A method and device for obtaining user travel tags

Abstract

The present application provides a method, device, electronic device and computer program product for obtaining user travel labels, and relates to the field of data analysis technology. The method includes: performing data cleaning on signaling data based on specific information; generating and classifying travel chains according to user trajectory data, and calculating the information entropy of each type of travel chain list; extracting a travel chain from each travel chain list and combining it with the corresponding information entropy to generate a travel chain information entropy list; when it is determined that the travel regularity index of the target user meets the standard, the travel chain information entropy list is matched with the map point of interest classification table to obtain the travel label of the target user. The embodiment of the present application cleans the acquired signaling data based on specific information, and filters the user's travel information based on the extracted user travel regularity characteristics, thereby matching travel labels that can accurately reflect the user's travel habits.

Description

Technical Field

[0001] The present application relates to the field of data analysis technology, and specifically to a method, device, electronic device, and computer program product for obtaining user travel tags. Background Art

[0002] A user's travel tags reflect their travel patterns and are helpful in areas such as urban planning, epidemic prevention, traffic scheduling, and precision marketing. Signaling data, or mobile phone signaling data, is captured and recorded by the operator's communication base stations when a mobile phone user makes a call, sends a text message, or moves around. This data includes the user's current latitude and longitude, the time they connect to or leave the base station, and other information. It features strong real-time performance, high accuracy, and wide coverage. Identifying user travel patterns through signaling data is currently the mainstream technology, but the technologies used vary, and the results can vary significantly.

[0003] Existing user travel identification solutions suffer from inaccurate identification results due to incomplete signaling data cleansing. This is because when users move near the intersection of multiple base stations, multiple signaling data records repeatedly appear, resulting in a "ping-pong effect" in their travel trajectories, which negatively impacts the identification of travel results. Furthermore, existing solutions simply match user travel tags directly to signaling data without evaluating the regular characteristics of signaling data, resulting in inaccurate user travel identification results. Summary of the Invention

[0004] The embodiments of the present application provide a method, device, electronic device, and computer program product for obtaining user travel tags, so as to solve the problem of inaccurate user travel identification in the prior art.

[0005] In a first aspect, an embodiment of the present application provides a method for obtaining a user travel tag, comprising:

[0006] Acquire signaling data of a target user within a preset time, perform data cleaning on the signaling data according to specific information in the signaling data, and generate target user trajectory data; wherein the specific information includes at least one of the time spent at a base station and the distance between adjacent base stations;

[0007] Generate a number of travel chains according to the user trajectory data, classify all travel chains into corresponding travel chain lists based on the initial points of each travel chain, and calculate the information entropy of each travel chain list;

[0008] Extracting the trip chain that appears the most times in each of the trip chain lists and combining it with the information entropy of the trip chain list to form a trip chain information entropy entry, and then combining all the trip chain information entropy entries into a trip chain information entropy list;

[0009] When it is determined according to the travel chain information entropy list that the travel regularity index of the target user exceeds the preset index threshold, each travel chain information entropy entry in the travel chain information entropy list is matched with the map interest point classification table to obtain the travel label of the target user.

[0010] In one embodiment, the performing data cleaning on the signaling data according to specific information in the signaling data includes:

[0011] Acquire signaling data of the target user within a preset time, and delete corresponding signaling record entries in the signaling data whose stay time at the base station is less than a preset time threshold;

[0012] Two signaling record entries whose distance between adjacent base stations is less than a preset distance threshold are merged into one signaling record entry.

[0013] In one embodiment, merging two signaling record entries whose distance between adjacent base stations is less than a preset distance threshold into one signaling record entry includes:

[0014] When it is determined based on the adjacent base station spacing information that the distance between two adjacent signaling record entries is less than a preset distance threshold, the corresponding signaling record entry with the smaller base station residence time is merged into the corresponding signaling record entry with the larger base station residence time based on the base station residence time information of the two signaling record entries;

[0015] Among them, the base station entry time of the merged signaling record entry is updated to the minimum value of the base station entry time in the two signaling record entries, the base station departure time of the merged signaling record entry is updated to the maximum value of the base station departure time in the two signaling record entries, and the base station stay time of the merged signaling record entry is updated to the sum of the base station stay time in the two signaling record entries.

[0016] In one embodiment, the travel tag of the target user includes travel habit information; wherein the travel habit information is generated based on base station access time information retained in the travel chain corresponding to the travel tag of the target user.

[0017] In one embodiment, matching each travel chain information entropy entry in the travel chain information entropy list with a map point of interest classification table to obtain the travel label of the target user includes:

[0018] The travel chain information entropy entries whose information entropy is less than a preset first entropy value are screened out from the travel chain information entropy list and matched with the map interest point classification table to obtain the travel label of the target user.

[0019] In one embodiment, the step of filtering out travel chain information entropy entries with information entropy less than a preset first entropy value from the travel chain information entropy list and matching the obtained entries with a map point of interest classification table to obtain the travel label of the target user includes:

[0020] A merging operation is performed on the trip chain information entropy list to obtain a classified information entropy list, wherein the merging operation includes merging multiple travel chain information entropy entries with the same trip chain in the trip chain information entropy list into one travel chain information entropy entry; wherein the information entropy of the merged travel chain information entropy entry is equal to the average information entropy of the multiple travel chain information entropy entries before the merging;

[0021] The travel chain information entropy entries whose information entropy is less than a preset first entropy value are screened out from the classified information entropy list and matched with the map interest point classification table to obtain the travel label of the target user.

[0022] In one embodiment, the travel regularity index is equal to the proportion of travel chain information entropy entries whose information entropy is less than a preset second entropy value to all travel chain information entropy entries in the travel chain information entropy list.

[0023] In a second aspect, an embodiment of the present application provides a device for obtaining a user travel tag, comprising:

[0024] a data cleaning module, configured to obtain signaling data of a target user within a preset time, clean the signaling data according to specific information in the signaling data, and generate target user trajectory data; wherein the specific information includes at least one of the time spent at a base station and the distance between adjacent base stations;

[0025] an entropy calculation module, configured to generate a plurality of travel chains based on the user trajectory data, classify all travel chains into corresponding travel chain lists based on the initial points of each travel chain, and calculate the information entropy of each travel chain list;

[0026] a list generation module, configured to extract the trip chain that appears the most times in each of the trip chain lists and combine it with the information entropy of the trip chain list to form a trip chain information entropy entry, and then combine all the trip chain information entropy entries into a trip chain information entropy list;

[0027] The label acquisition module is used to match each travel chain information entropy entry in the travel chain information entropy list with the map interest point classification table to obtain the travel label of the target user when it is determined that the travel regularity index of the target user exceeds the preset index threshold according to the travel chain information entropy list.

[0028] In a third aspect, an embodiment of the present application provides an electronic device, comprising a processor and a memory storing a computer program, wherein when the processor executes the program, the steps of the method for obtaining user travel tags described in the first aspect are implemented.

[0029] In a fourth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method for obtaining user travel tags described in the first aspect.

[0030] The user travel label acquisition method, device, electronic device and computer program product provided in the embodiments of the present application clean the acquired signaling data based on specific information and filter the user's travel information based on the extracted user travel pattern characteristics, thereby matching travel labels that can accurately reflect the user's travel habits. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the present application or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0032] Figure 1 This is one of the flow charts of the method for obtaining user travel tags provided in an embodiment of the present application;

[0033] Figure 2 This is the second flow chart of the method for obtaining user travel tags provided in an embodiment of the present application;

[0034] Figure 3 This is one of the structural diagrams of the user travel tag acquisition device provided in an embodiment of the present application;

[0035] Figure 4 It is a structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0036] To make the objectives, technical solutions, and advantages of this application more clear, the technical solutions in this application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this application.

[0037] It's important to note that existing methods for user travel identification don't completely clean data. Because users are likely to be located at the intersection of multiple base stations, even though they're in one location, their signaling data can appear repeatedly at multiple nearby base stations. This "ping-pong effect" can significantly impact travel identification results.

[0038] In addition, for a user (foreign tourist), there are often only one or a few user travel trajectories, which are not periodic. With this amount of data, the travel of individual users is likely to be accidental due to the user's own unobservable reasons. For example, accidental events such as meeting other tourists at a shopping mall near a tourist attraction cannot be used as effective trajectory features. If the time characteristics of the user's travel are not taken into account, the stop locations are directly extracted without considering the stop time. For example, some users like to go to the vegetable market to buy vegetables in the morning, while some users like to buy vegetables in the evening. This leads to insufficient interpretability of the results and poor application effect. At the same time, not all users' travel is regular and can be identified and interpreted. Therefore, if such users are not cleaned up, the overall recognition effect will be affected to a certain extent.

[0039] Furthermore, some existing solutions primarily focus on identifying tourists' identities and travel methods, but fail to extract relevant tags for user travel preferences, such as preferred destinations. For example, some solutions emphasize identifying the travel destinations of users who meet certain criteria within a large sample size, but fail to address the specific travel preferences of individual users. For example, some users may prefer to travel from their residential area to a scenic spot in the morning due to their personal preferences, while the vast majority of users do not. Therefore, the user's travel destination cannot be identified using existing identification rules, and there is no way to generate a travel tag for that user.

[0040] In order to solve the above technical problems, the embodiment of the present application proposes the following method for obtaining user travel tags.

[0041] Figure 1 Method for obtaining user travel tags. Figure 1 , an embodiment of the present application provides a method for obtaining a user travel tag, which may include:

[0042] S1. Acquire signaling data of a target user within a preset time, clean the signaling data according to specific information in the signaling data, and generate target user trajectory data; wherein the specific information includes at least one of the time spent at a base station and the distance between adjacent base stations;

[0043] S2. Generate several travel chains based on the user trajectory data, classify all travel chains into corresponding travel chain lists based on the initial points of each travel chain, and calculate the information entropy of each travel chain list;

[0044] S3. Extract the trip chain that appears the most times in each of the trip chain lists and combine it with the information entropy of the trip chain list to form a trip chain information entropy entry, and then combine all the trip chain information entropy entries into a travel chain information entropy list;

[0045] S4. When it is determined that the travel regularity index of the target user exceeds a preset index threshold according to the travel chain information entropy list, each travel chain information entropy entry in the travel chain information entropy list is matched with a map point of interest classification table to obtain a travel label of the target user.

[0046] It should be noted that the embodiment of the present application first performs data cleaning on the acquired signaling data based on specific information. Specifically, the signaling data may include one or more of the base station dwell time information and the adjacent base station spacing information. Through this data cleaning method, the problem of inaccurate recognition results caused by the "ping-pong effect" of trajectory data when the user is active near the intersection of multiple base stations is overcome. After data cleaning, the target user trajectory data is generated, wherein the format of the user trajectory data may include the following information: user ID, base station ID, base station access time, base station departure time, base station dwell time, base station latitude and longitude, etc.; in addition, the entries of the user trajectory data may be sorted according to the base station access time and divided according to natural days.

[0047] Then, two adjacent user trajectory data are combined to generate a travel chain. For example, the data with labels 1 and 2 form travel chain 1, the data with labels 2 and 3 form travel chain 2, and so on. The format of the travel chain can be "<key1,keyx> ”, where key1 is the ID of the starting base station of the trip chain, and keyx is the ID of the base station of the next trajectory record of this starting base station. After all travel chains are combined, they are classified based on the starting base station of each travel chain. Travel chains with the same starting base station are classified into the same travel chain list, and the information entropy of each travel chain list is calculated.

[0048] Then, the trip chain with the most occurrences in each trip chain list is extracted and combined with the information entropy of the trip chain list to form a trip chain information entropy entry. Then, all trip chain information entropy entries are combined into a trip chain information entropy list. If there are multiple trip chains in the same trip chain list that have the most occurrences, then all of them are extracted and combined into trip chain information entropy entries.

[0049] Then calculate the travel pattern index of the travel chain information entropy list to determine whether the user's travel data pattern meets the requirements. If so, perform travel label matching. If not, end the step.

[0050] The user travel label acquisition method provided in the embodiment of the present application cleans the acquired signaling data based on specific information, simplifies the user's travel trajectory, thereby effectively improving the accuracy of user travel identification; in addition, the embodiment of the present application obtains the user's travel regularity index based on information entropy, and filters the user's travel information according to the extracted user travel regularity index, so as to match travel labels that can accurately reflect the user's travel habits, thereby effectively improving the accuracy of user travel identification as a whole.

[0051] In one embodiment, the performing data cleaning on the signaling data according to specific information in the signaling data may include:

[0052] Acquire signaling data of the target user within a preset time, and delete corresponding signaling record entries in the signaling data whose stay time at the base station is less than a preset time threshold;

[0053] Two signaling record entries whose distance between adjacent base stations is less than a preset distance threshold are merged into one signaling record entry.

[0054] It should be noted that data cleaning methods may include: filtering out signaling records whose dwell time at a base station is less than a preset time threshold based on the signaling data; and merging two records whose dwell time is less than a preset threshold into a single record based on the distance between two adjacent base stations. It is understood that the order of these two data cleaning steps can be merging first and then filtering, or filtering first and then merging.

[0055] The user travel label acquisition method provided in the embodiment of the present application can overcome the factors affecting the travel identification results by performing data cleaning based on the base station residence time and the distance between adjacent base stations of the signaling data, thereby avoiding the "ping-pong effect" phenomenon in the user trajectory, thereby effectively improving the efficiency and accuracy of user travel identification.

[0056] In one embodiment, merging two signaling record entries whose distance between adjacent base stations is less than a preset distance threshold into one signaling record entry may include:

[0057] When it is determined based on the adjacent base station spacing information that the distance between two adjacent signaling record entries is less than a preset distance threshold, the corresponding signaling record entry with the smaller base station residence time is merged into the corresponding signaling record entry with the larger base station residence time based on the base station residence time information of the two signaling record entries;

[0058] Among them, the base station entry time of the merged signaling record entry is updated to the minimum value of the base station entry time in the two signaling record entries, the base station departure time of the merged signaling record entry is updated to the maximum value of the base station departure time in the two signaling record entries, and the base station stay time of the merged signaling record entry is updated to the sum of the base station stay time in the two signaling record entries.

[0059] It should be noted that when merging two records less than a preset threshold into one record based on the distance between two adjacent base stations, the above-mentioned method can be used. In the process of merging the two records, the record with the smaller base station dwell time can be deleted, and then the record with the larger base station dwell time can be retained and some information can be modified. For example, the base station entry time of the retained record can be modified to the base station entry time of the deleted record, the base station dwell time of the retained record can be modified to the sum of the base station dwell time of the two signaling records, and the longitude and latitude of the retained record can be modified to the midpoint longitude and latitude of the base station of the two signaling records.

[0060] The user travel tag acquisition method provided in the embodiment of the present application cleans the data according to the distance between adjacent base stations, merges two base stations with similar distances into one signaling record, thereby streamlining the data and avoiding adverse effects on the recognition results, thereby further improving the efficiency and accuracy of user travel recognition.

[0061] In one embodiment, the travel tag of the target user includes travel habit information; wherein the travel habit information is generated based on base station access time information retained in the travel chain corresponding to the travel tag of the target user.

[0062] It should be noted that when generating a trip chain, time information (base station access time information) can be included in the trip chain expression. This time can be the start or end time of the trip chain, or it can be calculated based on the access and departure time information of the two base stations in the trip chain according to certain rules. This time information is used to represent the time at which the user was during the trip chain. This time information can be retained throughout the calculation process (trip chain - trip chain list - trip chain information entropy entry) until the corresponding trip label is matched. In specific applications, this time information can be generalized, for example, matching the specific time to a certain time period (e.g., 0-6 am, 6-12 pm, etc.), and using this generalized time information as the user's travel habit information. It should be noted that this generalization operation can be performed at any point in the above calculation process, for example, it can be generalized into travel habit information when generating the trip chain, or it can be generalized when matching the user's travel label.

[0063] The user travel label acquisition method provided in the embodiment of the present application records the time of the user's travel chain to represent the user's travel habit information and reflects it in the user's travel label, thereby effectively improving the interpretability and applicability of the user travel identification results.

[0064] In one embodiment, matching each travel chain information entropy entry in the travel chain information entropy list with a map point of interest classification table to obtain the travel label of the target user includes:

[0065] The travel chain information entropy entries whose information entropy is less than a preset first entropy value are screened out from the travel chain information entropy list and matched with the map interest point classification table to obtain the travel label of the target user.

[0066] It should be noted that after determining that the user's travel regularity indicators meet the standards, the travel chain information entropy entries can be matched with the map point of interest classification table. The travel chain information entropy entries contain the IDs of two base stations. The map point of interest classification table (POI area classification table) stores the area type corresponding to each base station. During matching, each travel chain information entropy entry can be matched to generate a travel label. For example, the format can be <catering and shopping → health care>. It should be noted that all travel chain information entropy entries can be matched into travel labels, but in order to make the travel labels more representative, travel chain information entropy entries in the travel chain information entropy list with information entropy less than the preset entropy value can be selected for matching (the smaller the information entropy, the higher the certainty of the information, which is equivalent to higher regularity).

[0067] The user travel label acquisition method provided in the embodiment of the present application uses information entropy to match more regular travel chains as user travel labels, thereby effectively improving the interpretability and applicability of travel identification results, and further improving the accuracy of travel identification as a whole.

[0068] In one embodiment, the step of filtering out travel chain information entropy entries with information entropy less than a preset first entropy value from the travel chain information entropy list and matching the obtained entries with a map point of interest classification table to obtain the travel label of the target user includes:

[0069] A merging operation is performed on the trip chain information entropy list to obtain a classified information entropy list, wherein the merging operation includes merging multiple travel chain information entropy entries with the same trip chain in the trip chain information entropy list into one travel chain information entropy entry; wherein the information entropy of the merged travel chain information entropy entry is equal to the average information entropy of the multiple travel chain information entropy entries before the merging;

[0070] The travel chain information entropy entries whose information entropy is less than a preset first entropy value are screened out from the classified information entropy list and matched with the map interest point classification table to obtain the travel label of the target user.

[0071] It should be noted that before matching trip labels, the trip chain entropy list can be merged to merge multiple trip chain entropy entries with the same trip chain into one trip chain entropy entry. The entropy of the merged trip chain entropy entry is then updated to the average entropy of the multiple trip chain entropy entries before the merge. Trip label matching is then performed after the merge operation.

[0072] The user travel label acquisition method provided in the embodiment of the present application avoids the problem of multiple repeated travel labels in the travel identification results by streamlining the travel chain information entropy entries, thereby further improving the accuracy and effectiveness of travel identification.

[0073] In one embodiment, the travel regularity index is equal to the proportion of travel chain information entropy entries whose information entropy is less than a preset second entropy value to all travel chain information entropy entries in the travel chain information entropy list.

[0074] It should be noted that the travel regularity index can be represented by the ratio of travel chain information entropy entries in the travel chain information entropy list whose information entropy is less than the preset second entropy value to all travel chain information entropy entries. When it is judged that this travel regularity index exceeds the preset index threshold, travel label matching is performed. If the index does not meet the standard, travel label matching is not performed.

[0075] The user travel label acquisition method provided in this application determines whether to match travel labels based on the user's travel pattern indicators, so that the matched travel labels can more accurately reflect the user's actual travel habits, thereby effectively improving the accuracy, explainability and applicability of user travel identification as a whole.

[0076] See Figure 2 Based on the above solution, in order to better understand the method for obtaining user travel tags provided in the embodiment of the present application, the following is a detailed description:

[0077] In response to the technical problems raised by the above-mentioned background technology, an embodiment of the present application provides a method for obtaining user travel labels, which obtains user travel labels through mobile signaling data. Specifically, the stop data is obtained through signaling data, and the user's stop data is cleaned and merged to obtain a streamlined user trajectory; then, based on the user's daily user trajectory over a period of time, "travel chains" are constructed with two adjacent nodes, and the information entropy of each travel chain is calculated as a reflection of the regularity of the user's travel. The travel pattern that combines nodes and time is regarded as a strong pattern, and the calculation result without combining time is regarded as a weak pattern. Finally, the travel pattern label is output according to the travel pattern indicator, and the user's travel preference label is output in combination with POI data. According to the technical implementation content, it can be divided into three main steps:

[0078] (1) Extraction of residents’ travel trajectories based on signaling data

[0079] Telecommunications carriers' signaling data provides comprehensive coverage of user movement trajectories. This step first extracts user stop data over a period of time from the carrier's signaling data. This data is then cleaned daily to form a user trajectory. Specifically, trajectory cleaning includes removing invalid stops and merging adjacent stops. Stops are then sorted chronologically. By counting the user's longest stops, the user's daily stop loop is identified. This outputs the user's trajectory.

[0080] (2) Extraction of residents’ travel patterns based on information entropy

[0081] Based on the user trajectories found, adjacent stop points are grouped into travel chains. The travel chains are then classified based on the initial point of each link, and the information entropy of each type of travel chain is calculated. The information entropy results of travel chains that do not consider the "time period" factor can be considered weak regularities, while the information entropy results of travel chains that consider the "time period" factor can be considered strong regularities. A list of travel chains with strong and weak travel regularities is output.

[0082] (3) User travel label extraction based on travel patterns

[0083] Based on the user's travel chain-information entropy list, the user's travel pattern strength (index) is calculated and the user's travel pattern label is classified based on the pattern strength. The user's travel chain-information entropy list with weak travel pattern labels and strong travel pattern labels is matched to POIs through the map API program. The user's travel chain is replaced with the POI classification to form a travel classification-information entropy list. Finally, based on the information entropy strength, the user's travel preference label is output.

[0084] The following is a specific example of the embodiment of the present application:

[0085] 1. Extraction of Resident Travel Trajectory Based on Signaling Data

[0086] The operator's signaling data is generated from the communication records between residents' mobile phones and base stations. Since various programs in users' mobile phones use data all the time, the operator's signaling data fully records the residents' activities and stops. The signaling data sample is shown in Table 1:

[0087] Table 1 Signaling data sample

[0088] User ID Enter the base station Leave the base station Base station ID Base station latitude and longitude 136****0079 20210705102530 20210705105423 21**z7 (126.68296，45.77581) 136****0079 20210705105423 20210705114109 83**k2 (126.68312，45.77577)

[0089] Through the signaling data of residents, the user's activity trajectory can be cleaned up. The specific method is as follows:

[0090] Step 1: Use midnight as the time division and input the user's daily signaling data.

[0091] Step 2: Sort the time of entry into the base station in the user's signaling data in ascending order.

[0092] Step 3: Add a new column "Dwell time", dwell time = time of leaving base station - time of entering base station.

[0093] Step 4: Delete the signaling records with a stay time of less than 30 minutes.

[0094] Step 5: Calculate the distance L between base stations for each adjacent record in the table. If L is less than 500 meters, merge the records. Merge by merging the record with the shorter dwell time into the record with the longer dwell time. The entry time is the minimum of the two values, the exit time is the maximum of the two values, and the dwell time is the sum of the two values. The base station ID and longitude and latitude are the same as the one with the longest dwell time. After merging, delete the original record.

[0095]

[0096] in,<Lng1,Lat1> ,<Lng2,Lat2> are the latitude and longitude of the base stations in the two records respectively, and 0.000899 and 0.001141 are the coefficients for converting the latitude and longitude units into meters respectively.

[0097] Step 6: Traverse the base station IDs of the cleaned records, count the user's stay time at each base station, and find the base station with the longest stay. Then, based on this base station ID, find the record with the earliest access time in the daily signaling table and mark it as the starting point of the daily trajectory.

[0098] Step 7: Input the user's signaling data for 3 to 6 months, and repeat steps 1 to 6 to form a trajectory list TA;

[0099] The characteristic of this step is that the signaling data is cleaned and the "ping-pong effect" is removed by merging between base stations, and then converted into a user trajectory. The trajectory is directed and has a set starting point.

[0100] 2. Extraction of Residents’ Travel Patterns Based on Information Entropy

[0101] (1) Extraction of user weak travel patterns based on information entropy

[0102] The travel chain is derived from the extracted user trajectories and is the basis for calculating the residents' travel patterns. The method for extracting the user's weak travel patterns based on the travel chain is as follows:

[0103] Step 1: Input a user's daily movement trajectory data TA for a period of time.

[0104] Step 2: Count the base station key1 with the most occurrences of the starting point of the trajectory in the user's daily trajectory data.

[0105] Step 3: Find the base station keyx (x represents any base station) of the next trajectory record that appears in the base station key1 record from all daily trajectories to form a travel chain "<key1,keyx> ”.

[0106] Step 4: Form a travel chain list ST{<key1,key2> ,<key1,key3> ,<key1,keyx> …};

[0107] Step 5: Select base stations that have not been calculated and repeat steps 2 to 4 until a new list ST cannot be generated.

[0108] Step 6: Calculate the entropy value for all STs:

[0109] entropy(key1,keyx)=-∑P key1,keyx ln(P key1,keyx ) (2)

[0110] Pkey1,keyx For travel chain<key1,keyx> The probability of appearing in ST.

[0111] Step 7: Keep P in each type of ST key1,keyx The largest travel chain generates a travel chain-information entropy list:

[0112] SE{<key1,key2,entropy(key1,key2)> ,...}

[0113] (2) Extraction of user-driven travel patterns based on information entropy

[0114] Weak travel patterns do not take into account the time of the user's travel, so the key to extracting strong travel patterns is to consider the travel time. The method for extracting strong travel patterns of users combined with travel chains is as follows:

[0115] Step 1: Input a user's daily movement trajectory data TA for a period of time.

[0116] Step 2: Count the base station key1 with the most occurrences of the starting point of the trajectory in the user's daily trajectory data.

[0117] Step 3: Find the base station keyx (x represents any base station) of the next trajectory record that appears in the base station key1 record from all daily trajectories, and the access time T of keyx, to form a travel chain "<key1,keyx,G(T)> ”.

[0118] Where T is the time when the user accesses the base station, and the G(T) function is a generalization of the travel time. The generalization rules are as follows:

[0119]

[0120] Step 4: Form a travel chain list:

[0121] ST{<key1,key2,G(T)> ,<key1,key3,G(T)> ,<key1,keyx,G(T)> …}

[0122] Step 5: Select the base stations that have not been calculated and repeat steps 2 to 4 until a new list ST cannot be generated;

[0123] Step 6: Calculate the entropy value for all STs:

[0124] entropy(key1,keyx,G(T))=-∑P key1,keyx,G(T) ln(P key1,keyx,G(T) ) (4)

[0125] P key1,keyx,G(T)For travel chain<key1,keyx,G(T)> The probability of appearing in ST.

[0126] Step 7: Keep P in each type of ST key1,keyx The largest travel chain generates a travel chain-information entropy list:

[0127] SE{<key1,key2,G(T),entropy(key1,key2,G(T))> ,...}

[0128] 3. User travel label extraction based on travel patterns

[0129] The extraction of user travel tags is based on the user's travel chain - information entropy list SE. The specific rules are as follows:

[0130] Step 1: Calculate the user travel regularity index Sw:

[0131]

[0132] The denominator is the length of the list SE, and the numerator is the number of trip chains in the list SE that satisfy an information entropy less than 0.5 (or other thresholds).

[0133] Step 2: Match the user's travel pattern tag #TagSw based on the indicator Sw. The matching method is shown in Table 2:

[0134] Table 2 User travel pattern label extraction matching table

[0135] Sw [0,0.3] (0.3,0.5] (0.5,1.0] #TagSw Irregular travel Weak travel patterns Strong travel rules

[0136] Step 3: Based on the travel pattern tag #TagSw results, if the weak travel pattern is met, proceed to the next step, otherwise terminate.

[0137] Step 4: Remove the trip chains with entropy ≥ 0.5 (can be other values) in SE.

[0138] Step 5: The specific base station address (latitude and longitude,<Lng1,Lat1> ) Query the closest POI table (map point of interest classification table) through the map API interface, and use the classification of the POI to replace the travel chain.

[0139] Table 3 POI classification

[0140] Serial number 1 2 3 4 POI tags Dining and Shopping scenic spots Public Services Enterprise Company Serial number 5 6 7 8 POI tags Leisure and Entertainment School Science and Education Healthcare Accommodation

[0141] Such as: SE{<key1,key2,entropy(key1,key2)> ,...} is transformed into:

[0142] SEP{<catering and shopping, medical care, entropy(key1, key2)>, ...}

[0143] Step 6: After converting SE into SEP, the same trip chains are merged, that is, the information entropy of the same trip chains is averaged to form a trip classification-information entropy list SQ.

[0144] Step 7: Output the travel chain in SQ whose information entropy is less than 0.5 (or other values) as the travel label of the user.

[0145] The output format is <catering and shopping→health care>

[0146] Step 8: Repeat the above steps for all users.

[0147] Note that the steps for extracting user travel tags for strong and weak regularities are the same, and the output results are also similar. The difference is that the travel tags for strong regularities have a time range (travel habit information).

[0148] The following lists specific values ​​for illustration. An example of the process of extracting some travel tags for a user is as follows:

[0149] Part of the daily movement trajectory table TA of a certain user is shown in Table S1

[0150] Table S1 Part of the daily movement trajectory of a user TA

[0151]

[0152] According to the residents’ strong travel pattern extraction module based on information entropy, the user’s travel chain can be obtained:

[0153] ST{<21A2z7, 83B5k2, 2>, <21A2z7, 83B5k2, 2>},

[0154] ST{<83B5k2, 32Ta38, 3>, <83B5k2, 32Ta38, 3>, <83B5k2, 21A2z7, 3>},

[0155] ST{<32Ta38,83B5k2,3>}

[0156] According to formula (4) and related steps, the corresponding travel chain-information entropy list can be obtained:

[0157] SE{<21A2z7, 83B5k2, 2, 0>, <83B5k2, 32Ta38, 3, 0.6063>, <32Ta38, 83B5k2, 3, 0>}

[0158] According to formula (5) and the travel chain-information entropy list, the overall travel regularity index Sw = 2 / 3 = 0.67 is obtained. According to Table 2, the user's travel regularity tag #TagSw = strong travel regularity;

[0159] According to the map API (map point of interest classification table), the POI corresponding to the base station in the SE table is obtained:

[0160] Table S2 Base station address query map POI results

[0161] Base station ID Base station latitude and longitude POI classification 21A2z7 (126.68296，45.77581) Accommodation 83B5k2 (126.68312，45.77577) School Science and Education 32Ta38 (126.68323，45.77565) Dining and Shopping

[0162] Remove <83B5k2, 32Ta38, 3, 0.6063> and convert SE to SEP:

[0163] SEP{<Accommodation, School, Science and Education, 2, 0>, <Food and Beverage, School, Science and Education, 3, 0>}

[0164] Since merging is not required, the travel classification-information entropy SQ is:

[0165] SQ{<Accommodation, School, Science and Education, 2, 0>, <Food and Beverage, School, Science and Education, 3, 0>}

[0166] The output travel label of user 136****0079 is:

[0167] <Accommodation, School, Science and Education, 2>, <Food and Beverage, School, Science and Education, 3>

[0168] The travel preference is explained as follows: users usually go from the residential area to the school science and education area between 6-12 in the morning; users usually go from the dining and shopping area to the school science and education area between 12-18 in the afternoon.

[0169] The key points of the method for obtaining user travel tags provided in the embodiment of the present application include:

[0170] 1) Signaling data cleaning based on spatiotemporal relationships: When cleaning signaling data, we optimize the problem of multiple occurrences of signaling data under multi-base station coverage. We combine the time when the signaling data is generated with the spatial location of the base station and merge and clean adjacent signaling data. This cleaning takes into account the spatiotemporal characteristics of the signaling data.

[0171] 2) Calculation of user travel patterns based on information entropy: The user's "travel chain" is extracted through the user trajectory, and the user's "travel chain" is calculated through information entropy to obtain the user's travel habits. Then, travel patterns are extracted based on the intensity of the user's travel habits.

[0172] 3) User travel preference label extraction method combined with POI: After calculating the travel chain based on information entropy, the travel chain and POI are combined to convert the travel chain into the user's travel preference label.

[0173] Compared with the prior art, the embodiments of the present application have the following advantages:

[0174] 1) To address the "ping-pong effect" problem with travel trajectories in existing technologies, signaling data is cleaned. In addition to considering the duration of a user's stops, the present embodiment also combines similar signaling data based on the access time and geographic proximity of base stations in the signaling data. This approach, to a certain extent, addresses the issue of recurring user signaling data in areas covered by multiple base stations. It also streamlines the user's travel trajectory to a certain extent, reducing noise and improving the efficiency and accuracy of travel identification.

[0175] 2) Calculating patterns based on user travel time provides highly interpretable and applicable results. This embodiment of the application employs two methods for calculating user travel patterns: weak patterns and strong patterns. The strong pattern calculation considers the timeliness of user travel. Time-sensitive user travel identification can more accurately reflect the purpose of a user's travel and provide more information to explain their travel behavior.

[0176] 3) Ability to extract travel tags for individual users. This embodiment first calculates the user's travel patterns to determine whether the user has distinct travel characteristics. After further eliminating invalid users, it combines POI region classification to extract specific tags based on the user's travel preferences. This tag can reflect the individual's travel habits.

[0177] The following describes a user travel tag acquisition device provided in an embodiment of the present application. The user travel tag acquisition device described below and the user travel tag acquisition method described above can refer to each other.

[0178] See Figure 3 , an embodiment of the present application provides a device for obtaining a user travel tag, comprising:

[0179] Data cleaning module 1 is used to obtain signaling data of a target user within a preset time, clean the signaling data according to specific information in the signaling data, and generate target user trajectory data; wherein the specific information includes at least one of the time information of the stay at the base station and the distance between adjacent base stations;

[0180] Entropy calculation module 2, used to generate a plurality of travel chains based on the user trajectory data, classify all travel chains into corresponding travel chain lists based on the initial points of each travel chain, and calculate the information entropy of each travel chain list;

[0181] List generation module 3 is used to extract the trip chain that appears the most times in each trip chain list and combine it with the information entropy of the trip chain list to form a trip chain information entropy entry, and then combine all the trip chain information entropy entries into a trip chain information entropy list;

[0182] The label acquisition module 4 is used to match each travel chain information entropy entry in the travel chain information entropy list with the map interest point classification table to obtain the travel label of the target user when it is determined that the travel regularity index of the target user exceeds the preset index threshold according to the travel chain information entropy list.

[0183] In one embodiment, the data cleaning module 1 is specifically configured to:

[0184] Acquire signaling data of the target user within a preset time, and delete corresponding signaling record entries in the signaling data whose stay time at the base station is less than a preset time threshold;

[0185] Two signaling record entries whose distance between adjacent base stations is less than a preset distance threshold are merged into one signaling record entry.

[0186] In one embodiment, merging two signaling record entries whose distance between adjacent base stations is less than a preset distance threshold into one signaling record entry includes:

[0187] When it is determined based on the adjacent base station spacing information that the distance between two adjacent signaling record entries is less than a preset distance threshold, the corresponding signaling record entry with the smaller base station residence time is merged into the corresponding signaling record entry with the larger base station residence time based on the base station residence time information of the two signaling record entries;

[0188] Among them, the base station entry time of the merged signaling record entry is updated to the minimum value of the base station entry time in the two signaling record entries, the base station departure time of the merged signaling record entry is updated to the maximum value of the base station departure time in the two signaling record entries, and the base station stay time of the merged signaling record entry is updated to the sum of the base station stay time in the two signaling record entries.

[0189] In one embodiment, the travel tag of the target user includes travel habit information; wherein the travel habit information is generated based on base station access time information retained in the travel chain corresponding to the travel tag of the target user.

[0190] In one embodiment, the tag acquisition module 4 is specifically configured to:

[0191] The travel chain information entropy entries whose information entropy is less than a preset first entropy value are screened out from the travel chain information entropy list and matched with the map interest point classification table to obtain the travel label of the target user.

[0192] In one embodiment, the tag acquisition module 4 is further configured to:

[0193] A merging operation is performed on the trip chain information entropy list to obtain a classified information entropy list, wherein the merging operation includes merging multiple travel chain information entropy entries with the same trip chain in the trip chain information entropy list into one travel chain information entropy entry; wherein the information entropy of the merged travel chain information entropy entry is equal to the average information entropy of the multiple travel chain information entropy entries before the merging;

[0194] The travel chain information entropy entries whose information entropy is less than a preset first entropy value are screened out from the classified information entropy list and matched with the map interest point classification table to obtain the travel label of the target user.

[0195] In one embodiment, the travel regularity index is equal to the proportion of travel chain information entropy entries whose information entropy is less than a preset second entropy value to all travel chain information entropy entries in the travel chain information entropy list.

[0196] It can be understood that the above-mentioned device embodiment corresponds to the method embodiment of the present application. The user travel label acquisition device provided by the embodiment of the present application can implement the user travel label acquisition method provided by any method embodiment of the present application.

[0197] Figure 4 An example of a physical structure diagram of an electronic device is shown below. Figure 4 As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communication interface 420, and the memory 430 communicate with each other via the communication bus 440. The processor 410 may call a computer program in the memory 430 to execute the steps of the method for obtaining a user travel tag, for example, including:

[0198] Acquire signaling data of a target user within a preset time, perform data cleaning on the signaling data according to specific information in the signaling data, and generate trajectory data of the target user; wherein the specific information includes at least one of the time information of the stay at the base station and the distance between adjacent base stations;

[0199] Generate a number of travel chains according to the user trajectory data, classify all travel chains into corresponding travel chain lists based on the initial points of each travel chain, and calculate the information entropy of each travel chain list;

[0200] Extracting the trip chain that appears the most times in each of the trip chain lists and combining it with the information entropy of the trip chain list to form a trip chain information entropy entry, and then combining all the trip chain information entropy entries into a trip chain information entropy list;

[0201] When it is determined according to the travel chain information entropy list that the travel regularity index of the target user exceeds the preset index threshold, each travel chain information entropy entry in the travel chain information entropy list is matched with the map interest point classification table to obtain the travel label of the target user.

[0202] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0203] On the other hand, embodiments of the present application further provide a computer program product, which includes a computer program. The computer program may be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can perform the steps of the user travel tag acquisition method provided in the above embodiments, for example, including:

[0204] Acquire signaling data of a target user within a preset time, perform data cleaning on the signaling data according to specific information in the signaling data, and generate trajectory data of the target user; wherein the specific information includes at least one of the time information of the stay at the base station and the distance between adjacent base stations;

[0205] Generate a number of travel chains according to the user trajectory data, classify all travel chains into corresponding travel chain lists based on the initial points of each travel chain, and calculate the information entropy of each travel chain list;

[0206] Extracting the trip chain that appears the most times in each of the trip chain lists and combining it with the information entropy of the trip chain list to form a trip chain information entropy entry, and then combining all the trip chain information entropy entries into a trip chain information entropy list;

[0207] When it is determined according to the travel chain information entropy list that the travel regularity index of the target user exceeds the preset index threshold, each travel chain information entropy entry in the travel chain information entropy list is matched with the map interest point classification table to obtain the travel label of the target user.

[0208] On the other hand, an embodiment of the present application further provides a processor-readable storage medium, wherein the processor-readable storage medium stores a computer program, wherein the computer program is configured to cause a processor to execute the steps of the methods provided in the above embodiments, for example, including:

[0209] Acquire signaling data of a target user within a preset time, perform data cleaning on the signaling data according to specific information in the signaling data, and generate trajectory data of the target user; wherein the specific information includes at least one of the time information of the stay at the base station and the distance between adjacent base stations;

[0210] Generate a number of travel chains according to the user trajectory data, classify all travel chains into corresponding travel chain lists based on the initial points of each travel chain, and calculate the information entropy of each travel chain list;

[0211] Extracting the trip chain that appears the most times in each of the trip chain lists and combining it with the information entropy of the trip chain list to form a trip chain information entropy entry, and then combining all the trip chain information entropy entries into a trip chain information entropy list;

[0212] When it is determined according to the travel chain information entropy list that the travel regularity index of the target user exceeds the preset index threshold, each travel chain information entropy entry in the travel chain information entropy list is matched with the map interest point classification table to obtain the travel label of the target user.

[0213] The processor-readable storage medium can be any available medium or data storage device that can be accessed by the processor, including but not limited to magnetic storage (such as floppy disks, hard disks, magnetic tapes, magneto-optical disks (MO)), optical storage (such as CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (such as ROMs, EPROMs, EEPROMs, non-volatile memories (NANDFLASH), solid-state drives (SSDs)), etc.

[0214] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0215] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0216] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for obtaining user travel tags, characterized in that: include: Acquire signaling data of a target user within a preset time, perform data cleaning on the signaling data according to specific information in the signaling data, and generate trajectory data of the target user; wherein the specific information includes at least one of the time information of the stay at the base station and the distance between adjacent base stations; Generate a number of travel chains according to the user trajectory data, classify all travel chains into corresponding travel chain lists based on the initial points of each travel chain, and calculate the information entropy of each travel chain list; Extracting the trip chain that appears the most times in each of the trip chain lists and combining it with the information entropy of the trip chain list to form a trip chain information entropy entry, and then combining all the trip chain information entropy entries into a trip chain information entropy list; When it is determined according to the travel chain information entropy list that the travel regularity index of the target user exceeds the preset index threshold, each travel chain information entropy entry in the travel chain information entropy list is matched with the map interest point classification table to obtain the travel label of the target user.

2. The method for obtaining user travel tags according to claim 1, characterized in that: The performing data cleaning on the signaling data according to specific information in the signaling data includes: Acquire signaling data of the target user within a preset time, and delete corresponding signaling record entries in the signaling data whose stay time at the base station is less than a preset time threshold; Two signaling record entries whose distance between adjacent base stations is less than a preset distance threshold are merged into one signaling record entry.

3. The method for obtaining user travel tags according to claim 2, characterized in that: The step of merging two signaling record entries whose distance between adjacent base stations is less than a preset distance threshold into one signaling record entry includes: When it is determined based on the adjacent base station spacing information that the distance between two adjacent signaling record entries is less than a preset distance threshold, the corresponding signaling record entry with the smaller base station residence time is merged into the corresponding signaling record entry with the larger base station residence time based on the base station residence time information of the two signaling record entries; Among them, the base station entry time of the merged signaling record entry is updated to the minimum value of the base station entry time in the two signaling record entries, the base station departure time of the merged signaling record entry is updated to the maximum value of the base station departure time in the two signaling record entries, and the base station stay time of the merged signaling record entry is updated to the sum of the base station stay time in the two signaling record entries.

4. The method for obtaining user travel tags according to claim 1, characterized in that: The travel tag of the target user includes travel habit information; wherein the travel habit information is generated based on base station access time information retained in the travel chain corresponding to the travel tag of the target user.

5. The method for obtaining user travel tags according to claim 1, characterized in that: The step of matching each travel chain information entropy entry in the travel chain information entropy list with a map point of interest classification table to obtain the travel label of the target user includes: The travel chain information entropy entries whose information entropy is less than a preset first entropy value are screened out from the travel chain information entropy list and matched with the map interest point classification table to obtain the travel label of the target user.

6. The method for obtaining user travel tags according to claim 5, characterized in that: The step of filtering out the travel chain information entropy entries whose information entropy is less than a preset first entropy value from the travel chain information entropy list and matching the entries with the map interest point classification table to obtain the travel label of the target user includes: A merging operation is performed on the trip chain information entropy list to obtain a classified information entropy list, wherein the merging operation includes merging multiple travel chain information entropy entries with the same trip chain in the trip chain information entropy list into one travel chain information entropy entry; wherein the information entropy of the merged travel chain information entropy entry is equal to the average information entropy of the multiple travel chain information entropy entries before the merging; The travel chain information entropy entries whose information entropy is less than a preset first entropy value are screened out from the classified information entropy list and matched with the map interest point classification table to obtain the travel label of the target user.

7. The method for obtaining user travel tags according to claim 1, characterized in that: The travel regularity index is equal to the proportion of travel chain information entropy entries in the travel chain information entropy list whose information entropy is less than a preset second entropy value to all travel chain information entropy entries.

8. A device for obtaining user travel tags, characterized in that: include: a data cleaning module, configured to obtain signaling data of a target user within a preset time, clean the signaling data according to specific information in the signaling data, and generate target user trajectory data; wherein the specific information includes at least one of the time spent at a base station and the distance between adjacent base stations; an entropy calculation module, configured to generate a plurality of travel chains based on the user trajectory data, classify all travel chains into corresponding travel chain lists based on the initial points of each travel chain, and calculate the information entropy of each travel chain list; a list generation module, configured to extract the trip chain that appears the most times in each of the trip chain lists and combine it with the information entropy of the trip chain list to form a trip chain information entropy entry, and then combine all the trip chain information entropy entries into a trip chain information entropy list; The label acquisition module is used to match each travel chain information entropy entry in the travel chain information entropy list with the map interest point classification table to obtain the travel label of the target user when it is determined that the travel regularity index of the target user exceeds the preset index threshold according to the travel chain information entropy list.

9. An electronic device comprising a processor and a memory storing a computer program, characterized in that: When the processor executes the computer program, the steps of the method for obtaining user travel tags according to any one of claims 1 to 7 are implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method for obtaining user travel tags according to any one of claims 1 to 7 are implemented.

Images