Geographic position information recommendation processing method and device

Abstract

The invention relates to the field of traffic, and discloses a geographic position information recommendation processing method and a geographic position information recommendation processing device, which are used for solving the problem that after a driver receives an order, the driver cannot find the accurate position of a user and cannot timely receive the order to the user, and the method comprises the following steps of: acquiring the current position of user equipment by using a GPS (Global Positioning System) receiver; a mobile network base station connected with the equipment is determined, the position of the mobile network base station is acquired, the moving mode and direction of the equipment are detected to deduce position change, the position of the equipment is broadcasted to the equipment by utilizing a Bluetooth low-power-consumption beacon on the user equipment, preprocessing and analysis are carried out according to a large amount of collected data, and user preferences and activity ranges are obtained. According to the movement track data of the user, the activity range of the user is calculated, the server constructs a personalized recommendation model through a machine learning algorithm, generates address position recommendation meeting user requirements, establishes a feedback mechanism, and updates a data model algorithm in time, and the pick-up rate is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of transportation, and more specifically to a method and device for recommending and processing geographic location information. Background Art

[0002] In the online car-hailing system, geographic location information recommendation provides users and drivers with personalized pick-up and drop-off point recommendations, optimal route planning, dynamic pricing and discounts, and heat map recommendations for high-demand areas by analyzing the user's current location, historical behavior data, and real-time traffic conditions. This system uses big data and artificial intelligence technology to not only help users get a more convenient and efficient travel experience, but also improve the driver's order-taking efficiency and the overall operational efficiency of the platform.

[0003] The existing geolocation information recommendation is that after the driver accepts the order, he follows the GPS navigation to arrive near the user's departure point, and then communicates with the user online or by phone to determine the final pick-up location. However, this geolocation information recommendation method will cause the user and the driver to be unable to find each other in time, resulting in a lower pick-up rate and affecting driving safety.

[0004] In view of the above problems, the present invention proposes a solution. Summary of the invention

[0005] In order to overcome the above-mentioned defects in the prior art, the present invention provides a method and device for recommending and processing geographic location information to solve the problems existing in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for recommending geographic location information includes the following steps:

[0008] When the driver goes to pick up the driver, if the user chooses not to enter the pick-up location, the online ride-hailing platform collects geographic location information based on the order information, and uses the GPS receiver to obtain the current location of the user's device, determines the mobile network base station to which the device is connected based on the user's device, and obtains its location;

[0009] Utilize sensors on the user device to detect the device's movement pattern and direction to infer location changes, and utilize Bluetooth low energy beacons on the user device to broadcast its location to the device;

[0010] If the user chooses to manually enter the pick-up location, the user is allowed to manually enter their location;

[0011] Collect data from different channels, and the server performs pre-processing based on the large amount of data collected;

[0012] Analyze user preferences and activity range based on user stay points and movement path data;

[0013] According to the user's movement trajectory data, the user's activity range is calculated using the minimum enclosing rectangle;

[0014] The server uses machine learning algorithms to combine user behavior data and location data to build a personalized recommendation model;

[0015] Based on the constructed model and real-time data, continuous comparison is performed to generate address location recommendations that meet user needs;

[0016] By establishing a feedback mechanism, users can promptly feedback error information to the platform and update the data model algorithm in a timely manner.

[0017] Preferably, the server performs preprocessing steps based on the large amount of collected data:

[0018] Perform denoising and deviation correction operations on the data, convert data from different sources into a unified format, and standardize the data in the unified format;

[0019] Convert GPS coordinates and Wi-Fi access point information into geographic locations;

[0020] Fingerprinting technology is used to match movement patterns and location information to identify specific locations and establish probabilistic geographic areas.

[0021] Preferably, the steps of using fingerprint technology to match movement patterns and location information, identify specific locations, and establish probabilistic geographic areas are:

[0022] Collect the user's GPS coordinates, longitude and latitude, and timestamp, and collect the user's Wi-Fi access point data, including the access point's MAC address, signal strength, and timestamp;

[0023] Clean the data, remove noise and outliers, unify the timestamp format, and align GPS data with Wi-Fi data;

[0024] Systematically collect Wi-Fi signal strength and corresponding GPS coordinates in the target area to establish a fingerprint database. Each fingerprint record includes: longitude and latitude and a list of Wi-Fi access point signal strengths at the corresponding location, and store the collected fingerprint data in the database;

[0025] Convert the user's current Wi-Fi access point data into a signal strength vector, use the Euclidean distance measurement method to calculate the similarity between the user's current signal strength vector and each fingerprint in the fingerprint database, find the fingerprint that best matches the user's current signal strength vector, and extract the corresponding longitude and latitude;

[0026] Based on the matching results, the probability distribution of the user's location is calculated using a Gaussian mixture probability model. Based on the matched fingerprint and the signal characteristics of the surrounding environment, the probability distribution of the user's current location is calculated to generate a probabilistic geographic area.

[0027] Using a Bayesian filtering approach, the GPS coordinates and Wi-Fi signal data are fused, combined with the output of the fusion model, to update the user's location estimate and generate a more accurate probabilistic geographic area.

[0028] Preferably, the step of analyzing user preferences and activity range according to user stay points and movement path data is:

[0029] Collecting the user's geographic location data, which includes timestamp, longitude and latitude, and speed, and preprocessing the data, which includes cleaning the data, removing noise and outliers, sorting the raw data by time, and calculating the distance and time interval between each data point;

[0030] According to the longitude and latitude data, the user's movement trajectory is obtained by calculating the distance between two data points. The calculation formula is d = R arccos (sin (φ1) sin (φ2) + cos (φ1) cos (φ2) cos (λ2-λ1)), where R is the radius of the earth, φ1, φ2 are the latitudes of the two points, and λ1, λ2 are the longitudes of the two points;

[0031] Calculate the speed between each data point, the calculation formula is: Where d is the distance between two points, and Δt is the time difference between the two points;

[0032] When the speed is lower than the threshold and the duration exceeds the predetermined time, it is determined as a stop point, the stay time of each stop point is calculated, and the environmental information near the stop point is obtained using the geographic information system;

[0033] Based on the GPS signal strength and Wi-Fi signal information, combined with the location database, it is determined whether the user is indoors or outdoors, and POI data is used to identify whether the user is in a public place;

[0034] Combine the stay point data with the POI database to identify whether the user has entered a specific location and record the user's stay time and frequency at these locations;

[0035] Statistics on the user's stay time and visit frequency at various locations, and calculate the user's preference index for each location. The calculation formula is: Where P type Expressed as a preference index, The time a user spends at a specific type of place. The total dwell time of the user.

[0036] Preferably, the step of calculating the user's activity range using a minimum bounding rectangle based on the user's movement trajectory data is:

[0037] Acquire the user's movement trajectory data, which includes timestamp, longitude and latitude information, clean the data, remove noise and outliers, and arrange the location point data in chronological order;

[0038] Find the maximum and minimum longitude and latitude of all location points, and use the extracted boundary points to determine the four corner points of the minimum enclosing rectangle;

[0039] Calculate the width of the rectangle using the east-west direction of the rectangle as the width;

[0040] Take the north-south direction of the rectangle as the height, calculate the height of the rectangle, and multiply the width of the rectangle by the height of the rectangle to get the area;

[0041] Draw a minimum bounding rectangle on the map and mark all the user's location points.

[0042] Preferably, the steps of constructing a personalized recommendation model by combining user behavior data and geographic location data through a machine learning algorithm are:

[0043] Step 1: Collect historical user behavior data, geographic location data, and environmental data;

[0044] Step 2: Calculate the preference index, user activity range, and average stay time for each location type based on historical user behavior data, geographic location data, and environmental data;

[0045] Step 3: According to the recommendation task, define the target label as the recommended boarding point and construct label data, which is the actual selection of the user in the historical data;

[0046] Step 4: Divide the data set into training set and test set according to the set ratio;

[0047] Step 5: Use the random forest algorithm to train the model using the training set data and adjust the model parameters to optimize performance;

[0048] Step 6: Use the test set data to evaluate the model performance, calculate the model's accuracy, precision, recall rate and F1 value, conduct a comprehensive evaluation of the model performance to obtain the model performance index, and compare the model performance index with the preset threshold. If the model performance index is greater than the preset threshold, output the current model as the personalized recommendation model. If the model performance index is less than the preset threshold, return to step 5 for parameter optimization until the model performance index is greater than the preset threshold.

[0049] Preferably, the step of generating an address location recommendation that meets user needs is:

[0050] Obtain the user's current geographic location, collect real-time traffic condition data, the traffic condition data includes road congestion and traffic accidents, and obtain the user's time constraint information, the time constraint information includes the appointment time and the preferred arrival time;

[0051] Clean and organize data to ensure its accuracy and completeness, and convert real-time data into standard formats;

[0052] Collect users’ historical location data, places visited, search queries, social media following, and online reviews to generate user preference models;

[0053] Use collaborative filtering algorithms to find users with similar behaviors to the current user and recommend places visited by these users, and recommend places similar to places visited by the current user;

[0054] Recommend places that meet the user's needs in real time based on the user's current location, time constraints, and traffic conditions. Use user preference models and collaborative filtering results to personalize the recommended place list.

[0055] Use a Gaussian mixture model to generate a probability distribution of places that users may visit and determine the most likely recommended places;

[0056] Use the Dijkstra algorithm, combined with real-time traffic data, to calculate the optimal path from the user's current location to the recommended location and select the fastest route.

[0057] Preferably, a device for processing geographic location information recommendation includes:

[0058] A data collection module is used to collect order information and user information, and transmit the order information and user information to the location information recommendation module;

[0059] The model building module is used to collect historical user information and build a personalized recommendation model, and transmit the personalized recommendation model to the location information recommendation module;

[0060] The location information recommendation module is used to receive the order information, user information and personalized recommendation model transmitted by the data collection module and the model building module, import the information into the model to obtain the recommended location information, and transmit the recommended location information to the map display module;

[0061] The map display module is used to receive the recommended location information transmitted by the location information recommendation module and display it on the map.

[0062] Technical effects and advantages of the present invention:

[0063] The GPS receiver is used to obtain the current location of the user's device, determine the mobile network base station to which the device is connected, and obtain its location. The movement pattern and direction of the device are detected to infer location changes, and the Bluetooth low-power beacon on the user's device is used to broadcast its location to the device. Preprocessing and analysis are performed based on the large amount of collected data to obtain user preferences and activity ranges. The user's activity range is calculated based on the user's movement trajectory data. The server uses machine learning algorithms to build a personalized recommendation model to generate address location recommendations that meet user needs, establish a feedback mechanism, and update the data model algorithm in a timely manner, effectively improving the pick-up rate. It provides location-based recommendations, real-time traffic information, path planning, and optimal route suggestions, recommends popular destinations, attractions, and activities, and can assist in emergency response based on user interests and locations, providing location information of missing persons and evacuation route planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is the overall flow chart of the present invention.

[0065] Figure 2 This is the overall module diagram of the present invention. DETAILED DESCRIPTION

[0066] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. In addition, the forms of the various structures recorded in the following embodiments are only examples. The method and device for recommending geographic location information involved in the present invention are not limited to the various structures recorded in the following embodiments. All other implementations obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention.

[0067] The present invention provides a method for recommending geographic location information, comprising the following steps:

[0068] When the driver goes to pick up the user, the user can choose to manually enter the pick-up location or not enter the pick-up location. If the pick-up location is not entered, the online ride-hailing platform collects geographic location information based on the order information, and uses the GPS receiver to obtain the current location of the user's device, determines the mobile network base station to which the device is connected based on the user's device, and obtains its approximate location;

[0069] Leverage sensors on the user device to detect the device’s movement pattern and direction to infer location changes, and leverage Bluetooth Low Energy (BLE) beacons on the user device to broadcast its location to devices at specific locations;

[0070] If the user chooses to manually enter the pick-up location, the user is allowed to manually enter their location, such as by entering the pick-up location through a search bar or map interface;

[0071] Collect data from different channels. The server performs preprocessing based on the large amount of data collected. The preprocessing includes operations such as denoising and correction to remove outliers and invalid data, and convert data from different sources into a unified format. The unified format data is standardized to ensure the accuracy and consistency of the data, such as converting GPS coordinates into geodetic coordinates.

[0072] Convert GPS coordinates and Wi-Fi access point information into geographic locations, use fingerprinting technology to match movement patterns and location information, identify specific locations, and establish probabilistic geographic regions to represent the possible location range of users. Fingerprinting technology is a geographic positioning method that establishes a "fingerprint" database by pre-collecting and recording wireless signal characteristics (such as Wi-Fi signal strength) in a target area, and matches the real-time collected signal characteristics with the database during actual use to determine the specific location of the user;

[0073] Fingerprint technology can more accurately identify the specific location of users by collecting wireless signal data (such as Wi-Fi, Bluetooth signal strength, etc.) of devices at specific locations. Compared with GPS-based positioning, fingerprint technology is usually more accurate in indoor and urban environments and can provide high-precision geographic location information; by collecting and analyzing multiple location fingerprints, probabilistic geographic regions can be established. These areas are not just simple coordinate points, but have statistical probability distributions that represent possible locations of users. This probabilistic geographic region can better reflect the dynamics and variability of location information in the real world, and facilitate more refined location inference and service provision; by understanding the activity patterns and preferences of users in different geographic areas, service providers can provide personalized recommendations and services. For example, based on the user's activity habits in different locations, recommend nearby restaurants, stores or activities to improve user experience and satisfaction.

[0074] In this embodiment, it should be specifically explained that the steps of using fingerprint technology to match movement patterns and location information, identify specific locations, and establish probabilistic geographic areas are:

[0075] Collect the user's GPS coordinate data (latitude, longitude and timestamp), and collect the user's Wi-Fi access point data, including the access point's MAC address, signal strength (RSSI) and timestamp;

[0076] Clean the data, remove noise and outliers, unify the timestamp format, and align GPS data with Wi-Fi data;

[0077] Systematically collect Wi-Fi signal strength and corresponding GPS coordinates in the target area to establish a fingerprint database. Each fingerprint record includes: location (latitude and longitude) and a list of Wi-Fi access point signal strengths at the corresponding location. The collected fingerprint data is stored in the database for subsequent matching and query.

[0078] Convert the user's current Wi-Fi access point data into a signal strength vector S u =[RSSI1,RSSI2,...,RSSI n ], the Euclidean distance measurement method is used to calculate the similarity between the user's current signal strength vector and each fingerprint in the fingerprint database, and its expression is: Find the fingerprint that best matches the user's current signal strength vector and extract the corresponding geographic location (latitude and longitude);

[0079] Based on the matching results, the probability distribution of the user's location is calculated using a Gaussian mixture probability model. Based on the matched fingerprint and the signal characteristics of the surrounding environment, the probability distribution of the user's current location is calculated to generate a probabilistic geographic area.

[0080] Using methods such as Bayesian filtering or Kalman filtering, GPS coordinates and Wi-Fi signal data are fused to improve the accuracy of location estimation. The output of the fusion model is combined to update the user's location estimate and generate a more accurate probabilistic geographic area. The Bayesian filter is a type of statistical method for state estimation and prediction based on Bayesian theorem and probability theory. It is mainly used in dynamic systems to infer the probability distribution of the system's state changing over time by fusing sensor measurement data and system models.

[0081] Analyze user preferences and activity range based on user stay points and movement path data;

[0082] In this embodiment, it should be specifically explained that the steps of analyzing user preferences and activity ranges according to user stay points and movement path data are as follows:

[0083] Collecting the user's geographic location data, including timestamp, longitude and latitude, speed, etc., and preprocessing the data, including cleaning the data, removing noise and outliers, sorting the raw data by time, and calculating the distance and time interval between each data point;

[0084] According to the longitude and latitude data, the user's movement trajectory is obtained by calculating the distance between two data points. The calculation formula is d = R arccos (sin (φ1) sin (φ2) + cos (φ1) cos (φ2) cos (λ2-λ1)), where R is the radius of the earth, φ1, φ2 are the latitudes of the two points, and λ1, λ2 are the longitudes of the two points;

[0085] Calculate the speed between each data point, the calculation formula is: Where d is the distance between two points, and Δt is the time difference between the two points;

[0086] When the speed is lower than a certain threshold and the duration exceeds a certain time, it is determined as a stop point, and the stay time of each stop point is calculated. The geographic information system (GIS) or location service (such as POI database) is used to obtain the environmental information near the stop point, such as shops, restaurants, parks, etc.;

[0087] Based on information such as GPS signal strength and Wi-Fi signal, combined with the location database, determine whether the user is indoors or outdoors, and use POI data to identify whether the user is in a public place (such as a shopping mall, station, school, etc.);

[0088] Combine the stop point data with the POI database to identify whether the user has entered a specific place (such as a store or restaurant) and record the time and frequency of the user's stay in these places;

[0089] Statistics are collected on the time users spend at various places (such as restaurants, shops, public places, etc.) and the frequency of their visits, and the preference index of users for various places is calculated. The calculation formula is: Where P type Expressed as a preference index, The time a user spends at a specific type of place. The total dwell time of the user.

[0090] According to the user's movement trajectory data, the user's activity range is calculated using a minimum bounding rectangle, where the minimum bounding rectangle is the smallest rectangle that covers all trajectory points.

[0091] In this embodiment, it should be specifically explained that the steps of calculating the user's activity range using the minimum enclosing rectangle according to the user's movement trajectory data are as follows:

[0092] Acquire the user's movement trajectory data, which includes timestamp, longitude and latitude information, clean the data, remove noise and outliers, and arrange the location point data in chronological order;

[0093] Find the maximum and minimum longitude and latitude of all location points, and use the extracted boundary points to determine the four corner points of the minimum enclosing rectangle;

[0094] The width of the rectangle is calculated using the east-west direction of the rectangle as the width. The calculation formula is Width = max(λ) - min(λ), where Width represents the width of the rectangle, max(λ) represents the maximum longitude of the right boundary, and min(λ) represents the minimum longitude of the left boundary.

[0095] The height of the rectangle is calculated by taking the north-south direction of the rectangle as the height. The calculation formula is Height = max(φ)-min(φ), where Height represents the height of the rectangle, max(φ) represents the maximum latitude of the upper boundary, and min(φ) represents the minimum latitude of the lower boundary. The area of ​​the rectangle is calculated by multiplying the width of the rectangle by the height of the rectangle. The calculation formula is A = Width × Height;

[0096] Draw a minimum enclosing rectangle on the map to visually display the user's activity range, mark all the user's location points, and display the trajectory and range of the activity.

[0097] The server uses machine learning algorithms to combine user behavior data and location data to build a personalized recommendation model;

[0098] Machine learning algorithms can analyze and understand user behavior patterns, interests, and activity habits in different geographic locations. With this data, the system can generate accurate personalized recommendations, recommending content, products, or services that users may be interested in, thereby improving user experience and satisfaction; personalized recommendations can make it easier for users to find content that matches their interests and needs, thereby increasing the frequency and duration of their use of the platform or application. By enhancing user engagement and stickiness, user retention and loyalty can be effectively improved; personalized recommendation models based on user behavior and geographic location data can help marketing teams more accurately target user groups and provide customized marketing strategies and promotional activities. This precise marketing method usually leads to higher conversion rates and return on investment; machine learning algorithms can monitor and analyze user behavior data and feedback information in real time, and continuously optimize personalized recommendation models. Through continuous learning and adjustment, the system can more accurately predict changes in user needs and preferences, and maintain the timeliness and effectiveness of recommendations.

[0099] In this embodiment, it should be specifically explained that the steps of building a personalized recommendation model by combining the user's behavior data and geographic location data through a machine learning algorithm are as follows:

[0100] Step 1: Collect historical user behavior data, geographic location data, and environmental data. The behavior data includes ride history, stay time, and visit frequency, etc. The location data includes longitude and latitude and movement trajectory, and the environmental data includes nearby POI information;

[0101] Step 2: Calculate the preference index, user activity range, and average stay time for each location type based on historical user behavior data, geographic location data, and environmental data;

[0102] Step 3: According to the recommendation task, define the target label as the recommended boarding point and construct label data, which is the actual selection of the user in the historical data;

[0103] Step 4: Divide the data set into training set and test set according to a certain ratio (such as 80 / 20) to ensure the representativeness of training and test data;

[0104] Step 5: Use the random forest algorithm to train the model using the training set data and adjust the model parameters to optimize the performance. The random forest algorithm is an ensemble learning algorithm consisting of multiple decision trees. Each tree is trained independently and increases the difference by randomly selecting features. Finally, prediction is made by voting (classification) or averaging (regression). It has good generalization ability and anti-overfitting ability and is suitable for various classification and regression problems.

[0105] Step 6: Use the test set data to evaluate the model performance, calculate the model's accuracy, precision, recall rate and F1 value, conduct a comprehensive evaluation of the model performance to obtain the model performance index, and compare the model performance index with the preset threshold. If the model performance index is greater than the preset threshold, output the current model as the personalized recommendation model. If the model performance index is less than the preset threshold, return to step 5 for parameter optimization until the model performance index is greater than the preset threshold.

[0106] Based on the constructed model and real-time data, continuous comparison is performed to generate address location recommendations that meet user needs;

[0107] In this embodiment, it should be specifically explained that the steps of generating address location recommendations that meet user needs are:

[0108] Obtain the user's current geographic location, collect real-time traffic condition data, the traffic condition data includes road congestion and traffic accidents, and obtain the user's time constraint information, the time constraint information includes the appointment time and the preferred arrival time;

[0109] Clean and organize data to ensure its accuracy and completeness, and convert real-time data into a standard format for subsequent processing;

[0110] Collect users’ historical location data, places they’ve visited, search queries, social media followers, and online reviews to generate a user preference model, including the types of places and activity ranges that users prefer;

[0111] Use collaborative filtering algorithms to find other users with similar behaviors to the current user, and recommend places visited by these users. Recommend other places similar to places visited by the current user. The collaborative filtering algorithm is a type of technology used in recommendation systems. Its core idea is to use the historical behavior data of users and items to infer the interests of users and recommend items that they may be interested in to users. This recommendation method does not rely on the content information of the items, but makes recommendations based on the historical behavior of users and the behavior patterns of other users;

[0112] Recommend places that meet the user's needs in real time based on the user's current location, time constraints, and traffic conditions. Use user preference models and collaborative filtering results to personalize the recommended place list.

[0113] A Gaussian mixture model is used to generate a probability distribution of places that a user may visit, and the most likely recommended place is determined. The Gaussian mixture model is a probability model used to describe a mixture of multiple Gaussian distribution components. The Gaussian mixture model is a flexible model that can adapt to data sets of various shapes and sizes.

[0114] The Dijkstra algorithm is used to calculate the optimal path from the user's current location to the recommended location in combination with real-time traffic data, taking into account the user's time constraints and selecting the fastest or most convenient route. The Dijkstra algorithm solves the problem of finding the shortest path to all other vertices in the graph from a vertex (called the source point) in the graph. The algorithm is applicable to graphs with non-negative weights.

[0115] By establishing a feedback mechanism, users can promptly feedback erroneous information to the platform and update the data model algorithm in a timely manner to make the positioning information more accurate.

[0116] In this embodiment, it is necessary to specifically explain a device for processing geographic location information recommendation, the device comprising:

[0117] A data collection module is used to collect order information and user information, and transmit the order information and user information to the location information recommendation module;

[0118] The model building module is used to collect historical user information and build a personalized recommendation model, and transmit the personalized recommendation model to the location information recommendation module;

[0119] The location information recommendation module is used to receive the order information, user information and personalized recommendation model transmitted by the data collection module and the model building module, import the information into the model to obtain the recommended location information, and transmit the recommended location information to the map display module;

[0120] The map display module is used to receive the recommended location information transmitted by the location information recommendation module and display it on the map.

[0121] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0122] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for recommending geographic location information, characterized in that: The following steps are involved: When the driver goes to pick up the driver, if the user chooses not to enter the pick-up location, the online ride-hailing platform collects geographic location information based on the order information, and uses the GPS receiver to obtain the current location of the user's device, determines the mobile network base station to which the device is connected based on the user's device, and obtains its location; Utilize sensors on the user device to detect the device's movement pattern and direction to infer location changes, and utilize Bluetooth low energy beacons on the user device to broadcast its location to the device; If the user chooses to manually enter the pick-up location, the user is allowed to manually enter their location; Collect data from different channels, and the server performs pre-processing based on the large amount of data collected; Analyze user preferences and activity range based on user stay points and movement path data; According to the user's movement trajectory data, the user's activity range is calculated using the minimum enclosing rectangle; The server uses machine learning algorithms to combine user behavior data and location data to build a personalized recommendation model; Based on the constructed model and real-time data, continuous comparison is performed to generate address location recommendations that meet user needs; By establishing a feedback mechanism, users can promptly feedback error information to the platform and update the data model algorithm in a timely manner.

2. A method for recommending geographic location information according to claim 1, characterized in that: The server performs preprocessing steps based on the large amount of collected data: Perform denoising and deviation correction operations on the data, convert data from different sources into a unified format, and standardize the data in the unified format; Convert GPS coordinates and Wi-Fi access point information into geographic locations; Fingerprinting technology is used to match movement patterns and location information to identify specific locations and establish probabilistic geographic areas.

3. A method for recommending geographic location information according to claim 2, characterized in that: The steps of using fingerprint technology to match movement patterns and location information, identify specific locations, and establish probabilistic geographic areas are as follows: Collect the user's GPS coordinates, longitude and latitude, and timestamp, and collect the user's Wi-Fi access point data, including the access point's MAC address, signal strength, and timestamp; Clean the data, remove noise and outliers, unify the timestamp format, and align GPS data with Wi-Fi data; Systematically collect Wi-Fi signal strength and corresponding GPS coordinates in the target area to establish a fingerprint database. Each fingerprint record includes: longitude and latitude and a list of Wi-Fi access point signal strengths at the corresponding location, and store the collected fingerprint data in the database; Convert the user's current Wi-Fi access point data into a signal strength vector, use the Euclidean distance measurement method to calculate the similarity between the user's current signal strength vector and each fingerprint in the fingerprint database, find the fingerprint that best matches the user's current signal strength vector, and extract the corresponding longitude and latitude; Based on the matching results, the probability distribution of the user's location is calculated using a Gaussian mixture probability model. Based on the matched fingerprint and the signal characteristics of the surrounding environment, the probability distribution of the user's current location is calculated to generate a probabilistic geographic area. Using a Bayesian filtering approach, the GPS coordinates and Wi-Fi signal data are fused, combined with the output of the fusion model, to update the user's location estimate and generate a more accurate probabilistic geographic area.

4. A method for recommending geographic location information according to claim 1, characterized in that: The steps of analyzing user preferences and activity range according to user stay points and movement path data are as follows: Collecting the user's geographic location data, which includes timestamp, longitude and latitude, and speed, and preprocessing the data, which includes cleaning the data, removing noise and outliers, sorting the raw data by time, and calculating the distance and time interval between each data point; According to the longitude and latitude data, the user's movement trajectory is obtained by calculating the distance between two data points. The calculation formula is d = R arccos (sin (φ1) sin (φ2) + cos (φ1) cos (φ2) cos (λ2-λ1)), where R is the radius of the earth, φ1, φ2 are the latitudes of the two points, and λ1, λ2 are the longitudes of the two points; Calculate the speed between each data point, the calculation formula is: Where d is the distance between two points, and Δt is the time difference between the two points; When the speed is lower than the threshold and the duration exceeds the predetermined time, it is determined as a stop point, the stop time of each stop point is calculated, and the environmental information near the stop point is obtained using the geographic information system; Based on the GPS signal strength and Wi-Fi signal information, combined with the location database, it is determined whether the user is indoors or outdoors, and POI data is used to identify whether the user is in a public place; Combine the stay point data with the POI database to identify whether the user has entered a specific location and record the user's stay time and frequency at these locations; Statistics on the user's stay time and visit frequency at various locations, and calculate the user's preference index for each location. The calculation formula is: Where P type Expressed as a preference index, is the user’s stay time at a specific type of location, T stay,totali The total dwell time of the user.

5. A method for recommending geographic location information according to claim 1, characterized in that: The steps of calculating the user's activity range using the minimum enclosing rectangle according to the user's movement trajectory data are as follows: Acquire the user's movement trajectory data, which includes timestamp, longitude and latitude information, clean the data, remove noise and outliers, and arrange the location point data in chronological order; Find the maximum and minimum longitude and latitude of all location points, and use the extracted boundary points to determine the four corner points of the minimum enclosing rectangle; Calculate the width of the rectangle using the east-west direction of the rectangle as the width; Take the north-south direction of the rectangle as the height, calculate the height of the rectangle, and multiply the width of the rectangle by the height of the rectangle to get the area; Draw a minimum bounding rectangle on the map and mark all the user's location points.

6. A method for recommending geographic location information according to claim 1, characterized in that: The steps of building a personalized recommendation model by combining the user's behavior data and geographic location data through machine learning algorithms are as follows: Step 1: Collect historical user behavior data, geographic location data, and environmental data; Step 2: Calculate the preference index, user activity range, and average stay time for each location type based on historical user behavior data, geographic location data, and environmental data; Step 3: According to the recommendation task, define the target label as the recommended boarding point and construct label data, which is the actual selection of the user in the historical data; Step 4: Divide the data set into training set and test set according to the set ratio; Step 5: Use the random forest algorithm to train the model using the training set data and adjust the model parameters to optimize performance; Step 6: Use the test set data to evaluate the model performance, calculate the model's accuracy, precision, recall rate and F1 value, conduct a comprehensive evaluation of the model performance to obtain the model performance index, and compare the model performance index with the preset threshold. If the model performance index is greater than the preset threshold, output the current model as the personalized recommendation model. If the model performance index is less than the preset threshold, return to step 5 for parameter optimization until the model performance index is greater than the preset threshold.

7. A method for recommending geographic location information according to claim 1, characterized in that: The steps of generating address location recommendations that meet user needs are: Obtain the user's current geographic location, collect real-time traffic condition data, the traffic condition data includes road congestion and traffic accidents, and obtain the user's time constraint information, the time constraint information includes the appointment time and the preferred arrival time; Clean and organize data to ensure its accuracy and completeness, and convert real-time data into standard formats; Collect users’ historical location data, places visited, search queries, social media following, and online reviews to generate user preference models; Use collaborative filtering algorithms to find users with similar behaviors to the current user and recommend places visited by these users, and recommend places similar to places visited by the current user; Recommend places that meet the user's needs in real time based on the user's current location, time constraints, and traffic conditions. Use user preference models and collaborative filtering results to personalize the recommended place list. Use a Gaussian mixture model to generate a probability distribution of places that users may visit and determine the most likely recommended places; Use the Dijkstra algorithm, combined with real-time traffic data, to calculate the optimal path from the user's current location to the recommended location and select the fastest route.

8. A device for recommending geographic location information, characterized in that: The device comprises: A data collection module is used to collect order information and user information, and transmit the order information and user information to the location information recommendation module; The model building module is used to collect historical user information and build a personalized recommendation model, and transmit the personalized recommendation model to the location information recommendation module; The location information recommendation module is used to receive the order information, user information and personalized recommendation model transmitted by the data collection module and the model building module, import the information into the model to obtain the recommended location information, and transmit the recommended location information to the map display module; The map display module is used to receive the recommended location information transmitted by the location information recommendation module and display it on the map.

Images