Civil aviation ticket information monitoring method and system
By collecting and preprocessing data to train a convolutional neural network, and combining API interfaces to support and store real-time pricing data, real-time fare suggestions are generated, solving the problem of non-real-time fare setting and enabling accurate fare analysis and decision-making.
Patent Information
- Application Number
- CN202510972735.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-15
- Publication Date
- 2025-11-11
AI Technical Summary
Existing technologies make it difficult to analyze ticket prices in real time, resulting in poor pricing and failing to meet the accurate ticketing information needs of airlines and passengers.
By collecting historical external and internal data, preprocessing it, training a convolutional neural network, generating a machine learning model, and obtaining real-time pricing data through an API interface for support and storage, the machine learning model is finally used to generate real-time fare recommendations.
It enables real-time analysis of ticket prices, providing accurate fare recommendations based on dynamic data, improving flight resource allocation and service quality, and meeting passengers' demand for information transparency.
Smart Images
Figure CN120931309A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of fare decision-making technology, specifically to a method and system for monitoring civil aviation ticketing information. Background Technology
[0002] The aviation market is highly competitive, and airlines and ticketing platforms need to obtain more accurate ticketing information to optimize flight resource allocation, formulate reasonable pricing strategies, and improve service quality in order to enhance their competitiveness. At the same time, passengers have increasingly higher demands for the ticketing experience and information transparency, expecting to obtain accurate ticketing information in real time, such as ticket availability and price trends. Currently, due to the large amount of data required for pricing, it is difficult to analyze such a large volume of data, making real-time pricing analysis impossible and resulting in less effective pricing. Summary of the Invention
[0003] The purpose of this invention is to provide a method and system for monitoring civil aviation ticketing information to address the shortcomings in the prior art.
[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for monitoring civil aviation ticketing information, comprising the following steps: The pricing data is based on external and internal ticketing data collected from the Internet. The ticket prices for the corresponding timeline are determined based on the pricing data. The pricing data and corresponding ticket prices are preprocessed and stored in the database. A machine learning model is obtained by training a convolutional neural network using the pricing data and corresponding ticket prices. Real-time pricing data is obtained through API interface, and the real-time pricing data is protected until it is stored in the database. The pricing data is stored in a chain according to the collection time in the database. Based on the database, real-time pricing data is used to generate real-time fare suggestions through machine learning models.
[0005] In a preferred embodiment, the step of preprocessing and storing the pricing data and corresponding fares in a database, and training a machine learning model using the pricing data and corresponding fares to train a convolutional neural network includes: Based on the Internet, historical external and internal data are collected and combined to obtain pricing data. The historical external data includes competitive data, market data and dynamic data, while the historical internal data includes user account information and ticketing data. The pricing data is sorted according to timeline, and the ticket price corresponding to the pricing data is obtained based on the timeline. The pricing data and the corresponding timeline ticket prices are stored in a database. The convolutional neural network is trained using the pricing data and the corresponding timeline ticket prices to obtain a machine learning model.
[0006] In a preferred embodiment, the step of acquiring real-time pricing data based on an API interface, safeguarding the real-time pricing data until it is stored in a database, and storing the pricing data in a chain according to the acquisition time in the database includes: Configure the escort package according to the corresponding API interface, obtain real-time pricing data based on the API interface, put the real-time pricing data into the escort package for escort processing, and obtain escort data; The escort data is transmitted to the database via the API interface until it is transmitted to the database. The real-time pricing data from the escort information is sorted in a chain according to the collection time and stored in the database.
[0007] In a preferred embodiment, the step of configuring a support package for the corresponding API interface, obtaining real-time pricing data based on the API interface, and inputting the real-time pricing data into the support package for support processing to obtain support data includes: Configure the corresponding API interface with escort zones, multiple escort queues, and multiple jump packets as escort packets; Set the acquisition frequency, and obtain real-time pricing data through the API interface according to the acquisition frequency, and put the real-time pricing data into the protection zone; A redirect packet is randomly activated, which includes a first escort zone and a second escort zone. Real-time pricing data is stored in the first escort zone. An escort queue is activated, which consists of multiple escort points connected together. Each escort point includes an escort capsule and a jump point within the escort capsule. The multiple escort points in the escort queue are connected to the first escort zone to safeguard real-time pricing data.
[0008] In a preferred embodiment, the step of connecting multiple escort points in the escort convoy to the first escort zone to escort real-time pricing data includes: Multiple receiving terminals are set up for the first escort zone. The number of receiving terminals is the same as the number of escort points. There is a one-to-one connection between the escort points and the receiving terminals. Establish connections between multiple escort points to form an escort network; Connect multiple jump points in the escort points to the second escort zone.
[0009] In a preferred embodiment, the step of transmitting escort data through the channel between the API interface and the database until it is transmitted to the database includes: During the transmission of escort data through the channel, when an external network accesses the first escort zone, the external network is connected to the escort network through the receiving end; When an external network directly accesses the escort network or connects to the escort network through a receiving terminal, the external network is guided to the escort point through the escort network. When the external network accesses the jump point of the escort point, the external network is guided to the second escort zone through the jump point. The escort data is transmitted to the database, where real-time pricing data is extracted and stored.
[0010] In a preferred embodiment, the step of generating real-time fare suggestions based on real-time pricing data using a machine learning model from a database includes: When new real-time pricing data is stored in the database, the new real-time pricing data is fed into the machine learning model; Real-time fare recommendations are obtained by analyzing new real-time pricing data using machine learning models.
[0011] The present invention also provides a civil aviation ticketing information monitoring system, comprising: The training module is used to collect external and internal ticketing data from the Internet as pricing data, determine the ticket price for the corresponding timeline based on the pricing data, preprocess the pricing data and the corresponding ticket price and store them in the database, and train the convolutional neural network to obtain the machine learning model using the pricing data and the corresponding ticket price. The storage module is used to obtain real-time pricing data based on the API interface, safeguard the real-time pricing data, and store it in the database. In the database, the pricing data is stored in a chain according to the collection time. The fare suggestion module, connected to the training and storage modules, is used to generate real-time fare suggestions based on real-time pricing data from the database using a machine learning model.
[0012] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention acquires a large amount of historical pricing data and then trains a learning model to analyze dynamic pricing data, thereby obtaining real-time ticket price suggestions to assist in ticket price setting and decision-making. Attached Figure Description
[0013] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.
[0014] Figure 1 This is a flowchart of the method of the present invention.
[0015] Figure 2This is a system block diagram of the present invention. Detailed Implementation
[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0017] Example 1, please refer to Figure 1 As shown in the figure, the civil aviation ticketing information monitoring method described in this embodiment includes the following steps: S1. Collect external and internal ticketing data from the Internet as pricing data, determine the ticket price for the corresponding timeline based on the pricing data, preprocess the pricing data and the corresponding ticket price and store them in the database, and train the convolutional neural network with the pricing data and the corresponding ticket price to obtain a machine learning model. S2. Obtain real-time pricing data based on API interface, safeguard the real-time pricing data until it is stored in the database, and store the pricing data in a chain according to the collection time in the database; S3. Based on the database, real-time pricing data is used to generate real-time fare suggestions through machine learning models.
[0018] As described in steps S1-S3 above, airlines and ticketing platforms need to more accurately grasp ticketing information to enhance competitiveness, optimize flight resource allocation, formulate reasonable pricing strategies, and improve service quality. Simultaneously, passengers have increasingly higher demands for ticket purchasing experience and information transparency, expecting to obtain accurate ticketing information in real time, such as ticket availability and price trends. Currently, due to the large amount of data required for pricing, it is difficult to analyze such a large volume, making real-time price analysis impossible and resulting in suboptimal pricing effectiveness. This application, however, acquires a large amount of historical pricing data and then trains a learning model to analyze dynamic pricing data, thereby obtaining real-time pricing suggestions to assist in pricing decisions.
[0019] In one embodiment, step S1, which involves preprocessing and storing pricing data and corresponding fares in a database, and then training a machine learning model using the pricing data and corresponding fares to train a convolutional neural network, includes: S11. Collect historical external and internal data based on the Internet, and combine the historical external and internal data to obtain pricing data. The historical external data includes competitive data, market data and dynamic data, and the historical internal data includes user account information and ticketing data. S12. Sort the pricing data according to the timeline and obtain the ticket price corresponding to the pricing data based on the timeline; S13. Store the pricing data and the corresponding timeline ticket prices in the database, and train the convolutional neural network using the pricing data and the corresponding timeline ticket prices to obtain a machine learning model.
[0020] As described in steps S11-S13 above, the purpose of acquiring pricing data is to train a convolutional neural network. This acquisition process can utilize a combination of web scraping and API calls. For publicly available airline websites and OTA platforms, a web scraping program is written using the Scrapy framework to simulate browser requests and scrape real-time ticket prices, cabin availability, and promotional activities from other airlines and OTA platforms at a set collection frequency (e.g., once every 15 minutes). Simultaneously, for platforms with open API interfaces, Python's Requests library is used to call the interfaces according to the API documentation specifications to obtain data, such as obtaining ticket price fluctuation data for specific routes of competitors over the past year. Market data collection: Through cooperation with market research institutions, structured data files such as holiday / peak season travel trends and regional traffic data are obtained; web scraping technology is used to scrape macroeconomic indicator data, such as fuel prices and exchange rates, from industry information websites and statistical bureau websites. Dynamic data collection: A real-time data monitoring channel is established. For weather data, APIs from platforms like AccuWeather are used to obtain real-time weather conditions at the route's transit points and destinations. Flight punctuality data is obtained through interfaces with the Civil Aviation Administration of China (CAAC). Historical internal data, including user account information and ticketing data, is collected through internal system interfaces: User account information is extracted from the user management system database, including user registration information (such as name, age, and contact information), frequent flyer data (such as flight mileage and membership level), and ticketing preferences (such as preferred cabin class and travel time). An ETL tool (such as Apache NiFi) is used to establish a data extraction pipeline. User account information is extracted from the source database at a fixed time interval of 2 AM daily, converted to a unified format, and loaded into the database. It is stored chronologically according to historical pricing data and corresponding ticket prices, allowing us to understand the relationship between pricing data and ticket prices over time. The collected historical external and internal data underwent cleaning and preprocessing: Missing ticket price data in the competitive data was filled using time-series-based linear interpolation; anomalous traffic data (such as values significantly deviating from the normal range) in the market data were identified and removed using the 3σ principle; and incorrectly formatted email addresses and phone numbers in user account information were corrected using regular expression matching. Data standardization involved converting data of different formats and units to a unified standard. For ticket price data, a unified currency unit was required; for date and time data, it was uniformly converted to the ISO 8601 standard format; and for numerical data such as user age and flight mileage, the Z-score standardization method was used to map the data to a distribution with a mean of 0 and a standard deviation of 1.The convolutional neural network (CNN) was trained and optimized using pricing data and corresponding timeline ticket prices. Historical data was divided into training and test sets in an 8:2 ratio; the training set was used for model training, and the test set for evaluating model performance. Mean squared error (MSE) was used as the loss function to measure the error between predicted and actual ticket prices, with the goal of minimizing the loss function value. The Adam optimization algorithm was selected to update model parameters, automatically adjusting the learning rate based on changes in the loss function during training to improve training efficiency and convergence speed. Model evaluation and tuning: During training, the model performance was periodically evaluated using the test set, calculating metrics such as root mean squared error (RMSE) and mean absolute error (MAE). Based on the evaluation results, the hyperparameters of the CNN (such as the number of convolutional kernels, the number of neurons in fully connected layers, and the learning rate) were adjusted. The optimal hyperparameter combination was found through cross-validation, resulting in a high-performance machine learning model. This model, capable of training the CNN using historical data, facilitates subsequent airline ticket pricing and demonstrates good market adaptability.
[0021] In one embodiment, step S2, which involves acquiring real-time pricing data via an API interface, safeguarding the real-time pricing data until it is stored in a database, and performing chained storage of pricing data in the database according to the acquisition time, includes: S21. Configure the escort package according to the corresponding API interface, obtain real-time pricing data based on the API interface, put the real-time pricing data into the escort package for escort processing, and obtain escort data; S22. Transmit the escort data to the database via the API interface until it is transmitted to the database; S23. Store the real-time pricing data in the escort information in the database in a chained order according to the collection time.
[0022] In one embodiment, the step S21, which involves configuring a support package using the corresponding API interface, obtaining real-time pricing data based on the API interface, and inputting the real-time pricing data into the support package for support processing to obtain support data, includes: S211, corresponding API interface configuration escort zone (cloud server), multiple escort queues and multiple jump packets as escort packets; S212. Set the acquisition frequency, acquire real-time pricing data through the API interface according to the acquisition frequency, and put the real-time pricing data into the protection zone; S213. Randomly activate a jump packet, wherein the jump packet includes a first escort zone and a second escort zone (both the first escort zone and the second escort zone are data storage spaces within the escort zone), and store the real-time pricing data in the first escort zone; S214. Activate an escort queue, wherein the escort queue consists of multiple escort points (escort points are virtual machines) connected together. The escort points include escort capsules (escort capsules are storage spaces in the virtual machine) and jump points in the escort capsules. Connect the multiple escort points in the escort queue to the first escort zone to escort the real-time pricing data.
[0023] In one embodiment, step S214, which connects multiple escort points in the escort queue to the first escort zone to escort real-time pricing data, includes: S2141. Multiple receiving terminals (connecting ports) are set up for the first escort zone. The number of receiving terminals is the same as the number of escort points. There is a one-to-one connection between the escort points and the receiving terminals. S2142. Establish connections between multiple escort points to form an escort network; S2143. Connect multiple jump points in the escort point (the jump point is a virtual machine in the escort point, the escort point is an external virtual machine, and the jump point is a virtual machine inside the external virtual machine) to the second escort zone.
[0024] In one embodiment, step S22, which involves transmitting escort data through an API interface to a database until the data is transmitted to the database, includes: S221. During the transmission of escort data through the channel, when an external network accesses the first escort zone, the external network is connected to the escort network through the receiving end. S222. When an external network directly accesses the escort network or connects to the escort network through a receiving terminal, the external network is guided to the escort point through the escort network. When the external network accesses the jump point of the escort point, the external network is guided to the second escort zone through the jump point. S223, until the escort data is transmitted to the database, the real-time pricing data is extracted from the escort data and stored in the database.
[0025] As described in steps S21-S23 above, a protection package is configured for the corresponding API interface. After the API interface obtains real-time pricing data, the real-time pricing data is put into the protection zone for processing. Within the protection zone, there are multiple protection queues and multiple redirect packets. The redirect packets include a first protection zone and a second protection zone, both of which are data storage spaces within the protection zone. One redirect packet corresponds to one protection queue, and one protection queue includes multiple protection points. All protection queues and redirect packets are kept in reserve within the protection zone. A single real-time pricing data collection, based on the acquisition frequency (which can be once per minute or once every few seconds, acquiring the pricing data at the current moment), activates one protection queue. The system also includes a redirect packet that stores the collected real-time pricing data in the first escort zone. This packet connects multiple escort points in the escort queue to multiple receiving terminals within the first escort zone. The receiving terminals are pre-installed connection ports in the first escort zone, and their number is the same as the number of escort points. Once a receiving terminal is connected to an escort point, if an external network subsequently attempts to access the first escort zone to obtain or tamper with the real-time pricing data, two scenarios will occur. First, if the external network accesses a receiving terminal in the first escort zone, bypassing the escort network, the receiving terminal, being connected to the escort points, will redirect the external network to the connection point. The first scenario involves two connections: First, external networks accessing the escort network directly without bypassing it (the escort network is a network connection deployed between multiple escort points outside the first escort zone; when an external network accesses it, it is highly likely to be affected by the escort network and directly access the network deployment within the escort network, thus directly accessing the escort network itself, which can guide the external network to the escort point). In both scenarios, the final result is that the external network is handled by the escort point. When the external network's access attack is intense, it will continuously access and deeply access the escort point, thus attacking the redirection points within the escort point. The first protection zone will be activated when an external network is accessed, redirecting it to the second protection zone. This second protection zone, along with the protection network, protects the first protection zone and its pricing data. If the access is simply to a protection point, the redirection point will not be activated. The redirection point redirects the external network to the second protection zone to prevent aggressive network attacks from breaching the protection point and accessing the first protection zone. The protection queue and the second protection zone provide network protection for the first protection zone, ensuring the accuracy of the acquired pricing data and preventing errors in analysis results due to inaccurate pricing data, thus guaranteeing the accuracy of fare decisions.
[0026] In one embodiment, step S3, which generates real-time fare suggestions based on real-time pricing data from a database using a machine learning model, includes: S31. When new real-time pricing data is stored in the database, the new real-time pricing data is input into the machine learning model; S32. Analyze the new real-time pricing data using a machine learning model to obtain real-time fare suggestions.
[0027] As described in steps S31 and S32 above, the new real-time pricing data here refers to real-time pricing data acquired at a specific frequency, for example, every ten minutes. This real-time pricing data includes both external and internal data. Each acquired real-time pricing data is used as the new real-time pricing data. When the database stores new real-time pricing data, it is input into the machine learning model. This allows the machine learning model to analyze the new real-time pricing data and output a real-time ticket price suggestion. For example, if the ticket price was 500 yuan at the previous acquisition time, and due to market changes and changes in internal data at the current time, the machine learning model can analyze the data to obtain a current ticket price suggestion, such as 450 yuan. This provides accurate analysis results, enabling airline ticket prices to be dynamically adjusted to adapt to both external and internal market conditions.
[0028] Example 2, please refer to Figure 2 As shown in this embodiment, a civil aviation ticketing information monitoring system includes: The training module is used to collect external and internal ticketing data from the Internet as pricing data, determine the ticket price for the corresponding timeline based on the pricing data, preprocess the pricing data and the corresponding ticket price and store them in the database, and train the convolutional neural network to obtain the machine learning model using the pricing data and the corresponding ticket price. The storage module is used to obtain real-time pricing data based on the API interface, safeguard the real-time pricing data, and store it in the database. In the database, the pricing data is stored in a chain according to the collection time. The fare suggestion module, connected to the training and storage modules, is used to generate real-time fare suggestions based on real-time pricing data from the database using a machine learning model.
[0029] It should be noted that by acquiring a large amount of historical pricing data and then training the learning model, it is possible to analyze dynamic pricing data and obtain real-time ticket price suggestions, which can assist in ticket price setting and decision-making.
[0030] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A method for monitoring civil aviation ticketing information, characterized in that, Includes the following steps: The pricing data is based on external and internal ticketing data collected from the Internet. The ticket prices for the corresponding timeline are determined based on the pricing data. The pricing data and corresponding ticket prices are preprocessed and stored in the database. A machine learning model is obtained by training a convolutional neural network using the pricing data and corresponding ticket prices. Real-time pricing data is obtained through API interface, and the real-time pricing data is protected until it is stored in the database. The pricing data is stored in a chain according to the collection time in the database. Based on the database, real-time pricing data is used to generate real-time fare suggestions through machine learning models.
2. The civil aviation ticketing information monitoring method according to claim 1, characterized in that, The steps of preprocessing pricing data and corresponding ticket prices and storing them in a database, and then training a machine learning model using the pricing data and corresponding ticket prices to train a convolutional neural network include: Based on the Internet, historical external and internal data are collected and combined to obtain pricing data. The historical external data includes competitive data, market data, and dynamic data, while the historical internal data includes user account information and ticketing data. The pricing data is sorted according to timeline, and the ticket price corresponding to the pricing data is obtained based on the timeline. The pricing data and the corresponding timeline ticket prices are stored in a database. The convolutional neural network is trained using the pricing data and the corresponding timeline ticket prices to obtain a machine learning model.
3. The civil aviation ticketing information monitoring method according to claim 1, characterized in that, The steps of acquiring real-time pricing data based on the API interface, safeguarding the real-time pricing data until it is stored in the database, and storing the pricing data in a chain according to the acquisition time in the database include: Configure the escort package according to the corresponding API interface, obtain real-time pricing data based on the API interface, put the real-time pricing data into the escort package for escort processing, and obtain escort data; The escort data is transmitted to the database via the API interface until it is transmitted to the database. The real-time pricing data from the escort information is sorted in a chain according to the collection time and stored in the database.
4. The civil aviation ticketing information monitoring method according to claim 3, characterized in that, The steps of configuring the corresponding API interface escort package, obtaining real-time pricing data based on the API interface, and inputting the real-time pricing data into the escort package for escort processing to obtain escort data include: Configure the corresponding API interface with escort zones, multiple escort queues, and multiple jump packets as escort packets; Set the acquisition frequency, and obtain real-time pricing data through the API interface according to the acquisition frequency, and put the real-time pricing data into the protection zone; A redirect packet is randomly activated, which includes a first escort zone and a second escort zone. Real-time pricing data is stored in the first escort zone. An escort queue is activated, which consists of multiple escort points connected together. Each escort point includes an escort capsule and a jump point within the escort capsule. The multiple escort points in the escort queue are connected to the first escort zone to safeguard real-time pricing data.
5. The civil aviation ticketing information monitoring method according to claim 4, characterized in that, The step of connecting multiple escort points in the escort queue to the first escort zone to safeguard real-time pricing data includes: Multiple receiving terminals are set up for the first escort zone. The number of receiving terminals is the same as the number of escort points. There is a one-to-one connection between the escort points and the receiving terminals. Establish connections between multiple escort points to form an escort network; Connect multiple jump points in the escort points to the second escort zone.
6. The civil aviation ticketing information monitoring method according to claim 5, characterized in that, The step of transmitting escort data through the API interface to the database, until the data is transmitted to the database, includes: During the transmission of escort data through the channel, when an external network accesses the first escort zone, the external network is connected to the escort network through the receiving end; When an external network directly accesses the escort network or connects to the escort network through a receiving terminal, the external network is guided to the escort point through the escort network. When the external network accesses the jump point of the escort point, the external network is guided to the second escort zone through the jump point. The escort data is transmitted to the database, where real-time pricing data is extracted and stored.
7. The civil aviation ticketing information monitoring method according to claim 6, characterized in that, The step of generating real-time fare suggestions based on real-time pricing data using a machine learning model from a database includes: When new real-time pricing data is stored in the database, the new real-time pricing data is fed into the machine learning model; Real-time fare recommendations are obtained by analyzing new real-time pricing data using machine learning models.
8. A civil aviation ticketing information monitoring system, used to implement the civil aviation ticketing information monitoring method according to any one of claims 1-7, characterized in that, include: The training module is used to collect external and internal ticketing data from the Internet as pricing data, determine the ticket price for the corresponding timeline based on the pricing data, preprocess the pricing data and the corresponding ticket price and store them in the database, and train the convolutional neural network to obtain the machine learning model using the pricing data and the corresponding ticket price. The storage module is used to obtain real-time pricing data based on the API interface, safeguard the real-time pricing data, and store it in the database. In the database, the pricing data is stored in a chain according to the collection time. The fare suggestion module, connected to the training and storage modules, is used to generate real-time fare suggestions based on real-time pricing data from the database using a machine learning model.