Flight resource recommendation method and device

By constructing a pre-defined passenger cluster and considering the similarity of fellow passengers, a comprehensive calculation method for flight resource recommendation is used to solve the problem of low acceptance of recommended flight resources after flight cancellation, thereby improving server efficiency and passenger evacuation efficiency.

CN112200625BActive Publication Date: 2024-11-08TRAVELSKY TECHNOLOGY LIMITED
View PDF 8 Cites 0 Cited by

Patent Information

Application Number
CN202011056782.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-09-30
Publication Date
2024-11-08
Estimated Expiration
2040-09-30

AI Technical Summary

Technical Problem

When airlines recommend alternative itineraries to transit passengers after flight cancellations, the current methods are too simplistic, resulting in low acceptance of recommended flight resources, wasted server computing resources, and impact on flight resource utilization and airport passenger evacuation efficiency.

Method used

By constructing pre-defined passenger clusters, and based on passenger contextual characteristics and the similarity of fellow passengers, the acceptance of alternative itineraries is comprehensively calculated, and the flight resources that best meet the needs of passengers are recommended.

Benefits of technology

This improved the acceptance rate of flight resource recommendations, reduced server wastage, and enhanced flight resource utilization and airport passenger evacuation efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112200625B_ABST
    Figure CN112200625B_ABST
Patent Text Reader

Abstract

The flight resource recommendation method and apparatus provided by this invention are applied in the field of computer technology. The method sets up multiple preset passenger clusters and, based on the PNR (Passenger Name Notification) of the passenger to be recommended, determines the accompanying passengers. After obtaining the original itinerary of the passenger to be recommended, including canceled flight segments, multiple alternative itineraries are constructed. For each of the obtained alternative itineraries, a first acceptance rate, a second acceptance rate, and a similarity are calculated. Finally, based on the obtained first acceptance rate, second acceptance rate, and similarity, the alternative itineraries are pushed to the passenger to be recommended. This method comprehensively considers multiple factors, and the flight resources recommended to passengers are derived from a combination of factors, thus having higher acceptance rates. This helps reduce the ineffective calculations performed by airline servers, thereby improving the effective utilization rate of airline servers.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of computer technology, and in particular relates to a flight resource recommendation method and device. Background Art

[0002] In flight operations, flight cancellations often occur due to bad weather, air traffic control, flight equipment failure, ground service support, etc., which seriously affects passengers' travel arrangements and reduces their travel experience. Therefore, in order to improve the service quality of airlines and the travel experience of passengers, how to provide passengers with alternative itineraries after flight cancellations, that is, to recommend other flight resources that can complete the established travel arrangements, has always been a matter of great concern to airlines.

[0003] At present, after a flight cancellation occurs, when airlines provide alternative itineraries to transit passengers, the rules for recommending flight resources are relatively simple, such as only recommending flight resources to passengers based on passenger value, which leads to passengers being dissatisfied with the recommended flight resources and refusing to accept the alternative itineraries recommended by the airlines.

[0004] In practical applications, each protection plan requires the server to follow the existing flight resource recommendation method, call available flights for corresponding calculations, and then conclude that the alternative itineraries recommended by airlines are not well accepted, which means that the server has performed a lot of ineffective work and wasted a lot of computing resources. The effective utilization rate of server computing resources is low, and at the same time, it will also affect the utilization rate of flight resources and the evacuation efficiency of airport passengers. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide a flight resource recommendation method and device, which comprehensively evaluates alternative itineraries based on multiple factors, improves the acceptability of recommended flight resources, and thus improves the effective utilization rate of airline servers, the utilization rate of flight resources and the evacuation efficiency of airport passengers. The specific scheme is as follows:

[0006] In a first aspect, the present invention provides a flight resource recommendation method, comprising:

[0007] Obtaining the original itinerary of the passenger to be recommended, wherein the original itinerary includes the flight cancellation segment;

[0008] Constructing available flight combinations from the departure airport of the cancelled flight segment to the passenger destination to be recommended, and obtaining multiple alternative itineraries;

[0009] Determine a target preset passenger cluster to which the passenger to be recommended belongs among a plurality of preset passenger clusters, and calculate the first acceptance degree of each of the alternative itineraries according to the itinerary similarity between the passenger to be recommended and each sample passenger in the target preset passenger cluster, wherein the sample passengers in any of the preset passenger clusters correspond to similar passenger context characteristics;

[0010] Determining the companion passengers of the passenger to be recommended according to the passenger reservation record PNR of the passenger to be recommended, and calculating the second acceptance degree of each of the alternative itineraries according to the similarity between the itineraries of the passenger to be recommended and the companion passengers;

[0011] Calculating the similarity between each of the alternative itineraries and the original itinerary of the passenger to be recommended, and obtaining the alternative itinerary similarity corresponding to each of the alternative itineraries;

[0012] Based on the first acceptance degree, the second acceptance degree, and the similarity of the alternative itineraries, each of the alternative itineraries is pushed to the passenger to be recommended.

[0013] In a second aspect, the present invention provides a flight resource recommendation device, comprising:

[0014] An acquisition unit, configured to acquire an original itinerary of the passenger to be recommended, wherein the original itinerary includes a cancelled flight segment;

[0015] A construction unit, configured to construct a combination of available flights from the departure airport of the cancelled flight segment of the flight to the passenger destination to be recommended, so as to obtain a plurality of alternative itineraries;

[0016] A first calculation unit is used to determine, from among a plurality of preset passenger clusters, a target preset passenger cluster to which the passenger to be recommended belongs, and to calculate a first acceptance degree of each of the alternative itineraries according to a similarity between the passenger to be recommended and each sample passenger in the target preset passenger cluster, wherein the sample passengers in any of the preset passenger clusters correspond to similar passenger context features;

[0017] A second calculation unit is used to determine the companions of the passenger to be recommended according to the passenger reservation record PNR of the passenger to be recommended, and calculate the second acceptance degree of each of the alternative itineraries according to the similarity between the itineraries of the passenger to be recommended and the companions;

[0018] A third calculation unit is used to calculate the similarity between each of the alternative itineraries and the original itinerary of the passenger to be recommended, and obtain the similarity of the alternative itineraries corresponding to each of the alternative itineraries;

[0019] A recommendation unit is used to push each of the alternative itineraries to the passenger to be recommended based on the first acceptance degree, the second acceptance degree, and the similarity of the alternative itineraries.

[0020] Based on the above technical solution, the flight resource recommendation method provided by the present invention sets multiple preset passenger clusters, and the sample passengers in each preset passenger cluster correspond to similar passenger situational characteristics. At the same time, the companion passengers of the passenger to be recommended are determined based on the PNR of the passenger to be recommended. After obtaining the original itinerary of the passenger to be recommended, including the flight cancellation segment, a combination of available flights from the departure airport of the flight cancellation segment to the destination of the passenger to be recommended is constructed to obtain multiple alternative itineraries. Further, for the obtained alternative itineraries, based on the itinerary similarity between the passenger to be recommended and the sample passengers in the target preset passenger cluster, the first acceptance of the passenger to be recommended to each alternative itinerary is calculated; based on the itinerary similarity between the passenger to be recommended and the companion passengers, the second acceptance of the passenger to be recommended to each alternative itinerary, as well as the alternative itinerary similarity corresponding to each alternative itinerary are calculated, and finally, based on the obtained first acceptance, second acceptance, and alternative itinerary similarity, each alternative itinerary is pushed to the passenger to be recommended.

[0021] The flight resource recommendation method provided by the present invention comprehensively considers three factors: group characteristics represented by the target preset passenger clustering cluster, mutual influence between fellow passengers, and similarity between each alternative itinerary and the original itinerary of the passenger to be recommended. Compared with the prior art, the flight resources recommended to passengers by this method are derived from a comprehensive analysis of multiple factors, and thus have a higher degree of acceptance, which helps to reduce the situation where airline servers do useless work, thereby improving the effective utilization rate of airline servers, improving the utilization rate of flight resources and the evacuation efficiency of airport passengers. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale.

[0023] Figure 1 is a flow chart of a flight resource recommendation method provided by an embodiment of the present invention;

[0024] Figure 2 It is a structural block diagram of a flight resource recommendation device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0025] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0026] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0027] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0028] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0029] Optional, see Figure 1 , Figure 1 : is a flow chart of a flight resource recommendation method provided by an embodiment of the present invention. The process of the method may include:

[0030] S100: Obtain the original itinerary of the passenger to be recommended, where the original itinerary includes the flight cancellation segments.

[0031] According to different flight cancellation situations, the specific composition of the original itinerary of the recommended passenger will vary. In the flight resource recommendation method provided in the embodiment of the present invention, the recommended passenger refers to a transfer passenger. Therefore, the original itinerary of the recommended passenger includes not only the flight cancellation segment, but also the subsequent segment after the flight cancellation segment and the previous segment before the flight cancellation segment.

[0032] Optionally, the original itinerary of the passenger to be recommended can be obtained from the PNR information of the passenger to be recommended, and of course, can also be obtained from the passenger service system of the airline. The embodiment of the present invention does not limit the specific source of the original itinerary of the passenger to be recommended.

[0033] S110: Construct available flight combinations from the departure airport of the cancelled flight segment to the passenger destination to be recommended, and obtain multiple alternative itineraries.

[0034] Optionally, after obtaining the original itinerary of the passenger to be recommended, the flight cancellation segments in the original itinerary can be clearly known, and the flight cancellation segments include basic information such as the departure airport, the landing airport, and the departure time. The embodiment of the present invention specifically combines all available flights to the destination of the passenger to be recommended based on the departure airport value of the flight cancellation segments, thereby obtaining multiple alternative itineraries.

[0035] Specifically, all available flights within a preset time after the departure time of the cancelled flight segment can be obtained, for example, all available flights within 12 hours after the departure time, and then the available flights are combined according to the itinerary construction rules provided in the embodiment of the present invention to obtain multiple alternative itineraries. It can be imagined that the basic principle of combining the available flights is to enable the recommended passengers to reach their final destination after a limited number of transfers.

[0036] Optionally, the itinerary construction rules provided by the embodiment of the present invention may include: taking the original itinerary including segments as an example, the original itinerary includes I2-I1-CUR-O1-O2, where the CUR segment is the flight cancellation segment mentioned in the embodiment of the present invention, I2 and I1 are the previous segments, and O1 and O2 are the subsequent segments. On this basis, the construction rules of the alternative itinerary are:

[0037] 1. If both O1 and O2 exist in the subsequent flight segments and are international segments, then when constructing an alternative itinerary, you can modify the three segments CUR-O1-O2 in the original itinerary; otherwise, you can modify at most the three segments I1-CUR-O1 in the original itinerary, and at least only the CUR segment.

[0038] 2. The time difference between the previous and next flight segments of each alternative itinerary’s transit point must meet the MCT requirements of the transit airport.

[0039] 3. The flight arrangements in the alternative itinerary are not limited to the original airlines in the PNR of the recommended passenger. Other domestic / foreign airlines are also available.

[0040] Optionally, both the original itinerary and the constructed alternative itinerary may include the following information:

[0041] The characteristics of a single flight segment include: PNR creation date, selling / operating airline, flight number, departure-arrival airport pair, region pair, country pair, departure-arrival date pair, time pair, cabin class, special services (such as special meals), etc.

[0042] Flight information: departure time, arrival time, whether to stop, total travel time, original flight, cabin class, etc.

[0043] S120. Determine, among multiple preset passenger clusters, a target preset passenger cluster to which the passenger to be recommended belongs, and calculate a first acceptance degree of each alternative itinerary according to the itinerary similarity between the passenger to be recommended and each sample passenger in the target preset passenger cluster.

[0044] In order to assess the acceptability of each alternative itinerary by means of group characteristics, the recommendation method provided in the embodiment of the present invention is further provided with a plurality of preset passenger clusters, and the sample passengers in any preset passenger cluster correspond to similar passenger situational characteristics.

[0045] Optionally, the passenger situational features described in the embodiment of the present invention may include a passenger portrait and cancellation situational features, wherein the passenger portrait includes: preset passenger basic information and passenger preference information, specifically, the preset passenger basic information includes: passenger gender, age, nationality, native place (permanent residence), VIP level, frequent flyer information, etc.

[0046] Passenger preference information includes: departure month preference, departure date preference (working day / non-working day, holiday / three days before or after holiday / non-holiday), departure time preference, arrival date preference, arrival time preference, airline preference, aircraft model preference, number of days in advance to book tickets preference, flight duration preference, flight frequency preference, transit stop time preference, fare preference, discount rate preference, baggage preference, check-in preference, etc.

[0047] Among them, in the passenger portrait, some features are continuous variables, such as passenger age, and some features are discrete variables, such as passenger gender.

[0048] The cancellation scenario features include: flight cancellation features and preset market features corresponding to the flight cancellation features. Specifically,

[0049] Flight cancellation features include: number of days in advance of flight cancellation, cancellation type (ASM, SSM, SSIM), cancellation time (hourly / 2 hours scale), etc.

[0050] The preset duration characteristics corresponding to the flight cancellation characteristics include: the number of flights canceled on the day, the flight cancellation ratio, the number of flights deployed on the day (by time period), the average daily flight frequency of the airline, the average daily market flight frequency, etc.

[0051] Based on the above information, the following describes the process of obtaining each preset passenger cluster provided by the embodiment of the present invention. As mentioned above, each preset passenger cluster includes sample passengers with similar passenger situational characteristics. Therefore, in order to obtain the preset passenger cluster, a certain number of sample passengers need to be prepared first. Of course, each sample passenger corresponds to the passenger situational characteristics described above.

[0052] After obtaining the passenger context characteristics of each sample passenger, the passenger context characteristics of each sample passenger are subjected to data standardization and passenger clustering, and the high-dimensional coefficient user matrix is ​​converted into a low-dimensional user matrix.

[0053] The data standardization process uses the z-score method to normalize the above features, that is,

[0054]

[0055] Among them, x represents the characteristic value of any situational feature;

[0056] μ = mean(x), which represents the average value of the situational feature corresponding to x;

[0057] σ = std(x), which represents the standard deviation of the situational characteristics corresponding to x;

[0058] x' represents the normalized value of the context feature corresponding to x.

[0059] It should be noted that the normalization process of each passenger situational feature can be specifically implemented by referring to the z-score method in the prior art, which will not be described in detail here.

[0060] Furthermore, the K-means clustering algorithm with improved original cluster center is used to perform cluster analysis on sample passengers. The main algorithm steps are as follows:

[0061] Input: A dataset D to be clustered containing m n-dimensional samples, where D = {x i |i=1,2,3,...m},x i =(x i1 , x i2 , x i3 , ..., x in ) T ; Let the number of clusters be K; the maximum number of initial clusters within a cluster is C max , minimum support min Sup .

[0062] Output: K clusters that minimize the squared error.

[0063] Description: When selecting the initial cluster center of clustering, first use the greedy strategy to find the cluster with larger neighborhood radius and density of the data object; then continuously select the cluster with larger density and the farthest distance as the temporary initial cluster, and take the core data object with the highest support in the initial cluster as the center of the initial cluster.

[0064] Step 1: Select the initial cluster center:

[0065] Calculate the Euclidean distance d between each pair of sample objects in D. The calculation formula is: Traverse to find support > minimum support min Sup The core sample objects are sorted by density and the sample object c with the largest density is selected. 1 and the sample object c that is farthest away from it 2 As the initial cluster center, and iteratively select the sample object with the second largest density and the farthest distance from the selected initial cluster center from the remaining core data objects as the initial cluster center, until K initial cluster centers are selected to form the initial cluster center set C, where C = {c i |i=1, 2, 3, ...K}.

[0066] Step 2: Traverse the data set D, calculate and compare each x i To the initial cluster center c j The Euclidean distance d(x i , c j ), and x i Assign to the closest cluster;

[0067] Step 3: Recalculate the average value of each cluster as the new cluster center;

[0068] Step 4: Repeat steps 2 and 3 until the cluster center C no longer changes, and obtain multiple preset passenger clusters.

[0069] It should be noted that, in specific implementation, the clustering algorithm in the prior art may be referred to to cluster the preset sample passengers, and the embodiment of the present invention will not be described in detail here.

[0070] Based on the above multiple preset passenger clusters, when determining the target preset passenger cluster to which the passenger to be recommended belongs, it is first necessary to obtain the passenger context characteristics of the passenger to be recommended, and then calculate the distance between the passenger context characteristics of the passenger to be recommended and the center of each preset passenger cluster; finally, the preset passenger cluster with the smallest distance among the preset passenger clusters is used as the target preset passenger cluster to which the passenger to be recommended belongs.

[0071] It can be seen from the above content that the flight resource recommendation method proposed in the embodiment of the present invention introduces preset passenger clusters, evaluates the acceptability of alternative itineraries from the perspective of group characteristics, and at the same time, comprehensively considers the personal preference information of passengers, which is not available in the prior art.

[0072] Based on the above content, the calculation process of the first acceptance of each alternative itinerary is described below:

[0073] In the embodiment of the present invention, each sample passenger in each passenger sample cluster corresponds to an original itinerary and an actual itinerary. When calculating the first acceptance of each of the alternative itineraries based on the itinerary similarity between the passenger to be recommended and each sample passenger in the target preset passenger cluster, the similarity between the original itinerary of the passenger to be recommended and the original itinerary of each sample passenger in the target preset passenger cluster is first calculated to obtain the corresponding first original itinerary similarity. Specifically, the similarity between the original itinerary of the passenger to be recommended and the original itinerary of any sample passenger in the target preset passenger cluster, that is, the first original itinerary similarity, can be calculated using the following formula. It can be seen that the following similarity calculation formula is calculated based on the Pearson correlation coefficient method.

[0074]

[0075] Among them, X represents the original itinerary feature vector of the passenger to be recommended;

[0076] X' represents the original itinerary feature vector of the sample passenger;

[0077] X i represents the i-th feature in the feature vector of the original itinerary of the passenger to be recommended;

[0078] X i ' represents the i-th feature in the original itinerary feature vector of the sample passenger;

[0079] μ X represents the mean of the original itinerary feature vector of the passenger to be recommended,

[0080] μ X' represents the mean of the original itinerary feature vector of the sample passengers,

[0081] n represents the number of feature vectors of the original itineraries of the recommended passengers and the sample passengers.

[0082] Furthermore, for each alternative itinerary, the similarity between the alternative itinerary and the actual itinerary of each sample passenger in the target preset passenger cluster is calculated to obtain the corresponding first alternative itinerary similarity.

[0083] Specifically, the similarity between any alternative itinerary and the actual itinerary of any sample passenger in the target preset passenger cluster, that is, the similarity of the first alternative itinerary corresponding to each alternative itinerary, can be calculated using the following formula:

[0084]

[0085] Wherein, Y represents the alternative trip feature vector of any alternative trip;

[0086] Y' represents the actual itinerary feature vector of the sample passengers;

[0087] Y i The i-th feature in the alternative itinerary feature vector representing the alternative itinerary;

[0088] Y i ' represents the i-th feature in the actual itinerary feature vector of the sample passenger;

[0089] μ Y represents the mean of the alternative trip feature vectors of any alternative trip,

[0090] μ Y' represents the mean of the actual itinerary feature vector of the sample passengers,

[0091] m represents the number of feature vectors in the alternative itineraries and the actual itineraries of the sample passengers.

[0092] Finally, the first acceptance degree of each candidate itinerary is calculated according to the similarities of each first original itinerary and the similarities of each first candidate itinerary.

[0093] Specifically, firstly, the sample passengers in the target preset passenger cluster are arranged in descending order according to the first original itinerary similarity, so as to obtain the sample passenger ranking;

[0094] Then, based on the above sample passenger ranking, the first N sample passengers in the sample passenger ranking are taken as the first nearest neighbor set, where N is an integer greater than or equal to 1, expressed as U(μ 1 ,μ 2 ,...μ N ).

[0095] Based on the first original itinerary similarity and the first alternative itinerary similarity corresponding to each sample passenger in the first nearest neighbor set, the first acceptance of each alternative itinerary can be calculated respectively.

[0096] Specifically, for any alternative itinerary, its first acceptance degree can be expressed as:

[0097]

[0098] Among them, μ represents the passenger to be recommended.

[0099] It should be noted that, in this step and in subsequent steps, the travel feature vector of each mid-trip is composed of the following features:

[0100] Date features: departure-arrival date (month / day / week), departure-arrival time

[0101] Geographical characteristics: departure-arrival airport, city, province, region, country

[0102] Itinerary characteristics: connection time, main cabin, discount rate, flight distance, fare.

[0103] S130: Determine the companions of the passenger to be recommended according to the PNR of the passenger to be recommended, and calculate the second acceptance of each alternative itinerary according to the similarity between the itinerary of the passenger to be recommended and the companions.

[0104] The flight resource recommendation method provided in the embodiment of the present invention also takes into account the possible mutual influence between the itineraries of fellow travelers or the same social group. Therefore, in this step, the second acceptance degree of each alternative itinerary is calculated.

[0105] Optionally, the fellow passengers in this embodiment include passengers who booked tickets at the same OD (origin and destination) and passengers whose historical familiarity is higher than a preset threshold (i.e., the same social group). k The historical familiarity of can be expressed as:

[0106]

[0107] Among them, F(μ,γ k ) indicates historical familiarity;

[0108] Seg_cnt(μ,γ k ) indicates the number of flight segments that the recommended passenger and the accompanying passengers are traveling together;

[0109] Seg_cnt(μ) indicates the total number of flight segments of the passenger μ to be recommended, which refers to the total number of flight segments of the original itinerary of the passenger to be recommended.

[0110] It should be noted that the booking information of the same OD booker and the same social group can be determined based on the PNR of the recommended passenger. The specific acquisition process can be implemented with reference to the existing technology. The present invention does not specifically limit the screening process of the above-mentioned fellow passengers.

[0111] Similar to the sample passengers, the traveling companions also correspond to an original itinerary and an actual itinerary. The process of calculating the second acceptance of each alternative itinerary can be performed as follows:

[0112] The similarity between the original itinerary of the passenger to be recommended and the original itinerary of each traveling companion is calculated respectively to obtain the corresponding second original itinerary similarity. Specifically, it can be performed according to the following formula:

[0113]

[0114] Among them, X represents the original itinerary feature vector of the passenger to be recommended;

[0115] P' represents the original itinerary feature vector of the fellow passengers;

[0116] X i represents the i-th feature in the feature vector of the original itinerary of the passenger to be recommended;

[0117] P i ' represents the i-th feature in the original itinerary feature vector of the fellow passengers;

[0118] μ X represents the mean of the original itinerary feature vector of the passenger to be recommended,

[0119] μ P' represents the mean of the original itinerary feature vectors of fellow passengers,

[0120] n represents the number of feature vectors of the original itineraries of the recommended passengers and the accompanying passengers.

[0121] Furthermore, for each alternative itinerary, the similarity between the alternative itinerary and the actual itinerary of each fellow passenger is calculated to obtain the corresponding second alternative itinerary similarity. Specifically, the second alternative itinerary similarity can be calculated using the following formula:

[0122]

[0123] Wherein, Y represents the alternative trip feature vector of any alternative trip;

[0124] T' represents the actual itinerary feature vector of fellow passengers;

[0125] Y i The i-th feature in the alternative itinerary feature vector representing the alternative itinerary;

[0126] T i ' represents the i-th feature in the actual itinerary feature vector of the fellow passengers;

[0127] μ Y represents the mean of the alternative trip feature vectors of any alternative trip,

[0128] μ T' represents the mean of the actual itinerary feature vectors of fellow passengers,

[0129] m represents the number of feature vectors in the alternative itineraries and the actual itineraries of the sample passengers.

[0130] Finally, the second acceptance degree of each candidate itinerary is calculated according to the similarities of each second original itinerary and the similarities of each second candidate itinerary.

[0131] Specifically, firstly, the fellow passengers are arranged in descending order according to the similarity of the second original itinerary to obtain a ranking of the fellow passengers;

[0132] Then, the first M fellow travelers in the fellow travelers ranking are taken as the second nearest neighbor set, denoted as R(γ 1 ,γ 2 ,...,γ M ), where M is an integer greater than or equal to 1.

[0133] Based on the similarity of the second original itinerary and the second alternative itinerary corresponding to each fellow passenger in the second nearest neighbor set, the second acceptance of each alternative itinerary is calculated respectively. For any alternative itinerary, its second acceptance can be calculated according to the following formula:

[0134]

[0135] in,

[0136] k represents the kth fellow traveler in the second nearest neighbor set.

[0137] S140, respectively calculating the similarity between each alternative itinerary and the original itinerary of the passenger to be recommended, and obtaining the alternative itinerary similarity corresponding to each alternative itinerary.

[0138] Optionally, for any alternative itinerary, the similarity between the alternative itinerary and the original itinerary of the passenger to be recommended may be calculated according to the following formula:

[0139]

[0140] Among them, l is the variable dimension, is the indicator function, C i Indicates the original stroke, C j Indicates an alternative itinerary. i , C j If the vth variable measurement value does not exist, then is 0, otherwise it is 1; if C i , C j If the vth variable metric value is the same, then is 1; if C i , C j The vth variables in are all numerical variables and are different, then otherwise is 0.

[0141] S150: Pushing each alternative itinerary to the passenger to be recommended based on the first acceptance degree, the second acceptance degree, and the similarity of the alternative itineraries.

[0142] Optionally, after the above steps, the first acceptance degree, the second acceptance degree and the similarity of the alternative itinerary are obtained for each alternative itinerary. On this basis, for each alternative itinerary, the first acceptance degree, the second acceptance degree and the average value of the similarity of the alternative itinerary are calculated to obtain the final acceptance degree of the alternative itinerary. Specifically, it can be expressed as:

[0143]

[0144] Then, the alternative itineraries are pushed to the recommended passengers in the order of final acceptance from high to low.

[0145] In summary, the flight resource recommendation method provided by the present invention comprehensively considers three factors: the group characteristics represented by the target preset passenger clustering cluster, the mutual influence between fellow passengers, and the similarity between each alternative itinerary and the original itinerary of the recommended passenger. Compared with the prior art, the flight resources recommended to passengers by this method are derived from a comprehensive analysis of multiple factors, and therefore have a higher degree of acceptance, which helps to reduce the situation where airline servers do useless work, thereby improving the effective utilization rate of airline servers, improving the utilization rate of flight resources and the evacuation efficiency of airport passengers.

[0146] It should be noted that the flowcharts in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram may represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some implementations as replacements, the functions marked in the boxes may also occur in a sequence different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they may sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or the flowchart, and the combination of the boxes in the block diagram and / or the flowchart, can be implemented with a dedicated hardware-based system that performs the specified functions or operations, or can be implemented with a combination of dedicated hardware and computer instructions.

[0147] It should be noted that although the above content describes each operation in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. In certain circumstances, multitasking and parallel processing may be advantageous.

[0148] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0149] The flight resource recommendation device provided in an embodiment of the present invention is introduced below. The flight resource recommendation device described below can be considered as a functional module architecture that needs to be set in a central device to implement the flight resource recommendation method provided in an embodiment of the present invention; the content described below can be cross-referenced with the above.

[0150] Optional, see Figure 2 , Figure 2 A flight resource recommendation device provided by an embodiment of the present invention includes:

[0151] An acquisition unit 10 is used to acquire an original itinerary of the passenger to be recommended, wherein the original itinerary includes a flight cancellation segment;

[0152] A construction unit 20, configured to construct a combination of available flights from the departure airport of the cancelled flight segment to the to-be-recommended passenger destination, to obtain a plurality of alternative itineraries;

[0153] A first calculation unit 30 is used to determine, from among a plurality of preset passenger clusters, a target preset passenger cluster to which the passenger to be recommended belongs, and to calculate a first acceptance degree of each of the alternative itineraries according to a similarity between the passenger to be recommended and each sample passenger in the target preset passenger cluster, wherein the sample passengers in any of the preset passenger clusters correspond to similar passenger context features;

[0154] The second calculation unit 40 is used to determine the companions of the passenger to be recommended according to the passenger reservation record PNR of the passenger to be recommended, and calculate the second acceptance degree of each of the alternative itineraries according to the similarity between the itineraries of the passenger to be recommended and the companions;

[0155] The third calculation unit 50 is used to calculate the similarity between each of the alternative itineraries and the original itinerary of the passenger to be recommended, and obtain the similarity of the alternative itineraries corresponding to each of the alternative itineraries;

[0156] The recommendation unit 60 is configured to push each of the alternative itineraries to the passenger to be recommended based on the first acceptance level, the second acceptance level, and the similarity of the alternative itineraries.

[0157] Optionally, the recommendation unit 60 is configured to push each of the alternative itineraries to the passenger to be recommended based on the first acceptance degree, the second acceptance degree, and the similarity of the alternative itineraries, specifically including:

[0158] For each of the alternative itineraries, calculating the first acceptance degree, the second acceptance degree, and an average value of the similarity of the alternative itineraries to obtain a final acceptance degree of the alternative itinerary;

[0159] The alternative itineraries are pushed to the passenger to be recommended in the order of the final acceptance from high to low.

[0160] Optionally, the first calculation unit 30 is configured to determine, from among a plurality of preset passenger clusters, the target preset passenger cluster to which the passenger to be recommended belongs, specifically comprising:

[0161] Acquiring passenger situational characteristics of the passenger to be recommended;

[0162] Calculating the distances between the passenger situational features of the passenger to be recommended and the centers of each preset passenger cluster;

[0163] The preset passenger cluster with the smallest distance among the preset passenger clusters is used as the target preset passenger cluster to which the passenger to be recommended belongs.

[0164] Optionally, the passenger situational features include passenger portrait and cancellation situational features, wherein:

[0165] The passenger portrait includes: preset passenger basic information and passenger preference information;

[0166] The cancellation scenario characteristics include: flight cancellation characteristics and preset market characteristics corresponding to the flight cancellation characteristics.

[0167] Optionally, each sample passenger in each of the passenger sample clusters corresponds to an original itinerary and an actual itinerary, and the first calculation unit 30 is used to calculate the first acceptance of each of the alternative itineraries according to the itinerary similarity between the passenger to be recommended and each sample passenger in the target preset passenger cluster, specifically including:

[0168] Calculating the similarity between the original itinerary of the passenger to be recommended and the original itinerary of each sample passenger in the target preset passenger cluster respectively, to obtain a corresponding first original itinerary similarity;

[0169] For each of the alternative itineraries, respectively calculating the similarity between the alternative itinerary and the actual itinerary of each of the sample passengers in the target preset passenger cluster, to obtain a corresponding first alternative itinerary similarity;

[0170] The first acceptance degree of each of the alternative itineraries is calculated according to the similarities of each of the first original itineraries and the similarities of each of the first alternative itineraries.

[0171] Optionally, the first calculating unit 30 is configured to calculate the first acceptance of each of the alternative itineraries according to the similarities of each of the first original itineraries and the similarities of each of the first alternative itineraries, specifically including:

[0172] Arranging the sample passengers in the target preset passenger cluster in descending order of the similarity of the first original itinerary to obtain a sample passenger ranking;

[0173] The first N sample passengers in the sample passenger ranking are taken as the first nearest neighbor set, where N is an integer greater than or equal to 1;

[0174] Based on the first original itinerary similarity and the first alternative itinerary similarity corresponding to each of the sample passengers in the first nearest neighbor set, the first acceptance of each of the alternative itineraries is calculated respectively.

[0175] Optionally, the traveling companions include multiple ones, and each of the traveling companions corresponds to an original itinerary and an actual itinerary. The second calculating unit 40 is used to calculate the second acceptance of each of the alternative itineraries according to the itinerary similarity between the to-be-recommended passenger and the traveling companions, specifically including:

[0176] Calculating the similarity between the original itinerary of the passenger to be recommended and the original itinerary of each of the traveling passengers, respectively, to obtain a corresponding second original itinerary similarity;

[0177] For each of the alternative itineraries, respectively calculating the similarity between the alternative itinerary and the actual itinerary of each of the traveling companions, to obtain a corresponding second alternative itinerary similarity;

[0178] The second acceptance degree of each of the candidate itineraries is calculated according to the similarities of each of the second original itineraries and the similarities of each of the second candidate itineraries.

[0179] Optionally, the second calculating unit 40 is configured to calculate the second acceptance of each of the alternative itineraries according to the similarities of each of the second original itineraries and the similarities of each of the second alternative itineraries, specifically including:

[0180] Arranging the fellow passengers in descending order according to the similarity of the second original itinerary to obtain a ranking of the fellow passengers;

[0181] The first M fellow travelers in the order of the fellow travelers are used as the second nearest neighbor set, where M is an integer greater than or equal to 1;

[0182] Based on the second original itinerary similarity and the second alternative itinerary similarity corresponding to each of the fellow travelers in the second nearest neighbor set, the second acceptance of each of the alternative itineraries is calculated respectively.

[0183] Optionally, each of the preset passenger clusters is obtained by clustering and dividing based on passenger situational characteristics of multiple sample passengers using a clustering algorithm.

[0184] It should be noted that the units involved in the embodiments described in the present disclosure may be implemented by software or hardware. The name of a unit does not limit the unit itself in some cases. For example, the first calculation unit may also be described as a "unit for calculating the first acceptance degree".

[0185] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.

[0186] Although several specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Certain features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination.

[0187] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

Claims

1. A flight resource recommendation method, characterized in that: include: Obtaining the original itinerary of the passenger to be recommended, wherein the original itinerary includes the flight cancellation segment; Constructing available flight combinations from the departure airport of the cancelled flight segment to the passenger destination to be recommended, and obtaining multiple alternative itineraries; Determine a target preset passenger cluster to which the passenger to be recommended belongs among a plurality of preset passenger clusters, and calculate the first acceptance degree of each of the alternative itineraries according to the itinerary similarity between the passenger to be recommended and each sample passenger in the target preset passenger cluster, wherein the sample passengers in any of the preset passenger clusters correspond to similar passenger context characteristics; Determining the companion passengers of the passenger to be recommended according to the passenger reservation record PNR of the passenger to be recommended, and calculating the second acceptance degree of each of the alternative itineraries according to the similarity between the itineraries of the passenger to be recommended and the companion passengers; Calculating the similarity between each of the alternative itineraries and the original itinerary of the passenger to be recommended, and obtaining the alternative itinerary similarity corresponding to each of the alternative itineraries; Based on the first acceptance degree, the second acceptance degree, and the similarity of the alternative itineraries, pushing each of the alternative itineraries to the passenger to be recommended; The sample passengers in each of the target preset passenger clusters correspond to an original itinerary and an actual itinerary. The first acceptance degree of each of the alternative itineraries is calculated according to the itinerary similarity between the passenger to be recommended and each of the sample passengers in the target preset passenger clusters, including: Calculating the similarity between the original itinerary of the passenger to be recommended and the original itinerary of each sample passenger in the target preset passenger cluster respectively, to obtain a corresponding first original itinerary similarity; For each of the alternative itineraries, respectively calculating the similarity between the alternative itinerary and the actual itinerary of each of the sample passengers in the target preset passenger cluster, to obtain a corresponding first alternative itinerary similarity; Calculating first acceptance of each of the alternative itineraries respectively according to the similarities of each of the first original itineraries and the similarities of each of the first alternative itineraries; The traveling companions include a plurality of traveling companions, and each of the traveling companions corresponds to an original itinerary and an actual itinerary. The second acceptance degree of each of the alternative itineraries is calculated according to the itinerary similarity between the to-be-recommended passenger and the traveling companions, including: Calculating the similarity between the original itinerary of the passenger to be recommended and the original itinerary of each of the traveling passengers, respectively, to obtain a corresponding second original itinerary similarity; For each of the alternative itineraries, respectively calculating the similarity between the alternative itinerary and the actual itinerary of each of the traveling companions, to obtain a corresponding second alternative itinerary similarity; The second acceptance degree of each of the candidate itineraries is calculated according to the similarities of each of the second original itineraries and the similarities of each of the second candidate itineraries.

2. The flight resource recommendation method according to claim 1, characterized in that: The step of pushing each of the alternative itineraries to the passenger to be recommended based on the first acceptance degree, the second acceptance degree, and the similarity of the alternative itineraries includes: For each of the alternative itineraries, calculating the first acceptance degree, the second acceptance degree, and an average value of the similarity of the alternative itineraries to obtain a final acceptance degree of the alternative itinerary; The alternative itineraries are pushed to the passenger to be recommended in the order of the final acceptance from high to low.

3. The flight resource recommendation method according to claim 1, characterized in that: The step of determining, among a plurality of preset passenger clusters, a target preset passenger cluster to which the passenger to be recommended belongs comprises: Acquiring passenger situational characteristics of the passenger to be recommended; Calculating the distances between the passenger situational features of the passenger to be recommended and the centers of each preset passenger cluster; The preset passenger cluster with the smallest distance among the preset passenger clusters is used as the target preset passenger cluster to which the passenger to be recommended belongs.

4. The flight resource recommendation method according to any one of claims 1 to 3, characterized in that: The passenger situational features include passenger portrait and cancellation situational features, wherein: The passenger portrait includes: preset passenger basic information and passenger preference information; The cancellation scenario characteristics include: flight cancellation characteristics and preset market characteristics corresponding to the flight cancellation characteristics.

5. The flight resource recommendation method according to claim 1, characterized in that: The calculating the first acceptance degree of each of the alternative itineraries according to the similarities of each of the first original itineraries and the similarities of each of the first alternative itineraries respectively includes: Arranging the sample passengers in the target preset passenger cluster in descending order of the similarity of the first original itinerary to obtain a sample passenger ranking; The first N sample passengers in the sample passenger ranking are taken as the first nearest neighbor set, where N is an integer greater than or equal to 1; Based on the first original itinerary similarity and the first alternative itinerary similarity corresponding to each of the sample passengers in the first nearest neighbor set, the first acceptance of each of the alternative itineraries is calculated respectively.

6. The flight resource recommendation method according to claim 1, characterized in that: The calculating the second acceptance of each of the alternative itineraries according to the similarities of each of the second original itineraries and the similarities of each of the second alternative itineraries respectively includes: Arranging the fellow passengers in descending order according to the similarity of the second original itinerary to obtain a ranking of the fellow passengers; The first M fellow travelers in the order of the fellow travelers are used as the second nearest neighbor set, where M is an integer greater than or equal to 1; Based on the second original itinerary similarity and the second alternative itinerary similarity corresponding to each of the fellow travelers in the second nearest neighbor set, the second acceptance of each of the alternative itineraries is calculated respectively.

7. The flight resource recommendation method according to claim 1, characterized in that: Each of the preset passenger clusters is obtained by clustering and dividing based on passenger situational characteristics of a plurality of sample passengers using a clustering algorithm.

8. A flight resource recommendation device, characterized in that: include: An acquisition unit, configured to acquire an original itinerary of the passenger to be recommended, wherein the original itinerary includes a cancelled flight segment; A construction unit, configured to construct a combination of available flights from the departure airport of the cancelled flight segment of the flight to the passenger destination to be recommended, so as to obtain a plurality of alternative itineraries; A first calculation unit is used to determine, from among a plurality of preset passenger clusters, a target preset passenger cluster to which the passenger to be recommended belongs, and to calculate a first acceptance degree of each of the alternative itineraries according to a similarity between the passenger to be recommended and each sample passenger in the target preset passenger cluster, wherein the sample passengers in any of the preset passenger clusters correspond to similar passenger context features; A second calculation unit is used to determine the companions of the passenger to be recommended according to the passenger reservation record PNR of the passenger to be recommended, and calculate the second acceptance degree of each of the alternative itineraries according to the similarity between the itineraries of the passenger to be recommended and the companions; A third calculation unit is used to calculate the similarity between each of the alternative itineraries and the original itinerary of the passenger to be recommended, and obtain the similarity of the alternative itineraries corresponding to each of the alternative itineraries; A recommendation unit, configured to push each of the alternative itineraries to the passenger to be recommended based on the first acceptance degree, the second acceptance degree, and the similarity of the alternative itineraries; The sample passengers in each of the target preset passenger clusters correspond to an original itinerary and an actual itinerary. The first calculation unit calculates the first acceptance of each of the alternative itineraries according to the itinerary similarity between the passenger to be recommended and each of the sample passengers in the target preset passenger clusters, including: Calculating the similarity between the original itinerary of the passenger to be recommended and the original itinerary of each sample passenger in the target preset passenger cluster respectively, to obtain a corresponding first original itinerary similarity; For each of the alternative itineraries, respectively calculating the similarity between the alternative itinerary and the actual itinerary of each of the sample passengers in the target preset passenger cluster, to obtain a corresponding first alternative itinerary similarity; Calculating first acceptance of each of the alternative itineraries respectively according to the similarities of each of the first original itineraries and the similarities of each of the first alternative itineraries; The traveling companions include a plurality of traveling companions, and each of the traveling companions corresponds to an original itinerary and an actual itinerary. The second calculating unit calculates the second acceptance of each of the alternative itineraries according to the similarity between the itineraries of the to-be-recommended passenger and the traveling companions, including: Calculating the similarity between the original itinerary of the passenger to be recommended and the original itinerary of each of the traveling passengers, respectively, to obtain a corresponding second original itinerary similarity; For each of the alternative itineraries, respectively calculating the similarity between the alternative itinerary and the actual itinerary of each of the traveling companions, to obtain a corresponding second alternative itinerary similarity; The second acceptance degree of each of the candidate itineraries is calculated according to the similarities of each of the second original itineraries and the similarities of each of the second candidate itineraries.

Citation Information

Patent Citations

  • Point-to-point traffic plan pushing method, system and equipment, and storage medium

    CN107944586A

  • Flight shipping space resource allocation method and device

    CN110334959A

  • Passenger travel state determination method and device and server

    CN110751395A

  • Passenger protection method and system

    CN110889610A

  • Flight recommendation method and system based on combined recommendation

    CN111581505A