Prediction Method and Device for Travel Purposes of Passengers, Storage Medium and Electronic Device

Abstract

The present invention provides a method and device for predicting the travel purpose of passengers, a storage medium, and an electronic device. The method includes: collecting passenger data in an information system, processing the passenger data, and performing feature extraction processing on each obtained sample data to obtain respective feature data; constructing respective prediction models, and using each feature data to train each prediction model. After the training of each prediction model is completed, the prediction model with the highest prediction accuracy is used as the target prediction model, and the target prediction model is used to perform prediction processing on the passenger to be predicted, so as to obtain the prediction result of the travel purpose of the passenger to be predicted. By classifying and extracting features from the passenger data, feature data is obtained, and the feature data is used to construct a target prediction model that can predict the travel purpose of passengers, thereby providing prediction data on the travel purpose of passengers for airlines and providing important data basis for civil aviation companies to adjust flight services, filling the gap in the industry.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and particularly relates to a method and device for predicting the purpose of passenger travel, a storage medium, and an electronic device. Background Art

[0002] With the continuous development of computer technology, passenger travel data can be saved, and a large amount of passenger travel data is stored in the information system of airlines. The massive ticket data in the information system has become an important reference data for airlines to analyze market development and provide services for passengers.

[0003] The purposes of passengers' travel are different. For example, travel for business purposes and travel for tourism purposes. Different travel purposes have different demands for flights. So far, the analysis of passengers' travel purposes in the aviation industry is still blank. In order to provide better services for passengers, predicting the travel purposes of passengers has become an urgent problem to be solved. Summary of the Invention

[0004] In view of this, the present invention provides a method and device for predicting the purpose of passenger travel, a storage medium, and an electronic device. The present invention can predict the travel purposes of passengers, filling the industry gap, and providing prediction data on the travel purposes of passengers for airlines, providing an important data basis for airlines to adjust flight services.

[0005] To achieve the above object, the embodiments of the present invention provide the following technical solutions:

[0006] The first aspect of the present invention discloses a method for predicting the purpose of passenger travel, including:

[0007] Collect passenger data in a preset information system;

[0008] Process the passenger data to obtain a sample data set, where the sample data set includes sample data of business travel type and sample data of non-business travel type;

[0009] Perform feature extraction processing on each sample data to obtain various feature data of each sample data;

[0010] Based on various prediction algorithms, construct a prediction model corresponding to each prediction algorithm, and use each feature data to train each prediction model. After each prediction model is trained, obtain the prediction accuracy of each prediction model;

[0011] Take the prediction model with the highest prediction accuracy as the target prediction model, and use the target prediction model to perform prediction processing on the to-be-predicted passenger to obtain the prediction result of the travel purpose of the to-be-predicted passenger.

[0012] Compared with the prior art, the present invention has the following advantages:

[0013] The present invention provides a method and device for predicting the travel purpose of passengers, a storage medium and an electronic device. The method includes: collecting passenger data in an information system, processing the passenger data to obtain a sample data set; performing feature extraction processing on each sample data to obtain various feature data of each sample data; based on various prediction algorithms, constructing a prediction model corresponding to each prediction algorithm, and using each feature data to train each prediction model. After each prediction model is trained, obtaining the prediction accuracy of each prediction model; taking the prediction model with the highest prediction accuracy as the target prediction model, and using the target prediction model to perform prediction processing on the to-be-predicted passenger to obtain the prediction result of the travel purpose of the to-be-predicted passenger. After classifying and processing the data in the information system, the present invention obtains a sample data set, extracts various feature data using the sample data in the sample data set, constructs a prediction model using each feature data, and trains the prediction model. After the training is completed, the prediction model with the highest prediction accuracy is taken as the target prediction model, so as to predict the travel purpose of the to-be-traveled passenger using the target prediction model. Using the present invention, the travel purpose of the to-be-traveled passenger can be predicted, and the relevant prediction data can be fed back to the civil aviation company or the enterprise company, so that the civil aviation company can timely know the travel purpose of the passenger, and then provide more high-quality and appropriate services for the passenger. In addition, the future development trend of civil aviation passenger transportation can be timely understood, providing important data basis for the civil aviation company to adjust flight services, and filling the gap in the industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0015] Figure 1 It is a flowchart of a method for predicting the travel purpose of passengers provided by an embodiment of the present invention;

[0016] Figure 2 It is a flowchart of a method for processing passenger data to obtain a sample data set provided by an embodiment of the present invention;

[0017] Figure 3 It is an application example diagram for classifying ticket data provided by an embodiment of the present invention;

[0018] Figure 4A flowchart of a method for extracting feature data provided by an embodiment of the present invention;

[0019] Figure 5 An associated example diagram of a data table provided by an embodiment of the present invention;

[0020] Figure 6 A schematic structural diagram of a prediction device for passenger travel purposes provided by an embodiment of the present invention;

[0021] Figure 7 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. Detailed implementation manners

[0022] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0023] In this application, the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, the element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or device including the said element.

[0024] With the improvement of people's living standards and the continuous development of the civil aviation industry, passenger friends are more inclined to choose to take an airplane as a means of transportation for their travel and business trips. Compared with other means of transportation such as trains, airplanes have the advantages of being fast, convenient, and having considerate services. In 2020, the total transport turnover, passenger traffic, and cargo and mail traffic completed by the civil aviation industry were 79.85 billion ton-kilometers, 420 million person-times, and 6.766 million tons respectively. The passenger traffic of China's civil aviation has ranked second in the world for 15 consecutive years.

[0025] At the same time, the information system of civil aviation has been continuously improved with the rapid development of civil aviation business. The information system of civil aviation includes the ICS system, CRS system, and departure system, which have accumulated the travel records of passengers. If big data technology cannot be fully utilized to mine and analyze these massive data, it is indeed a huge waste.

[0026] At present, some large enterprises in the world have their own business travel platforms, which enable online declaration and approval, speeding up the work process and improving work efficiency. If it is possible to successfully predict whether a certain ticket is for business travel, then the business travel volume in the entire Chinese market can be calculated, and issues such as which cities are popular business travel destinations can be known, providing data support for business travel companies. Airlines can understand the changes in the business travel market through business travel data, and then adjust their marketing strategies to improve passengers' consumption enthusiasm. Combining time series knowledge, civil aviation companies can understand the characteristics of passenger transport changes, use machine learning algorithms to fit the data, establish a prediction model describing the development trend of civil aviation passenger transport, and make reasonable predictions in the future.

[0027] At present, the aviation industry does not have a technology for predicting the purpose of passengers' travel. In order to fill the industry gap and provide data basis for airlines on the purpose of passengers' travel, the present invention provides a method and device for predicting the purpose of passengers' travel, a storage medium and an electronic device. The present invention can be used in many general or special computing device environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multi-processor devices, distributed computing environments including any of the above devices or equipment, and so on.

[0028] An embodiment of the present invention provides a method for predicting the purpose of passengers' travel. The execution subject of this method can be a processor or a server of an aviation system. Refer to Figure 1 , which is a flowchart of the method for predicting the purpose of passengers' travel provided by an embodiment of the present invention, and is specifically described as follows:

[0029] S101. Collect passenger data in a preset information system.

[0030] The information system is a data warehouse in the civil aviation system for storing passengers' travel data. Among them, the passenger data includes but is not limited to passenger reservation record (PNR) data and ticket issuance data. Among them, the PNR data includes but is not limited to passengers' basic information, such as age, gender, encrypted ID number, PNR number, etc.; the ticket issuance data includes but is not limited to passengers' ticket face information, such as flight number, departure place, destination, departure date, ticket price and other data.

[0031] S102. Process the passenger data to obtain a sample data set.

[0032] The sample data set provided by the embodiment of the present invention includes sample data of business travel types and sample data of non-business travel types.

[0033] Refer to Figure 2 , which is a flowchart of the method for processing passenger data to obtain a sample data set provided by an embodiment of the present invention, and is specifically described as follows:

[0034] S201. Determine the identity information of each passenger and each passenger based on the business trip data in a preset business trip system.

[0035] The business trip system provided by the embodiments of the present invention is a system for an airline company and an enterprise user to sign a cooperation agreement for enterprise employees to book tickets during business trips. The ticket orders completed in this system are all business trip-related ticket orders. Therefore, the data in this system are all business trip-related data. It should be noted that the relevant ticket booking data in this system also exists in the information system. Further, when an enterprise employee uses the business trip system to book tickets, the enterprise employee can be regarded as a passenger.

[0036] The business trip data includes the passenger identity data and ticket booking data when each passenger books a ticket. Further, the ticket booking data includes, but is not limited to, data such as flight number, departure place, destination, departure date, ticket price, etc. The passenger identity data includes, but is not limited to, the ID information of the passenger, such as ID card number, name, contact information, etc.

[0037] Parse the business trip data to determine each passenger who uses the business trip system to book tickets, and determine the identity information of each passenger. Among them, the identity information includes, but is not limited to, information such as ID card number, name, and gender.

[0038] S202. Based on the identity information of each passenger, obtain each ticket data of each said passenger within a preset time period from the passenger data.

[0039] After obtaining the identity information of each passenger, based on the identity information of each passenger, extract each ticket data of the passenger within a preset time period from the passenger data, so as to obtain the travel records of each passenger within this time period.

[0040] It should be noted that the passenger data is the data in the information system, and each ticket data of each passenger extracted from the passenger data is data that has not been classified and tagged. Preferably, the time period can be set according to actual needs. For example, the time period can be in years, or in months or days. Exemplarily, if the time period is 2020, then each ticket data extracted is data generated when the ticket booking behavior occurred in 2020.

[0041] Preferably, by extracting each ticket data of each passenger within the time period from the passenger data, invalid data can be effectively eliminated, so as to obtain valid each ticket data.

[0042] S203. For each ticket data of each passenger, obtain the ticket number in the ticket data, and determine whether the ticket number exists in the business travel system. If it does not exist, determine the ticket data to which the ticket number belongs as non-business travel sample data. If it exists, determine the ticket data to which the ticket number belongs as business travel sample data.

[0043] It should be noted that after obtaining the respective ticket data of each passenger, classification and tagging operations need to be performed on the respective ticket data. One method is as follows:

[0044] For each ticket data, obtain the ticket number in the ticket data, and determine whether the ticket number exists in the business travel system. If it exists, determine the ticket data to which the ticket number belongs as business travel sample data. If it does not exist, determine the ticket data to which the ticket number belongs as non-business travel sample data. It should be noted that when determining the ticket data as business travel sample data or non-business travel sample data, the allocation and tagging operations of the ticket data are completed.

[0045] Furthermore, when determining whether a ticket number exists in the business travel system, the business travel data in the business travel system can be parsed, and it can be determined whether there is a ticket number in the parsed data that is the same as the ticket number. If it exists, it can be determined that there is a ticket number in the business travel system. If it does not exist, it can be determined that there is no ticket number in the business travel system.

[0046] S204. Combine the respective ticket data of each passenger into a sample data set.

[0047] Combine the ticket data that has completed classification and tagging into a sample data set.

[0048] Refer to Figure 3 , which is an application example diagram for classifying ticket data provided by the invention embodiment. As shown in the figure, obtain the ID information of the passengers who purchase tickets in the business travel system, query the respective ticket data of these passengers throughout the year in the information system. For each ticket data of each passenger, when the ticket number in the ticket data is in the business travel system, the ticket data can be determined as business travel ticket data. When the ticket number in the ticket data is not in the business travel system, the ticket data can be determined as non-business travel ticket data.

[0049] In the method provided by the invention embodiment, when classifying and tagging the respective ticket data, screening processing of the data is also performed simultaneously. Screening the data can effectively remove invalid data, and perform classification and tagging processing on the valid data obtained after screening, so as to improve the accuracy of the data.

[0050] S103. Perform feature extraction processing on each sample data to obtain the respective feature data of each sample data.

[0051] After obtaining each sample data, it is necessary to extract the respective feature data of each sample data. The method flowchart for extracting the respective feature data from the sample data can refer to Figure 4 , for each sample data, it can be based on Figure 4 The method flow shown to extract the respective feature data. The description of Figure 4 is as follows:

[0052] S401. Obtain the ticket number in the sample data and determine the respective data tables associated with the ticket number.

[0053] Parse the sample data to obtain the ticket number in the sample data. It should be noted that the ticket number can be the ticket number of the air ticket in the sample data; further, the ticket number can be saved in the passenger ticket information table. When determining the respective data tables associated with the ticket number, the respective data tables associated with the passenger ticket information table can be determined, specifically as Figure 5 shown, Figure 5 This is an example diagram of the association of data tables provided by an embodiment of the present invention. The data tables include but are not limited to the passenger information table, air ticket price table, air ticket itinerary table, and flight distance static table, etc.

[0054] S402. Based on the respective data tables, obtain the data parameters of the respective data fields required.

[0055] Based on the respective data tables, obtain the data parameters of the respective data fields required, specifically, such as the data parameters in data fields such as the passenger's date of birth, international / domestic flag, group / single flag, departure time, time interval, PNR number, total fare, mileage, departure city, arrival city, etc.

[0056] S403. Process the data parameters of the respective data fields to obtain the respective feature data.

[0057] Processing the data parameters of the respective data fields can obtain the respective feature data. It should be noted that the feature data can include two types of feature data. One is the derived feature data, and the other is the non-derived feature data. Among them, the non-derived feature data can be the data parameters in the data field, and the derived feature data can be the data obtained after performing arithmetic processing on the data parameters in the data field.

[0058] Each characteristic data includes but is not limited to gender, group / dispersed flag, whether it takes off on a holiday, what day of the week the take-off date is, whether it is a tourist city, passenger age, whether it is a foreign airline, the number of flight segments, unit fare, etc.; Exemplarily, non-derived characteristic data can be gender, group / dispersed flag, etc.; Derived characteristic data can be whether it takes off on a holiday, what day of the week the take-off date is, whether it is a tourist city, passenger age, whether it is a foreign airline, the number of flight segments, unit fare, etc.

[0059] In the method provided by the embodiments of the present invention, by processing each sample data, each characteristic data of each sample data can be extracted, thereby removing dirty data and redundancy in the sample data and obtaining more accurate data.

[0060] S104. Based on each prediction algorithm, construct a prediction model corresponding to each prediction algorithm, and use each characteristic data to train each prediction model. After each prediction model is trained, obtain the prediction accuracy of each prediction model.

[0061] In the method provided by the embodiments of the present invention, the prediction algorithm can be a machine learning algorithm, such as: decision tree, KNN (KNN, K-Nearest Neighbor, proximity algorithm), random forest and other algorithms.

[0062] When using each prediction algorithm to construct a prediction model, it should be noted that at least one prediction model is constructed for each prediction algorithm. Input each characteristic data into each prediction model to train each prediction model. It should be noted that the process of training the prediction model is as follows:

[0063] For each prediction model, input each characteristic data into the prediction model, and based on the data processing rules in the prediction model, continuously process each characteristic data until the training result of the prediction model meets the preset end condition, then the training of the prediction model is completed.

[0064] The data processing rules and end conditions of the prediction models using different prediction algorithms are different. Exemplarily, when the prediction algorithm used by the prediction model is KNN, the data processing rule is to perform clustering processing on each sample data based on each characteristic data of each sample data, and the end condition is that the number of clustering processes reaches the preset number of times; when the prediction algorithm used by the prediction model is a decision tree, it is necessary to perform recursive processing on each sample data according to information entropy, information gain and each characteristic data until the recursion ends.

[0065] Exemplarily, the present invention takes the decision tree, KNN and random forest as examples for illustration, which are specifically as follows:

[0066] 1). Decision tree

[0067] Decision trees are a common type of machine learning method. In 1986, Quinlan proposed the ID3 algorithm. After years of development, its improved algorithm, the C4.5 decision tree classification algorithm, is now widely used in industry. This algorithm is a decision tree algorithm developed after optimizing and improving the deficiencies of the ID3 classification algorithm.

[0068] Decision trees make branch condition judgments based on a tree structure to reach the decision result at the leaf node, so they are named decision trees. A decision tree is a tree structure. The judgment of attributes is completed by each internal node, the output result is given by each branch, and the finally reached leaf node represents the classification result.

[0069] So, how to generate a decision tree from data? Several concepts need to be introduced.

[0070] (1) Information entropy: Generally, the entropy value calculation formula for a random sample is as follows:

[0071]

[0072] Among them, Ent(D) means information entropy, y represents the number of categories, and p k is the proportion of the k-th category of samples in the sample set D;

[0073] If a sample set has two types of samples, each accounting for 1 / 2, its value calculated according to the formula is:

[0074]

[0075] (2) Information gain: We calculate the information entropy of dv according to the formula. Since the branch node with a large number of samples has a greater influence, weights need to be assigned to different branches to obtain the information gain formula:

[0076]

[0077] Among them, Gain(D, m) is the information gain, v is a certain attribute of the sample, and D is the sample set;

[0078] A decision tree is constructed by using the above information entropy and information gain.

[0079] 2) KNN

[0080] KNN is a classic classification method proposed by Cover T and Hart P in 1967. Its working principle is as follows: There is a sample data set, also known as the training sample set, and each data in the sample set has a label, that is, we know the corresponding relationship between each data in the sample set and the classification it belongs to. After inputting new data without a label, each feature of the new data is compared with the corresponding features of the data in the sample set, and then the algorithm finds several samples closest to the sample to be classified. Generally speaking, we only select the first k most similar samples in the sample data set, which is the origin of k in the k-nearest neighbor algorithm. Usually, k is an integer not greater than 20. Finally, the classification that appears most frequently among the k most similar samples is selected as the classification of the new data.

[0081] 3), Random Forest

[0082] Mainly proposed by statistician Leo Breiman in the 1990s, random forest is a type of ensemble learning. The so-called ensemble learning means that a classifier may be simple and the classification accuracy is not high, but by constructing multiple classifiers to complete the classification task, a relatively high accuracy can often be obtained.

[0083] Boosting can be understood as an enhancement method, which is a machine learning algorithm that reduces the error generated in supervised learning. By learning a series of weak classifiers and then combining them into a strong classifier. The most representative is the AdaBoost algorithm: During the training process, the samples are initialized with the same weights. After N rounds of training, the weights of the samples that fail in training are increased, and the learning pays more attention to the error samples, and finally multiple prediction functions are obtained.

[0084] Simply put, random forest is to build multiple decision trees, and then these multiple decision trees make decisions together on which class the current sample belongs to. First, assume that the size of the training set is M. For each tree, N training samples are randomly drawn from M with replacement for training, so the training sets of each tree are different. Second, the number of sample features is K. Set a constant i < K, and randomly select i feature subsets from the K features. Each time the tree performs node splitting, the optimal one is selected from these K features; then repeat the above step to make each tree grow well without pruning; finally, build multiple decision trees, and then there will be multiple classification results for the sample to be classified. Generally, the voting method is used to determine the output category.

[0085] Exemplarily, Table 1 shows the data of the accuracy of the prediction models using different prediction algorithms, and the specific content is shown in Table 1:

[0086]

[0087]

[0088] Table 1

[0089] It should be noted that the prediction model can also be referred to as a machine learning model.

[0090] As can be seen from Table 1, the target prediction model is a prediction model with an accuracy rate of 76.76.

[0091] S105. Take the prediction model with the maximum prediction accuracy as the target prediction model, and use the target prediction model to perform prediction processing on the to-be-predicted passenger to obtain the prediction result of the travel purpose of the to-be-predicted passenger.

[0092] After obtaining the target prediction model, the target prediction model can be used to perform prediction processing on the to-be-predicted passenger, so that the prediction result of the travel purpose of the to-be-predicted passenger can be obtained. It should be noted that the prediction result of the travel purpose includes the predicted travel purpose information of the passenger, such as the travel purpose is business travel, or the travel purpose is tourism travel.

[0093] Preferably, when using the method provided by the present invention to predict the travel purpose of a passenger, there are two travel purposes of the passenger, one is business travel, and the other is tourism travel.

[0094] It should be noted that when the present invention uses the target prediction model to perform prediction on the to-be-predicted passenger, the specific process is as follows:

[0095] Obtain the ticket data of the to-be-predicted passenger;

[0096] Perform feature extraction processing on the ticket data to obtain each to-be-predicted feature data of the to-be-predicted passenger;

[0097] Input each to-be-predicted feature data into the target prediction model, trigger the target prediction model to process the to-be-predicted feature data, and obtain the prediction result of the travel purpose of the to-be-predicted passenger.

[0098] It should be noted that the ticket data includes, but is not limited to, the ticket information of the to-be-predicted passenger and the passenger information, etc. The extraction method of the to-be-predicted feature data can refer to the description of the extraction of feature data in the above text.

[0099] After extracting each to-be-predicted feature data, input each to-be-predicted feature data into the target prediction model, so that after the target prediction model processes each to-be-predicted feature data, the prediction result of the travel purpose of the to-be-predicted passenger is obtained.

[0100] In the method provided by the embodiments of the present invention, passenger data in the information system is collected, and the passenger data is processed to obtain a sample data set; feature extraction processing is performed on each sample data to obtain respective feature data of each sample data; based on respective prediction algorithms, a prediction model corresponding to each prediction algorithm is constructed, and each prediction model is trained using each feature data. After the training of each prediction model is completed, the prediction accuracy of each prediction model is obtained; the prediction model with the maximum prediction accuracy is used as the target prediction model, and the target prediction model is used to perform prediction processing on the to-be-predicted passenger to obtain the prediction result of the travel purpose of the to-be-predicted passenger. After classifying and processing the data in the information system, the present invention obtains a sample data set, extracts respective feature data using the sample data in the sample data set, constructs a prediction model using each feature data, and trains the prediction model. After the training is completed, the prediction model with the maximum prediction accuracy is used as the target prediction model, so as to predict the travel purpose of the to-be-traveled passenger using the target prediction model. Using the present invention, the travel purpose of the to-be-traveled passenger can be predicted, and the relevant prediction data can be fed back to civil aviation companies or enterprise companies, so that civil aviation companies can timely learn about the travel purpose of passengers, and then provide more high-quality and appropriate services for passengers. In addition, the future development trend of civil aviation passenger transportation can be timely understood, providing important data basis for civil aviation companies to adjust flight services, and filling the gap in the industry.

[0101] Furthermore, the present invention provides an application example for illustration. When the present invention is implemented in real life, it can be achieved through the following 5 steps, which are specifically described as follows:

[0102] 1. Analyze and collect civil aviation passenger data;

[0103] 2. Sort out classical machine learning classification algorithms;

[0104] 3. Classify and label ticket data;

[0105] 4. Extract features from ticket data;

[0106] 5. Perform algorithm modeling, select the best model, predict the to-be-predicted passengers, and give prediction results.

[0107] The above steps are described in detail as follows:

[0108] 1. Analyze and collect civil aviation passenger data;

[0109] For the data mining task of classification, good training data is necessary. The data comes from the data warehouse of the civil aviation company and mainly includes two parts of information. One is PNR (Passenger Name Record) data, which contains passengers' basic information such as age, gender, encrypted ID number, and PNR number. The other is ticket issuance data, which contains passengers' ticket face information such as flight number, departure place, destination, departure date, ticket price, etc.

[0110] By repeatedly discussing with business personnel and the data characteristics related to the purpose of travel, and combining the existing data situation in the database, some data fields are selected. The important data fields are as follows:

[0111] Ticket number, passenger gender, passenger ID number, office number, passenger date of birth, international / domestic flag, group / individual flag, ticket issuance date, booking date, departure date, departure time, time interval, number of passengers in PNR, total fare, mileage, departure city, arrival city, etc., 30 fields in total. Among them, number fields such as ticket number and passenger ID number are only used to identify which ticket and which passenger. They are not helpful for the machine learning model. A number cannot help us identify whether it is a business trip or a non-business trip. Similarly, the ticket issuance date and departure date cannot be input into the machine learning model. Therefore, the fields of ticket number, passenger ID number, office number, departure date, ticket issuance date, and booking date are deleted.

[0112] Generally speaking, tourists will book tickets in advance, while business travelers will book tickets a few days in advance temporarily. Using the departure date - booking date, the time interval is obtained.

[0113] It should be noted that the data fields can be selected according to actual needs.

[0114] 2. Sorting out classical machine learning classification algorithms;

[0115] For the convenience of research, the purpose of passengers' travel can be divided into two types: business travel and non-business travel. Business travel refers to travel with official nature, such as business activities, meetings, negotiations, etc. Non-business travel includes leisure travel for personal reasons such as tourism and visiting relatives. Thus, this project is a classification task, so classification algorithms need to be used for modeling. We sort out the common algorithms, understand the advantages and disadvantages of the algorithms, so as to apply them to the project of travel purpose. The classical classification algorithms include knn, decision tree, random forest, and svm.

[0116] 3. Classifying and tagging ticket data;

[0117] After obtaining the passenger data in the information system, the data has not been classified and tagged yet. And for the classification task, it is necessary to classify the data. The classification idea is as follows:

[0118] (1) Business trip data: The civil aviation company has developed a business trip system. A cooperation agreement is signed between the business trip system and enterprise users, requiring employees of the cooperating enterprises to book tickets through this system when they have business trip ticket booking needs in the company. Therefore, this can ensure that the flight tickets booked in the business trip system are all of the business trip category, and thus business trip labeled data is obtained.

[0119] (2) Non-business trip data: After obtaining the business trip labeled data, the passenger ID numbers of business trip flight tickets are obtained. The annual travel records of these passengers are retrieved from the data warehouse. Tickets with ticket numbers in the business trip system are business trip flight tickets, and those without ticket numbers in the business trip system mean that they are tickets booked for the non-business trip travel of passengers, which are determined as non-business trip flight tickets. In this way, we obtain the business trip and non-business trip flight tickets of the same batch of passengers.

[0120] 4. Extract features from the ticket data;

[0121] Feature extraction is very important for real data mining projects. In actual work, feature extraction aims to remove dirty data and redundancy in the original data, encode the data so that the algorithm can model and mine the processed data. Good feature engineering can make the existing data reach the upper limit of the machine learning accuracy.

[0122] Derived features are also added through the original sample data: whether it takes off on a holiday, the day of the week of the take-off date, whether it is a tourist city, the passenger age, whether it is a foreign airline, the number of flight segments, and the unit ticket price.

[0123] Since machine learning algorithms can only recognize numbers and cannot perform calculations on strings, string variables need to be encoded into numbers. For example, for gender, female is marked as 0 and male is marked as 1; for the group / individual identification, non-group is marked as 0 and group is marked as 1. It should be noted that machine learning algorithms can be understood as the prediction algorithms in the above text.

[0124] Another is to add derived features. The sample data contains the departure date of the air ticket, but the date cannot be directly used for calculation. However, it can be used to determine whether this day is a holiday. If a passenger departs on a holiday, it is very likely for tourism, and if on a working day, it is more likely for business trips. The sample data also includes the departure location and destination of the air ticket. If the landing is at a tourist destination, it is very likely for travel. By collecting the top ten popular domestic and international cities of large tourism companies, then querying the popular cities announced by the International Tourism Association, and finally comprehensively comparing to obtain the popular tourist cities, and then coding. If the destination is a tourist city, it is marked as 1, otherwise 0. The passenger's age is the departure date minus the birth year. There is a certain relationship between age and business trips. If the passenger is less than 20 years old, then it is very likely for travel. Finally, the unit price of the ticket can also be calculated, where the unit price of the ticket = total ticket price / mileage. Since the flight distance of each air ticket is different, the ticket prices cannot be directly compared. Generally, tourists will buy discounted tickets, which are cheaper than business trips. After converting to the unit price, the ticket prices can be compared.

[0125] Exemplarily, Table 2 contains the coding rules for some data fields, and the specific content is as follows:

[0126] Field Name Coding Rule Gender 0: Female; 1: Male Group Dissolution Indicator Non-group: 0, Group: 1 Whether Departing on Holiday No: 0, Yes: 1 Day of the Week of Departure Date Sunday - Saturday Whether a Tourist City No: 0, Yes: 1 Passenger Age Departure Date - Birth Year Whether a Foreign Airline No: 0, Yes: 1 Number of Flight Segments Number of Flight Segments Taken by the Passenger Unit Fare Total Fare / Mileage

[0127] Table 2

[0128] 5. Conduct algorithm modeling, select the best model, predict the passengers to be predicted, and give the prediction results;

[0129] Model the data processed by feature engineering, respectively use KNN, decision tree, and random forest for modeling, and perform algorithm tuning to obtain the best model of the algorithm. Then compare the accuracies of each algorithm, and select the prediction model with the highest accuracy as the best model, which is the target prediction model. Use the target prediction model to predict the passengers to be predicted and give the prediction results.

[0130] With the improvement of people's living standards and the continuous development of the civil aviation industry, passengers are more inclined to choose airplanes as their means of transportation for travel and business trips. Compared with other means of transportation such as trains, airplanes have the advantages of being fast, convenient, and having considerate services. The passenger transportation volume of China's civil aviation has ranked second in the world for 15 consecutive years. At the same time, due to the continuous development of big data technology, more and more massive ticket data can be stored. In this context, it is urgent to analyze and model civil aviation data. Whether it is airlines, travel agencies, or tourism platform companies, they all hope to know the travel purposes of passengers, which can provide important basis for their business. This patent uses machine learning classification algorithms for modeling and proposes a method for judging passengers' travel purposes, which has high accuracy. It can provide important data support for the civil aviation industry and can be further analyzed using the results of travel purposes to obtain the travel portraits of national civil aviation passengers.

[0131] Corresponding to Figure 1 the method shown, an embodiment of the present invention provides a prediction device for passengers' travel purposes. This device can be applied to the civil aviation system and is used to support Figure 1 the implementation of the method shown in real life. The structural schematic diagram of this device is as Figure 6 shown and is specifically described as follows:

[0132] A collection unit 601 for collecting passenger data in a preset information system;

[0133] A processing unit 602 for processing the passenger data to obtain a sample data set, where the sample data set includes sample data of business travel types and non-business travel types;

[0134] A feature extraction unit 603 for performing feature extraction processing on each sample data to obtain various feature data of each sample data;

[0135] A training unit 604 for constructing a prediction model corresponding to each prediction algorithm based on various prediction algorithms and using each feature data to train each prediction model. After each prediction model is trained, obtain the prediction accuracy of each prediction model;

[0136] A prediction unit 605 for using the prediction model with the highest prediction accuracy as the target prediction model and using the target prediction model to perform prediction processing on the to-be-predicted passenger to obtain the travel purpose prediction result of the to-be-predicted passenger.

[0137] In the device provided by the embodiment of the present invention, passenger data in the acquisition information system is collected, and the passenger data is processed to obtain a sample data set; feature extraction processing is performed on each sample data to obtain various feature data of each sample data; based on various prediction algorithms, a prediction model corresponding to each prediction algorithm is constructed, and each prediction model is trained using each feature data. After the training of each prediction model is completed, the prediction accuracy of each prediction model is obtained; the prediction model with the highest prediction accuracy is used as the target prediction model, and the target prediction model is used to perform prediction processing on the to-be-predicted passenger to obtain the prediction result of the travel purpose of the to-be-predicted passenger. After classifying and processing the data in the information system, the present invention obtains a sample data set, extracts various feature data using the sample data in the sample data set, constructs a prediction model using each feature data, and trains the prediction model. After the training is completed, the prediction model with the highest prediction accuracy is used as the target prediction model, so as to predict the travel purpose of the to-be-traveled passenger. Using the present invention, the travel purpose of the to-be-traveled passenger can be predicted, and the relevant prediction data can be fed back to the civil aviation company or enterprise company, so that the civil aviation company can timely know the travel purpose of the passenger, and then provide more high-quality and appropriate services for the passenger. In addition, the future development trend of civil aviation passenger transportation can be timely understood, providing important data basis for the civil aviation company to adjust flight services, and filling the gap in the industry.

[0138] In an embodiment of the present application, based on the foregoing solution, the processing unit 602 may be configured to:

[0139] A first determination subunit, configured to determine each passenger and the identity information of each passenger based on the travel data in a preset travel system;

[0140] A first acquisition subunit, configured to obtain each ticket data of each passenger within a preset time period from the passenger data based on the identity information of each passenger;

[0141] A judgment subunit, configured to, for each ticket data of each passenger, obtain the ticket number in the ticket data, and judge whether the ticket number exists in the travel system. If not, determine the ticket data to which the ticket number belongs as non-travel sample data, and if so, determine the ticket data to which the ticket number belongs as travel sample data;

[0142] A composition subunit, configured to compose the respective ticket data of each passenger into a sample data set.

[0143] In an embodiment of the present application, based on the foregoing solution, the feature extraction unit 603 may be configured to:

[0144] A second determination subunit, configured to, for each of the sample data, obtain the ticket numbers in the sample data and determine the respective data tables associated with the ticket numbers;

[0145] A second acquisition subunit, configured to obtain the data parameters of the respective required data fields based on the respective data tables;

[0146] A first processing subunit, configured to process the data parameters of the respective data fields to obtain respective feature data.

[0147] In an embodiment of the present application, based on the foregoing solution, the training unit 604 may be configured as follows:

[0148] A training subunit, configured to, for each of the prediction models, input the respective feature data into the prediction model and continuously process the respective feature data based on the data processing rules in the prediction model until the training result of the prediction model meets a preset end condition, at which point the training of the prediction model is completed.

[0149] In an embodiment of the present application, based on the foregoing solution, the prediction unit 605 may be configured as follows:

[0150] A third acquisition subunit, configured to obtain the ticket data of the passenger to be predicted;

[0151] A first processing subunit, configured to perform feature extraction processing on the ticket data to obtain respective to-be-predicted feature data of the passenger to be predicted;

[0152] A triggering subunit, configured to input the respective to-be-predicted feature data into the target prediction model, trigger the target prediction model to process the to-be-predicted feature data, and obtain a prediction result of the travel purpose of the passenger to be predicted.

[0153] An embodiment of the present invention further provides a storage medium, where the storage medium includes stored instructions, and when the instructions are running, the device where the storage medium is located is controlled to execute the method for predicting the travel purpose of a passenger as described above.

[0154] An embodiment of the present invention further provides an electronic device, a schematic structural diagram of which is as Figure 7 shown, and specifically includes a memory 701 and one or more instructions 702, where one or more instructions 702 are stored in the memory 701 and are configured to be executed by one or more processors 703 to perform the method for predicting the travel purpose of a passenger as described above.

[0155] The specific implementation processes and their derivative methods of the above respective embodiments are all within the protection scope of the present invention.

[0156] Although the subject matter has been described in language specific to structural features and / or methodological logical acts, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. On the contrary, the specific features and acts described above are merely example forms for implementing the claims.

[0157] Although a number of specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the disclosure of the present application. Certain features described in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, the various features described in the context of a single embodiment can also be implemented separately or in any suitable sub-combination in multiple embodiments.

[0158] The above description is only a preferred embodiment of the disclosure of the present application and an explanation of the technical principles applied. Those skilled in the art should understand that the scope of the disclosure involved in the present application is not limited to the technical solutions formed by the 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 disclosure concept. For example, the technical solutions formed by mutually replacing the above features with the technical features (but not limited to) having similar functions disclosed in the present application.

[0159] In the detailed implementation part, the present application repeats all the content protected in the form of claims in the following form:

[0160] According to one or more embodiments disclosed in the present application, Figure 1 A method for predicting the purpose of passenger travel is provided, including:

[0161] Collect passenger data in a preset information system;

[0162] Process the passenger data to obtain a sample data set, where the sample data set includes sample data of business travel type and non-business travel type;

[0163] Perform feature extraction processing on each of the sample data to obtain respective feature data of each of the sample data;

[0164] Based on various prediction algorithms, construct a prediction model corresponding to each prediction algorithm, and use each of the feature data to train each prediction model. After each prediction model is trained, obtain the prediction accuracy of each prediction model;

[0165] Take the prediction model with the highest prediction accuracy as the target prediction model, and use the target prediction model to perform prediction processing on the to-be-predicted passenger to obtain the prediction result of the travel purpose of the to-be-predicted passenger;

[0166] Among them, the training of each prediction model using the respective feature data includes:

[0167] For each prediction model, input the respective feature data into the prediction model, and based on the data processing rules in the prediction model, continuously process the respective feature data until the training result of the prediction model meets the preset end condition, then complete the training of the prediction model;

[0168] Among them, the use of the target prediction model to perform prediction processing on the to-be-predicted passenger to obtain the prediction result of the travel purpose of the to-be-predicted passenger includes:

[0169] Obtain the ticket data of the to-be-predicted passenger;

[0170] Perform feature extraction processing on the ticket data to obtain the respective to-be-predicted feature data of the to-be-predicted passenger;

[0171] Input the respective to-be-predicted feature data into the target prediction model, trigger the target prediction model to process the to-be-predicted feature data, and obtain the prediction result of the travel purpose of the to-be-predicted passenger.

[0172] According to one or more embodiments disclosed in the present application, Figure 2 A method for processing passenger data to obtain a sample data set is provided, including:

[0173] Based on the travel data in a preset travel system, determine the identity information of each passenger and each passenger;

[0174] Based on the identity information of each passenger, obtain the respective ticket data of each passenger within a preset time period from the passenger data;

[0175] For each ticket data of each passenger, obtain the ticket number in the ticket data, and determine whether the ticket number exists in the travel system. If it does not exist, determine the ticket data to which the ticket number belongs as non-travel type sample data. If it exists, determine the ticket data to which the ticket number belongs as travel type sample data;

[0176] Form the respective ticket data of each passenger into a sample data set.

[0177] According to one or more embodiments disclosed in the present application, Figure 4A method for extracting feature data is provided, including:

[0178] For each of the sample data, obtain the ticket number in the sample data and determine each data table associated with the ticket number;

[0179] Based on each of the data tables, obtain the data parameters of each required data field;

[0180] Process the data parameters of each of the data fields to obtain each feature data.

[0181] Each embodiment in this specification is described in a progressive manner. For the same or similar parts between the embodiments, reference can be made to each other. The key point of each embodiment is to illustrate the differences from other embodiments. In particular, for a system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple. For the relevant parts, reference can be made to the partial description of the method embodiment. The systems and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative work.

[0182] Those skilled in the art can further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0183] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting the travel purpose of passengers, characterized in that, it includes: Collect passenger data in a preset information system; The information system is a data warehouse in the civil aviation system for storing the travel data of passengers; Process the passenger data to obtain a sample data set, where the sample data set contains sample data of business travel type and non-business travel type; Perform feature extraction processing on each sample data to obtain various feature data of each sample data; Based on various prediction algorithms, construct a prediction model corresponding to each prediction algorithm, and use each feature data to train each prediction model. After each prediction model is trained, obtain the prediction accuracy of each prediction model; Take the prediction model with the highest prediction accuracy as the target prediction model, and use the target prediction model to perform prediction processing on the passengers to be predicted to obtain the prediction result of the travel purpose of the passengers to be predicted; Among them, the processing of the passenger data to obtain a sample data set includes: Based on the business travel data in a preset business travel system, determine the identity information of each passenger and each passenger; the business travel system is a system in which the civil aviation company has signed a cooperation agreement with enterprise users for enterprise employees to book tickets when on business trips; Based on the identity information of each passenger, obtain all the ticket data of each passenger within a preset time period from the passenger data; For each ticket data of each passenger, obtain the ticket number in the ticket data, and determine whether the ticket number exists in the business travel system. If it does not exist, determine the ticket data to which the ticket number belongs as non-business travel type sample data. If it exists, determine the ticket data to which the ticket number belongs as business travel type sample data; Form a sample data set with the various ticket data of each passenger; Among them, the performing feature extraction processing on each sample data to obtain various feature data of each sample data includes: For each sample data, obtain the ticket number in the sample data, and determine the various data tables associated with the ticket number; each data table includes a passenger information table, a ticket price table, a ticket itinerary table, and a flight distance static table; Based on each data table, obtain the data parameters of the various data fields required; Process the data parameters of each data field to obtain various feature data.

2. The method according to claim 1, characterized in that, the training of each prediction model using each feature data includes: For each prediction model, input each feature data into the prediction model, and continuously process each feature data based on the data processing rules in the prediction model until the training result of the prediction model meets the preset end condition, and then complete the training of the prediction model.

3. The method according to claim 1, characterized in that, the performing prediction processing on the passengers to be predicted using the target prediction model to obtain the prediction result of the travel purpose of the passengers to be predicted includes: Obtain the ticket data of the passenger to be predicted; Perform feature extraction processing on the ticket data to obtain various feature data to be predicted of the passenger to be predicted; Input each of the feature data to be predicted into the target prediction model, trigger the target prediction model to process the feature data to be predicted, and obtain the prediction result of the travel purpose of the passenger to be predicted.

4. A prediction device for the travel purpose of a passenger, Characterized in that, Comprising: A collection unit, configured to collect passenger data in a preset information system; The information system is a data warehouse in the civil aviation system for storing the travel data of passengers A processing unit, configured to process the passenger data to obtain a sample data set, where the sample data set includes sample data of the business travel type and sample data of the non-business travel type; A feature extraction unit, configured to perform feature extraction processing on each sample data to obtain various feature data of each sample data; A training unit, configured to construct a prediction model corresponding to each prediction algorithm based on various prediction algorithms, and use each feature data to train each prediction model. After each prediction model is trained, obtain the prediction accuracy of each prediction model; A prediction unit, configured to use the prediction model with the highest prediction accuracy as the target prediction model, and use the target prediction model to perform prediction processing on the passenger to be predicted to obtain the prediction result of the travel purpose of the passenger to be predicted; Wherein, the processing unit includes: A first determination subunit, configured to determine each passenger and the identity information of each passenger based on the business travel data in a preset business travel system; the business travel system is a system in which the civil aviation company has signed a cooperation agreement with enterprise users for enterprise employees to book tickets when on business trips; A first acquisition subunit, configured to obtain each ticket data of each passenger within a preset time period from the passenger data based on the identity information of each passenger; A judgment subunit, configured to, for each ticket data of each passenger, obtain the ticket number in the ticket data, and judge whether the ticket number exists in the business travel system. If not, determine the ticket data to which the ticket number belongs as sample data of the non-business travel type. If it exists, determine the ticket data to which the ticket number belongs as sample data of the business travel type; A composition subunit, configured to form the sample data set from each ticket data of each passenger; Wherein, the feature extraction unit includes: A second determination subunit, configured to, for each sample data, obtain the ticket number in the sample data, and determine each data table associated with the ticket number; each data table includes a passenger information table, a ticket price table, a ticket itinerary table, and a flight distance static table; A second acquisition subunit, configured to obtain the data parameters of each required data field based on each data table; A first processing subunit, configured to process the data parameters of each data field to obtain each feature data.

5. A storage medium, Characterized in that, The storage medium includes stored instructions, wherein when the instructions are running, the device where the storage medium is located is controlled to execute the prediction method for the purpose of passenger travel described in any one of claims 1-3.

6. An electronic device, characterized in that it includes a memory, and one or more instructions, wherein the one or more instructions are stored in the memory and are configured to be executed by one or more processors to execute the prediction method for the purpose of passenger travel described in any one of claims 1-3.

Images