A government enterprise travel civil aviation big data trip intelligent dynamic planning method and system

Abstract

The application discloses a kind of government enterprise travel civil aviation big data's itinerary wisdom dynamic planning method and system, first according to the demand of first user, initial travel route planning based on price, hobby, travel time, travel type and the like content is carried out, and initial travel path and civil aviation route are recommended;Meanwhile, the second user feature is obtained, and the initial travel path and civil aviation route are recommended;The route similarity of first user and second user is calculated, and part of the route is merged into the same route according to the similarity;The time when first user is disturbed by fan users and forced to communicate is detected, and the travel type is recorded;The travel log or social media dynamic published by user on social media is detected;Destination planning is carried out again, and the route is adjusted, and travel surrounding service is recommended according to the route;According to the demand of privacy protection, the information content of travel log on social media is automatically adjusted, and the goal of protecting the route privacy of government enterprise high-level user is reached.

Description

Technical Field

[0001] This invention relates to the field of intelligent device technology, and in particular to a method and system for intelligent dynamic planning of travel itineraries based on big data from government, enterprise, business travel, and civil aviation. Background Technology

[0002] Business travel often requires multiple companions or large-scale group interactions to achieve optimal results. However, many high-ranking government and corporate executives, as well as certain celebrities, require privacy protection for their travel information. For example, many news figures, celebrities, government and corporate officials, and the ever-emerging livestreaming influencers on the internet all need public recognition but sometimes desire a private life free from disturbance. Especially when multiple people are involved and privacy needs to be protected, a mechanism is needed to meet the travel and business needs of these special customers while also satisfying their privacy requirements. On the other hand, these privacy-conscious customers may not be traveling alone; they may be accompanied by family or friends. These companions also have their own needs, their privacy is not sensitive, and they will not be frequently disturbed. Therefore, automated route planning should automatically adjust different flight routes for these companions to ensure they can participate in shared activities and avoid inappropriate participation. During business trips, it's common to take photos or post them on social media platforms like WeChat Moments and Weibo. This is normal for ordinary people, but for influencers or celebrities, it's crucial to be mindful of whether their posts infringe on their location privacy and cause disruption to their lives. Therefore, route planning apps not only need to recommend relevant travel destinations but also need to have different restrictions and masking capabilities for different travel routes and flights. Summary of the Invention

[0003] This invention provides a method for intelligent dynamic planning of travel itineraries based on big data from government and enterprise business travel and civil aviation, mainly including:

[0004] Based on the needs of the first user, initial travel route planning is performed based on factors such as price, preferences, travel time, and travel type, recommending initial travel routes and civil aviation routes; simultaneously, the characteristics of the second user are obtained, and initial travel routes and civil aviation routes are recommended; the first user is a celebrity user with privacy protection needs, and the second user refers to the first user's family or partners; the similarity between the passenger's and partner's routes is calculated, and the overlapping parts of the passenger's and partner's routes are merged into the same route based on the similarity;

[0005] Based on the initial travel route and past routes, during the travel process, the system detects the time when the first user is disturbed and forced to communicate by fans, and records the travel type; further, it detects the user's travel logs or social media updates posted on social media.

[0006] Based on the probability that the first user may be recognized and disturbed, the destination planning is re-engineered. Travel projects and routes that are highly similar to those of the first and second users but are easily disturbed are adjusted, and flight routes are also adjusted. Based on privacy protection needs, the information on mobile phone location and travel logs on social media is automatically adjusted to protect the user's flight route privacy.

[0007] Further, optionally, the initial travel route planning based on factors such as price, preferences, travel time, and travel type includes:

[0008] Based on users' travel booking needs, we conduct initial travel route planning based on factors such as price, preferences, travel time, and travel type, and recommend relevant civil aviation routes based on big data.

[0009] Collect user information provided by users when purchasing airline tickets, collect it after obtaining the user's authorization and consent, and build a big data database of user information, including: gender, age, consumption level, and historical travel destination data;

[0010] Establishing a user classification model includes: acquiring data on gender, age, consumption level, place of residence, and historical travel destinations from the user's big data information as a training set, and using preset flight routes and scenic spot routes suitable for user information as annotation values. The annotation content mainly includes establishing a user feature-scenic spot type mapping table, including: statistically analyzing the user characteristics of tourists received by each type of scenic spot nationwide, and identifying user features that appear more than a preset value as the best scenic spots for that type, used to construct the initial annotations for the scenic spots;

[0011] A flight route classification model is trained using a random forest algorithm. User information data is input into the model to obtain routes and destinations corresponding to a specific user. Based on the route details, user profiles, and the search origin and destination, as well as travel time, relevant air routes are recommended, and fares are calculated.

[0012] Further optionally, the step of calculating the route similarity between the first user and the second user, and merging some routes into identical routes based on the similarity, includes:

[0013] First, a relationship prediction model is established to predict the relationships of users who cannot be matched as relatives. Specifically, this includes: obtaining information about users who purchased airline tickets based on the user information big data and extracting features, including name, gender, and age; collecting user feature data with confirmed relationships as a training set, including relationships such as colleagues, lovers, classmates, and unknown; classifying the training set using the SVM algorithm to obtain the relationship prediction model; predicting the pairwise relationships between second users under the same ticket purchase ID using the relationship prediction model; and predicting the relationship between second users under the same ticket purchase ID based on the same ID used when purchasing tickets.

[0014] Furthermore, based on the gender, age, consumption level, place of residence, and historical travel data of the first and second users, and the geographical location of the recommended travel attractions, the various attractions are connected geographically, routes are drawn, and similarity is calculated based on Euclidean distance for the points and edges of the drawn routes. Similar attractions and routes are merged together as recommended attractions for the first and second users to travel together.

[0015] Further optionally, based on the initial travel route and past routes, during the travel process, the step of detecting the time when the first user is disturbed and forced to communicate by fan users, and recording the travel type, includes:

[0016] By recording the time when the first user was disturbed and forced to communicate after being recognized by fans during a business trip, and comparing this time with the type of business trip, a mapping table is established between the disturbance time and the type of business trip. This table is used to statistically analyze which types and methods of business trips are more likely to be disturbed.

[0017] Further optionally, the detection of user travel logs or social media activity posted on social media includes:

[0018] It can acquire content posted by users on social media during their travels, detect whether the text contains place names or brands with geographical indications, identify whether there are landmark buildings in social media images, and determine whether users have exposed their own geographical location information.

[0019] Further optionally, the step of replanning the destination based on the probability that the first user may be recognized and disturbed, adjusting travel items and routes that are highly similar to the first user and the second user but are easily disturbed, and adjusting flight routes, includes:

[0020] The extracted user features are used to calculate the sensitivity of a user in a user group to different types of privacy. Specifically, this includes: inputting the training samples into a neural network for model training based on pre-stored privacy-sensitive feature data of the first user to obtain privacy information type labels; acquiring the user's personal privacy category identifier, adding the personal privacy category identifier to the user's feature values, determining the user's privacy type based on the feature vector and the personal privacy category identifier, establishing a user feature-privacy type lookup table, and obtaining the user's sensitive privacy type through the lookup table. Privacy types include: personal affairs, personal information, and accompanying persons.

[0021] Then, based on the sensitive privacy type, it is determined whether the travel destination is suitable for exposing the first user's identity or related privacy information. Based on the determination result, it is recommended whether the first user and the second user can participate in the travel project simultaneously. The determination method is based on the user's privacy type and makes preset and automatic judgments based on whether the travel project will expose that privacy type.

[0022] The system identifies which types of services frequently lead to the first user's identity being recognized. For probabilistic travel projects, it recommends that the first user avoid these projects to protect their privacy. When the first user is unable to participate in a project, the system automatically deletes the flight route and issues a refund.

[0023] Further, optionally, the recommendation of travel-related services based on flight routes includes:

[0024] Based on privacy protection requirements, we analyze and filter each item in the categories of personal care, health check-ups, outdoor sports, hotel sports, car rental services, and travel-related services, retaining only those suitable for the second user and removing those suitable for the first user.

[0025] Further, optionally, the automatic adjustment of mobile phone location and travel log information based on privacy protection requirements includes:

[0026] Measuring privacy types and the sensitivity of exposed privacy mainly includes: determining the type of privacy exposed and measuring the sensitivity of privacy. The privacy sensitivity types specifically include: Let p = [p1, p2, ..., pn] be the user's privacy type, which includes: personal affairs, personal information, and accompanying persons; the privacy sensitivity is represented as: sv = [sv1, sv2, ..., svn], where svi represents the sensitivity to privacy type pi, which is divided into 5 levels: A++ very sensitive, A+ relatively sensitive, A sensitive, B+ moderately sensitive, and B insensitive.

[0027] Based on the privacy type and corresponding sensitivity level as labeling values, log information that may expose personal information is used as training data to train a privacy protection sensitivity model. Specifically, this includes: using pre-stored user privacy sensitivity feature data as training sample labeling values, and using pre-collected address information, travel attraction information, and easily identifiable unit or content information mentioned on user social networks or travel logs as training feature values, which are then input into the neural network for model training; when a user's travel log is detected, the model automatically determines whether the log contains relevant geographical location information and easily identifiable unit or content information. When it is identified as the first privacy content that the user needs to protect, and the sensitivity level is B+ or higher, the information is automatically masked or cropped.

[0028] The masking or cropping includes: hiding or changing place names when the invention can identify common place names or multiple place names; blurring or cropping landmark buildings in photos; using OCR technology to determine recruitment areas where place names appear and automatically cropping or blurring the photos of areas where place names appear; and masking flight information in travel logs to protect the privacy of users' travel itineraries.

[0029] This invention discloses a smart dynamic planning system for travel data in government and enterprise business trips and civil aviation, the system comprising:

[0030] The initial route recommendation module is used to initially recommend routes suitable for user characteristics based on user information;

[0031] The route adjustment and merging module is used to recommend routes suitable for both users to travel together, based on information about the user and their companion.

[0032] The privacy protection monitoring module is used to monitor the leakage of users' travel log information during business trips, as well as the time and extent to which they are disturbed by fans.

[0033] A privacy-preserving route adjustment module is used to change a user's travel route;

[0034] The travel log correction module is used to protect the privacy of travel logs as well.

[0035] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0036] Through the design of this invention, civil aviation users with privacy protection needs can be automatically recommended travel routes with privacy protection value when traveling. The invention also takes into account the interests and needs of friends and family members during the trip, and determines which routes can be joined together and which routes require better routes and cannot be joined together. Furthermore, the invention optimizes the geographical location and pictures of the user's travel logs that are likely to expose their privacy. Attached Figure Description

[0037] Figure 1 This is a structural diagram of an embodiment of a method and system for intelligent dynamic planning of travel itineraries based on big data in government and enterprise business travel and civil aviation, according to the present invention.

[0038] Figure 2 This is a structural diagram of another embodiment of the intelligent dynamic planning method and system for government and enterprise business travel and civil aviation big data according to the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0040] Figure 1 This is a diagram illustrating the intelligent dynamic planning method and system structure for government and enterprise business travel and civil aviation big data based on the present invention. Figure 1 As shown in the figure, the intelligent dynamic planning method and system for government and enterprise business travel and civil aviation big data in this embodiment may specifically include:

[0041] Based on the needs of the first user, an initial travel route is planned based on factors such as price, preferences, travel time, and travel type, and an initial travel route and civil aviation route are recommended. At the same time, the characteristics of the second user are obtained, and an initial travel route and civil aviation route are recommended. The first user is a celebrity user with privacy protection needs, and the second user refers to the first user's family or partner. The route similarity between the first user and the second user is calculated, and some routes are merged into the same route based on the similarity.

[0042] Based on the initial travel route and past routes, during the travel process, detect the time when the first user is disturbed and forced to communicate by fans, and record the travel type; detect the travel logs or social media updates posted by the user on social media;

[0043] Based on the probability of the first user being recognized and disturbed, destination planning is re-engineered. Travel activities and routes with high similarity between the first and second users but prone to disruption are adjusted, and flight routes are modified. Nearby services are recommended based on these routes. To protect user privacy, mobile phone location and travel log information on social media are automatically adjusted. Specifically, this includes:

[0044] Step 101: Based on the user's travel ticketing needs, conduct initial travel route planning based on factors such as price, preferences, travel time, and travel type, and recommend relevant civil aviation routes based on big data.

[0045] Collect user information provided by users when purchasing airline tickets. Collect the information after obtaining the user's authorization and consent, and build a big data database of user information. The user information includes: gender, age, consumption level, and historical travel destination data.

[0046] Establishing a user classification model includes: acquiring data on gender, age, consumption level, place of residence, and historical travel destinations from the user's big data information as a training set; and using preset flight routes and tourist attractions suitable for the user information as annotation values. The annotation content mainly includes establishing a mapping table between user characteristics and tourist attraction types. For example, this includes: statistically analyzing the user characteristics of tourists visiting each type of tourist attraction nationwide, and identifying user characteristics that appear more than a preset value as the best tourist attractions for that type, used to construct the initial annotations for the tourist attractions.

[0047] A route classification model is trained using a random forest algorithm. User information data is input into the model to obtain routes and destinations corresponding to a specific user. These routes can include, but are not limited to, short-term business trips, group trips, family trips, and wedding trips. For example, for family trips, navigation routes might include Chimelong Water Park, Whampoa Military Academy, Guangzhou Tower, Xinghai Concert Hall, Shenzhen Youth Palace, and Zhuhai Ocean Aquarium, involving multiple routes suitable for each family member. The routes may be short-haul flights, ensuring that the needs of the entire family are met.

[0048] Based on the travel route and the user's personal information, the system recommends relevant air routes and calculates fares by searching the origin and destination as well as the travel time.

[0049] Step 102: Calculate the route similarity between passengers and their companions, and merge the overlapping parts of the passengers' and companions' routes into identical routes based on the similarity score, including:

[0050] Since business trips may involve trips with family, friends, or colleagues, targeted matching and analysis are needed for different groups of people traveling together. The first user refers to celebrity users with privacy concerns, while the second user refers to the first user's family, partners, or colleagues.

[0051] First, a relationship prediction model is established to predict relationships for users who cannot be matched as relatives. Specifically, this includes: obtaining information about users who purchased airline tickets based on the user information big data and extracting features, including name, gender, and age; collecting user feature data with confirmed relationships as a training set, including relationships such as colleagues, lovers, friends, and unknown; classifying the training set using the SVM algorithm to obtain the relationship prediction model; predicting the relationship between the first and second users, or between multiple users, under the same ticket purchase ID using the relationship prediction model; and predicting the relationship between multiple users under the same ticket purchase ID using the same ID used when purchasing tickets.

[0052] The relationships mentioned include users who are related by blood and their specific relationship types, which include: husband and wife, father and son, father and daughter, mother and son, mother and daughter, grandparents and grandchildren, etc.

[0053] For example, if two user profiles cannot be matched as kinship, but after feature extraction they are identified as a male and a female, both 22 years old, then the relationship prediction model predicts they are a couple.

[0054] Based on the user characteristics-attraction type mapping table, the preferred attraction types of the first user and their companions (i.e., the second user) are determined. Specific types include, for example, amusement parks, natural parks, museums, beaches, and large shopping malls. For instance, boys aged 5-10 prefer amusement parks, while people over 60 years old prefer museums.

[0055] The system presets demand weight values ​​for the first and second users, obtains the user characteristics of each user, and retrieves the preferred attraction types for each user through a user characteristic-attraction type mapping table. Based on the demand weight values ​​of each user, the attraction type preferred by the user with the highest demand weight value is selected as the first attraction type, and the attraction type preferred by the user with the second highest demand weight value is selected as the second attraction type. If two or more users have the same attraction type preference, their demand weight values ​​are superimposed, and the specific attraction with the higher rating in the first attraction type is selected as the first attraction. The specific location of the attraction in the first attraction is obtained. Centered on the first attraction, a second attraction is selected through a path planning model, ensuring that the distance between the first and second attractions does not exceed a set threshold, and that the second attraction conforms to the second attraction type. The first and second attractions are then used as the preliminary travel route determined by the first and second users.

[0056] Furthermore, based on the gender, age, consumption level, place of residence, and historical travel data of the first and second users, and the geographical location of the recommended travel attractions, the various attractions are connected geographically, routes are drawn, and similarity is calculated based on Euclidean distance for the points and edges of the drawn routes. Similar attractions and routes are merged together as recommended attractions for the first and second users to travel together.

[0057] Step 103: Based on the initial travel route and past routes, during the travel process, detect the time the first user is disturbed and forced to communicate by fan users, and record the travel type, including:

[0058] By recording the time a user is recognized by fans during a business trip and the resulting disturbances and forced interactions, a mapping table is established between the disturbance time and the type of business trip. This table is used to statistically analyze which types and methods of business trips are more likely to result in disturbances. For example, the table records the times a user was recognized by fans and engaged in activities such as interaction or signing autographs at Chimelong Water Park. This table records the correspondence between these activities and the disturbance time. Linear regression can be performed based on this table to predict which types of activities are more likely to result in disturbances and the duration of such disturbances. The system can statistically determine what types of activities the first user can participate in and what types of activities will result in disturbances to the first user. This information can then be used as a basis for recommending activities.

[0059] Step 104: Detect the user's travel logs or social media posts on social media.

[0060] It can acquire content posted by users on social media during their travels, detect whether the text contains place names or brands with geographical indications, identify whether there are landmark buildings in social media images, and determine whether users have exposed their own geographical location information.

[0061] For example, if a user posts "Chimelong Water Park in Guangzhou is so much fun!" along with a photo of the entrance after a business trip, this content is easily recognizable by followers, making it more likely to lead to privacy breaches and intrusions. Some travel destinations are difficult to identify, while some landmarks are easily identifiable; therefore, monitoring travel logs is necessary to protect user privacy.

[0062] Step 105: Based on the probability that the first user may be recognized and disturbed, re-plan the destination, adjust the travel items and routes that are highly similar to the first user and the second user but are easily disturbed, and adjust the flight routes, including:

[0063] The extracted user features are used to calculate the sensitivity of a user in a user group to different types of privacy. Specifically, this includes: inputting the training samples into a neural network for model training based on pre-stored privacy-sensitive feature data of the first user to obtain privacy information type labels; acquiring the user's personal privacy category identifier, adding the personal privacy category identifier to the user's feature values, determining the user's privacy type based on the feature vector and the personal privacy category identifier, establishing a user feature-privacy type lookup table, and obtaining the user's sensitive privacy type through the lookup table. The privacy information of the first user corresponds to celebrities, government and business officials, internet celebrities, stars, etc., all of whom have their own important privacy needs. Therefore, different people have different priorities regarding their privacy content. These privacy types include: personal affairs, personal information, and accompanying persons. For example, personal affairs refer to things one wants to do or travel programs one wants to participate in, but doesn't want others to know.

[0064] Then, based on the sensitive privacy type, it is determined whether the travel destination is suitable for exposing the first user's identity or related privacy information. Based on the determination result, it is recommended whether the first user and the second user can participate in the travel project simultaneously. The determination method is based on the user's privacy type and makes preset and automatic judgments based on whether the travel project will expose that privacy type.

[0065] First, it's necessary to characterize and create a mapping table to identify the significant privacy risks posed by surrounding services along the defined travel route. Then, based on flight routes, recommend surrounding services. According to privacy protection needs, analyze and filter each surrounding service item—personal care, health checkups, outdoor sports, hotel activities, car rentals, etc.—retaining only those suitable for the second user and removing those suitable for the first user. Surrounding services also include value-added services, consumption, accommodation, dining, and flight routes. Determine the types of privacy to be prioritized based on the frequency and types of privacy exposure frequently encountered in these travel activities. For example, train tickets, shopping malls, car rentals, hotels, VIP business services, water parks, etc., all expose some information about the first user. Therefore, it's necessary to pre-define the exposure risks and types for various activities and create a risk table. Among the services mentioned above, we can analyze which types of services frequently lead to the identification of the first user. For example, participating in a water park requires users to be completely dressed or unable to conceal their identity. Such occasions make it very easy for celebrities or internet celebrities to expose their identities and attract a large number of fans. Therefore, for some entertainment projects, we can allow only the second user to participate while the first user avoids them, thus protecting their privacy and allowing them to continue participating in some travel projects.

[0066] When a project is unable to be participated in by the first user, the system will automatically delete the flight route and refund the ticket for that user.

[0067] Step 105: Based on privacy protection needs, automatically adjust the information on mobile phone location and travel logs on social media to protect the user's flight privacy.

[0068] Due to differences in identity and status, exposing personal information can sometimes pose risks. Publishing travel logs containing private information also carries risks to an individual's life. Therefore, it is necessary to design and measure privacy exposure risks, mainly including: determining the type and sensitivity of the exposed privacy, and measuring the privacy sensitivity. The privacy sensitivity type specifically includes: Let p = [p1, p2, ..., pn] be the user's privacy type, such as personal affairs privacy, identity information privacy, or other various privacy needs. The privacy sensitivity can be represented as: sv = [sv1, sv2, ..., svn], where svi represents the sensitivity of the privacy type pi, which is divided into 5 levels: A++ Very Sensitive, A+ Relatively Sensitive, A Sensitive, B+ Moderately Sensitive, and B Insensitive. The privacy type and corresponding sensitivity level are used as label values. Log information that may expose personal information is used as training data to train a privacy protection sensitivity model.

[0069] Specifically, this involves using pre-stored user privacy-sensitive feature data as training sample labels, and pre-collected address information, travel attraction information, and easily identifiable information from users' social networks or travel logs as training feature values, which are then input into the neural network for model training. When a user's travel log is detected, the model automatically determines whether the log contains relevant geographical location information and easily identifiable information from organizations or content. If it identifies information as the primary privacy content that needs protection for the user, and the sensitivity level is B+ or higher, it automatically masks or cropps the information.

[0070] The masking or cropping includes hiding or altering place names when the invention can identify common place names or multiple place names. It also involves blurring or cropping landmarks in photos to remove areas containing place names or landmarks. Additionally, it masks flight information to protect the privacy of users' travel itineraries.

[0071] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

[0072] The program used to implement the information control of this invention can be written in one or more programming languages ​​or a combination thereof to perform computer program code for carrying out the operations of this invention. The programming languages ​​include object-oriented programming languages—such as Java, Python, and C++—and conventional procedural programming languages—such as C or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0073] In the several embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and other division methods may be used in actual implementation.

[0074] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0075] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units. The integrated unit implemented as a software functional unit can be stored in a computer-readable storage medium. This software functional unit, stored in a storage medium, includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods described in the various embodiments of the present invention.

[0076] The aforementioned storage media include: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media that can store program code.

Claims

1. A method for intelligent dynamic planning of travel itineraries based on big data from government and enterprise business travel and civil aviation, characterized in that, The method includes: Based on the needs of the first user, initial travel route planning is performed based on price, interests, travel time, and travel type, and initial travel routes and civil aviation routes are recommended. At the same time, the characteristics of the second user are obtained, and initial travel routes and civil aviation routes are recommended. The first user is a government or enterprise executive or celebrity user with privacy protection needs, and the second user refers to the first user's family or partner. The route similarity between the first user and the second user is calculated, and some routes are merged into the same route based on the similarity. Based on the initial travel route and past routes, during the travel process, detect the time when the first user is disturbed and forced to communicate by fans, and record the travel type; detect the travel logs or social media updates posted by the user on social media; Based on the probability that the first user might be recognized and disturbed, destination planning is re-engineered. Travel items and routes that are highly similar to the first user and the second user but are likely to cause disturbance are adjusted, and flight routes are also adjusted. This includes: identifying travel items that frequently lead to the first user's identity being identified and recommending that the first user avoid these travel items to protect privacy; recommending travel-related services based on flight routes; and automatically adjusting the content of travel logs on social media based on privacy protection needs to protect the user's flight route privacy. This includes: when a user's travel log is detected, the model automatically determines whether the log contains relevant geographical location information and easily identifiable information about the user's organization or content. When it is identified as privacy content that the first user needs to protect, and the sensitivity is B+ or higher, the information is automatically masked or cropped.

2. The method according to claim 1, wherein, The initial travel route planning based on price, preferences, travel time, and travel type includes: Based on users' travel booking needs, initial travel route planning is carried out based on price, preferences, travel time, and travel type, and relevant civil aviation routes are recommended based on big data. Collect user information provided by users when purchasing airline tickets, collect it after obtaining the user's authorization and consent, and build a big data database of user information, including: gender, age, consumption level, and historical travel destination data; Establishing a user classification model includes: acquiring data on gender, age, consumption level, place of residence, and historical travel destinations from the user's big data information as a training set; using preset flight routes and scenic spot routes suitable for user information as annotation values; the annotation content mainly includes establishing a mapping relationship table between user characteristics and scenic spot types, including statistically analyzing the user characteristics of tourists received by each type of scenic spot nationwide, and using the characteristics of users whose occurrence frequency exceeds a preset value as the best scenic spot for that scenic spot type to construct the initial annotation of scenic spots; A route classification model is trained using a random forest algorithm. User information data is input into the route classification model to obtain routes and attractions corresponding to specific users. Based on the route content of the travel route and the user's personal information characteristics, relevant civil aviation routes are recommended by searching the origin and destination as well as the travel time, and the route price is calculated.

3. The method according to claim 1, wherein, The calculation of route similarity between the first user and the second user, and the merging of some routes into identical routes based on the similarity, includes: First, a relationship prediction model is established to predict the relationship between users who cannot be matched as relatives. Specifically, this includes: obtaining information about users who purchase airline tickets based on big data of user information and extracting features, including name, gender, and age; collecting user feature data with confirmed relationships as a training set, including relationships such as colleagues, lovers, friends, and unknown; classifying the training set using the SVM algorithm to obtain the relationship prediction model; predicting the relationship between second users under the same ticket purchase ID using the same ID used when purchasing tickets; Furthermore, based on the gender, age, consumption level, place of residence, historical travel destinations, and geographical location of the recommended travel attractions for the first and second users, the attractions are connected by geographical routes. Routes are drawn based on these routes, and similarity calculations based on Euclidean distance are performed on the points and edges of the drawn routes. Similar attractions and routes are merged together as recommended attractions for the first and second users to travel together.

4. The method according to claim 1, wherein, Based on the initial travel route and past routes, the system detects the time the first user is disturbed and forced to communicate with fans during the trip, and records the travel type, including: By recording the time when the first user was disturbed and forced to communicate after being recognized by fans during a business trip, and comparing this time with the type of business trip, a mapping table is established between the disturbance time and the type of business trip. This table is used to statistically analyze which types and methods of business trips are more likely to be disturbed.

5. The method according to claim 1, wherein, The detection of users' travel logs or social media activity posted on social media includes: It acquires content posted by users on social media during their business trips, detects whether place names or brands with geographical indications appear in the text, identifies whether landmark buildings are present in social media images, and determines whether users have exposed their own geographical location information.

6. The method according to claim 1, wherein, Based on the probability that the first user might be recognized and disturbed, the destination planning is re-performed, adjusting travel items and routes that are highly similar to the first user and the second user but are easily disturbed, and adjusting flight routes, including: Extracting user features and calculating the sensitivity of a user in a user group to different types of privacy specifically includes: inputting training samples into a neural network for model training based on pre-stored user privacy sensitivity feature data to obtain privacy information type labels; obtaining the user's personal privacy category identifier, adding the personal privacy category identifier to the user's feature values ​​respectively, determining the user's privacy type based on the feature vector and the personal privacy category identifier, establishing the user feature-privacy type lookup table, and obtaining the user's sensitive privacy type through the lookup table; privacy types include: personal affairs, personal information, and accompanying persons; Then, based on the sensitive privacy type, it is determined whether the travel destination is suitable for exposing the first user's identity or related privacy information. Based on the judgment result, it is recommended whether the first user and the second user can participate in the travel project at the same time. The judgment method is based on the user's privacy type and whether the travel project will expose that privacy type. The system identifies travel activities that frequently lead to the first user's identity being recognized, and recommends that the first user avoid these activities to protect their privacy. When the first user is unable to participate in a particular activity, the system automatically deletes the flight route and issues a refund.

7. The method according to claim 1, wherein, The recommended travel-related services based on flight routes include: Based on privacy protection requirements, we analyze and filter each item in the categories of personal care, health check-ups, outdoor sports, hotel sports, car rental services, and travel-related services, retaining only those suitable for the second user and removing those suitable for the first user.

8. The method according to claim 1, wherein, The automatic adjustment of travel log content on social media to protect users' flight privacy includes: Measuring privacy types and the sensitivity of exposed privacy mainly includes: determining the type of privacy exposed and measuring its sensitivity. Specifically, privacy types include: Let p = [p1, p2, ..., pn] be the user's privacy type, which includes: personal affairs, personal information, and accompanying persons; privacy sensitivity is represented as: sv = [sv1, sv2, ..., svn], where svi represents the sensitivity to privacy type pi, which is divided into 5 levels: A++ very sensitive, A+ relatively sensitive, A sensitive, B+ moderately sensitive, and B insensitive. Based on the privacy type and corresponding sensitivity level as labeling values, log information that may expose personal information is used as training data to train a privacy protection sensitivity model. Specifically, this includes: using pre-stored user privacy sensitivity feature data as training sample labeling values, and using pre-collected address information, travel attraction information, and easily identifiable unit or content information mentioned on the user's social networks or travel logs as training feature values, which are then input into the neural network for model training; when a user's travel log is detected, the model automatically determines whether the log contains relevant geographical location information and easily identifiable unit or content information. When it is identified as the first user's privacy content that needs protection, and the sensitivity level is B+ or higher, the information is automatically masked or cropped. The masking or cropping includes: hiding or changing place names when the published content can identify common place names or multiple place names; blurring or cropping landmarks in photos; using OCR technology to determine the areas in photos where place names appear and automatically cropping or blurring those areas; and masking flight information in travel logs to protect the privacy of users' travel itineraries.

9. A smart dynamic planning system for travel data in government and enterprise business trips and civil aviation, executing the smart dynamic planning method as described in any one of claims 1-8, characterized in that, The system includes: The initial route recommendation module is used to initially recommend routes suitable for user characteristics based on user information; The route adjustment and merging module is used to recommend routes suitable for both users to travel together, based on information about the user and their companion. The privacy protection monitoring module is used to monitor the leakage of users' travel log information during business trips, as well as the time and extent to which they are disturbed by fans. A privacy-preserving route adjustment module is used to change a user's travel route; The travel log correction module is used to protect the privacy of travel logs as well.

Images