Cold and hot air route classification method and system, electronic equipment and storage medium
By collecting and processing route data in real time, using the K-mean clustering algorithm to automatically discover similarities and differences between routes, the precise classification of hot and hot routes is achieved, and the problem of lack of real-time and dynamics of hot and hot routes in the existing technology is solved, and the operation and management efficiency of airlines is improved.
Patent Information
- Application Number
- CN202411887646.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-19
- Publication Date
- 2025-04-22
AI Technical Summary
It is difficult for the existing technology to accurately distinguish hot and hot routes, lack real-time and dynamic adjustment capabilities, and it is difficult to meet changes in market demand.
By collecting the original data of route flights and passengers in real time, performing statistics, summary and abnormal detection, and building a complete data set. Then, based on the K-mean clustering algorithm, the feature data is divided into different clusters, and the popularity and characteristics of each cluster are evaluated through statistical features and visual analysis to achieve the division of hot and hot routes.
It has realized the automation and accurate classification of hot and hot routes, improved the operation and management efficiency of airlines, and ensured the comprehensiveness and accuracy of classification results.
Smart Images
Figure CN119862441A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of civil aviation air transport management, and in particular relates to a hot and cold air route classification method, system, electronic equipment and storage medium. Background Art
[0002] In the field of air transportation, routes can be divided into hot and cold routes based on factors such as demand, passenger volume and profit. Hot routes refer to those routes with strong demand, large passenger flow and high profits. They often connect popular tourist destinations or commercial centers. In contrast, cold routes refer to routes with relatively low demand, limited passenger flow and low profits. They may serve remote areas or secondary cities. Airlines will formulate corresponding flight schedules, pricing strategies, marketing activities and service levels based on the popularity of the route. For hot routes, airlines often increase flight frequency, provide more options and services, and carry out promotional marketing activities; for cold routes, airlines may adjust flight schedules, flexibly price, and take other measures to increase the attractiveness of the route.
[0003] Therefore, the classification of hot and cold routes has important practical significance for both airlines and travelers; it helps both parties to have a deeper understanding of the market demand and development trends of routes, so as to make more reasonable business and travel decisions. In addition, the classification of hot and cold routes also helps airlines to conduct data analysis, gain an in-depth understanding of market demand, passenger travel habits and trends, and thus provide strong data support for route planning and operational decisions; at the same time, it can also help travelers better choose flights and plan travel routes, understand which routes may provide more service options, and which routes may require advance planning and preparation.
[0004] At present, how to accurately distinguish between hot and cold routes is of great significance for rationally allocating resources, optimizing flight arrangements, and improving passenger satisfaction. However, most traditional methods rely on historical data analysis and manual experience for judgment, lacking real-time and dynamic adjustment capabilities, and are difficult to meet changes in market demand. Therefore, it is urgent to develop a method that can accurately classify hot and cold routes in real time. Summary of the invention
[0005] In view of the shortcomings of the prior art, the present invention provides a method for classifying hot and cold routes, as well as a system, electronic equipment and storage medium for implementing the method, and improves the efficiency and accuracy of route management through intelligent means. The specific technical scheme is as follows:
[0006] In one aspect of the present invention, a method for classifying hot and cold routes is provided, the method comprising:
[0007] Step 1: Collect raw data of routes, flights and passengers in real time, perform statistics, aggregation and anomaly detection on the collected raw data, process missing values and outliers in the raw data, and build a complete and comprehensive data set;
[0008] Step 2: Convert, extract and select features based on the statistically summarized data in the data set to obtain feature data, and input the feature data into the feature training model for training;
[0009] Step 3: Input the trained feature data into the K-means clustering algorithm for clustering, and set the value of K to obtain K different clusters;
[0010] Step 4: Through statistical characteristics and visual analysis of the K different clusters obtained through cluster analysis, comprehensively evaluate the popularity and characteristics of each cluster to achieve the division of hot and cold routes.
[0011] Furthermore, in step 2, the data statistically summarized in the data set is converted, feature extracted and selected to obtain feature data, which also includes the following steps:
[0012] Step 201: Filtering feature data with a high correlation with hot and cold routes from the statistically summarized data by means of dimensionality reduction;
[0013] Step 202: Clarify the selected feature data to ensure the quality and consistency of the data;
[0014] Step 203: transform and map the cleaned feature data, and uniformly encode the discrete feature data to ensure that the proportions between different features are consistent.
[0015] Furthermore, in step 4, the K different clusters obtained through cluster analysis are analyzed by statistical features and visualization, specifically:
[0016] First, one or two popular and unpopular routes are manually annotated as examples;
[0017] Then, the labeled examples are trained together with all the clusters for classification;
[0018] Finally, all clusters are labeled and classified according to the affiliation of the labeled examples.
[0019] Another aspect of the present invention, based on the same inventive concept as the above disclosed content, provides a hot and cold air route classification system, the system comprising a data collection and sorting module, a feature engineering module, a cluster analysis module and a group analysis and division module, wherein:
[0020] The data collection and collation module is used to collect the original data of routes, flights and passengers from various systems in real time, perform statistics, aggregation and anomaly detection on the collected original data, process missing values and anomalies in the original data, and build a complete and comprehensive data set;
[0021] The feature engineering module is used to convert, extract and select features from the statistically summarized data in the data set to obtain feature data, and input the feature data into the feature training model for training;
[0022] The cluster analysis module is used to input the trained feature data into the K-means clustering algorithm for clustering, and set the value of K to obtain K different clusters;
[0023] The group analysis and division module is used to comprehensively evaluate the popularity and characteristics of each cluster through statistical characteristics and visual analysis of the K different clusters obtained through cluster analysis, and to realize the division of hot and cold routes.
[0024] In another aspect of the present invention, based on the same inventive concept as the above-disclosed contents, there is provided an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes all steps of the method for classifying hot and cold routes as described above.
[0025] In another aspect of the present invention, a computer storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, it is used to implement all the steps of the method for classifying hot and cold routes as described above.
[0026] Compared with the prior art, the present invention has the following advantages:
[0027] The present invention comprehensively considers multiple factors and automatically discovers the similarities and differences between routes through the K-means clustering algorithm, providing visual analysis results for optimizing operational strategies, ensuring the comprehensiveness and accuracy of classification results, facilitating management personnel to monitor and analyze route conditions, and realizing automatic and accurate classification of hot and cold routes, thereby improving the operational management efficiency of airlines.
[0028] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0030] Figure 1 It is a flow chart of a method for classifying hot and cold routes.
[0031] Figure 2 It is a structural diagram of a hot and cold route classification system. DETAILED DESCRIPTION
[0032] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0033] Embodiment 1:
[0034] In this embodiment, a method for classifying hot and cold routes is provided. Figure 1 , Figure 1 The present invention is a flowchart of a method for classifying hot and cold routes, the method comprising the following steps:
[0035] Step S1: collect raw data of routes, flights and passengers in real time, perform statistics, aggregation and anomaly detection on the collected raw data, process missing values and outliers in the raw data, and build a complete and comprehensive data set;
[0036] Furthermore, the original data of the routes, flights and passengers include the aircraft models, flight numbers, flight inquiries, actual transportation volume, trip revenue, ticket prices, route passenger flow, whether the starting point and end point are transportation hubs, the throughput of the city airport, actual passenger flow and passenger travel demand and other original data.
[0037] Furthermore, the above-mentioned statistics, aggregation and anomaly detection of the collected raw data, processing of missing values and outliers in the raw data, and construction of a complete and comprehensive data set also include:
[0038] Based on the collected original data of routes, flights and passengers, through statistical analysis, the data are sorted and divided into two date dimensions: weekdays and holidays, and the characteristics and patterns of passenger behavior are statistically summarized.
[0039] Further, in one embodiment, taking the departure-arrival data in Table 1 as an example, the data is divided according to the two date dimensions of weekdays and holidays, and the characteristics and rules of passenger behavior are statistically summarized. The sorted data is shown in Table 1: The original data table of the date dimension after statistical summary:
[0040] Table 1: The original data table of date dimension after statistical aggregation
[0041]
[0042] Step S2: converting, extracting and selecting features based on the statistically summarized data in the data set to obtain feature data, and inputting the feature data into a feature training model for training;
[0043] Furthermore, the method of converting, extracting and selecting features based on the statistically summarized data in the data set to obtain feature data further includes the following steps:
[0044] Step S201: filtering feature data with a high correlation with hot and cold routes from the statistically summarized data by means of dimensionality reduction;
[0045] Furthermore, the screening is performed on characteristic data with a high correlation with hot and cold routes, specifically, characteristic data such as the actual civil aviation transport volume, itinerary revenue, ticket price, passenger travel demand, and airport passenger flow at the starting and ending points of the route.
[0046] Step S202: clarify the selected feature data to ensure the quality and consistency of the data;
[0047] Furthermore, the screened feature data is clarified, including filling in missing data, removing abnormal data, and identifying and processing duplicate data.
[0048] Step S203: transform and map the cleaned feature data, and uniformly encode the discrete feature data to ensure that the proportions between different features are consistent.
[0049] Furthermore, the cleaned feature data is transformed and mapped, and the discrete feature data is uniformly encoded, specifically:
[0050] Use scikit-learn's regularization function to improve the distribution and representation of feature data by standardizing, normalizing, and logarithmic transformation, thereby achieving data scaling, which can make the feature training model converge faster and improve performance;
[0051] Onehot encoding is used to uniformly encode discrete feature data.
[0052] Furthermore, in one embodiment, according to the above step S2, the feature data in Table 1 is subjected to feature screening and sorting, and the following Table 2 is obtained: a date dimension feature data table after sorting and screening.
[0053] Table 2: Date dimension feature data table after sorting and filtering
[0054]
[0055] Step S3: input the trained feature data into the K-means clustering algorithm for clustering, and set the value of K to obtain K different clusters;
[0056] Furthermore, the trained feature data is input into a K-means clustering algorithm for clustering. Specifically, the K-means clustering algorithm automatically discovers patterns and structures in the data, and divides the feature data into K different clusters, i.e., K route groups, according to the similarities between routes, each cluster representing a group of routes with similar features;
[0057] Furthermore, in one embodiment, taking the data in Table 1 and Table 2 in the above text as an example, the original data of the date dimension after sorting and selection in Table 2 in the above text is input into the K-means clustering algorithm for clustering. Since the date dimension consists of two dimensions, namely weekdays and holidays, the value of K is set to 2. The K-means clustering algorithm will automatically discover the patterns and structures in the data, and divide the feature data into two different clusters according to the similarity between the routes. The same clusters are represented by the same numbers. The specific clustering results are shown in Table 3: Clustering result analysis table.
[0058] Table 3: Clustering results analysis table
[0059]
[0060] Step S4: Through statistical characteristics and visual analysis of the K different clusters obtained through cluster analysis, comprehensively evaluate the popularity and characteristics of each cluster, and divide the hot and cold routes.
[0061] Furthermore, the K different clusters obtained through cluster analysis are analyzed through statistical features and visualization, specifically:
[0062] First, one or two popular and unpopular routes are manually annotated as examples;
[0063] Then, the labeled examples are trained together with all the clusters for classification;
[0064] Finally, all clusters are labeled and classified according to the affiliation of the labeled examples.
[0065] Furthermore, the comprehensive evaluation of the popularity and characteristics of each cluster is to evaluate the popularity and characteristics of each cluster by counting key indicator characteristics such as the income, passenger demand, and average ticket price of each cluster; these key indicator characteristics can help us quickly understand the basic attributes of each cluster and provide basic data for subsequent analysis.
[0066] Further, in one embodiment, taking Table 3: Clustering result analysis table above as an example: the route from Beijing to Shanghai is marked as hot route sample 1, and the route from Qingdao to Kashgar is marked as cold route sample 2, and training classification is performed together with all data; the cluster where hot route sample 1 is located is considered to be a hot route set, and the cluster where cold route sample 2 is located is considered to be a cold route set. The specific result division is shown in Table 4: Hot and cold route classification result division table.
[0067] Table 4: Classification results of hot and cold routes
[0068]
[0069] Embodiment 2:
[0070] In this embodiment, a hot and cold route classification system is provided. Figure 2 , Figure 2 The present invention is a structural diagram of a hot and cold route classification system, which includes: a data collection and sorting module, a feature engineering module, a cluster analysis module, and a group analysis and division module, wherein:
[0071] The data collection and collation module is used to collect the original data of routes, flights and passengers from various systems in real time, perform statistics, aggregation and anomaly detection on the collected original data, process missing values and anomalies in the original data, and build a complete and comprehensive data set;
[0072] The feature engineering module is used to convert, extract and select features from the statistically summarized data in the data set to obtain feature data, and input the feature data into the feature training model for training;
[0073] The cluster analysis module is used to input the trained feature data into the K-means clustering algorithm for clustering, and set the value of K to obtain K different clusters;
[0074] The group analysis and division module is used to comprehensively evaluate the popularity and characteristics of each cluster through statistical characteristics and visual analysis of the K different clusters obtained through cluster analysis, and to realize the division of hot and cold routes.
[0075] Embodiment three:
[0076] In this embodiment, an electronic device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes all steps of the hot and cold air route classification method as described above.
[0077] Embodiment 4:
[0078] In this embodiment, a computer storage medium is further provided, on which a computer program is stored. When the computer program is executed by the processor, it is used to implement all the steps of the hot and cold air routes classification method as described above.
[0079] In addition, in the description of the present invention, the explanations for the terms are:
[0080] Unsupervised learning: Unsupervised learning is a method in the field of machine learning. Its main feature is that it does not require labeled target variables when training the model. In unsupervised learning, the model attempts to discover hidden patterns, structures, and relationships from the data rather than predicting or classifying specific outputs;
[0081] Clustering: Clustering algorithms can help identify patterns and clusters in data by dividing the data into different groups or clusters so that the data points within each group are similar to each other, but the data points between different groups are different.
[0082] Dimensionality reduction: By retaining the most important information in the data and mapping the data into a lower dimensional space, dimensionality reduction techniques help reduce the complexity and noise of the data while maintaining the structure and characteristics of the data;
[0083] Cluster: The process or result of grouping samples in a data set according to their similarities. It can be understood as a collection of tightly clustered points in the data space, which are more similar to each other under some metric. In clustering algorithms, our goal is to divide the samples in the data set into different clusters so that the similarity between samples in the same cluster is high, while the similarity between samples in different clusters is low.
[0084] Regularization: By penalizing the complexity of the model, the model can be made to focus more on the important features in the training data;
[0085] One-hot encoding: Encodes data to ensure that the distance between data is equal and there is no order between adjacent information.
[0086] Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent substitutions for some of the technical features therein; and these modifications or substitutions do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for classifying hot and cold routes, characterized by: Collect raw data of routes, flights and passengers in real time, perform statistics, aggregation and anomaly detection on the collected raw data, process missing values and outliers in the raw data, and build a complete and comprehensive data set; Based on the statistically summarized data in the data set, conversion, feature extraction and selection are performed to obtain feature data, and the feature data is input into the feature training model for training; Input the trained feature data into the K-means clustering algorithm for clustering, and set the value of K to obtain K different clusters; Through statistical characteristics and visual analysis of the K different clusters obtained through cluster analysis, the popularity and characteristics of each cluster are comprehensively evaluated to achieve the division of hot and cold routes.
2. A method for classifying hot and cold routes according to claim 1, characterized in that: The original data of the routes, flights and passengers, including the aircraft type, flight number, flight enquiries, actual transport volume, itinerary revenue, ticket prices, route passenger flow, whether the starting point and end point are transportation hubs, the throughput of the city airport, actual passenger flow and passenger travel demand; The statistics, aggregation and anomaly detection of the collected raw data also include: sorting and dividing according to the two date dimensions of weekdays and holidays, and statistically summarizing the characteristics and patterns of passenger behavior.
3. A method for classifying hot and cold routes as claimed in claim 1, characterized in that: The method converts, extracts and selects the statistically summarized data in the data set to obtain feature data, and further includes: By reducing the dimension, we can select the characteristic data with high correlation with hot and cold routes from the statistical data. Clarify the selected feature data to ensure data quality and consistency; The cleaned feature data is transformed and mapped, and the discrete feature data is uniformly encoded to ensure the consistency of the proportions between different features.
4. A method for classifying hot and cold routes as claimed in claim 3, characterized in that: The screening of characteristic data with a high correlation with hot and cold routes is specifically characteristic data such as actual civil aviation transport volume, itinerary revenue, ticket price, passenger travel demand, passenger flow at the starting and ending airports, etc. The screening feature data is clarified, including filling missing data, removing abnormal data, identifying and processing duplicate data; The transformation and mapping of the cleaned feature data is specifically: Use scikit-learn's regularization function to improve the distribution and representation of feature data through standardization, normalization, and logarithmic transformation, thereby achieving data scaling. Onehot encoding is used to uniformly encode discrete feature data.
5. A method for classifying hot and cold routes as claimed in claim 1, characterized in that: The step of inputting the trained feature data into a K-means clustering algorithm for clustering includes: The K-means clustering algorithm is based on which patterns and structures in the data are automatically discovered, and the characteristic data are divided into K different clusters, i.e. K route groups, according to the similarities between the routes, each cluster representing a group of routes with similar characteristics.
6. A method for classifying hot and cold routes as claimed in claim 1, characterized in that: The K different clusters obtained through cluster analysis are subjected to statistical features and visual analysis to comprehensively evaluate the popularity and characteristics of each cluster, thereby realizing the division of hot and cold routes, and further includes: Manually mark one or two popular and unpopular routes as examples; Perform classification training on the labeled examples together with all clusters; All clusters are labeled and classified according to the affiliation of the labeled samples.
7. A method for classifying hot and cold routes as claimed in claim 6, characterized in that: The comprehensive evaluation of the popularity and characteristics of each cluster is performed by counting key indicator characteristics such as the revenue, passenger demand, and average ticket price of each cluster to evaluate the popularity and characteristics of each cluster.
8. A hot and cold air route classification system, characterized in that: The system includes a data collection and sorting module, a feature engineering module, a clustering analysis module, and a group analysis and division module, wherein: The data collection and collation module is used to collect the original data of routes, flights and passengers from various systems in real time, perform statistics, aggregation and anomaly detection on the collected original data, process missing values and anomalies in the original data, and build a complete and comprehensive data set; The feature engineering module is used to convert, extract and select features from the statistically summarized data in the data set to obtain feature data, and input the feature data into the feature training model for training; The cluster analysis module is used to input the trained feature data into the K-means clustering algorithm for clustering, and set the value of K to obtain K different clusters; The group analysis and division module is used to comprehensively evaluate the popularity and characteristics of each cluster through statistical characteristics and visual analysis of the K different clusters obtained through cluster analysis, and to realize the division of hot and cold routes.
9. An electronic device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the computer program is executed by the processor, the processor executes the hot and cold air routes classification method according to any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the hot and cold air routes classification method according to any one of claims 1 to 7 is implemented.