College student identification method and device, server and computer readable storage medium

Abstract

This application proposes a method, apparatus, server, and computer-readable storage medium for identifying college students, relating to the field of computer technology. It acquires signaling data of the user to be identified within a preset time range and usage data of multiple target applications; the preset time range is a short duration sufficient to characterize the spatiotemporal rhythms of college students; it constructs a spatiotemporal state sequence corresponding to the user to be identified based on the signaling data, and performs spatiotemporal context fusion on the usage data of each target application to obtain an application behavior score corresponding to the user to be identified; it determines a comprehensive score for the user to be identified based on the pre-constructed college student behavior modality, the spatiotemporal state sequence corresponding to the user to be identified, and the application behavior score; and it determines whether the user to be identified is a college student based on the comprehensive score. This improves the effective and accurate identification of newly added college students in a short period.

Description

Technical Field

[0001] This application relates to the field of computer technology, and more specifically, to a method, device, server, and computer-readable storage medium for identifying college students. Background Technology

[0002] Identifying university students is a fundamental and crucial technological requirement for refined urban governance, campus security and prevention, educational big data analysis, and precise public services and business outreach to young people. Accurate student identification helps optimize the allocation of educational resources, improve traffic and security management around campuses, and support real-time response capabilities for business scenarios such as enrollment monitoring, new student orientation, and career guidance.

[0003] In related technologies, when identifying college students, it is generally necessary to rely on the long-term residence patterns of users within the geographical fence of the college. By using signaling or location data that lasts for several months or even spans the winter and summer vacations, the stability of their nighttime residence locations, the rhythm of their weekday activities, and their periodic travel patterns are statistically analyzed, and then combined with the age information on their ID cards for auxiliary judgment.

[0004] However, this method implicitly assumes that users have been stably present in the campus environment for a sufficiently long time, allowing their behavior patterns to fully exhibit typical student rhythms. In reality, however, a large number of users are short-term newcomers—such as newly enrolled students, exchange students, summer interns, or visiting faculty and students temporarily residing on campus. In the initial stages, they have not yet developed stable residency patterns that can be captured by long-term statistical methods, making them often difficult to identify effectively. This identification lag directly restricts the timeliness of many practical applications. For example, orientation services cannot promptly deliver electronic campus cards and registration instructions to genuine new students, and regional-level crowd control systems struggle to accurately assess the student load on campus before large-scale events. Summary of the Invention

[0005] In view of this, the purpose of this application is to provide a method, device, server and computer-readable storage medium for identifying college students, so as to improve the effective and accurate identification of newly added college students in a short period of time.

[0006] To achieve the above objectives, the technical solutions adopted in the embodiments of this application are as follows: Firstly, this application provides a method for identifying college students, the method comprising: Acquire signaling data of the user to be identified within a preset time range and usage data of multiple target applications; the preset time range is a short duration sufficient to characterize the spatiotemporal rhythm of college students; Based on the signaling data, a spatiotemporal state sequence corresponding to the user to be identified is constructed, and the usage data of each target application is fused in a spatiotemporal context to obtain the application behavior score corresponding to the user to be identified. The spatiotemporal state sequence represents the behavioral trajectory of the user to be identified within the preset time range, and the application behavior score represents the degree of overlap between the application usage behavior of the user to be identified and the application usage behavior of college students. The comprehensive score of the user to be identified is determined based on the pre-constructed behavioral modality of college students, the spatiotemporal state sequence corresponding to the user to be identified, and the application behavior score. Whether the user to be identified is a college student is determined based on the user's overall score.

[0007] In an optional implementation, constructing the spatiotemporal state sequence corresponding to the user to be identified based on the signaling data includes: The signaling data is divided into multiple consecutive time slices according to a preset time interval; wherein each time slice corresponds to a time slice index; For each time slice, the spatial information of the user to be identified is determined based on the signaling data corresponding to the time slice, and the mobility entropy of the user to be identified is calculated based on the signaling data corresponding to the time slice and the adjacent time slices of the time slice. The spatiotemporal state sequence is constructed based on the spatial information, time slice index, and movement entropy corresponding to each time slice.

[0008] In an optional implementation, calculating the mobility entropy of the user to be identified based on the time slice and signaling data corresponding to adjacent time slices includes: Based on the signaling data corresponding to the time slice and the adjacent time slices, calculate the probability that the user to be identified appears in each spatial location of the target university; The movement entropy is calculated based on the probability of each of the aforementioned spatial locations.

[0009] In an optional implementation, the usage data includes time data and spatial data, wherein the time data represents the usage time of the application and the spatial data represents the usage location of the application; The step of performing spatiotemporal context fusion on the usage data of each of the target applications to obtain the application behavior score corresponding to the user to be identified includes: The spatiotemporal context fusion of the usage data of each of the target applications is performed using the following formula:

[0010] in, Characterizes application behavior scores. The preset weights characterizing application k Characterizing the time data of application k, Characterizes a predefined time set. Spatial data representing application k, Characterize the AOI range of the target university. Characterizing the time activation function, Characterization space indicator function.

[0011] In an optional implementation, determining the comprehensive score of the user to be identified based on the pre-constructed behavioral modality of college students, the spatiotemporal state sequence corresponding to the user to be identified, and the application behavior score includes: The nighttime spatial location of the user to be identified is determined based on the spatiotemporal state sequence corresponding to the user to be identified. The overall score of the user to be identified is determined by the following formula:

[0012] in, Characterizing the overall score, Characterizing the spatiotemporal state sequence, Characterizing the behavioral modality of the college students, Characterizing the distance dynamic time warping algorithm, Characterizes application behavior scores. Characterizes the nighttime spatial location of the user to be identified. Functions for calculating discreteness The hyperparameters are characterized separately.

[0013] In an optional implementation, determining whether the user to be identified is a college student based on the user's comprehensive score includes: The comprehensive score is normalized using the Sigmoid function based on preset bias parameters to obtain the comprehensive probability. If the overall probability exceeds a preset probability threshold, then the user to be identified is determined to be a college student.

[0014] In an optional implementation, the method further includes: Acquire college student signaling data, divide the college student signaling data into multiple consecutive time slices according to a preset time interval, and construct the college student behavior modality for each time slice based on the college student signaling data corresponding to the time slice.

[0015] Secondly, this application provides a college student identification device, the device comprising: The acquisition module is used to acquire signaling data of the user to be identified and usage data of multiple target applications within a preset time range; the preset time range is a short duration sufficient to characterize the spatiotemporal rhythm of college students. The construction module is used to construct the spatiotemporal state sequence corresponding to the user to be identified based on the signaling data, and to perform spatiotemporal context fusion on the usage data of each target application to obtain the application behavior score corresponding to the user to be identified. The spatiotemporal state sequence represents the behavioral trajectory of the user to be identified within the preset time range, and the application behavior score represents the degree of overlap between the application usage behavior of the user to be identified and the application usage behavior of college students. The determination module is used to determine the comprehensive score of the user to be identified based on the pre-constructed behavioral modality of college students, the spatiotemporal state sequence corresponding to the user to be identified, and the application behavior score; The determining module is further configured to determine whether the user to be identified is a college student based on the comprehensive score of the user to be identified.

[0016] Thirdly, this application provides a server including a processor and a memory, the memory storing a computer program executable by the processor, the processor being able to execute the computer program to implement the method described in any of the foregoing embodiments.

[0017] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any of the foregoing embodiments.

[0018] The college student identification method, device, server, and computer-readable storage medium provided in this application transform the signaling data of the user to be identified into a spatiotemporal state sequence characterizing its short-cycle behavioral patterns. Simultaneously, it integrates the contextual features of usage data from multiple target applications in the temporal and spatial dimensions to generate an application behavior score reflecting the degree of matching between the user's behavior and the college student population. Based on this, a comprehensive score is calculated using the spatiotemporal state sequence, a pre-constructed college student behavior modality, and the application behavior score. Finally, the comprehensive score is compared with a preset decision threshold to determine the user's college student identity. The entire process no longer relies solely on dwell time but focuses on modeling the consistency between the short-term micro-scale behavioral rhythms of the user and application semantics. It leverages the unique spatiotemporal rhythms of college students to effectively distinguish them from surrounding residents, commuters, and other interfering groups, even with significantly shortened data cycles, thus achieving effective and accurate identification of newly added college students in a short period.

[0019] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A block diagram of a server provided in an embodiment of this application is shown; Figure 2 This paper illustrates a flowchart of a college student identification method provided in an embodiment of this application. Figure 3 This paper illustrates another flowchart of the college student identification method provided in an embodiment of this application; Figure 4 The diagram shows a functional block diagram of a college student identification device provided in an embodiment of this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can be arranged and designed in various different configurations.

[0023] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0024] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0025] Figure 1 Please refer to the block diagram of the server provided in the embodiments of this application. Figure 1 A server includes a memory, a processor, and a communication module. These components are electrically connected directly or indirectly to enable data transmission or interaction. For example, they can be electrically connected via one or more communication buses or signal lines.

[0026] Memory is used to store computer programs or data that can be executed by a processor. Memory can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc.

[0027] The processor is used to read / write data or computer programs stored in the memory, and execute the computer program to implement the college student identification method provided in the embodiments of this application.

[0028] The communication module is used to establish communication connections between the server and other communication terminals via the network, and to send and receive data via the network.

[0029] It should be understood that, Figure 1 The structure shown is only a schematic diagram of the server structure; the server may also include components such as... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown. Figure 1 The components shown can be implemented using hardware, software, or a combination thereof.

[0030] The following is based on the above. Figure 1 The server in this application is the executing entity. The college student identification method provided in this embodiment is illustrated with a flowchart. Specifically, Figure 2 For a flowchart illustrating a college student identification method provided in this application, please refer to [link / reference]. Figure 2 The method includes: Step S20: Obtain signaling data of the user to be identified within a preset time range and usage data of multiple target applications.

[0031] The preset duration range is a short duration sufficient to characterize the spatiotemporal rhythms of college students.

[0032] This embodiment can be understood as follows: by collecting mobile phone signaling data and frequently used APP usage records of the user to be identified within a short but sufficient period of time to reflect their daily routine, this embodiment determines whether the user is a college student.

[0033] In this embodiment, the preset duration range is a specially selected period sufficient to reflect the typical life rhythm of college students, such as 7-14 days. This is because college students' daily activities often exhibit a highly repetitive "dormitory-classroom-canteen" rhythm, and even observing only one week reveals identifiable and stable behavioral waveforms. Specifically, the server will synchronously acquire the user's signaling data and usage data of multiple target applications within these 7 days, thereby providing basic behavioral materials for subsequent modeling.

[0034] Optionally, the target application refers to applications that are highly or lowly related to the learning and life of college students on campus, such as academic affairs systems, campus card apps, postgraduate entrance examination apps, maternal and infant related apps, etc.

[0035] Step S21: Construct a spatiotemporal state sequence corresponding to the user to be identified based on signaling data, and perform spatiotemporal context fusion on the usage data of each target application to obtain the application behavior score corresponding to the user to be identified.

[0036] Among them, the spatiotemporal state sequence represents the behavioral trajectory of the user to be identified within a preset time range, and the application behavior score represents the degree of overlap between the application usage behavior of the user to be identified and the application usage behavior of college students.

[0037] Step S22: Determine the comprehensive score of the user to be identified based on the pre-constructed behavioral modality of college students, the spatiotemporal state sequence corresponding to the user to be identified, and the application behavior score.

[0038] In this embodiment, the server can convert short-term signaling data of users into behavior sequences with spatiotemporal labels, dynamically weight APP usage behavior according to the university scenario, and combine it with the similarity between the behavior and the standard student modality to output a comprehensive score that reflects the overall student characteristic strength of the user.

[0039] In this embodiment, the college student behavior modality refers to the standard behavioral trajectory generated in advance based on known college students.

[0040] Step S23: Determine whether the user to be identified is a college student based on the comprehensive score of the user to be identified.

[0041] Understandably, since this comprehensive score reflects the overall student characteristic strength of the user to be identified, it can be used to directly determine whether the user is a college student. In this embodiment, college students refer to natural persons with formal student status who are enrolled in full-time regular institutions of higher learning (including undergraduate institutions, higher vocational colleges, etc.).

[0042] The college student identification method, device, server, and computer-readable storage medium provided in this application transform the signaling data of the user to be identified into a spatiotemporal state sequence characterizing its short-cycle behavioral patterns. Simultaneously, it integrates the contextual features of usage data from multiple target applications in the temporal and spatial dimensions to generate an application behavior score reflecting the degree of matching between the user's behavior and the college student population. Based on this, a comprehensive score is calculated using the spatiotemporal state sequence, a pre-constructed college student behavior modality, and the application behavior score. Finally, the comprehensive score is compared with a preset decision threshold to determine the user's college student identity. The entire process no longer relies solely on dwell time but focuses on modeling the consistency between the short-term micro-scale behavioral rhythms of the user and application semantics. It leverages the unique spatiotemporal rhythms of college students to effectively distinguish them from surrounding residents, commuters, and other interfering groups, even with significantly shortened data cycles, thus achieving effective and accurate identification of newly added college students in a short period.

[0043] In this embodiment, in order to facilitate the subsequent identification of the user to be identified as a college student, the server can obtain the signaling data of college students with known identities in advance, and construct the college student behavior modality based on the signaling data of college students.

[0044] This embodiment can be understood as using real behavioral data of a batch of college students with known identities to depict the daily activity profile of a typical high-achieving student, which serves as a comparison benchmark for subsequently identifying whether other users are college students.

[0045] In this embodiment, the server first filters out a group of highly confident seed users by combining long-term dwell patterns with specific application usage patterns. For example, users who frequently use campus-specific applications such as "Campus Card APP", "Academic Affairs System", and "Postgraduate Entrance Examination Helper" are considered to be highly likely to be college students. The signaling data of these seed users is obtained as the signaling data of college students with known identities. This college student signaling data is divided into multiple consecutive time slices according to a preset time interval. For each time slice, a college student behavior modality is constructed based on the college student signaling data corresponding to the time slice.

[0046] Optionally, the preset time interval can be set according to the actual application, for example, to 15 minutes.

[0047] In one possible implementation, the server can retrieve signaling data from these seed users during a normal teaching cycle (e.g., 7-14 days), dividing the day evenly into 96 time slices, that is, each time slice is 15 minutes long, represented as... Based on this, for each time slice Based on the signaling data of college students corresponding to that time slice, the distribution of all seed users connected to base stations within that time slice is statistically analyzed. Then, the distribution probability of these users in campus functional areas such as dormitory base stations, teaching area base stations, and canteen area base stations is calculated, forming a standard spatial feature vector set that changes over time. .

[0048] It should be noted that this probability distribution essentially refers to what percentage of a group of known college student seed users connected to a base station in a specific campus functional area within a certain time slice.

[0049] In practice, the server first identifies several physical areas closely related to universities, such as the coverage areas of dormitory base stations, teaching building cluster base stations, canteen area base stations, and library base stations, and assigns a unique identifier to each area. Then, for a specific time slot (e.g., 8:00-8:15 AM), it counts how many records from all seed users reported during this time period fell on dormitory base stations, how many fell on teaching building base stations, and how many fell on canteen base stations. Finally, the number of records falling on each area is divided by the total number of valid records during that time period to obtain a set of probability values ​​that add up to 1.

[0050] For example, if 1000 signaling records from seed users are collected between 8:00 and 8:15 AM, with 620 from teaching area base stations, 210 from dormitory area base stations, 130 from cafeteria base stations, and the remaining 40 from off-campus base stations, then the distribution probability for the teaching area is 0.62, for the dormitory area it is 0.21, and for the cafeteria area it is 0.13. This set of values ​​constitutes the standard spatial feature vector for that time slice. It should be noted that this probability does not indicate that a particular student is "definitely" in a certain location, but rather reflects the spatial clustering tendency of the entire student group at that moment—a statistically significant regularity.

[0051] Understandably, this set of vectors constitutes the "behavioral modality of college students," which essentially describes the spatiotemporal rhythm of college students throughout the day, indicating "when they are most likely to be in which place."

[0052] For example, around 8:00 AM, this modality will show a high probability value at the location corresponding to the base station in the teaching area; around noon, the probability will increase near the base station in the canteen area; and after 10:00 PM, the probability of the base station in the dormitory area will increase significantly again. This shows that this modality is not a static map, but a dynamic probability curve with time labels, providing a quantifiable and comparable behavioral benchmark for subsequent short-cycle identification.

[0053] Optionally, it's important to consider the significant differences in behavioral patterns among university students of different grades. For example, freshmen, having just entered university, primarily attend classes in the teaching buildings and spend their free time familiarizing themselves with the environment in the dormitories, with relatively low frequency of visits to the library or laboratories. Senior students, on the other hand, spend less time on campus due to internships, graduation projects, or exam preparation, resulting in decreased stability in the dormitories at night and a higher proportion of connections to off-campus base stations. If a uniform behavioral modality for all students is used to construct a unified model, it might simply average the high-frequency occurrences of freshmen in the teaching areas with the low-frequency occurrences of seniors, leading to modal distortion. This would be particularly detrimental to identifying newly enrolled users who haven't yet developed stable long-term behaviors.

[0054] Therefore, to obtain more accurate behavioral modalities of college students, the server can first group seed users by grade level. Taking college students as an example, the server can obtain signaling data from freshmen, sophomores, juniors, and seniors, and construct behavioral modalities for each grade. Based on this, a grade weight is assigned to each grade's behavioral modal, reflecting the reference value of that grade's student behavior for identifying short-term new users. Since freshmen are in the early stages of identity establishment, their behavioral rhythms best represent the characteristics of "typical freshmen" and have the highest similarity to the short-term new user group to be identified; therefore, the grade weight corresponding to the freshman behavioral modal is set to a higher value. Senior students, whose behavior has significantly deviated from the norm on campus, exhibiting high mobility and unstable residence, have their grade weight correspondingly lowered. Finally, when calculating the standard spatial feature vector for a certain period, the distribution probability of college students in each grade during that period is weighted and averaged according to their corresponding grade weights, thus obtaining an optimized college student behavioral modal that takes into account both grade characteristics and the identification target.

[0055] The following section provides a possible implementation method for constructing the spatiotemporal state sequence corresponding to the user to be identified based on signaling data.

[0056] Specifically, in Figure 2 On this basis, Figure 3 For another flowchart illustrating the college student identification method provided in this application embodiment, please refer to [link / reference]. Figure 3 The construction of the spatiotemporal state sequence corresponding to the user to be identified based on the signaling data in step S21 above can also be obtained through the following steps: Step S21-1: Divide the signaling data into multiple consecutive time slices according to a preset time interval. Each time slice corresponds to a time slice index.

[0057] Optionally, the preset time interval can be set according to actual needs, for example, to 15 minutes.

[0058] This embodiment can be understood as follows: the user's mobile phone signaling data within a short observation window is divided into continuous time blocks of fixed duration, and each block is labeled with a sequential number starting from 1. This number is the time block index, which represents the specific position of the time block in the entire observation period. For example, the first block corresponds to 0:00-0:15 AM, the second block corresponds to 0:15-0:30 AM, and so on.

[0059] It is precisely by relying on these numbered time slices that the server is able to organize scattered signaling records into a temporally ordered sequence of actions. It should be noted that this number can be used not only for counting and marking, but also as a time coordinate basis for identifying the rhythms of university life such as "early class period", "lunch break period", and "evening self-study period".

[0060] Step S21-2: For each time slice, determine the spatial information of the user to be identified based on the signaling data corresponding to the time slice, and calculate the mobility entropy of the user to be identified based on the signaling data corresponding to the time slice and the adjacent time slices.

[0061] In this embodiment, for each time slice, the server determines the corresponding spatial location of the user based on the base station ID connected to the user within that time slice, such as whether it falls within the coverage area of ​​the dormitory area base station, the teaching area base station, or the canteen area base station.

[0062] At the same time, the server will also retrieve the signaling data of the time slice and the time slice before and after it, count the probability of the user appearing in various functional areas of the campus during these three time periods, and substitute it into the motion entropy formula to calculate a value. The smaller this value is, the more regular the user's movement is in the vicinity of that time period, for example, it is more in line with the behavior characteristics of college students going back and forth between "dormitory-classroom-canteen".

[0063] It should be understood that spatial information answers "where is the person?", while mobile entropy answers "whether the movement is orderly?". Together, they constitute a key dimension for characterizing the short-term behavioral rhythms of users.

[0064] Step S21-3: Construct a spatiotemporal state sequence based on the spatial information, time slice index, and movement entropy corresponding to each time slice.

[0065] In this embodiment, since signaling data from multiple days has been collected, the server can process the signaling data for each day separately to obtain multiple time slices corresponding to each day. In this case, the spatiotemporal state sequence... Each of them It can be represented as day i, and each Each time slice can correspond to multiple time slice indices. Taking a preset time interval of 15 minutes as an example, there can be 96 time slices per day, which can be bound to indices 1-96 respectively.

[0066] In this embodiment, the spatiotemporal state sequence Each of them Each time slice contains time slice index information, spatial information, and movement entropy.

[0067] In one possible implementation, the server can calculate the probability of the user to be identified appearing in various spatial locations within the target university based on the signaling data corresponding to the time slice and adjacent time slices, and then calculate the movement entropy based on the probability of each spatial location.

[0068] Understandably, the probability of a user appearing in various spatial locations actually refers to the probability distribution of the user appearing in different locations at different time slices. The server can calculate the probability distribution corresponding to each spatial location i by statistically analyzing the signaling data corresponding to multiple time slices. The movement entropy can then be calculated using the following formula:

[0069] Understandably, the various spatial locations within the target university can be defined based on the location of the base stations, such as the coverage area of ​​the dormitory area, the coverage area of ​​the teaching building complex, the coverage area of ​​the canteen area, the coverage area of ​​the library, and so on.

[0070] In one possible implementation, the application's usage data can include time data and spatial data, with the time data representing the time the application is used and the spatial data representing the location where the application is used.

[0071] Based on this, the server can perform application spatiotemporal context fusion on the usage data of each target application using the following formula:

[0072] in, Characterizes application behavior scores. The preset weights characterizing application k Characterizing the time data of application k, Characterizes a predefined time set. Spatial data representing application k, The AOI (Area of ​​Interest) range representing the target university. Characterizing the time activation function, The characterization space indicator function is 0.

[0073] Optionally, this preset time set refers to a set of business-significant time intervals pre-defined based on the typical work and rest patterns of universities; for example, 8:00 AM to 12:00 PM, 2:00 PM to 6:00 PM on weekdays, or 11:00 PM to 1:00 AM. It should be noted that this time set is not a fixed clock time, but rather an active time period template strongly associated with the university setting, which can be configured according to the season, semester schedule, or specific campus characteristics.

[0074] In this embodiment, the server can first obtain the time data and spatial data of the user's use of each target application during the observation period. The time data refers to the specific time when the user opens or uses the APP, and the spatial data refers to the user's geographical location at that time. Then, the server can calculate the contribution score of each APP separately. This score is obtained by multiplying the following three parts.

[0075] First, there is the inherent weight set by the app itself. This weight depends on how relevant the app is to the learning and life of college students. Apps with a higher degree of relevance have a higher default weight. For example, learning apps like "Kaoyanbang" have a higher weight. On the other hand, apps with a lower degree of relevance to the lives of college students have a lower default weight, or even a negative weight. For example, apps like "Muyingzhijia" that are clearly not relevant to the scenario have a negative weight.

[0076] Secondly, there's the time activation function, which automatically adjusts the weight based on the app type. For example, learning apps are amplified during weekday daytime hours, while gaming apps are activated only late at night in dormitories. Specifically, the time activation function determines whether a user's use of an app falls within a preset time set—if it does, the app's weight is activated according to the rules (e.g., learning apps have double the weight during daytime hours, while gaming apps are activated only late at night in dormitories); if it doesn't, the app's time dimension contribution is zero.

[0077] Third is the spatial indicator function, which is only allowed to score when the spatial location of the user to be identified is within the AOI range of the target university; it is not included when used outside the university.

[0078] Finally, the weighted scores of all apps are summed up to obtain the user's application behavior score. Obviously, this score is not simply a count of which apps were used, but rather a comprehensive consideration of three key dimensions: "what was used, when was it used, and where was it used."

[0079] In this embodiment, after obtaining the spatiotemporal state sequence and application score corresponding to the user to be identified, the server can calculate the comprehensive score of the user to be identified by combining the pre-constructed behavioral modality of college students. Specifically, the server can first determine the nighttime spatial location of the user to be identified based on the spatiotemporal state sequence, and then determine the comprehensive score of the user to be identified using the following formula:

[0080] in, Characteristic comprehensive score, Characterizing the spatiotemporal state sequence, Characterizing the behavioral modalities of college students Characterizing the dynamic time warping algorithm, Characterizes application behavior scores. Characterize the nighttime spatial location of the user to be identified. Functions for calculating discreteness The hyperparameters are characterized separately.

[0081] In this embodiment, the server can first extract the nighttime spatial location of the user to be identified from the spatiotemporal state sequence, that is, to statistically analyze the distribution of the base station locations connected to the user during a specific time period each night (e.g., from 22:00 to 6:00 the next day), forming a set of nighttime location points. .

[0082] Based on this, the server can calculate the overall score according to the above formula. The first item This is used to measure the degree of temporal and spatial matching between the entire behavioral trajectory of the user to be identified and the standard student modality. Since college students have a regular "dormitory → classroom → canteen → dormitory" rhythm in their daily lives, while residents or visitors often exhibit a chaotic or single-point static pattern, the DTW (Dynamic Time Warping) algorithm can be used to tolerate minor misalignments such as early morning class time deviations. The smaller the distance, the more the trajectory resembles that of a student.

[0083] Second item This is used to implement the inverse weighting of the application behavior score. The closer the APP behavior is to students' habits (such as high-frequency use of learning APPs and occurrence on campus, and reasonable active time periods), the lower the score of this item.

[0084] Third item This value is used to reflect the dispersion of a user's location at night. Students usually stay stably at base stations in the dormitory area, with little location change and low dispersion. However, people moving around the area may wander randomly inside and outside the campus, resulting in a loose distribution and high dispersion of their location at night. The larger this value is, the less likely it is to be identified as a student.

[0085] The three scores are evaluated by the parameters. The weighted scores are summed to form the final composite score. It's important to note that these hyperparameters are not manually set, but rather obtained automatically through training on labeled student and non-student samples using models such as logistic regression or support vector machines, combined with grid search or Bayesian optimization methods. Therefore, this composite score is essentially a quantitative result that organically integrates behavioral rhythmicity, app semantic rationality, and nighttime stability.

[0086] Next, the server can normalize the comprehensive score using the Sigmoid function based on preset bias parameters to obtain the comprehensive probability. In this case, if the comprehensive probability exceeds the preset probability threshold, the user to be identified is determined to be a college student.

[0087] It should be noted that since the overall score is an uncalibrated numerical value, it may be positive or negative and has inconsistent dimensions, making it unsuitable for direct comparison across users or scenarios. Therefore, the Sigmoid function can be introduced to perform non-linear normalization processing, and its mathematical expression is:

[0088] in, Characterizing the overall probability, Characteristic comprehensive score, The bias function, also known as the decision threshold, determines the center position of the normalization curve on the composite score axis—for example, when... Given a certain empirical value (e.g., 2.5), what is the normalized overall probability of a user whose overall score is exactly 2.5? It's 0.5.

[0089] The higher the overall score, The closer the score is to 1, the more it resembles a college student; the lower the overall score, the more... The closer it gets to 0. It should be noted that this Sigmoid transformation is not a simple scaling, but rather simulates the business intuition in real-world judgment that "a small increase in score leads to a significant increase in probability." For example, when a student's behavioral characteristics just reach a critical level, the confidence in recognition will increase rapidly.

[0090] Based on this, the server can obtain the normalized comprehensive probability. Compare with another preset probability threshold (e.g., 0.85): If If the value is greater than 0.85, meaning the server is more than 85% confident that a user is a college student, then the user is considered a college student; otherwise, the user is not considered a college student.

[0091] It should be understood that both the bias parameter and the probability threshold can be flexibly adjusted according to the actual deployment scenario. For example, the probability threshold can be appropriately lowered during the new student enrollment season to improve the recall rate, while it can be raised in the precision marketing scenario to ensure the accuracy rate.

[0092] To further improve the accuracy of identifying university students, a whitelist and blacklist mechanism can be introduced, along with a multiple-judgment strategy, to enhance the stability and reliability of the identification results. Specifically, the server can perform multiple identifications on each user, for example, using signaling data from the past 7 days for several consecutive periods to determine if the user is a university student. If a user is consistently identified as a university student within multiple observation windows (e.g., three independent 7-day periods), the user is officially added to the whitelist, confirming their university student status. Conversely, if the same user fails to meet the preset judgment threshold within multiple observation windows (e.g., three independent 7-day periods), i.e., is consistently identified as a non-university student, the user is added to the blacklist.

[0093] The core purpose of this mechanism is to reduce the random errors that may arise from single, short-term judgments through repeated verification over time, especially for new users or short-term residents whose behavioral patterns are not yet stable, avoiding misjudgments due to abnormal behavior on a few days. Furthermore, the model can be re-optimized using consistently long-term judgment results, enabling the rapid identification method for university students to continuously adapt and evolve. In this embodiment, users in the whitelist not only serve as the final identification result output but can also be simultaneously used as high-quality seed users to dynamically update the spatial distribution probability corresponding to each time slice in the university student behavioral modality, and participate in a new round of hyperparameter training, thereby continuously strengthening the model's ability to depict the behavioral texture of real university students.

[0094] Users on the blacklist will no longer be included in the candidate set for subsequent college student identification processes, nor will they participate in the construction and updating of the "standard student spatiotemporal modality," thereby effectively avoiding the interference of abnormal behavior samples on the model's generalization performance and ensuring the robustness and business availability of the overall identification system.

[0095] To perform the corresponding steps in the above embodiments and various possible methods, an implementation of a college student identification device is given below. Optionally, the college student identification device can adopt the above-described... Figure 1 The server's device structure is shown. For further details, please refer to... Figure 4 , Figure 4This is a functional block diagram of a college student identification device provided in an embodiment of this application. It should be noted that the basic principle and technical effects of the college student identification device provided in this embodiment are the same as those in the above embodiments. For the sake of brevity, any parts not mentioned in this embodiment can be referred to the corresponding content in the above embodiments. The college student identification device includes: an acquisition module, a construction module, and a determination module.

[0096] This acquisition module is used to acquire signaling data of the user to be identified and usage data of multiple target applications within a preset time range; the preset time range is a short duration sufficient to characterize the spatiotemporal rhythm of college students.

[0097] Understandably, this acquisition module can also be used to perform the above step S20.

[0098] This module is used to construct the spatiotemporal state sequence corresponding to the user to be identified based on signaling data, and to perform spatiotemporal context fusion on the usage data of each target application to obtain the application behavior score corresponding to the user to be identified.

[0099] Among them, the spatiotemporal state sequence represents the behavioral trajectory of the user to be identified within a preset time range, and the application behavior score represents the degree of overlap between the application usage behavior of the user to be identified and the application usage behavior of college students.

[0100] Understandably, this building block can also be used to perform step S21 described above.

[0101] The determination module is used to determine the comprehensive score of the user to be identified based on the pre-constructed behavioral modalities of college students, the spatiotemporal state sequence corresponding to the user to be identified, and the application behavior score.

[0102] Understandably, this determining module can also be used to perform step S22 described above.

[0103] The determination module is also used to determine whether a user is a college student based on the user's overall score.

[0104] Understandably, this determining module can also be used to perform step S23 described above.

[0105] Optionally, the construction module is further configured to divide signaling data into multiple consecutive time slices according to a preset time interval; wherein each time slice corresponds to a time slice index; for each time slice, the spatial information of the user to be identified is determined according to the signaling data corresponding to the time slice, and the mobility entropy of the user to be identified is calculated based on the time slice and the signaling data corresponding to the adjacent time slices; and a spatiotemporal state sequence is constructed based on the spatial information, time slice index and mobility entropy corresponding to each time slice.

[0106] Understandably, this building block can also be used to perform steps S21-1 to S21-3 as described above.

[0107] Optionally, the building module is also used to calculate the probability that the user to be identified appears in each spatial location of the target university based on the signaling data corresponding to the time slice and the adjacent time slices; and to calculate the movement entropy based on the probability.

[0108] Optionally, this building module is also used to perform application spatiotemporal context fusion on the usage data of each target application using the following formula:

[0109] in, Characterizes application behavior scores. The preset weights characterizing application k Characterizing the time data of application k, Characterizes a predefined time set. Spatial data representing application k, Characterize the AOI range of the target university. Characterizing the time activation function, Characterization space indicator function.

[0110] Optionally, the building module is also used to determine the nighttime spatial location of the user to be identified based on the spatiotemporal state sequence corresponding to the user to be identified; The overall score of the user to be identified is determined using the following formula:

[0111] in, Characteristic comprehensive score, Characterizing the spatiotemporal state sequence, Characterizing the behavioral modalities of college students Characterizing the distance dynamic time warping algorithm, Characterizes application behavior scores. Characterize the nighttime spatial location of the user to be identified. Functions for calculating discreteness The hyperparameters are characterized separately.

[0112] Optionally, the determining module is further configured to normalize the comprehensive score using the Sigmoid function based on preset bias parameters to obtain a comprehensive probability; if the comprehensive probability exceeds a preset probability threshold, then the user to be identified is determined to be a college student.

[0113] Optionally, the construction module is also used to acquire college student signaling data, divide the college student signaling data into multiple consecutive time slices according to a preset time interval, and construct college student behavior modalities based on the college student signaling data corresponding to each time slice.

[0114] Optionally, the above modules can be stored in the form of software or firmware. Figure 1 The memory shown is either stored in or embedded in the server's operating system (OS), and can be used by... Figure 1 The processor executes the commands. Meanwhile, the data and program code required to execute these modules can be stored in memory.

[0115] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the college student identification method provided in this application.

[0116] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0117] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0118] If a function is implemented as a software module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0119] The above are merely preferred embodiments of this application and are not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for identifying college students, characterized in that, The method includes: Acquire signaling data of the user to be identified within a preset time range and usage data of multiple target applications; the preset time range is a short duration sufficient to characterize the spatiotemporal rhythm of college students; Based on the signaling data, a spatiotemporal state sequence corresponding to the user to be identified is constructed, and the usage data of each target application is fused in a spatiotemporal context to obtain the application behavior score corresponding to the user to be identified. The spatiotemporal state sequence represents the behavioral trajectory of the user to be identified within the preset time range, and the application behavior score represents the degree of overlap between the application usage behavior of the user to be identified and the application usage behavior of college students. The comprehensive score of the user to be identified is determined based on the pre-constructed behavioral modality of college students, the spatiotemporal state sequence corresponding to the user to be identified, and the application behavior score. Whether the user to be identified is a college student is determined based on the user's overall score.

2. The method according to claim 1, characterized in that, The step of constructing the spatiotemporal state sequence corresponding to the user to be identified based on the signaling data includes: The signaling data is divided into multiple consecutive time slices according to a preset time interval; wherein each time slice corresponds to a time slice index; For each time slice, the spatial information of the user to be identified is determined based on the signaling data corresponding to the time slice, and the mobility entropy of the user to be identified is calculated based on the signaling data corresponding to the time slice and the adjacent time slices of the time slice. The spatiotemporal state sequence is constructed based on the spatial information, time slice index, and movement entropy corresponding to each time slice.

3. The method according to claim 2, characterized in that, The calculation of the mobility entropy of the user to be identified based on the time slice and the signaling data corresponding to the adjacent time slices includes: Based on the signaling data corresponding to the time slice and the adjacent time slices, calculate the probability that the user to be identified appears in each spatial location of the target university; The movement entropy is calculated based on the probability of each of the aforementioned spatial locations.

4. The method according to claim 1, characterized in that, The usage data includes time data and spatial data, where the time data represents the usage time of the application and the spatial data represents the usage location of the application. The step of performing spatiotemporal context fusion on the usage data of each of the target applications to obtain the application behavior score corresponding to the user to be identified includes: The spatiotemporal context fusion of the usage data of each of the target applications is performed using the following formula: in, Characterizes application behavior scores. The preset weights characterizing application k Characterizing the time data of application k, Characterizes a predefined time set. Spatial data representing application k, Characterize the AOI range of the target university. Characterizing the time activation function, Characterization space indicator function.

5. The method according to claim 1, characterized in that, The step of determining the comprehensive score of the user to be identified based on the pre-constructed behavioral modality of college students, the spatiotemporal state sequence corresponding to the user to be identified, and the application behavior score includes: The nighttime spatial location of the user to be identified is determined based on the spatiotemporal state sequence corresponding to the user to be identified. The overall score of the user to be identified is determined by the following formula: in, Characterizing the overall score, Characterizing the spatiotemporal state sequence, Characterizing the behavioral modality of the college students, Characterizing the distance dynamic time warping algorithm, Characterizes application behavior scores. Characterizes the nighttime spatial location of the user to be identified. Functions for calculating discreteness The hyperparameters are characterized separately.

6. The method according to claim 1, characterized in that, The step of determining whether the user to be identified is a college student based on the user's comprehensive score includes: The comprehensive score is normalized using the Sigmoid function based on preset bias parameters to obtain the comprehensive probability. If the overall probability exceeds a preset probability threshold, then the user to be identified is determined to be a college student.

7. The method according to claim 1, characterized in that, The method further includes: Acquire college student signaling data, divide the college student signaling data into multiple consecutive time slices according to a preset time interval, and construct the college student behavior modality for each time slice based on the college student signaling data corresponding to the time slice.

8. A college student identification device, characterized in that, The device includes: The acquisition module is used to acquire signaling data of the user to be identified and usage data of multiple target applications within a preset time range; the preset time range is a short duration sufficient to characterize the spatiotemporal rhythm of college students. The construction module is used to construct the spatiotemporal state sequence corresponding to the user to be identified based on the signaling data, and to perform spatiotemporal context fusion on the usage data of each target application to obtain the application behavior score corresponding to the user to be identified. The spatiotemporal state sequence represents the behavioral trajectory of the user to be identified within the preset time range, and the application behavior score represents the degree of overlap between the application usage behavior of the user to be identified and the application usage behavior of college students. The determination module is used to determine the comprehensive score of the user to be identified based on the pre-constructed behavioral modality of college students, the spatiotemporal state sequence corresponding to the user to be identified, and the application behavior score; The determining module is further configured to determine whether the user to be identified is a college student based on the comprehensive score of the user to be identified.

9. A server, characterized in that, It includes a processor and a memory, the memory storing a computer program executable by the processor, the processor being able to execute the computer program to implement the method of any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-7.

Images