Route passenger volume prediction method, device, electronic device and storage medium

Abstract

The present invention provides a route passenger volume prediction method, device, electronic device and storage medium, which relates to the field of data processing technology. The method comprises: predicting a first passenger volume prediction value of the route to be predicted in the period to be predicted according to the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted; obtaining at least one associated route associated with the route to be predicted according to the historical passenger volume of the route to be predicted and the departure airport; predicting a second passenger volume prediction value of the route to be predicted in the period to be predicted according to the historical passenger volume of each associated route, the passenger flow correlation between each associated route and the route to be predicted, the passenger preference corresponding to each associated route, and the passenger preference corresponding to the route to be predicted; obtaining a final passenger volume prediction value of the route to be predicted in the period to be predicted according to the first passenger volume prediction value and the second passenger volume prediction value. The present invention can improve the prediction accuracy of route passenger volume.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, device, electronic device and storage medium for predicting the volume of airline passengers. Background Art

[0002] As a hub of aviation network transportation, the understanding and prediction of international passenger volume will greatly affect the allocation and decision-making of airport-related resources, and then affect the inbound and outbound travel experience of passengers. Therefore, the prediction of international passenger volume will have outstanding practical significance for airport operations and the development of the civil aviation industry.

[0003] In related technologies, a large amount of historical travel data is usually considered, that is, the periodicity and trend characteristics of historical passenger flow to predict passenger volume. However, there are many factors that affect passenger volume. It is difficult to accurately predict passenger volume by relying solely on historical travel data.

[0004] Therefore, there is an urgent need for a route passenger volume prediction method, device, electronic device and storage medium to solve the above problems. Summary of the invention

[0005] The present invention provides a route passenger volume prediction method, device, electronic device and storage medium, which are used to solve the defect that the existing technology only relies on historical travel data and is difficult to accurately predict the passenger volume, so as to improve the prediction accuracy of the passenger volume.

[0006] The present invention provides a method for predicting the passenger volume of an airline route, comprising:

[0007] According to the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted, predicting a first passenger volume forecast value of the route to be predicted in the period to be predicted;

[0008] Acquire at least one associated route associated with the route to be predicted according to the historical passenger volume and the departure airport of the route to be predicted;

[0009] According to the historical volume of each of the associated routes, the passenger flow correlation between each of the associated routes and the route to be predicted, the passenger preference corresponding to each of the associated routes, and the passenger preference corresponding to the route to be predicted, a second passenger volume forecast value of the route to be predicted in the period to be predicted is forecasted;

[0010] According to the first passenger volume prediction value and the second passenger volume prediction value, a final passenger volume prediction value of the to-be-predicted route in the to-be-predicted period is obtained.

[0011] According to a route passenger volume prediction method provided by the present invention, the method predicts a first passenger volume prediction value of the route to be predicted in the period to be predicted based on the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted, including:

[0012] In the data storage system, the passenger volume of the route to be predicted in multiple periods before the period to be predicted is collected to obtain the historical passenger volume of the route to be predicted;

[0013] Inputting the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted into a first prediction model to obtain the first passenger volume prediction value;

[0014] Among them, the first prediction model is obtained by iteratively optimizing the seasonally adjusted autoregressive moving average model based on the historical passenger volume of the sample route in multiple periods before the preset historical period, the characteristic data of the preset historical period, and the actual passenger volume of the sample route in the preset historical period.

[0015] According to a route passenger volume prediction method provided by the present invention, obtaining at least one associated route associated with the route to be predicted based on the historical passenger volume and the departure airport of the route to be predicted includes:

[0016] In the data storage system, all target routes having the same departure airport as the route to be predicted are obtained;

[0017] In the data storage system, the passenger volume of each target route in multiple periods before the period to be predicted is collected to obtain the historical passenger volume of each target route;

[0018] Calculating the similarity between the historical passenger volume of each target route and the historical passenger volume of the route to be predicted respectively, and obtaining the passenger flow correlation between each target route and the route to be predicted;

[0019] At least one associated route is determined from at least one target route according to the passenger flow correlation between each of the target routes and the route to be predicted.

[0020] According to a route passenger volume prediction method provided by the present invention, the method predicts a second passenger volume prediction value of the route to be predicted in the period to be predicted based on the historical number of trips of each of the associated routes, the passenger flow correlation between each of the associated routes and the route to be predicted, the passenger preference corresponding to each of the associated routes, and the passenger preference corresponding to the route to be predicted, including:

[0021] Calculating the similarity between the historical passenger volume of each of the associated routes and the historical passenger volume of the route to be predicted respectively, and obtaining the passenger flow correlation between each of the associated routes and the route to be predicted;

[0022] Dividing the passenger preference degree corresponding to each of the associated routes by the passenger preference degree corresponding to the route to be predicted, to obtain a relative preference degree corresponding to each of the associated routes;

[0023] Inputting the historical volume of each of the associated routes, the passenger flow correlation between each of the associated routes and the route to be predicted, and the relative preference into a second prediction model to obtain a second passenger volume prediction value of the route to be predicted in the period to be predicted;

[0024] The second prediction model is obtained by training the initial neural network model based on the historical travel volume of each associated route of the sample route in the previous period of the preset historical period, the passenger flow correlation between each associated route of the sample route and the sample route, the relative preference corresponding to each associated route of the sample route, and the actual passenger volume of the sample route in the preset historical period.

[0025] According to a route passenger volume prediction method provided by the present invention, the passenger preference corresponding to each of the associated routes and the routes to be predicted is obtained based on the following steps:

[0026] Using a preset time period, the city area to which each of the routes belongs, and travel-related semantic words as keywords, crawling at least one related social media data corresponding to each of the routes from a social media platform;

[0027] Extracting the passenger preference parameters corresponding to each of the routes from the associated social media data;

[0028] Calculating the passenger preference degree corresponding to each of the routes according to the passenger preference parameters;

[0029] The preset time period is determined according to the period to be predicted.

[0030] According to a route passenger volume prediction method provided by the present invention, the passenger preference parameters corresponding to each route are extracted from each of the related social media data, including:

[0031] Extracting the number of interests and the number of followers corresponding to each city in each of the routes from the associated social media data;

[0032] According to the number of interests and the number of concerns corresponding to each city in each of the routes, the passenger preference parameters corresponding to each of the routes are obtained.

[0033] According to a route passenger volume prediction method provided by the present invention, obtaining a final passenger volume prediction value of the route to be predicted in the period to be predicted according to the first passenger volume prediction value and the second passenger volume prediction value comprises:

[0034] Inputting the first passenger volume prediction value and the second passenger volume prediction value into a weighted fusion model to obtain the final passenger volume prediction value;

[0035] The weighted fusion model is obtained by optimizing the initial fusion model with the optimization goal of minimizing the error between the final passenger volume prediction value of the sample route in the preset historical period and the actual passenger volume of the sample route in the preset historical period;

[0036] The final passenger volume prediction value of the sample route in the preset historical period is obtained by performing a fusion prediction on the first passenger volume prediction value of the sample route in the preset historical period and the second passenger volume prediction value of the sample route in the preset historical period based on the initial fusion model.

[0037] The present invention also provides a route passenger volume prediction device, comprising:

[0038] A first prediction module is used to predict a first passenger volume prediction value of the route to be predicted in the period to be predicted based on the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted;

[0039] A first acquisition module, configured to acquire at least one associated route associated with the route to be predicted according to the historical passenger volume and the departure airport of the route to be predicted;

[0040] A second prediction module is used to predict a second passenger volume forecast value of the to-be-predicted route in the to-be-predicted period according to the historical passenger volume of each of the associated routes, the passenger flow correlation between each of the associated routes and the to-be-predicted route, the passenger preference degree corresponding to each of the associated routes, and the passenger preference degree corresponding to the to-be-predicted route;

[0041] The second acquisition module is used to obtain the final passenger volume forecast value of the to-be-forecasted route in the to-be-forecasted period according to the first passenger volume forecast value and the second passenger volume forecast value.

[0042] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the above-mentioned methods for predicting passenger volume on an air route is implemented.

[0043] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for predicting passenger volume of an airline as described in any one of the above is implemented.

[0044] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned methods for predicting passenger volume on an air route.

[0045] The route passenger volume prediction method, device, electronic device and storage medium provided by the present invention predict the initial passenger volume of the route to be predicted based on the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted, introduce the historical travel passenger volume similarity matrix of each route, obtain the historical trip volume of the associated routes associated with the historical passenger volume of the route to be predicted and the passenger flow correlation between each associated route and the route to be predicted, and combine with social media data to obtain the passenger preference corresponding to each associated route and the passenger preference corresponding to the route to be predicted, so as to correct the initial passenger volume prediction value based on the historical trip volume of the associated routes and the passenger flow correlation between each associated route and the route to be predicted, as well as the passenger preference corresponding to each associated route and the passenger preference corresponding to the route to be predicted, thereby improving the accuracy of the final passenger volume prediction, so as to solve the problem that the prior art only relies on historical trip data for prediction and ignores other influencing factors, resulting in poor prediction accuracy in the final prediction effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 This is one of the flow charts of the route passenger volume prediction method provided by the present invention;

[0048] Figure 2 This is the second flow chart of the route passenger volume prediction method provided by the present invention;

[0049] Figure 3 It is a structural schematic diagram of the route passenger volume prediction device provided by the present invention;

[0050] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] As the demand for air transport continues to pay attention to passenger volume forecasting, its demand forecasting theories and methods emerge in an endless stream, including regression models, gravity models, time series models, neural networks, gray models, etc. For example, in the prior art, by obtaining the historical passenger volume feature data of the product to be predicted and the date feature data on the day to be predicted, calculating the interval day, and inputting the feature data and the normalized date feature data into the trained passenger volume forecasting model, the passenger volume forecast value with specific time characteristics is obtained.

[0053] In the above method, the periodicity and trend characteristics of the passenger volume data from the same origin to destination (OD) or airport in the time series are mainly used to predict the passenger volume of the same OD or airport in the future, which can meet the actual application needs to a certain extent. Although this model provides a method for passenger volume prediction in the route or airport dimension, since the prediction model only considers a large amount of historical row data, that is, the periodicity and trend of historical passenger flow, it lacks consideration of the special nature of the predicted date or the traffic attraction brought to the destination by accidental events, as well as the impact of similar related routes on the traffic of the predicted route, resulting in the problem of poor prediction accuracy in the final prediction effect of the existing technology.

[0054] By conducting a statistical comparison and analysis of the outbound passenger volume on different routes of an airport, it is found that there is a certain correlation between the traffic between routes. When the passenger volume of a certain route increases, the passenger volume of routes with strong correlation will also increase significantly. In other words, there may be similar passenger flow patterns between different ODs. In addition, social media has been proven to be a very credible and reliable data source to assist in passenger flow forecasting in recent years, and it can reflect passengers' current travel intentions and needs to a certain extent. Therefore, when predicting passenger volume, considering these two types of information to correct the passenger flow forecast value based on historical travel data to supplement the forecast will help improve the accuracy of the final forecast, and can to a certain extent solve the technical defect that the existing technology cannot predict the passenger volume of newly opened routes.

[0055] Therefore, in order to solve the problem that the existing prediction technology only considers a large amount of historical travel data, that is, the periodicity and trend influence of historical passenger flow, and lacks consideration of the traffic attraction brought to the destination by the particularity of the prediction day or the occurrence of accidental events, as well as the influence of similar related routes on the predicted route traffic, resulting in certain deficiencies in the final prediction effect of the existing technology, the present invention proposes a route passenger volume prediction method, which introduces social media data and considers the influence of the correlation between the destination passenger volumes of different routes on the basis of making an initial passenger volume prediction based on historical passenger volume and characteristic data of the period to be predicted, and corrects the prediction result to achieve the combination of the overall time characteristics of the period to be predicted and the characteristics of the passenger volume evolution trend, and considers the influence of accidental events on the tourist destination, thereby greatly improving the prediction accuracy and reducing the dependence of the original prediction technology on a large amount of historical data.

[0056] Figure 1 This is one of the flow charts of the route passenger volume prediction method provided by the present invention; Figure 1 As shown, the method includes:

[0057] Step 110, predicting a first passenger volume forecast value of the route to be predicted in the period to be predicted based on the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted;

[0058] Optionally, when it is necessary to predict the passenger volume of an airline route, the historical passenger volume of the airline route to be predicted and the characteristic data of the period to be predicted are first obtained.

[0059] The routes to be predicted here are the routes for which passenger volume prediction is required, which can be outbound international routes or other types of routes, such as routes to a certain country or a certain region. Outbound international routes refer to routes that take off from an airport in one country and go to other countries or regions.

[0060] The historical passenger volume of each route is the actual passenger volume generated in multiple periods before the period to be predicted, which can be obtained by collection in a data storage system or by inference, and this embodiment does not specifically limit this.

[0061] The so-called method of inferring the actual passenger volume can be based on the historical passenger volume and planned number of seats sold of the associated routes of each route on the date to be inferred, the historical passenger volume and planned number of seats sold of the associated routes of each route on the same historical date as the date to be inferred in historical years, the characteristic data of the date to be inferred and the planned number of seats sold of each route on the date to be inferred, to infer the historical passenger volume of each route on the date to be inferred.

[0062] The period to be predicted here may be the date for which passenger volume prediction is required, which may be a period time divided into days, weeks or months. The following takes days as the unit for period division as an example to describe the following method in detail.

[0063] The so-called multiple periods before the period to be predicted can be selected from multiple historical periods close to the period to be predicted, or can be selected from historical periods in the historical year with the same date as the period to be predicted, so as to make subsequent predictions more accurate. For example, the actual passenger volume generated in the three periods (such as three days) before the period to be predicted and the passenger volume in the historical period with the same date as the period to be predicted in the historical year can be used as the historical passenger volume.

[0064] The feature data of the period to be predicted here is obtained by extracting time features through a time feature extraction algorithm; the so-called time feature extraction algorithm includes but is not limited to an extraction algorithm for extracting date data for the period to be predicted, and a feature extraction algorithm for determining whether the period to be predicted is a weekend, a day of the week, a holiday, a compensatory day, a winter or summer vacation, and whether the three days before / after the period to be predicted are holidays.

[0065] Optionally, after obtaining the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted, the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted can be combined to predict the first passenger volume forecast value of the route to be predicted in the period to be predicted.

[0066] Here, the prediction method of the first passenger volume prediction value can be to obtain it by logically deducing the probability model of the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted; or it can be obtained by inputting the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted into a prediction model trained based on the historical passenger volume of the sample route in multiple periods before the preset historical period, the characteristic data of the preset historical period, and the actual passenger volume of the sample route in the preset historical period. This embodiment does not specifically limit this.

[0067] Step 120, acquiring at least one associated route associated with the route to be predicted based on the historical passenger volume and the departure airport of the route to be predicted;

[0068] Optionally, since there may be similar passenger flow patterns between different ODs, in order to further improve the accuracy of route passenger volume prediction, on the basis of obtaining the first passenger volume prediction value in step 110, it is necessary to obtain at least one associated route associated with the route to be predicted, so as to introduce relevant passenger information of the associated route to participate in the final prediction of route passenger volume, thereby improving the accuracy of route passenger volume prediction; the specific steps of obtaining the associated route are as follows:

[0069] The historical passenger volume and departure airport of the route to be predicted are obtained, and combined with the historical passenger volume and departure airport of the route to be predicted, at least one associated route associated with the route to be predicted is screened in the data storage system to obtain, so as to consider the impact of similar passenger flow patterns between the associated route and the route to be predicted on the passenger volume prediction of the route to be predicted, thereby improving the accuracy of the route passenger volume prediction.

[0070] Here, the method for determining associated routes includes combining the historical passenger volume and departure airport of the route to be predicted, as well as the historical passenger volume and departure airport of each route in the data storage system, to calculate the association matrix between the route to be predicted and each route in the data storage system; selecting at least one associated route with a higher degree of association based on the overall association matrix; or first screening out at least one target route with a higher degree of association with the departure airport of the route to be predicted from the data storage system, and then further screening out at least one associated route with a higher degree of association with the historical passenger volume of the route to be predicted from the at least one target route. This embodiment does not specifically limit this.

[0071] Step 130, predicting a second passenger volume forecast value of the to-be-forecasted route in the to-be-forecasted period based on the historical passenger volume of each of the associated routes, the passenger flow correlation between each of the associated routes and the to-be-forecasted route, the passenger preference degree corresponding to each of the associated routes, and the passenger preference degree corresponding to the to-be-forecasted route;

[0072] Optionally, after obtaining the associated routes, the actual passenger volume generated by each associated route in multiple periods before the period to be predicted can be obtained, thereby obtaining the historical passenger volume of each associated route; the similarity between the historical passenger volume of each associated route and the historical passenger volume of the route to be predicted is calculated to obtain the passenger flow correlation between each associated route and the route to be predicted; the social media data disclosed by the social media platform is analyzed and calculated to obtain the preference of each passenger for each associated route and the route to be predicted, thereby obtaining the passenger preference corresponding to each associated route and the route to be predicted.

[0073] Next, the passenger volume of the to-be-predicted route in the period to be predicted is predicted by combining the historical trip volume of each associated route, the passenger flow correlation between each associated route and the to-be-predicted route, the passenger preference corresponding to each associated route, and the passenger preference corresponding to the to-be-predicted route to obtain the second passenger volume prediction value of the to-be-predicted route in the period to be predicted.

[0074] Here, the historical volume of each associated route may be the volume of passengers actually transported to the respective destination cities by each associated route in at least one period before the period to be predicted, such as the period before or two periods before the period to be predicted.

[0075] Here, the prediction method of the second passenger volume prediction value can be to use a regression model to establish a relationship model between the passenger volume of the route to be predicted and the historical travel volume, passenger flow correlation and passenger preference of each related route, so as to obtain the second passenger volume prediction value of the route to be predicted in the period to be predicted based on the regression model fitting; or, the historical travel volume of each related route, the passenger flow correlation between each related route and the route to be predicted, the passenger preference corresponding to each related route, and the passenger preference corresponding to the route to be predicted are input into a pre-trained neural network model, and the pre-trained neural network model is used to predict the second passenger volume prediction value of the route to be predicted in the period to be predicted. This embodiment does not make any specific limitation on this.

[0076] Step 140: Obtain a final passenger volume forecast value for the to-be-forecasted route in the to-be-forecasted period based on the first passenger volume forecast value and the second passenger volume forecast value.

[0077] Optionally, after obtaining the first passenger volume prediction value and the second passenger volume prediction value, the first passenger volume prediction value and the second passenger volume prediction value may be combined to predict a final passenger volume prediction value of the route to be predicted in the period to be predicted.

[0078] Here, the method for obtaining the final passenger volume prediction value may be to average, weighted add, or fuse the prediction values ​​based on a fusion model, etc., the first passenger volume prediction value and the second passenger volume prediction value to obtain the final passenger volume prediction value of the route to be predicted in the prediction period.

[0079] The route passenger volume prediction method provided in the present embodiment makes an initial passenger volume prediction for the route to be predicted based on the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted, introduces a historical travel passenger volume similarity matrix of each route, obtains the historical trip volume of the associated routes associated with the historical passenger volume of the route to be predicted and the passenger flow correlation between each associated route and the route to be predicted, and obtains the passenger preference corresponding to each associated route and the passenger preference corresponding to the route to be predicted in combination with social media data, so as to correct the initial passenger volume prediction value based on the historical trip volume of the associated routes and the passenger flow correlation between each associated route and the route to be predicted, as well as the passenger preference corresponding to each associated route and the passenger preference corresponding to the route to be predicted, thereby improving the accuracy of the final passenger volume prediction, so as to solve the problem that the prior art only relies on historical trip data for prediction and ignores other influencing factors, resulting in poor prediction accuracy in the final prediction effect.

[0080] In some implementations, in step 110, predicting a first passenger volume forecast value of the route to be predicted in the period to be predicted based on the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted includes:

[0081] In the data storage system, the passenger volume of the route to be predicted in multiple periods before the period to be predicted is collected to obtain the historical passenger volume of the route to be predicted;

[0082] Inputting the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted into a first prediction model to obtain the first passenger volume prediction value;

[0083] Among them, the first prediction model is obtained by iteratively optimizing the seasonally adjusted autoregressive moving average model based on the historical passenger volume of the sample route in multiple periods before the preset historical period, the characteristic data of the preset historical period, and the actual passenger volume of the sample route in the preset historical period.

[0084] Optionally, the step of obtaining the first passenger volume prediction value in step 110 specifically includes:

[0085] The historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted are input into the trained first prediction model to obtain the first passenger volume prediction value (hereinafter also referred to as the initial passenger volume prediction value) y0 output by the first prediction model.

[0086] Here, the first forecasting model is SARIMA (Seasonal Autoregressive Integrated Moving Average); the core idea of ​​SARIMA is to use the autoregressive algorithm to link the historical sequence values ​​and the current sequence values ​​through a certain model, use historical samples for training, and then derive appropriate model parameters, and comprehensively consider endogenous and exogenous factors to improve the forecasting accuracy. The calculation formula of the first forecasting model is as follows:

[0087] Y(t)=c+Φ1Y(t-1)+…+ΦpY(tp)+θ1ε(t-1)+…+θ q ε(tq)+ΦsY(ts)+…+Φ Ps Y(t-Ps)+ε(t);

[0088] Among them, Y(t) represents the observed value of period t, c represents a constant, {Φ1,…,Φp} represents the autoregressive coefficient, {θ1,…,θ q} represents the moving average coefficient, {Φs,…,Φ Ps} represents the seasonal autoregressive coefficient, and ε(t) represents the white noise error term. P, q, Ps and each coefficient are model parameters obtained through iterative optimization training based on the historical passenger volume of the sample route in multiple periods before the preset historical period, the characteristic data of the preset historical period, and the actual passenger volume of the sample route in the preset historical period.

[0089] The sample routes here include but are not limited to routes with the same departure airport or destination airport as the route to be predicted, the route to be predicted, and the same route as the route to be predicted; the historical passenger volume of the sample routes can be extracted based on the actual flight record data of the sample routes; the so-called actual flight record data includes but is not limited to a preset historical period, multiple periods adjacent to the preset historical period, and the actual number of passengers carried, planned number of seats sold, departure and destination, and longitude and latitude of each flight on the route on the same date in the historical year.

[0090] The method provided in this embodiment accurately captures the characteristics of trends, seasonality, random fluctuations, etc. in the passenger volume time series based on the seasonally adjusted autoregressive moving average model, and combines historical passenger volume and characteristic data to predict the passenger volume of the route to be predicted, so as to more accurately predict the passenger volume of the route.

[0091] In some embodiments, in step 120, obtaining at least one associated route associated with the route to be predicted based on the historical passenger volume and the departure airport of the route to be predicted includes:

[0092] In the data storage system, all target routes having the same departure airport as the route to be predicted are obtained;

[0093] In the data storage system, the passenger volume of each target route in multiple periods before the period to be predicted is collected to obtain the historical passenger volume of each target route;

[0094] Calculating the similarity between the historical passenger volume of each target route and the historical passenger volume of the route to be predicted respectively, and obtaining the passenger flow correlation between each target route and the route to be predicted;

[0095] At least one associated route is determined from at least one target route according to the passenger flow correlation between each of the target routes and the route to be predicted.

[0096] Optionally, the step of obtaining the associated route in step 120 specifically includes:

[0097] The departure airport of each route in the data storage system is compared with the departure airport of the route to be predicted, so as to obtain all target routes having the same departure airport as the route to be predicted.

[0098] Next, in the data storage system, the passenger volume of each target route in multiple periods before the period to be predicted is collected to obtain the historical passenger volume of each target route. The similarity between the historical passenger volume of each target route and the historical passenger volume of the route to be predicted is calculated to obtain the passenger flow correlation between each target route and the route to be predicted.

[0099] The so-called similarity calculation can be implemented by using similarity calculations such as the Pearson coefficient or the Euclidean distance. For example, the calculation formula for calculating the passenger flow correlation between each target route and the route to be predicted using the Pearson coefficient is:

[0100]

[0101] Where x and y represent the historical passenger volume of the target route and the historical passenger volume of the route to be predicted, respectively; σ x and σ y They represent the standard deviation of x and y respectively; cov represents covariance; and E represents variance.

[0102] Next, for the route to be predicted, at least one associated route with a high passenger flow correlation with the route to be predicted is selected.

[0103] Here, the number of associated routes can be configured according to actual needs, such as selecting 2 or 3 routes with the highest passenger flow correlation from high to low as associated routes, that is, selecting 2 or 3 routes with higher passenger flow correlation with the route to be predicted as associated routes.

[0104] The method provided in this embodiment can accurately and efficiently obtain at least one associated route associated with the route to be predicted by performing multi-level screening based on departure airports and historical passenger volumes, so as to subsequently improve the efficiency and accuracy of route passenger volume prediction.

[0105] In some embodiments, in step 130, based on the historical volume of each of the associated routes, the passenger flow correlation between each of the associated routes and the route to be predicted, the passenger preference corresponding to each of the associated routes, and the passenger preference corresponding to the route to be predicted, a second passenger volume forecast value of the route to be predicted in the period to be predicted is predicted, including:

[0106] Calculating the similarity between the historical passenger volume of each of the associated routes and the historical passenger volume of the route to be predicted, respectively, to obtain the passenger flow correlation between each of the associated routes and the route to be predicted;

[0107] Dividing the passenger preference degree corresponding to each of the associated routes by the passenger preference degree corresponding to the route to be predicted, to obtain a relative preference degree corresponding to each of the associated routes;

[0108] Inputting the historical volume of each of the associated routes, the passenger flow correlation between each of the associated routes and the route to be predicted, and the relative preference into a second prediction model to obtain a second passenger volume prediction value of the route to be predicted in the period to be predicted;

[0109] The second prediction model is obtained by training the initial neural network model based on the historical travel volume of each associated route of the sample route in the previous period of the preset historical period, the passenger flow correlation between each associated route of the sample route and the sample route, the relative preference corresponding to each associated route of the sample route, and the actual passenger volume of the sample route in the preset historical period.

[0110] Optionally, the step of obtaining the second passenger volume prediction value in step 130 specifically includes:

[0111] In the data storage system, the number of passengers actually transported to the respective destination cities by each associated route in at least one period before the period to be predicted, such as the period before or two periods before the period to be predicted, is collected to obtain the historical number of passengers for each associated route; and in the data storage system, the number of passengers for each associated route in multiple periods before the period to be predicted is collected to obtain the historical number of passengers for each target route.

[0112] The similarity calculation is performed on the historical passenger volume of each associated route and the historical passenger volume of the route to be predicted to obtain the passenger flow correlation between each associated route and the route to be predicted. The so-called similarity calculation can refer to the calculation steps of the passenger flow correlation between the target route and the route to be predicted, which will not be repeated here.

[0113] The passenger preference for each associated route and the route to be predicted is obtained through the social media platform. The passenger preference for each associated route is divided by the passenger preference for the route to be predicted to obtain the relative preference of passengers for each associated route compared to the route to be predicted in a specific historical period.

[0114] Next, the historical travel volume of at least one associated route with a high passenger flow correlation with the route to be predicted, the passenger flow correlation between each associated route and the route to be predicted, and the relative preference are input into the pre-trained second prediction model to predict the second passenger volume forecast value y1 of the route to be predicted in the prediction period.

[0115] The second prediction model here can be trained based on the following steps:

[0116] The historical trip volume of each associated route of the sample route in the previous period of the preset historical period, the passenger flow correlation between each associated route of the sample route and the sample route, and the relative preference corresponding to each associated route of the sample route are taken as samples, and the actual passenger volume of the sample route in the preset historical period is taken as a label. The initial neural network model is iteratively trained to obtain a second prediction model which can accurately predict the passenger volume forecast value according to the historical trip volume, passenger flow correlation and relative preference.

[0117] The method provided in this embodiment makes a secondary prediction of the passenger volume by introducing the historical travel volume of each associated route, the passenger flow correlation between each associated route and the route to be predicted, and the passenger preference of each associated route relative to the route to be predicted. The passenger volume is predicted by comprehensively considering the correlation between the passenger flow patterns of the routes, which can effectively improve the accuracy of the final prediction.

[0118] In some embodiments, the passenger preference corresponding to each of the associated routes and the routes to be predicted is obtained based on the following steps:

[0119] Using a preset time period, the city area to which each of the routes belongs, and travel-related semantic words as keywords, crawling at least one related social media data corresponding to each of the routes from a social media platform;

[0120] Extracting the passenger preference parameters corresponding to each of the routes from the associated social media data;

[0121] Calculating the passenger preference degree corresponding to each of the routes according to the passenger preference parameters;

[0122] The preset time period is determined according to the period to be predicted.

[0123] Optionally, the passenger preference corresponding to each route is obtained based on the following steps:

[0124] A large amount of social media data is crawled from the public API (Application Programming Interface) of the social media platform; these data at least include specific time periods and target cities for each route, as well as travel-related semantic words such as tourism and relaxation. A large amount of social media data is filtered and screened to extract unrecognizable and invalid data, and retain data including retention time period (T), follow cities (C), published content, published time, number of interests (L), number of followers (F), etc.

[0125] Next, the preset time period, the urban area to which each of the routes belongs, and travel-related semantic words are used as keywords to crawl social media data matching the keywords from the social media platform as the associated social media data corresponding to each route.

[0126] Next, passenger preference parameters are extracted from each related social media data to obtain passenger preference parameters corresponding to each route.

[0127] In some embodiments, the step of obtaining the passenger preference parameters corresponding to each route specifically includes:

[0128] In each associated social media data, the number of interests and the number of attentions associated with the target city of each route are extracted, so that the number of interests and the number of attentions in the social media data containing the target city of each route are used as the passenger preference parameters corresponding to each route.

[0129] Next, the passenger preference parameters are used as calculation parameters to calculate the passenger preference degree corresponding to each route.

[0130] Here, the passenger preference corresponding to each route can be calculated according to the following method: first, based on the number of interests and attentions corresponding to each city in each route extracted from the single social media data, the single passenger preference corresponding to each route under the single social media data is calculated; then, the single passenger preference corresponding to each route under all the single social media data is added together to obtain the passenger preference corresponding to each route.

[0131] For example, the passenger preference corresponding to a single route can be calculated based on the following formula:

[0132] P i (C=c j , t in T) = (L i *3+F i ) / 4;

[0133] P TOTAL (C=c j , t in T) = ∑P i (C=c j , t in T);

[0134] Among them, c j is the destination city in the jth specific route; P i For a single containing a specific city c j Social media data showing travelers' preferences for specific cities j The passenger preference of the i-th social media data; T is the retention period; C is the concerned city recorded in the i-th social media data; t is the release time recorded in the i-th social media data; L i is the number of interests corresponding to the city in the i-th social media data, that is, the number of interests corresponding to each city in each route; F i is the number of followers of the city in the ith social media data, that is, the number of followers of each city in each route; P TOTAL For all the columns containing a specific city c j The social media data of travelers at a specific time point in a specific city c j The overall preference of each route, that is, the passenger preference of each route.

[0135] The method provided in this embodiment can accurately obtain the passenger preferences corresponding to each route by introducing social media data, so as to reflect the passengers' travel intentions and needs to a certain extent. Therefore, by comprehensively considering the passenger preferences to predict the passenger volume, the accuracy of the final prediction can be effectively improved.

[0136] In some embodiments, in step 140, obtaining the final passenger volume forecast value of the to-be-forecasted route in the to-be-forecasted period according to the first passenger volume forecast value and the second passenger volume forecast value includes:

[0137] Inputting the first passenger volume prediction value and the second passenger volume prediction value into a weighted fusion model to obtain the final passenger volume prediction value;

[0138] The weighted fusion model is obtained by optimizing the initial fusion model with the optimization goal of minimizing the error between the final passenger volume prediction value of the sample route in the preset historical period and the actual passenger volume of the sample route in the preset historical period;

[0139] The final passenger volume prediction value of the sample route in the preset historical period is obtained by performing a fusion prediction on the first passenger volume prediction value of the sample route in the preset historical period and the second passenger volume prediction value of the sample route in the preset historical period based on the initial fusion model.

[0140] Optionally, the step of obtaining the final passenger volume prediction value in step 140 includes:

[0141] The first passenger volume prediction value and the second passenger volume prediction value are input into the weighted fusion model, and the first passenger volume prediction value and the second passenger volume prediction value are weightedly added to obtain the final passenger volume prediction value of the route to be predicted in the prediction period; the specific calculation formula of the weighted fusion model is:

[0142] y final =w0y0+w1y1;

[0143] Among them, y final is the final passenger volume prediction value; y0 and y1 are the first passenger volume prediction value and the second passenger volume prediction value respectively; w0 and w1 are the model parameters of the weighted fusion model.

[0144] Here, the weighted fusion model is trained based on the following steps:

[0145] The first passenger volume prediction values ​​of different sample routes in the preset historical period and the second passenger volume prediction values ​​of different sample routes in the preset historical period are input into the initial fusion model to obtain the final passenger volume prediction values ​​of different sample routes in the preset historical period output by the fusion prediction of the initial fusion model; the average absolute error between the final passenger volume prediction values ​​of different sample routes in the preset historical period and the actual passenger volume of different sample routes in the preset historical period is calculated, and the model parameters of the initial fusion model are iteratively optimized with minimizing the average absolute error as the optimization goal, so as to obtain a weighted fusion model that can accurately output the final passenger volume prediction value based on the first passenger volume prediction value and the second passenger volume prediction value.

[0146] The method provided in this embodiment can accurately obtain the final passenger volume prediction value by combining a first passenger volume prediction value obtained by combining the historical passenger volume and characteristic data of the route to be predicted, and a second passenger volume prediction value predicted based on the historical travel volume of each related route, the passenger flow correlation between each related route and the route to be predicted, the passenger preference corresponding to each related route, and the passenger preference corresponding to the route to be predicted, and performing a fusion operation.

[0147] Figure 2 The second flow chart of the route passenger volume prediction method provided by the present invention; Figure 2 As shown in the figure, the complete process of the route passenger volume prediction method includes:

[0148] Step 210, obtaining actual flight record data of each flight in an outbound international route of a certain airport;

[0149] Step 220, preliminary route passenger volume prediction: obtain the historical passenger volume of the route to be predicted from the actual flight record data, input the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted into the trained first prediction model, and obtain the first passenger volume prediction value y0;

[0150] After obtaining the first passenger volume forecast value y0 of the route to be forecasted through the above steps, new information and parameters are introduced on the basis of the first passenger volume forecast value y0 to perform a secondary forecast, and the first passenger volume forecast value y0 and the second passenger volume forecast value y1 are combined to correct and obtain a final more accurate forecast value; the specific steps include the following:

[0151] Step 230, using the Pearson coefficient to calculate the passenger flow correlation between different routes;

[0152] Step 240, obtaining social media data from the social media platform, filtering, screening and analyzing the social media data to obtain the passenger preference parameters corresponding to each route, and then calculating the passenger preference degree corresponding to each route;

[0153] Step 250, selecting two associated routes that are consistent with the departure airport of the route to be predicted and have a high correlation with the passenger flow of the route to be predicted;

[0154] Step 260, calculating the passenger relative preference of each associated route relative to the route to be predicted;

[0155] Step 270, secondary prediction: input the historical passenger volume of each associated route with a high correlation with the passenger flow of the route to be predicted on the day before the predicted date, the passenger flow correlation between each associated route and the route to be predicted, and the relative preference of passengers of each associated route over the route to be predicted into the trained second prediction model to predict the second passenger volume prediction value y1;

[0156] Step 280: weighted fusion of the first passenger volume prediction value y0 and the second passenger volume prediction value y1 to obtain a final passenger volume prediction value.

[0157] The method provided in this embodiment proposes a method of introducing social media data and considering the influence of the correlation of destination passenger volumes of different routes on the basis of obtaining a first passenger volume prediction value based on historical passenger volume prediction, and correcting the prediction result, combining the overall time characteristics of the day to be predicted and the characteristics of the passenger volume evolution trend, and considering the impact of accidental events on the tourist destination, thereby greatly improving the prediction accuracy and reducing the dependence of the original prediction technology on a large amount of historical passenger volume, and solving the problem that the original prediction technology ignores the linkage effect between different routes and accidental events, resulting in poor prediction accuracy and poor universality.

[0158] The route passenger volume prediction device provided by the present invention is described below. The route passenger volume prediction device described below and the route passenger volume prediction method described above can be referenced to each other.

[0159] Figure 3 A schematic diagram of the structure of the route passenger volume prediction device provided by the present invention; Figure 3 As shown, the device comprises:

[0160] The first prediction module 310 is used to predict a first passenger volume prediction value of the route to be predicted in the period to be predicted based on the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted;

[0161] The first acquisition module 320 is used to acquire at least one associated route associated with the route to be predicted according to the historical passenger volume and the departure airport of the route to be predicted;

[0162] The second prediction module 330 is used to predict the second passenger volume forecast value of the to-be-predicted route in the to-be-predicted period according to the historical passenger volume of each of the associated routes, the passenger flow correlation between each of the associated routes and the to-be-predicted route, the passenger preference corresponding to each of the associated routes, and the passenger preference corresponding to the to-be-predicted route;

[0163] The second acquisition module 340 is used to acquire the final passenger volume prediction value of the to-be-predicted route in the to-be-predicted period according to the first passenger volume prediction value and the second passenger volume prediction value.

[0164] The route passenger volume prediction device provided in the present embodiment makes an initial passenger volume prediction for the route to be predicted based on the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted, introduces the historical travel passenger volume similarity matrix of each route, obtains the historical trip volume of the associated routes associated with the historical passenger volume of the route to be predicted and the passenger flow correlation between each associated route and the route to be predicted, and obtains the passenger preference corresponding to each associated route and the passenger preference corresponding to the route to be predicted in combination with the social media data, so as to correct the initial passenger volume prediction value based on the historical trip volume of the associated routes and the passenger flow correlation between each associated route and the route to be predicted, as well as the passenger preference corresponding to each associated route and the passenger preference corresponding to the route to be predicted, thereby improving the accuracy of the final passenger volume prediction, so as to solve the problem that the prior art only relies on historical trip data for prediction and ignores other influencing factors, resulting in poor prediction accuracy in the final prediction effect.

[0165] The device provided by the present invention is used to execute the above-mentioned method embodiments. Please refer to the above-mentioned embodiments for the specific processes and detailed contents, which will not be repeated here.

[0166] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 communicate with each other through the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the route passenger volume prediction method, which includes: predicting the first passenger volume prediction value of the route to be predicted in the period to be predicted according to the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted; obtaining at least one associated route associated with the route to be predicted according to the historical passenger volume and the departure airport of the route to be predicted; predicting the second passenger volume prediction value of the route to be predicted in the period to be predicted according to the historical passenger volume of each associated route, the passenger flow correlation between each associated route and the route to be predicted, the passenger preference corresponding to each associated route, and the passenger preference corresponding to the route to be predicted; obtaining the final passenger volume prediction value of the route to be predicted in the period to be predicted according to the first passenger volume prediction value and the second passenger volume prediction value.

[0167] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0168] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the route passenger volume prediction method provided by the above-mentioned methods, which includes: predicting a first passenger volume prediction value of the route to be predicted in the period to be predicted based on the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted; obtaining at least one associated route associated with the route to be predicted based on the historical passenger volume and the departure airport of the route to be predicted; predicting a second passenger volume prediction value of the route to be predicted in the period to be predicted based on the historical travel volume of each associated route, the passenger flow correlation between each associated route and the route to be predicted, the passenger preference corresponding to each associated route, and the passenger preference corresponding to the route to be predicted; and obtaining a final passenger volume prediction value of the route to be predicted in the period to be predicted based on the first passenger volume prediction value and the second passenger volume prediction value.

[0169] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the route passenger volume prediction method provided by the above-mentioned methods, the method comprising: predicting a first passenger volume prediction value of the route to be predicted in the period to be predicted based on the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted; obtaining at least one associated route associated with the route to be predicted based on the historical passenger volume and the departure airport of the route to be predicted; predicting a second passenger volume prediction value of the route to be predicted in the period to be predicted based on the historical passenger volume of each associated route, the passenger flow correlation between each associated route and the route to be predicted, the passenger preference corresponding to each associated route, and the passenger preference corresponding to the route to be predicted; obtaining a final passenger volume prediction value of the route to be predicted in the period to be predicted based on the first passenger volume prediction value and the second passenger volume prediction value.

[0170] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0171] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0172] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting passenger volume on an airline route, characterized in that: include: According to the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted, a first passenger volume prediction value of the route to be predicted in the period to be predicted is predicted; the historical passenger volume is the actual passenger volume generated in multiple periods before the period to be predicted; Acquire at least one associated route associated with the route to be predicted according to the historical passenger volume and the departure airport of the route to be predicted; According to the historical volume of each of the associated routes, the passenger flow correlation between each of the associated routes and the route to be predicted, the passenger preference corresponding to each of the associated routes, and the passenger preference corresponding to the route to be predicted, a second passenger volume forecast value of the route to be predicted in the period to be predicted is predicted; the passenger flow correlation is obtained by calculating the similarity between the historical passenger volume of each of the associated routes and the historical passenger volume of the route to be predicted; The passenger preference corresponding to each of the associated routes and the passenger preference corresponding to the to-be-predicted route are both obtained by analyzing and calculating the social media data disclosed by the social media platform; Obtaining a final passenger volume forecast value of the to-be-forecasted route in the to-be-forecasted period according to the first passenger volume forecast value and the second passenger volume forecast value; The step of predicting a second passenger volume forecast value of the route to be predicted in the period to be predicted based on the historical passenger volume of each of the associated routes, the passenger flow correlation between each of the associated routes and the route to be predicted, the passenger preference degree corresponding to each of the associated routes, and the passenger preference degree corresponding to the route to be predicted, comprises: Calculating the similarity between the historical passenger volume of each of the associated routes and the historical passenger volume of the route to be predicted, respectively, to obtain the passenger flow correlation between each of the associated routes and the route to be predicted; Dividing the passenger preference degree corresponding to each of the associated routes by the passenger preference degree corresponding to the route to be predicted, to obtain a relative preference degree corresponding to each of the associated routes; Inputting the historical volume of each of the associated routes, the passenger flow correlation between each of the associated routes and the route to be predicted, and the relative preference into a second prediction model to obtain a second passenger volume prediction value of the route to be predicted in the period to be predicted; The second prediction model is obtained by training the initial neural network model based on the historical travel volume of each associated route of the sample route in the previous period of the preset historical period, the passenger flow correlation between each associated route of the sample route and the sample route, the relative preference corresponding to each associated route of the sample route, and the actual passenger volume of the sample route in the preset historical period.

2. The route passenger volume prediction method according to claim 1, characterized in that: The step of predicting a first passenger volume forecast value of the route to be predicted in the period to be predicted based on the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted includes: In the data storage system, the passenger volume of the route to be predicted in multiple periods before the period to be predicted is collected to obtain the historical passenger volume of the route to be predicted; Inputting the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted into a first prediction model to obtain the first passenger volume prediction value; Among them, the first prediction model is obtained by iteratively optimizing the seasonally adjusted autoregressive moving average model based on the historical passenger volume of the sample route in multiple periods before the preset historical period, the characteristic data of the preset historical period, and the actual passenger volume of the sample route in the preset historical period.

3. The method for predicting airline passenger volume according to claim 1, characterized in that: The acquiring, according to the historical passenger volume and the departure airport of the route to be predicted, at least one associated route associated with the route to be predicted comprises: In the data storage system, all target routes having the same departure airport as the route to be predicted are obtained; In the data storage system, the passenger volume of each target route in multiple periods before the period to be predicted is collected to obtain the historical passenger volume of each target route; Calculating the similarity between the historical passenger volume of each target route and the historical passenger volume of the route to be predicted respectively, and obtaining the passenger flow correlation between each target route and the route to be predicted; At least one associated route is determined from at least one target route according to the passenger flow correlation between each of the target routes and the route to be predicted.

4. The method for predicting airline passenger volume according to any one of claims 1 to 3, characterized in that: The passenger preference corresponding to each of the associated routes and the routes to be predicted is obtained based on the following steps: Using a preset time period, the city area to which each of the routes belongs, and travel-related semantic words as keywords, crawling at least one related social media data corresponding to each of the routes from a social media platform; Extracting the passenger preference parameters corresponding to each of the routes from the associated social media data; Calculating the passenger preference degree corresponding to each of the routes according to the passenger preference parameters; The preset time period is determined according to the period to be predicted.

5. The method for predicting airline passenger volume according to claim 4, characterized in that: Extracting the passenger preference parameters corresponding to each route from each of the associated social media data includes: Extracting the number of interests and the number of followers corresponding to each city in each of the routes from the associated social media data; According to the number of interests and the number of concerns corresponding to each city in each of the routes, the passenger preference parameters corresponding to each of the routes are obtained.

6. The method for predicting passenger volume of an airline according to any one of claims 1 to 3, characterized in that: The obtaining, according to the first passenger volume prediction value and the second passenger volume prediction value, a final passenger volume prediction value of the to-be-predicted route in the to-be-predicted period includes: Inputting the first passenger volume prediction value and the second passenger volume prediction value into a weighted fusion model to obtain the final passenger volume prediction value; The weighted fusion model is obtained by optimizing the initial fusion model with the optimization goal of minimizing the error between the final passenger volume prediction value of the sample route in the preset historical period and the actual passenger volume of the sample route in the preset historical period; The final passenger volume prediction value of the sample route in the preset historical period is obtained by performing a fusion prediction on the first passenger volume prediction value of the sample route in the preset historical period and the second passenger volume prediction value of the sample route in the preset historical period based on the initial fusion model.

7. A route passenger volume prediction device, characterized in that: include: A first prediction module is used to predict a first passenger volume prediction value of the route to be predicted in the period to be predicted based on the historical passenger volume of the route to be predicted and the characteristic data of the period to be predicted; the historical passenger volume is the actual passenger volume generated in multiple periods before the period to be predicted; A first acquisition module, configured to acquire at least one associated route associated with the route to be predicted according to the historical passenger volume and the departure airport of the route to be predicted; A second prediction module is used to predict a second passenger volume forecast value of the route to be predicted in the prediction period according to the historical passenger volume of each of the associated routes, the passenger flow correlation between each of the associated routes and the route to be predicted, the passenger preference corresponding to each of the associated routes, and the passenger preference corresponding to the route to be predicted; the passenger flow correlation is obtained by calculating the similarity between the historical passenger volume of each of the associated routes and the historical passenger volume of the route to be predicted; The passenger preference corresponding to each of the associated routes and the passenger preference corresponding to the to-be-predicted route are both obtained by analyzing and calculating the social media data disclosed by the social media platform; A second acquisition module, configured to acquire a final passenger volume forecast value of the to-be-forecasted route in the to-be-forecasted period according to the first passenger volume forecast value and the second passenger volume forecast value; The second prediction module is specifically used for: Calculating the similarity between the historical passenger volume of each of the associated routes and the historical passenger volume of the route to be predicted, respectively, to obtain the passenger flow correlation between each of the associated routes and the route to be predicted; Dividing the passenger preference degree corresponding to each of the associated routes by the passenger preference degree corresponding to the route to be predicted, to obtain a relative preference degree corresponding to each of the associated routes; Inputting the historical volume of each of the associated routes, the passenger flow correlation between each of the associated routes and the route to be predicted, and the relative preference into a second prediction model to obtain a second passenger volume prediction value of the route to be predicted in the period to be predicted; The second prediction model is obtained by training the initial neural network model based on the historical travel volume of each associated route of the sample route in the previous period of the preset historical period, the passenger flow correlation between each associated route of the sample route and the sample route, the relative preference corresponding to each associated route of the sample route, and the actual passenger volume of the sample route in the preset historical period.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the route passenger volume prediction method according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the route passenger volume prediction method according to any one of claims 1 to 6 is implemented.

Images