresponding to the bus to be predicted to determine the meteorological station matching the bus to be predicted is specifically: Obtaining the distances between the bus to be predicted and several meteorological stations and sorting them, screening the distances of the meteorological stations according to a preset rule, and preliminarily matching the meteorological station combination corresponding to the bus to be predicted; Acquiring meteorological data corresponding to the historical load data from the meteorological station combination; The Pearson correlation coefficient method and the Spearman correlation coefficient method are used to analyze the correlation between the historical load data corresponding to the bus to be predicted and the meteorological data corresponding to the meteorological station combination, and obtain the meteorological index with the greatest correlation with the historical load data; wherein the meteorological data includes single meteorological data, multi-day accumulated meteorological data and comprehensive meteorological index; Based on the meteorological index, a meteorological station matching the bus to be predicted is determined.

4. The bus load prediction method based on Bayes-XGBoost according to claim 1 is characterized in that: The formula of the XGBoost model is specifically: Where ξ is the objective function of the XGBoost model, is the loss function of the XGBoost model, ∑ k φ(f k ) is the regularization term of the XGBoost model, T is the number of decision trees, α is the penalty coefficient of T, β is the regularization penalty coefficient, and ω is the column vector of leaf nodes.

5. The bus load prediction method based on Bayes-XGBoost according to claim 1, characterized in that: The XGBoost model is optimized by the Bayesian optimization algorithm to obtain the bus load prediction model of the bus to be predicted, which is specifically: Presetting the fitness function of the Bayesian optimization algorithm and initializing the parameters and ranges of the fitness function; Based on the fitness function and the parameters, the fitness value is iteratively calculated until a termination condition is met, and a fitness value representing a hyperparameter is output; wherein, in each iteration, the fitness value is evaluated by a probabilistic proxy model, and the parameters of the Bayesian optimization algorithm and the probabilistic proxy model are updated based on the evaluation result; The hyperparameters are input into the XGBoost model for training to obtain a bus load prediction model for the bus to be predicted.

6. The bus load prediction method based on Bayes-XGBoost according to claim 1 is characterized in that: The step of inputting a time to be predicted into the bus load prediction model and predicting the load of the bus to be predicted at the time to be predicted is specifically as follows: Confirming the time to be predicted, and acquiring actual meteorological data and forecast meteorological data corresponding to the bus to be predicted based on the time to be predicted; The actual meteorological data and the forecast meteorological data corresponding to the time to be predicted and the bus to be predicted are input into the bus load prediction model to predict the load of the bus to be predicted at the time to be predicted.

7. A bus load forecasting system based on Bayes-XGBoost, characterized in that: include: Acquisition module, analysis module, optimization module and prediction module; The acquisition module is used to acquire the historical load data of the bus to be predicted; The analysis module is used to perform correlation analysis on the historical load data and meteorological data corresponding to the bus to be predicted, and determine the meteorological station that matches the bus to be predicted; The optimization module is used to optimize the XGBoost model through a Bayesian optimization algorithm to obtain a bus load prediction model for the bus to be predicted, wherein the XGBoost model is obtained by training based on the historical load data of the bus to be predicted, the meteorological data corresponding to the meteorological station, and the time characteristics corresponding to the meteorological data; The prediction module is used to input a time to be predicted, and predict the load of the bus to be predicted at the time to be predicted based on the bus load prediction model.

8. The bus load prediction system based on Bayes-XGBoost according to claim 7, characterized in that: The acquisition module includes: a first acquisition unit and a calculation unit; The first acquisition unit is used for actual load data of the bus to be predicted in a preset time period; The calculation unit is used for the maximum load value, the minimum load value and the load change rate in the actual load data.

9. The bus load prediction system based on Bayes-XGBoost according to claim 7, characterized in that: The analysis module includes: a second acquisition unit, a third acquisition unit, an analysis unit and a matching unit; The second acquisition unit is used to acquire the distances between the bus to be predicted and a plurality of meteorological stations and sort them, screen the distances of the meteorological stations according to a preset rule, and preliminarily match the meteorological station combination corresponding to the bus to be predicted; The third acquisition unit is used to acquire meteorological data corresponding to the historical load data from the meteorological station combination; The analysis unit is used to analyze the correlation between the historical load data corresponding to the bus to be predicted and the meteorological data corresponding to the meteorological station combination by using the Pearson correlation coefficient method and the Spearman correlation coefficient method, so as to obtain the meteorological index with the greatest correlation with the historical load data; wherein the meteorological data includes single meteorological data, multi-day accumulated meteorological data and comprehensive meteorological index; The matching unit is used to determine a meteorological station that matches the bus to be predicted based on the meteorological index.

10. The bus load prediction system based on Bayes-XGBoost according to claim 7, characterized in that: The formula of the XGBoost model in the optimization module is specifically: Where ξ is the objective function of the XGBoost model, is the loss function of the XGBoost model, ∑ k φ(f k ) is the regularization term of the XGBoost model, T is the number of decision trees, α is the penalty coefficient of T, β is the regularization penalty coefficient, and ω is the column vector of leaf nodes.

11. The bus load prediction system based on Bayes-XGBoost according to claim 7, characterized in that: The optimization module includes: an initialization unit, an update unit and a training unit; The initialization unit is used to preset the fitness function of the Bayesian optimization algorithm and initialize the parameters and ranges of the fitness function; The updating unit is used to iteratively calculate the fitness value based on the fitness function and the parameters until a termination condition is met, and output the fitness value representing the hyperparameter; wherein, in each iteration, the fitness value is evaluated by a probabilistic proxy model, and the parameters of the Bayesian optimization algorithm and the probabilistic proxy model are updated based on the evaluation results; The training unit is used to input the hyperparameters into the XGBoost model for training to obtain a bus load prediction model for the bus to be predicted.

12. The bus load prediction system based on Bayes-XGBoost according to claim 7, characterized in that: The prediction module includes: a determination unit and a prediction unit; The determining unit is used to confirm the time to be predicted, and obtain the actual meteorological data and the forecast meteorological data corresponding to the bus to be predicted based on the time to be predicted; The prediction unit is used to input the actual meteorological data and the forecast meteorological data corresponding to the time to be predicted and the bus to be predicted into the bus load prediction model to predict the load of the bus to be predicted at the time to be predicted.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the bus load prediction method based on Bayes-XGBoost as described in any one of claims 1 to 6.

Citation Information

Cited By

  • Bus parameter determination method and device, equipment, storage medium and product

    CN121618646A