A data processing method for predicting the occurrence time of key events in flight turnaround

Abstract

The present invention provides a data processing method for predicting the occurrence time of key flight turnaround events, comprising: obtaining a dataset to be processed; the dataset to be processed includes at least one event association information; inputting the dataset to be processed into multiple target occurrence time prediction models to obtain multiple event occurrence time prediction results; weightedly fusing the multiple event occurrence time prediction results to obtain a predicted value of the event occurrence time corresponding to each event association information in the dataset to be processed; and adjusting the predicted value of the event occurrence time corresponding to each event association information based on a preset reference database to obtain a target predicted value of the event occurrence time corresponding to each event association information. The present invention can make the predicted event occurrence time as accurate as possible.

Description

Technical Field

[0001] The present invention relates to the field of computer technology applications, and in particular to a data processing method for predicting the occurrence time of key flight turnaround events. Background Art

[0002] In some application scenarios, it's necessary to predict the occurrence time of fixed events executed by target objects so that corresponding response strategies can be formulated based on the predicted occurrence time. For example, accurate predictions are needed for the occurrence time of each of the many links involved in airport flight operations support, such as flight takeoff and landing, ground services, and resource allocation, in order to improve airport operational efficiency and passenger experience.

[0003] Patent document CN112836905B discloses a method for predicting the time of an event. This patent document collects historical data of basic flight information and time-series message information to train a deep learning neural network model to obtain an event time prediction model. Although this document can improve the accuracy of event time prediction compared to rule-based expert systems, since this document uses a model to predict the time of an event, the accuracy of the event time prediction is still not accurate enough. Summary of the Invention

[0004] In view of the above technical problems, the technical solution adopted by the present invention is:

[0005] An embodiment of the present invention provides a data processing method for predicting the occurrence time of a key flight turnaround event. The key flight turnaround event occurrence time is the occurrence time of the key flight turnaround event experienced by an event-related object from the time it arrives at a specified geographical area to the time it leaves the specified geographical area. The method includes the following steps:

[0006] S100, obtaining a data set to be processed that currently needs to be processed; the data set to be processed includes at least one event association information, and the event association information includes basic information of an event association object, weather data, and the number of users of the event association object.

[0007] S200: Inputting the data set to be processed into multiple target occurrence time prediction models to obtain multiple event occurrence time prediction results.

[0008] S300 , performing weighted fusion on multiple event occurrence time prediction results to obtain a predicted value of the event occurrence time corresponding to each event association information in the data set to be processed.

[0009] S400, based on the preset reference database, adjust the predicted value of the event occurrence time corresponding to each event association information to obtain the target predicted value of the event occurrence time corresponding to each event association information. The preset reference database stores multiple reference data, each reference data includes the corresponding event association information and the real time of the event occurrence.

[0010] The present invention has at least the following beneficial effects:

[0011] An embodiment of the present invention provides a data processing method for predicting the occurrence time of key events in flight turnaround. The event occurrence time is obtained based on the fusion weighted result of the prediction results of multiple target occurrence time prediction models. Each target occurrence time prediction model is used to predict the occurrence time of the corresponding event, and the influence of different factor combinations on the occurrence time of different events is considered during the training process, which enables the model to learn the complex relationship between events, and uses historical data to correct the prediction results of the model, which can make the predicted event occurrence time more accurate than the results predicted by existing event occurrence time prediction methods.

[0012] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0014] Figure 1 A flowchart of a data processing method for predicting the occurrence time of key flight turnaround events provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0016] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used herein in the specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0017] It should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the steps as sequential processes, many of the steps can be performed in parallel, concurrently, or simultaneously. In addition, the order of the steps can be rearranged. A process can be terminated when its operation is completed, but can also have additional steps not included in the accompanying drawings. A process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0018] An embodiment of the present invention provides a data processing method for predicting the occurrence time of key flight turnaround events. The occurrence time of key flight turnaround events is the occurrence time of the key flight turnaround events experienced by event-related objects from arriving at a specified geographical area to leaving the specified geographical area.

[0019] In a specific application scenario, the event-related object may be a flight, the designated geographical area may be an airport, and the key flight turnaround events experienced by the event-related object from arriving at the designated geographical area to leaving the designated geographical area may be flight operation support events. Flight operation support events refer to events that occur in the flight operation support process, which may include wheel block time, cabin door opening time, passenger disembarkation start time, passenger disembarkation completion time, cleaning start time, cleaning completion time, refueling start time, refueling completion time, catering start time, catering end time, boarding start time, boarding end time, cabin door closing time, wheel block removal time, expected take-off time and other events.

[0020] Furthermore, in an embodiment of the present invention, the data processing method for predicting the occurrence time of key flight turnaround events provided by the embodiment of the present invention may include the following steps: Figure 1 Steps shown:

[0021] S100, obtaining a data set to be processed that currently needs to be processed; the data set to be processed includes at least one event association information, and the event association information includes basic information of an event association object, weather data, and the number of users of the event association object.

[0022] In an embodiment of the present invention, when the event-related object is a flight, the basic information of the event-related object may include the aircraft number, aircraft model, route, planned departure date, operating status, planned parking space, and actual arrival time. The operating status may include normal, early, and delayed.

[0023] In an embodiment of the present invention, the weather data in the event association information may be weather data for a set time period after the event association object is scheduled to arrive at a specified geographical area. In one exemplary embodiment, the set time period may be three hours. The weather data includes at least parameters such as temperature, wind speed, humidity, and weather characteristics. The weather characteristics may include light rain, moderate rain, thunderstorm, heavy rain, heavy fog, sandstorm, etc.

[0024] In the embodiment of the present invention, the number of users of the event association object refers to the number of users who use the event association object. In the case where the event association object is a flight, the users may be passengers.

[0025] S200: Inputting the data set to be processed into multiple target occurrence time prediction models to obtain multiple event occurrence time prediction results.

[0026] S300, performing weighted fusion on multiple event occurrence time prediction results to obtain a predicted value of the event occurrence time corresponding to each event association information in the data set to be processed;

[0027] In the embodiment of the present invention, the predicted value t of the vth event occurrence time corresponding to each event association information is v pre The following conditions are met: v pre =w1×t 1v pre +w2×t 2v pre +……+w z ×t zv pre +……+w Q ×t Qv pre , v ranges from 1 to Q, Q is the number of event occurrence times that need to be predicted, that is, the number of flight turnaround key event occurrence times that need to be predicted, w z is the weight of the prediction model for the occurrence time of the zth target, t zv pre The predicted value of the vth event occurrence time predicted by the zth target occurrence time prediction model, where z ranges from 1 to Q.

[0028] Furthermore, w z The following conditions must be met:

[0029] w z =AC z / (AC1+AC2+……+AC z +……+AC Q ), AC zThe prediction accuracy of the z-th target occurrence time prediction model for the corresponding event occurrence time.

[0030] S400 , based on a preset reference database, adjusting the predicted value of the event occurrence time corresponding to each event association information to obtain a target predicted value of the event occurrence time corresponding to each event association information.

[0031] Furthermore, S400 specifically includes:

[0032] S401, for each event association information, obtain the similarity between the event association information and the event association information corresponding to each reference data in the preset reference database, and obtain a similarity set corresponding to the event association information; wherein, the preset reference database stores multiple reference data, and each reference data includes corresponding event association information and the real time when the event occurred.

[0033] In the embodiment of the present invention, the similarity may be cosine similarity.

[0034] S402 : taking the reference data corresponding to the similarity set corresponding to each event association information and having a similarity greater than or equal to a set similarity threshold as the association data of the event association information.

[0035] In the embodiment of the present invention, the similarity threshold may be set to an empirical value, for example, 0.9.

[0036] Those skilled in the art know that if the similarity set corresponding to a certain event-related information does not have a similarity greater than or equal to the set similarity threshold, then the event-related information does not have associated data, which also means that there is no need to correct the event occurrence time corresponding to the event-related information.

[0037] S403, adjusting the predicted value of the event occurrence time corresponding to each event association information based on the real time of the event occurrence time corresponding to the associated data of each event association information, to obtain a target predicted value of the event occurrence time corresponding to each event association information.

[0038] Specifically, the target prediction value t of the vth event occurrence time corresponding to each event association information is v The following conditions must be met:

[0039] t v =t v pre +SUM(w e ×△t ev ) e=1……H .

[0040] Among them, w e is the weight of the e-th associated data corresponding to the event association information, we It is positively correlated with the similarity between the event association information and the event association information in the corresponding e-th association data. Specifically, it can be equal to the similarity between the event association information and the event association information in the corresponding e-th association data. The value of e ranges from 1 to H, and H is the number of association data of the event association information. ev t v pre The difference between the real time of the corresponding associated data, specifically t v pre The difference between the real time of the event association information and the occurrence time of the vth event in the corresponding eth association data.

[0041] Furthermore, in an embodiment of the present invention, the multiple target occurrence time prediction models can be obtained through the following steps:

[0042] S10, based on the basic data set, training multiple pre-built initial time prediction models to obtain an initial time prediction model for predicting the occurrence time of each event.

[0043] S20, adjusting the hyperparameter combination of the initial time prediction model of each event occurrence time, obtaining the optimal hyperparameter combination of the initial time prediction model of each event occurrence time, and obtaining the intermediate time prediction model of each event occurrence time.

[0044] S30 , based on the initial input features and the intermediate occurrence time prediction model of each event occurrence time, obtain the associated combined features of each event occurrence time.

[0045] S40, adding the associated combined features of each event occurrence time to the initial input features to obtain the target input features of each event occurrence time, and based on the target input features of each event occurrence time and the basic data set, training the intermediate time prediction model of each event occurrence time to obtain the trained intermediate time prediction model of each event occurrence time as the target occurrence time prediction model of each event occurrence time.

[0046] In the embodiment of the present invention, the multiple target occurrence time prediction models may be completely different, or may have some identical models.

[0047] Furthermore, S10 may specifically include:

[0048] S11, preprocessing the acquired basic data set to obtain a preprocessed basic data set as a training data set.

[0049] In an embodiment of the present invention, the basic data set includes a plurality of basic data within a set historical time period, and each basic data includes corresponding event association information and the real time when the event occurred.

[0050] In an embodiment of the present invention, the preprocessing may include: filling missing data in each basic data to obtain basic data after filling; and removing duplicate data and abnormal values ​​in the basic data after filling.

[0051] In an embodiment of the present invention, if the missing rate of a certain basic data in the basic data set is less than a set missing rate, the data is filled, otherwise, the data is deleted. The set missing rate can be set based on actual needs, for example, it can be 0.8.

[0052] In the embodiment of the present invention, the data may be filled using an existing data filling method, for example, using mean or median filling.

[0053] In the embodiment of the present invention, a standard score method may be used to remove outliers in the basic data.

[0054] In this embodiment of the present invention, the training dataset includes an input feature dataset and an output feature dataset. The input feature dataset is a dataset generated based on input features. The input features may include aircraft number, aircraft model, route, planned departure date, operating status, planned parking bay, actual arrival time, weather data, and number of users. The output feature dataset is a dataset generated based on the actual time at which an event occurs.

[0055] S12: Using the training data set, train each initial time prediction model to obtain multiple trained time prediction models, which are used as multiple first time prediction models respectively.

[0056] In an embodiment of the present invention, the initial time prediction model may be an existing data analysis model, such as a linear regression model, a random forest model, a long short-term memory network model, etc. The hyperparameters of the initial time prediction model may be randomly initialized hyperparameters.

[0057] Any method known to those skilled in the art for training each initial time prediction model using the training data set to obtain multiple trained time prediction models falls within the scope of the present invention. For example, in an exemplary embodiment of the present invention, training can be performed using a time series cross-validation approach, such as a 5-fold cross-validation. Each time training is performed using the divided training set data, the loss can be calculated using a mean square error loss function.

[0058] S13, obtaining the prediction performance evaluation value of each first time prediction model for each event occurrence time, and for each event occurrence time, using the first time prediction model with the highest prediction performance evaluation value for the event occurrence time as the initial time prediction model for the event occurrence time.

[0059] In an embodiment of the present invention, the prediction performance evaluation value includes a prediction stability evaluation value that characterizes the prediction stability of the model and at least one prediction error evaluation value that characterizes the prediction error of the model. A larger prediction stability evaluation value indicates a more stable model. A smaller prediction error evaluation value indicates a more accurate model.

[0060] In an exemplary embodiment, the prediction performance evaluation value satisfies the following conditions: P=W F ×PF+(1-SUM(W a ×PC a ) a=1……d ), where PF is the predicted stability evaluation value, W F is the weight of PF, PC a is the a-th prediction error evaluation value, a ranges from 1 to d, d is the number of prediction accuracy evaluation values, W a For PC a The weight, W F >W a In one exemplary embodiment, all prediction accuracy evaluation values ​​are weighted equally. In one exemplary embodiment, W F >0.5, W a <0.5, in one embodiment, W F =0.7, W a =0.3. SUM(W a ×PC a ) a=1……d Indicates the pair W1×PC1 to W d ×PC d Perform sum operation, i.e. SUM(W a ×PC a ) a=1……d =(W1×PC1)+……+(W a ×PC a )+……+(W d ×PC d ).

[0061] In an embodiment of the present invention, the prediction error evaluation value may include mean square error, mean absolute error, root mean square error, mean absolute percentage error, etc.

[0062] In this embodiment of the present invention, the prediction stability evaluation value of each first time prediction model for the j-th event occurrence time satisfies the following conditions:

[0063] PF j =1 / (1+(1 / (A1 / ((1 / N))×SUM(|F ij -Y ij |) i=1……N )))), where A1 is the intermediate quantity, A1=1 / (N-1)×SUM((|F （k+1）j -Y （k+1）j |-|F kj -Y kj |) k=1……N-1 ), SUM(|F ij -Y ij |) i=1……N ) indicates the value of |F 1j -Y 1j |To|F Nj -Y Nj | Perform sum operation, SUM((|F （k+1）j -Y （k+1）j |-|F kj -Y kj |) k=1……N-1 ) indicates the value of (|F 2j -Y 2j |-|F 1j -Y 1j |) to (|F Nj -Y Nj |-|F （N-1）j -Y （N-1）j |) to perform sum operation, F ij Y represents the predicted time of the jth event occurrence time obtained by each first-time prediction model based on the i-th training data in the training dataset, ij is the real time of the occurrence of the jth event corresponding to the i-th training data, i ranges from 1 to N, and N is the number of training data. kj Y represents the predicted time of the jth event occurrence time obtained by each first-time prediction model based on the kth training data in the training dataset, kj F kj The true value of , that is, the true time of the jth event corresponding to the kth training data, the value of k ranges from 1 to N-1, the value of j ranges from 1 to Q, and the initial value of j is 1.

[0064] Furthermore, in S20, the intermediate time prediction model of the j-th event occurrence time is obtained by the following steps:

[0065] S201, for the initial time prediction model of the j-th event occurrence time, randomly generate m hyperparameter combinations.

[0066] In the embodiment of the present invention, m can be set based on actual needs and is a value greater than 1.

[0067] S202, using the training data set to train the initial time prediction model of the jth event occurrence time with the hth hyperparameter combination, to obtain the prediction performance evaluation value P of the initial time prediction model of the jth event occurrence time under the hth hyperparameter combination jh , the value of h ranges from 1 to m, and the initial value of h is 1.

[0068] In the embodiment of the present invention, P jh =W F ×PF ih +(1-SUM(W a ×PC ih-a ) a=1……d ), PF ih is the prediction stability evaluation value of the initial time prediction model for the jth event occurrence time under the hth hyperparameter combination, PC ih-a is the ath prediction error evaluation value of the initial time prediction model for the jth event occurrence time under the hth hyperparameter combination. In this embodiment of the present invention, when training the initial time prediction model for the jth event occurrence time with the hth hyperparameter combination using the training dataset, the loss of the model may be the cross entropy loss of the jth event occurrence time.

[0069] S203, taking the hyperparameter combination corresponding to the largest of the m prediction performance evaluation values ​​corresponding to the initial time prediction model of the j-th event occurrence time as the optimal hyperparameter combination of the initial time prediction model of the j-th event occurrence time, and obtaining the intermediate time prediction model of each event occurrence time.

[0070] In this embodiment of the present invention, when adjusting the parameters of the optimal model for each event occurrence time, the prediction stability evaluation value, which represents the level of prediction reliability, serves as an important reference indicator to guide the parameter adjustment direction. When adjusting different parameters, the model is trained multiple times for each parameter combination. In addition to observing the conventional prediction error indicators, the corresponding prediction stability evaluation value is also calculated, thereby improving the prediction performance of the selected prediction model.

[0071] Furthermore, in S30, the associated combination features of the j-th event occurrence time are obtained by the following steps:

[0072] S301, obtaining an associated feature set corresponding to the initial input feature.

[0073] In an embodiment of the present invention, the associated feature set is a set formed by the classification features corresponding to each feature in the initial input features, that is, each associated feature in the associated feature set is a classification feature. The classification feature corresponding to each feature is determined based on the feature type contained in the feature. For example, for the feature machine number, according to the basic data set, it can be known that g types of machine numbers are included, then the classification features corresponding to the machine number include g classification features, corresponding to g types of machine numbers respectively. For another example, for the feature temperature, according to the basic data set, the temperature can be divided into three temperature types: high, medium and low. Then the classification features corresponding to the temperature include three classification features: high temperature, medium temperature and low temperature. For another example, for the feature number of users, according to the basic data set, the number of users can be divided into three number types: many, standard and few. Then the classification features corresponding to the number of users include three classification features: many users, standard users and few users.

[0074] S302. Based on the associated feature set, multiple associated feature combinations are obtained; wherein any associated feature combination includes a first combination and a second combination, the first combination includes at least one associated feature, the second combination includes at least one associated feature, and the first combination and the second combination do not have the same associated features. For example, a certain associated feature combination is: (heavy rain, small number of users), and another example is: (heavy rain + high wind speed, small number of users).

[0075] S303, based on the intermediate time prediction model of the occurrence time of the j-th event, obtain the association weight of each association feature combination; wherein the association weight corresponding to each association feature combination satisfies the following condition: I(X1, X2)=log2[P(X1, X2) / P(X1)P(X2)], wherein I(X1, X2) is the association weight of the association feature combination (X1, X2), X1 is the first combination in the association feature combination (X1, X2), X2 is the second combination in the association feature combination (X1, X2), P(X1, X2) is the joint weight of the association feature combination (X1, X2), P(X1) is the edge weight of X1, and P(X2) is the edge weight of X2.

[0076] In an embodiment of the present invention, P(X1, X2)=N(X1, X2) / TN(X1, X2), where N(X1, X2) is the number of data containing the associated feature combination (X1, X2) in the prediction result of the intermediate time prediction model of the j-th event occurrence time, indicating a decrease in prediction performance when the associated feature combination (X1, X2) is used as an input feature, and TN(X1, X2) is the total number of data in the preprocessed basic data set containing the associated feature combination (X1, X2).

[0077] In an embodiment of the present invention, P(X1) satisfies the following condition: P(X1)=TP(X1) / TN(X1), where TP(X1) is the number of data in which the prediction performance of the data containing the first combination X1 is reduced in the prediction result of the intermediate time prediction model of the j-th event occurrence time when the associated feature combination (X1, X2) is used as the input feature, and TN(X1) is the total number of data containing the first combination in the preprocessed basic data set.

[0078] In an embodiment of the present invention, P(X2) satisfies the following condition: P(X2)=TP(X2) / TN(X2), where TP(X2) is the number of data in which the prediction performance of the data containing the second combination X2 is reduced in the prediction result of the intermediate time prediction model of the j-th event occurrence time when the associated feature combination (X1, X2) is used as the input feature, and TN(X2) is the total number of data containing the second combination in the preprocessed basic data set.

[0079] In this embodiment of the present invention, N(X1, X2), TN(X1), and TN(X2) can be obtained by the following steps:

[0080] S1, add the data corresponding to N(X1, X2) to the data containing (X1, X2) in the input feature data set to obtain a new input feature data set;

[0081] S2, using the new input feature dataset and the output feature dataset to train the initial time prediction model for the j-th event occurrence time, and obtain corresponding training results, wherein the training results include the prediction results of the event occurrence time corresponding to the new input feature dataset;

[0082] S3, based on the training results and the corresponding real results, obtain prediction performance evaluation values ​​for the occurrence times of j events corresponding to the new input feature dataset, and use the prediction performance evaluation value corresponding to the data containing (X1, X2) in the input feature dataset as the first prediction performance evaluation value to be processed, use the prediction performance evaluation value corresponding to the data containing X1 in the input feature dataset as the second prediction performance evaluation value to be processed, and use the prediction performance evaluation value corresponding to the data containing X2 in the input feature dataset as the third prediction performance evaluation value to be processed;

[0083] S4, traverse the u-th pending prediction performance evaluation value. For a certain u-th pending prediction performance evaluation value traversed, if the comparison result of the u-th pending prediction performance evaluation value and the corresponding reference prediction performance evaluation value indicates that the prediction performance corresponding to the u-th pending prediction performance evaluation value is lower than the prediction performance corresponding to the reference prediction performance evaluation value, set the u-th counter Cu=Cu+1, and the initial value of Cu is 0; the reference prediction performance evaluation value is the prediction performance evaluation value of the j-th event occurrence time obtained by the intermediate time prediction model of the j-th event occurrence time; u=1, 2, 3.

[0084] In an embodiment of the present invention, if the prediction stability evaluation value and the prediction error evaluation value in the u-th to-be-processed prediction performance evaluation value are respectively smaller than the prediction stability evaluation value and the prediction error evaluation value in the corresponding reference prediction performance evaluation value, it means that the prediction performance corresponding to the u-th to-be-processed prediction performance evaluation value is lower than the prediction performance corresponding to the corresponding reference prediction performance evaluation value, indicating that the prediction performance of the model has declined after the associated feature combination is added.

[0085] In the embodiment of the present invention, if the prediction performance of the model decreases, it means that the added feature combination has an impact on the prediction performance of the model, which can be used as an influencing feature of the event occurrence time.

[0086] S5, set N(X1, X2) = C1, TN(X1) = C2, TN(X2) = C3.

[0087] S304: The association feature combination with the maximum association weight is used as the association combination feature of the occurrence time of the j events.

[0088] In summary, the data processing method for predicting the occurrence time of key flight turnaround events provided by the embodiment of the present invention is based on the weighted fusion of the prediction results of multiple target occurrence time prediction models. Each target occurrence time prediction model is used to predict the corresponding event occurrence time, and the influence of different factor combinations on the occurrence time of different events is considered during the training process, which enables the model to learn the complex relationship between events. In addition, the reliability and error of the model are comprehensively considered during the training process, which makes the prediction performance of the model more reliable. In addition, the prediction results of the model are corrected using historical data, which can make the predicted event occurrence time more accurate than the results predicted by the existing event occurrence time prediction method.

[0089] An embodiment of the present invention also provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are configured to execute the method described in the embodiment of the present invention.

[0090] An embodiment of the present invention further provides a computer-readable storage medium storing computer-executable instructions, wherein the computer instructions are used to execute the method described in the embodiment of the present invention.

[0091] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in the present invention can be achieved. This is not limited herein.

[0092] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A data processing method for predicting the occurrence time of key flight turnaround events, characterized in that: The occurrence time of the key flight turnaround event is the occurrence time of the key flight turnaround event experienced by the event-related object from the time of arriving at the designated geographical area to the time of leaving the designated geographical area. The method includes the following steps: S100, obtaining a data set to be processed that currently needs to be processed; the data set to be processed includes at least one event association information, the event association information including basic information of the event association object, weather data, and the number of users of the event association object; S200, inputting the data set to be processed into multiple target occurrence time prediction models to obtain multiple event occurrence time prediction results; S300, performing weighted fusion on multiple event occurrence time prediction results to obtain a predicted value of the event occurrence time corresponding to each event association information in the data set to be processed; S400, adjusting the predicted value of the event occurrence time corresponding to each event association information based on a preset reference database to obtain a target predicted value of the event occurrence time corresponding to each event association information, wherein the preset reference database stores a plurality of reference data, each reference data including corresponding event association information and the actual time of the event occurrence; The multiple target occurrence time prediction models are obtained by the following steps: S10, based on the basic data set, training multiple pre-built initial time prediction models to obtain an initial time prediction model for predicting the occurrence time of each event; S20, adjusting the hyperparameter combination of the initial time prediction model for each event occurrence time, obtaining the optimal hyperparameter combination of the initial time prediction model for each event occurrence time, and obtaining the intermediate time prediction model for each event occurrence time; S30, obtaining associated combined features of each event occurrence time based on the initial input features and the intermediate occurrence time prediction model of each event occurrence time; S40, adding the associated combined features of each event occurrence time to the initial input features to obtain the target input features of each event occurrence time, and training an intermediate time prediction model for each event occurrence time based on the target input features of each event occurrence time and the basic data set, obtaining the trained intermediate time prediction model for each event occurrence time as the target occurrence time prediction model for each event occurrence time; In S30, the associated combined features of the j-th event occurrence time are obtained through the following steps: S301, obtaining an associated feature set corresponding to the initial input feature; the associated feature set is a set formed by the classification features corresponding to each feature in the initial input feature; S302, based on the associated feature set, obtaining multiple associated feature combinations; wherein any associated feature combination includes a first combination and a second combination, the first combination includes at least one associated feature, the second combination includes at least one associated feature, and the first combination and the second combination do not have the same associated feature; S303: Based on the intermediate time prediction model of the j-th event occurrence time, obtain the association weight of each association feature combination; wherein the association weight corresponding to each association feature combination satisfies the following condition: I(X1, X2) = log2[P(X1, X2) / P(X1)P(X2)], wherein I(X1, X2) is the association weight of the association feature combination (X1, X2), X1 is the first combination in the association feature combination (X1, X2), X2 is the second combination in the association feature combination (X1, X2), P(X1, X2) is the joint weight of the association feature combination (X1, X2), P(X1) is the edge weight of X1, and P(X2) is the edge weight of X2; S304: The association feature combination with the maximum association weight is used as the association combination feature of the occurrence time of the j events.

2. The method according to claim 1, characterized in that In S300, the predicted value t of the vth event occurrence time corresponding to each event association information is v pre The following conditions are met: v pre =w1×t 1v pre +w2×t 2v pre +……+w z ×t zv pre +……+w Q ×t Qv pre , v ranges from 1 to Q, Q is the number of event occurrence times that need to be predicted, w z is the weight of the prediction model for the occurrence time of the zth target, t zv pre The predicted value of the vth event occurrence time predicted by the zth target occurrence time prediction model, where z ranges from 1 to Q.

3. The method according to claim 2, characterized in that w z The following conditions must be met: w z =AC z / (AC1+AC2+……+AC z +……+AC Q ), AC z The prediction accuracy of the z-th target occurrence time prediction model for the corresponding event occurrence time.

4. The method according to claim 2, characterized in that S400 specifically includes: S401, for each event association information, obtaining the similarity between the event association information and the event association information corresponding to each reference data in a preset reference database, and obtaining a similarity set corresponding to the event association information; S402, taking reference data corresponding to a similarity greater than or equal to a set similarity threshold in a similarity set corresponding to each event association information as association data of the event association information; S403, adjusting the predicted value of the event occurrence time corresponding to each event association information based on the real time of the event occurrence time corresponding to the associated data of each event association information, to obtain a target predicted value of the event occurrence time corresponding to each event association information.

5. The method according to claim 4, characterized in that The target prediction value t of the vth event occurrence time corresponding to each event association information v The following conditions must be met: t v =t v pre +SUM(w e ×△t ev ) e=1……H ; Among them, w e is the weight of the e-th associated data corresponding to the event association information, w e It is positively correlated with the similarity between the event association information and the event association information in the corresponding e-th association data. The value of e ranges from 1 to H, and H is the number of association data of the event association information. ev t v pre The difference between the real time of the event association information and the occurrence time of the vth event in the corresponding eth association data.

6. The method according to claim 1, characterized in that In S20, the intermediate time prediction model of the j-th event occurrence time is obtained by the following steps: S201, for the initial time prediction model of the j-th event occurrence time, randomly generate m hyperparameter combinations; j ranges from 1 to Q, and the initial value of j is 1, and Q is the number of event occurrence times to be predicted; S202, using the training data set to train the initial time prediction model of the jth event occurrence time with the hth hyperparameter combination, to obtain the prediction performance evaluation value P of the initial time prediction model of the jth event occurrence time under the hth hyperparameter combination jh , the value of h ranges from 1 to m, and the initial value of h is 1; S203, taking the hyperparameter combination corresponding to the largest of the m prediction performance evaluation values ​​corresponding to the initial time prediction model of the j-th event occurrence time as the optimal hyperparameter combination of the initial time prediction model of the j-th event occurrence time, and obtaining the intermediate time prediction model of the j-th event occurrence time.

7. The method according to claim 1, characterized in that in, P(X1, X2) = N(X1, X2) / TN(X1, X2), where N(X1, X2) is the number of data that show a decrease in prediction performance in the prediction results of the intermediate time prediction model for the j-th event time when the associated feature combination (X1, X2) is used as the input feature. TN(X1, X2) is the total number of data in the preprocessed basic data set that contains the associated feature combination (X1, X2). P(X1) satisfies the following condition: P(X1)=TP(X1) / TN(X1), where TP(X1) is the number of data with decreased prediction performance in the data containing the first combination X1 in the prediction results of the intermediate time prediction model for the j-th event occurrence time when the associated feature combination (X1, X2) is used as the input feature, and TN(X1) is the total number of data containing the first combination in the preprocessed basic data set; P(X2) satisfies the following conditions: P(X2)=TP(X2) / TN(X2), where TP(X2) is the number of data with decreased prediction performance in the data containing the second combination X2 in the prediction results of the intermediate time prediction model of the j-th event occurrence time when the associated feature combination (X1, X2) is used as the input feature, and TN(X2) is the total number of data with the second combination in the preprocessed basic data set.

Images