Digital marketing management system and method for telecommunication service

Abstract

The invention relates to the technical field of Internet of Things, in particular to a digital marketing management system and method for telecommunication services, and the system comprises a user behavior monitoring module, a path track extraction module, a dynamic work and rest modeling module, a behavior cycle analysis module and a user portrait generation module. According to the method, through group feature modeling of cellular communication data, accurate calculation of a behavior trend change rate, dynamic description of group behavior features and extraction of a high-frequency region node sequence, analysis of a behavior track is converted from a single-point discrete state to a path association mode, and recognition of a group activity mode is realized; based on group residence duration statistics and cross analysis of multi-dimensional behavior indexes, a time distribution model is constructed, a stable behavior interval and a typical mode are identified, the analysis capability of group behavior characteristics is improved, and in combination with holiday and festival characteristic identification and hot spot area distribution characteristics, marketing strategy matching is optimized, and the accuracy of a grouping recommendation strategy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things, and in particular to a digital marketing management system and method for telecommunication services. Background Art

[0002] The Internet of Things technology field includes systems that realize device interconnection, information collection and analysis through intelligent devices, network communication and data processing technology. The core content of this technology field includes the perception layer, network layer and application layer. The perception layer is mainly composed of devices such as sensors, RFID and cameras, which are responsible for collecting environmental data and device status information in the physical world; the network layer realizes communication between devices and servers through wired or wireless communication technologies such as cellular networks, Wi-Fi, Bluetooth, LoRa, etc.; the application layer relies on cloud computing, big data analysis and artificial intelligence technology to complete data storage, analysis and application services, such as smart homes, industrial automation and smart cities. The systematic development of the overall technology field of the Internet of Things involves multiple links such as data collection, transmission, processing and application, aiming to build an intelligent and automated digital world and improve resource utilization and operation efficiency.

[0003] Among them, the digital marketing management system for telecommunications services refers to a system that realizes precision marketing and customer relationship management based on the telecommunications network environment, combined with data analysis and intelligent computing capabilities. The system mainly involves cluster analysis of group behavior characteristics, precise positioning of target customer groups, dynamic generation and optimization of marketing content, and execution of multi-channel marketing plans. Specific methods include feature mining based on group behavior statistical models, using machine learning models to predict group preference trends, generating classified marketing content through natural language processing technology, and distributing it through channels such as SMS, social media, and APP push. In addition, the system also optimizes subsequent marketing strategies and improves the intelligence level of marketing management through real-time collection and analysis of feedback data.

[0004] Existing technologies rely on static data statistics and lack the calculation of time series distribution characteristics, resulting in user feature characterization being limited to fixed patterns. Behavior trajectory analysis stays at discrete location points and fails to associate continuous state migration, affecting the identification of group behavior patterns. The modeling of time distribution characteristics does not combine multi-dimensional statistical indicators, making it difficult to accurately determine typical behavior periods. Behavior prediction relies on fixed time windows and does not consider periodic change offsets, reducing the ability to track irregular behavior patterns. Marketing strategies are based on rule matching and fail to combine dynamic behavior feature optimization, affecting accuracy and user interaction effects. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a digital marketing management system and method for telecommunication services.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: A digital marketing management system for telecommunication services comprises:

[0007] The user behavior monitoring module obtains cellular communication records, collects event types, signaling strength, communication frequency, and dwell time, sorts them according to timestamps, and calculates the mean maximum fluctuation rate by partition to obtain the event intensity change rate value within the user's continuous segment; The path trajectory extraction module determines the trajectory state based on the event intensity change rate value, collects the base station number access state timestamp, sorts and generates a node sequence, counts the switching frequency, and obtains a high-frequency path node sequence; The dynamic work and rest modeling module detects the stable residence time based on the high-frequency path node sequence, collects the residence data of multiple regions in the same period, counts the frequency of downlink requests and uplink events, merges the frequency values, determines the stable interval, screens the repeated time periods, sorts out the intra-day distribution, accumulates the number of days, builds the work and rest model, and obtains the distribution value of the core work and rest period; The behavior cycle analysis module calls the work and rest time of consecutive days according to the distribution value of the core work and rest period, detects the period offset, filters the offset period, sorts the work and rest nodes, and obtains the dynamic prediction result of the work and rest cycle offset.

[0008] As a further solution of the present invention, the event intensity change rate value within the user's continuous segment includes dynamic indicators, extreme value indicators, and fluctuation indicators; the high-frequency path node sequence includes core nodes, timing associations, and switching identifiers; the core work and rest period distribution value includes main period identifiers, distribution characteristics, and concentration indicators; the work and rest cycle offset dynamic prediction results include period offsets, offset amplitudes, and dynamic trends.

[0009] As a further solution of the present invention, the user behavior monitoring module includes: The cellular communication data collection submodule obtains cellular communication record data, collects event type, signaling strength, communication frequency and dwell time, and sorts all data according to timestamps to obtain time-series cellular communication data; The event signal analysis submodule calculates the mean and maximum values ​​of the signaling strength, communication frequency and dwell time in the same time interval based on the time-series cellular communication data, divides the multiple data into time series, identifies the change trend in the differentiated intervals, and obtains the interval signal statistical characteristics; The continuous segment intensity change calculation submodule calls the interval signal statistical features, divides the data according to the timestamp, and uses the formula: ; Calculate the signal strength change rate in the continuous segment to obtain the user's continuous segment signal change rate value; in, Represents the rate of change of signal strength, Representative The signal strength at each time point, represents the total number of data points in the time window of the selected segment, Represents the total duration of the time window, Represents the difference in signal strength at adjacent time points, Represents a time point index.

[0010] As a further solution of the present invention, the path trajectory extraction module includes: The trajectory state determination submodule calculates the rate change difference between adjacent time points based on the event intensity change rate value, determines whether the trajectory state has changed, and if the rate change exceeds the set threshold, marks it as a trajectory switching point, calls the switching point set, matches the corresponding timestamp, and establishes a trajectory switching time sequence; The node sequence generation submodule calls the trajectory switching time sequence, extracts the base station number and access status in chronological order, matches the timestamps and sorts them to generate the initial node sequence, filters the base station numbers of continuous access, removes duplicate and abnormal data, and constructs the trajectory node sequence; The high-frequency path analysis submodule calls the trajectory node sequence, counts the base station switching frequency, and calculates the connection strength between nodes based on the frequency, using the formula: ; Obtain the influence weights between nodes through calculation, filter out high-frequency nodes above the threshold, and establish a high-frequency path node sequence; in, Represents the node influence weight, represents the number of switching times, p , q represents the node index, and r represents the other node indexes.

[0011] As a further solution of the present invention, the dynamic work and rest modeling module includes: The residence period screening submodule detects the stable residence time of multiple regions based on the high-frequency path node sequence, collects the residence data of multiple regions in the same period, calculates the residence time proportion values ​​of multiple regions, and screens the period whose residence time proportion exceeds the set threshold to obtain the period with high residence proportion; The stable interval judgment submodule counts the downlink request frequency and the uplink event frequency of the corresponding period based on the high residence ratio period, calculates the combined frequency value, judges whether the period belongs to the stable interval, and selects the period that appears repeatedly in multiple differentiated days to obtain the stable residence interval distribution; The core work and rest construction submodule sorts out the distribution of multiple time periods within a day based on the stable residence interval distribution, accumulates the number of days with the same time period, and uses the formula: ; The core work and rest period distribution values ​​are obtained by calculation and sorted into the time intervals within the day to obtain the core work and rest period distribution; in, Represents the distribution value of the core work and rest period, represents the cumulative duration of the stable residence period on the sth day, represents the combined frequency value of the s-th day period, Represents the total duration of stay on day s. represents the number of repetitions in the s-th day period, Represents the statistical day, and s represents the day index.

[0012] As a further solution of the present invention, the behavior cycle analysis module includes: The time period offset screening submodule calls the work and rest time of consecutive days based on the core work and rest time period distribution value, calculates the time period intervals of multiple time points, obtains the work and rest frequency change values ​​of differentiated time periods, screens the data points with significant time period changes, extracts the offset amplitude and direction of multiple work and rest time points, and calculates the standardized value and cumulative offset rate of the offset period to obtain the offset period data; The work and rest node arrangement submodule calls the offset period data, detects the work and rest time points that appear frequently, calculates the offset trend of adjacent time points, determines the stability of the time nodes, selects the time points with a lower fluctuation range, re-divides the work and rest cycle nodes, and obtains the work and rest cycle stable nodes; The cycle shift prediction submodule calls the work and rest cycle stability node, calculates the shift trend and fluctuation range of multiple time points, and establishes a cycle shift calculation model using the formula: ; Calculate the cycle shift composite index, obtain the shift trend vector, analyze the shift change rate in combination with past data, and generate a dynamic prediction result for the work and rest cycle shift; in, represents the periodic shift composite index, represents the offset of day t, represents the offset direction coefficient on the tth day, represents the stability coefficient on day t, Represents the total number of days in the statistical period. represents the expected value of the offset, represents the regularization factor, and t represents the cycle day index.

[0013] As a further solution of the present invention, the system further includes: The user profile generation module calls the dynamic prediction result of the work and rest cycle offset, identifies the holiday mode label, collects high-frequency path nodes and core dwell features, counts the spatiotemporal distribution, generates user behavior profiles, and obtains a dynamic behavior profile parameter set; The dynamic behavior profile parameter set includes pattern labels, path characteristics, and time and space parameters.

[0014] As a further solution of the present invention, the user portrait generation module includes: The work and rest activity variability analysis submodule calls the work and rest cycle deviation dynamic prediction result to calculate the activity deviation value of the user in the differentiated time period using the formula: ; Obtain the daily activity variability index through calculation and obtain the distribution parameters of work-rest deviation; in, represents the index of daytime activity variability, Representative Activity records at a certain point in time, Represents the total number of time points measured, Represents the average activity time at all time points, represents the standard deviation of all time points, and u represents the time point index; The holiday pattern recognition submodule compares the difference between the user's work and rest schedule on weekdays and holidays based on the work and rest schedule deviation distribution parameters, counts the behavior change trends of differentiated time periods, calculates the work and rest schedule deviations of multiple time periods, selects holiday behavior feature values ​​according to the work and rest schedule deviation mean and standard deviation, selects the time period with the most obvious changes, and obtains holiday pattern labels; The spatiotemporal behavior statistics submodule calls the holiday mode label, collects users' high-frequency path nodes and core dwelling locations, counts the dwelling time of multi-path nodes, calculates the dwelling time ratio of differentiated areas, calculates the spatiotemporal activity distribution according to the discrete degree of dwelling time, generates user behavior portraits, and obtains a dynamic behavior portrait parameter set.

[0015] A digital marketing management method for telecommunication services, the digital marketing management method for telecommunication services being implemented based on the digital marketing management system for telecommunication services, comprising the following steps: S1: Obtain cellular communication records, collect event type, signaling strength, communication frequency, and dwell time, sort them according to timestamps, calculate the mean, maximum value, and fluctuation rate by partition, and obtain the event intensity change rate value within the user's continuous segment; S2: Based on the event intensity change rate value in the continuous segment of the user, determine the trajectory state, collect the base station number access state timestamp, sort and generate a node sequence, count the switching frequency, and obtain a high-frequency path node sequence; S3: Call the high-frequency path node sequence, detect the stable residence time, collect the residence data of multiple regions in the same period, count the frequency of downlink requests and uplink events, merge the frequency values, filter the stable interval according to the residence time, filter the repeated time period, sort out the intra-day distribution, accumulate the number of days, and establish the distribution value of the core work and rest period; S4: according to the core work and rest period distribution value, call the work and rest time of consecutive days, detect the period offset, filter the offset period, sort the work and rest nodes, and obtain the dynamic prediction result of the work and rest cycle offset; S5: Call the dynamic prediction result of the work and rest cycle offset, combine it with the telecommunications service push data, filter the active time period of the target user, match the adaptive marketing push time, and obtain the optimized value of the user marketing reach time period.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by modeling the group characteristics of cellular communication data, the rate of change of behavior trends can be accurately calculated and the group behavior characteristics can be dynamically characterized. The extraction of high-frequency regional node sequences transforms the analysis of behavior trajectories from a single-point discrete state to a path association mode, thereby achieving accurate identification of group activity patterns. Based on the cross-analysis of group residence time statistics and multi-dimensional behavior indicators, a time distribution model is constructed to identify stable behavior intervals and typical patterns, thereby improving the ability to analyze group behavior characteristics. Dynamic prediction of periodic offsets enables effective tracking of irregular behavior trends. Combining holiday feature recognition with hot spot area distribution characteristics, marketing strategy matching is optimized to improve the accuracy of group recommendation strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a system flow chart of the present invention; Figure 2 This is a flow chart of the user behavior monitoring module of the present invention; Figure 3 This is a flow chart of the path trajectory extraction module of the present invention; Figure 4 This is a flow chart of the dynamic work and rest modeling module of the present invention; Figure 5 This is a flow chart of the behavior cycle analysis module of the present invention; Figure 6 This is a flow chart of the user portrait generation module of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0019] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0020] The entire process of data acquisition, calculation, inference and application involved in this invention strictly complies with current laws, regulations and ethical norms to ensure that the data source is legal, transparent and effectively authorized, and does not infringe on personal privacy or third-party rights; the calculation and inference process adopts an algorithm model that meets industry standards, and uses secure encryption and anonymization technology to ensure the security and privacy of data processing, avoiding any form of biased or discriminatory output; the final application scenario is clearly limited to the legal and compliant field, fully assessing and avoiding potential social risks, ensuring that technological achievements serve the public welfare and social development needs, and all links have passed independent ethical review and legal compliance verification to fully safeguard social morality and public interests.

[0021] Embodiment 1 See also Figure 1 The present invention provides a technical solution: a digital marketing management system for telecommunication services comprising: The user behavior monitoring module obtains cellular communication records, collects event types, signaling strength, communication frequency, and dwell time, sorts them according to timestamps, and calculates the mean maximum fluctuation rate by partition to obtain the event intensity change rate value within the user's continuous segment; The path trajectory extraction module determines the trajectory status based on the event intensity change rate value, collects the base station number access status timestamp, sorts and generates a node sequence, counts the switching frequency, and obtains a high-frequency path node sequence; The dynamic work and rest modeling module detects the stable residence time based on the high-frequency path node sequence, collects the residence data of multiple regions in the same period, counts the frequency of downlink requests and uplink events, merges the frequency values, determines the stable interval, filters the repeated time periods, sorts out the intra-day distribution, accumulates the number of days, builds the work and rest model, and obtains the distribution value of the core work and rest period; The behavior cycle analysis module calls the work and rest time of consecutive days according to the distribution value of the core work and rest period, detects the period offset, filters the offset period, sorts the work and rest nodes, and obtains the dynamic prediction result of the work and rest cycle offset; The user portrait generation module calls the dynamic prediction results of the work and rest cycle offset, identifies the holiday pattern labels, collects high-frequency path nodes and core dwelling features, counts the spatiotemporal distribution, generates user behavior portraits, and obtains the dynamic behavior portrait parameter set.

[0022] The event intensity change rate values ​​within the user's continuous segment include dynamic indicators, extreme value indicators, and fluctuation indicators. The high-frequency path node sequence includes core nodes, time series associations, and switching identifiers. The core work and rest period distribution values ​​include main period identifiers, distribution characteristics, and concentration indicators. The dynamic prediction results of the work and rest cycle offset include cycle offset, offset amplitude, and dynamic trend. The dynamic behavior portrait parameter set includes pattern labels, path characteristics, and time and space parameters.

[0023] See also Figure 2 , the user behavior monitoring module includes: The cellular communication data collection submodule obtains cellular communication record data, collects event type, signaling strength, communication frequency and dwell time, and sorts all data according to timestamps to obtain time-series cellular communication data; It is necessary to call multiple collection devices to monitor the user's cellular communication behavior, including base station equipment, terminal equipment and signal relay equipment. These devices record the user's communication behavior data at different time points, including event type, signaling strength, communication frequency and residence time. During the data collection process, each base station will record the timestamp when the user accesses, and store the signaling strength, communication frequency and residence time information, and use the synchronous clock to time align the data of different base stations. After data collection, the system will sort the data according to the timestamp to ensure the integrity of the time series. For example, at a certain time T1, base station A records the user's signaling strength as -85dBm, the residence time is 120s, and the communication frequency is 3 times / minute, while base station B records the signaling strength as -80dBm, the residence time is 100s, and the communication frequency is 4 times / minute at time T2. By comparing the data of multiple base stations, the communication behavior of users in different areas can be restored, and finally the complete time series cellular communication data can be obtained.

[0024] The event signal analysis submodule calculates the mean and maximum values ​​of the signaling strength, communication frequency and dwell time in the same time interval based on the time series cellular communication data, and divides the multiple data into time series to identify the change trend in the differentiated intervals and obtain the interval signal statistical characteristics; First, the system filters the data in the selected time interval, extracts all signaling strength values, communication frequency values, and dwell time values ​​in the time period, and then calculates the mean and maximum values ​​of these data. For example, assuming that in a certain time interval T1-T2, the collected signaling strength data is [-85,-83,-80,-78,-75]dBm, the communication frequency data is [3,5,4,6,7] times / minute, and the dwell time data is [120,110,95,130,140] seconds, then the calculated mean values ​​are signaling strength -80.2dBm, communication frequency -66.6dBm, and dwell time -66.6dBm. The maximum values ​​are -75dBm signaling strength, 7 times / minute communication frequency, and 140 seconds residence time. Then, the data is divided into multiple sub-intervals according to the time series, and the change trends of adjacent intervals are calculated. For example, the time interval is divided into T1-T3, T3-T5 and other sub-intervals, and the mean changes of each interval are calculated. If the mean signaling strength of a certain interval decreases by more than 10%, the communication frequency increases by more than 30%, and the residence time decreases by more than 20%, then the interval is identified as a significant signal change interval, and finally the interval signal statistical characteristics are obtained.

[0025] The continuous segment intensity change calculation submodule calls the interval signal statistical characteristics, divides the data according to the timestamp, and uses the formula: ; Calculate the signal strength change rate in the continuous segment to obtain the user's continuous segment signal change rate value; in, Represents the rate of change of signal strength, Representative The signal strength at each time point, represents the total number of data points in the time window of the selected segment, Represents the total duration of the time window, Represents the difference in signal strength at adjacent time points, Represents a time point index.

[0026] First, obtain all signal strength values ​​in the selected time interval, and calculate the signal strength change ΔSSG between adjacent time points. At the same time, count the total duration of the time window ΔTWD, and calculate the average value of all signal strength values ​​in the time period, using the formula: ; Assuming that within a certain time window T1-T5, the signal strength measurement values ​​are -85dBm, -83dBm, -80dBm, -78dBm, and -75dBm, respectively, ΔSSG is calculated as follows: ; The total time window duration ΔTWD = 10 minutes, the total number of data points NTP = 5, then the summation term is calculated as follows: ; ; ; ; The result shows that the signal strength change rate during this time period is 1.9dB / minute. If the baseline value is set to 1.5dB / minute, the current change rate exceeds the baseline value, indicating that the signal change during this time period is more drastic and can be used to identify changes in user mobility status.

[0027] Table 1. Sample table of collected data

[0028] As shown in Table 1, the collected signaling strength, communication frequency and dwell time data are used to calculate the signal change trend and signal strength change rate.

[0029] See also Figure 3 ,The path trajectory extraction module includes: The trajectory state determination submodule calculates the rate change difference between adjacent time points based on the event intensity change rate value, and determines whether the trajectory state has changed. If the rate change exceeds the set threshold, it is marked as a trajectory switching point, the switching point set is called, and the corresponding timestamp is matched to establish a trajectory switching time series; First, obtain the event intensity change rate value of the device at different time points, calculate the rate change of adjacent time points, and obtain the difference set of rate changes. The rate calculation formula can be expressed as: ; in, Indicates time The rate at Indicates time The intensity of the event, For example, if a device The event intensity at is 120, The event intensity at is 135, and the time interval is 2 seconds, then the rate ; For rate change, define: ; If the rate at the previous time point is 6.5, then , each value in the calculated rate change set is equal to the set threshold To compare, for example, to set a threshold , if the calculated rate changes , then mark the current time point is the trajectory switching point, and the time point Add to the switching point set and match the timestamp, establish a time series in the database, and obtain the complete trajectory switching time series.

[0030] The node sequence generation submodule calls the trajectory switching time sequence, extracts the base station number and access status in chronological order, matches the timestamps and sorts them to generate the initial node sequence, filters the base station numbers that are continuously accessed, removes duplicate and abnormal data, and constructs the trajectory node sequence; For example, if the device frequently switches between base stations A and B in a short period of time (such as switching 5 times within 3 seconds), it can be determined that the switch is affected by external factors, such as signal interference or equipment failure. At this time, this part of the data needs to be eliminated; for the sorting of base station numbers, they are arranged in ascending order according to the timestamps. For example, the time series is {12:00, 12:02, 12:05, 12:10}, and the corresponding base station number sequence is {A, B, B, C}, and then redundant data is removed. For example, if the device is connected to base station B between 12:02 and 12:05, then one B is retained, and finally the trajectory node sequence {A, B, C} is constructed.

[0031] The high-frequency path analysis submodule calls the trajectory node sequence, counts the base station switching frequency, and calculates the connection strength between nodes based on the frequency using the formula: ; Obtain the influence weights between nodes through calculation, filter out high-frequency nodes above the threshold, and establish a high-frequency path node sequence; in, Represents the node influence weight, represents the number of switching times, p , q represents the node index, and r represents the other node indexes.

[0032] formula: ; in and represents the number of handovers between nodes p and q. If p, q, and r represent base stations A, B, and C respectively, the number of handovers is , , ,but: ; For influence weight greater than the threshold Node pairs (for example, setting ), filter out qualified high-frequency nodes, such as {A, B}, and finally construct a high-frequency path node sequence {A, B}.

[0033] Table 2 High frequency path analysis data table: As shown in Table 2 , the connection strength of AB in the high-frequency path analysis is the highest and meets the screening threshold, so A and B form a high-frequency path node sequence.

[0034] See also Figure 4 ,The dynamic work and rest modeling module includes: The residence period screening submodule detects the stable residence time of multiple regions based on the high-frequency path node sequence, collects the residence data of multiple regions in the same period, calculates the residence time ratio of multiple regions, and screens the period whose residence time ratio exceeds the set threshold to obtain the period with high residence ratio; First, the stable residence time in multiple areas is detected. During this process, the system traverses the data set, calculates the residence time of each location point, and determines whether it exceeds the set residence time threshold. The threshold is set by referring to the historical behavior data statistics. If the threshold is 10 minutes, the points with a residence time of less than 10 minutes will be eliminated, and only the high residence time points will be retained. The residence areas in adjacent time periods are further screened to see if they are continuous. If the continuous residence time meets the set requirements, the time periods are merged and the residence information of the area is recorded. Subsequently, after obtaining the residence data in multiple areas, the system calculates the residence time ratio in multiple areas. The ratio is calculated using the formula: ; in, Representative area The percentage of residence is the total residence time in the area, The sum of the dwell time in all areas is then filtered by the dwell ratio, and the threshold is set to 0.3, that is, the area with a dwell time ratio greater than 30% will be marked as a high dwell area. Combined with the example, assuming that the user stays in areas A, B, and C for 100 minutes, 50 minutes, and 30 minutes respectively, and the total dwell time is 180 minutes, then the dwell ratio of each area is: ; According to the set threshold of 0.3, only area A is screened into the period of high residence ratio, and finally the time period of high residence ratio is obtained and the relevant data is recorded.

[0035] The stable interval judgment submodule counts the downlink request frequency and uplink event frequency of the corresponding period based on the high residence ratio period, calculates the combined frequency value, determines whether the period belongs to the stable interval, and selects the period that appears repeatedly in multiple differentiated days to obtain the stable residence interval distribution; First, traverse the selected dwell time period, count the number of user downlink requests and uplink events in the period, and calculate the combined frequency value. The calculation method of the combined frequency value is as follows: ; in, Representative period The combined frequency value of is the frequency of downlink requests during this period, is the frequency of uplink events in this period. Assuming that the number of downlink requests in a period is 40 and the number of uplink events is 60, the combined frequency value is: ; Subsequently, the system sets a benchmark value for the merge frequency, and sets a threshold of 50 with reference to historical data statistics. That is, if the merge frequency value of a certain period exceeds 50, it is judged to belong to the stable interval, and the time periods that repeat in multiple differentiated days are continuously screened. The window sliding method is used to calculate whether the stable residence period of each day matches other days. If it repeats in at least 70% of the days, the period is finally judged to be a stable residence interval. If the statistical days are 10 days, a period that appears for more than 7 days will be selected into the stable residence interval, and the interval information will be stored.

[0036] The core work and rest construction submodule is based on the stable residence interval distribution, sorting out the distribution of multiple time periods within the day, accumulating the number of days with the same time period, and using the formula: ; The core work and rest period distribution values ​​are obtained by calculation and sorted into the time intervals within the day to obtain the core work and rest period distribution; in, Represents the distribution value of the core work and rest period, represents the cumulative duration of the stable residence period on the sth day, represents the combined frequency value of the s-th day period, Represents the total duration of stay on day s. represents the number of repetitions in the s-th day period, Represents the statistical day, and s represents the day index.

[0037] The formula is as follows: ; in, For the The cumulative duration of the stable residence period of the day, Representative The combined frequency value for the day period, For the Total duration of stay for the whole day, For the The number of recurrences of the day period, Represents the number of statistical days. Assuming that in the statistical data of 5 days, the daily cumulative residence time, combined frequency value, total residence time for the whole day, and the number of repeated occurrences are as follows: Table 3 Core work and rest parameter data table Enter the calculation formula: ; Calculate each term: ; The final distribution value of the core work and rest period is 123.29. After sorting out the time intervals within each day, the final distribution result of the core work and rest period is obtained.

[0038] See also Figure 5 , the behavior cycle analysis module includes: The time period offset screening submodule calls the work and rest time of consecutive days based on the core work and rest time period distribution value, calculates the time period intervals of multiple time points, obtains the change value of the work and rest frequency of differentiated time periods, screens the data points with significant time period changes, extracts the offset amplitude and direction of multiple work and rest time points, and calculates the standardized value and cumulative offset rate of the offset period to obtain the offset period data; First, call the work and rest time records for multiple consecutive days, convert the daily work and rest time points into the standard time format, extract the start time, end time and intermediate key time nodes of each day's work and rest, calculate the time difference between the corresponding time points of each adjacent two days, and obtain the offset of each time point. For example, if a person wakes up at 6:30, 6:45, 6:40, 6:50, 6:35, 6:55, and 6:30 for 7 consecutive days, then calculate the daily offset: the second day is offset by 15 minutes compared to the first day, the third day is offset by -5 minutes compared to the second day, and so on, to obtain the time offset array within the entire cycle. , standardize the time offset array to eliminate the influence of time units on the calculation, then perform statistics on the offset data at different time points, obtain the offset mean, standard deviation and cumulative offset rate of each time period, set the screening threshold for significant changes, for example, set 2 times the standard deviation as the judgment standard for significant changes, if the standard deviation is calculated to be 10 minutes, then the time point with an absolute value of the offset greater than 20 minutes is regarded as a significant change point, filter out the data at these time points and record them, for example, the calculated offsets of 6:55 and 6:30 are +20 minutes and -25 minutes respectively, which meet the standard of significant offset, then these two time points will be screened and recorded, and used as feature points of work and rest period offset, so that the time point data with significant changes in the entire work and rest period can be obtained, and finally the offset period data set is formed.

[0039] The work and rest node sorting submodule calls the offset period data, detects the high-frequency work and rest time points, calculates the offset trend of adjacent time points, determines the stability of the time nodes, selects the time points with a lower fluctuation range, re-divides the work and rest cycle nodes, and obtains the stable nodes of the work and rest cycle; First, extract the frequency of occurrence of all time points in the statistical period, calculate the cumulative number of occurrences of high-frequency time points, and set the frequency threshold. For example, if a time point appears 25 times in a 30-day statistical period, the frequency of occurrence of the time point is , if the high-frequency threshold is set to 0.8, then this time point is considered to be a high-frequency work and rest time point and is marked. Then the deviation trend of adjacent high-frequency time points is calculated. First, the average deviation amplitude of adjacent high-frequency time points is calculated, that is, For example, if the offset data at 6:30 are [-5, 10, -10, 5, -5], then the average offset amplitude is Minutes, the overall offset trend vector is calculated by combining multiple high-frequency time points, the offsets of adjacent time points are compared to determine whether it is stable, and the stability threshold of the offset fluctuation range is set. For example, if the offset fluctuation range of a time point is less than 5 minutes, the time point is considered to be relatively stable, and the time points with a lower offset fluctuation range are screened. The work and rest cycle is re-divided according to these stable time points. For example, if the offset fluctuation ranges of 6:30, 12:00, and 22:30 are all less than 5 minutes, these time points can be used as new work and rest cycle stable nodes, and finally the stable work and rest cycle time points are output.

[0040] The cycle shift prediction submodule calls the work and rest cycle stability node, calculates the shift trend and fluctuation range of multiple time points, and establishes a cycle shift calculation model using the formula: ; Calculate the cycle shift composite index, obtain the shift trend vector, analyze the shift change rate in combination with past data, and generate a dynamic prediction result for the work and rest cycle shift; in, represents the periodic shift composite index, represents the offset of day t, represents the offset direction coefficient on the tth day, represents the stability coefficient on day t, Represents the total number of days in the statistical period. represents the expected value of the offset, represents the regularization factor, and t represents the cycle day index.

[0041] First, set the parameter values: Set the number of days in the statistical period , the daily offset Using the above calculation method, such as taking the value array , offset direction coefficient Assign values ​​according to the sign of the offset, for example, a positive offset is set to 1, a negative offset is set to -1, and the stability coefficient Using the stability score at each time point of the day, if the scoring criteria are set as follows: fluctuation range <5 minutes, score 1; fluctuation range 5-10 minutes, score 0.8; fluctuation range >10 minutes, score 0.5, then the stability score array at a certain time point may be , calculate the expected offset Minutes, set the regularization factor , put it into the formula: ; Calculate the specific value: ; ; ; ; ; ; ; Sum .

[0042] The result shows that the deviation fluctuation of this period is large. The deviation change rate is calculated by combining the previous data. The shift change rate of the previous and next cycles is calculated. If the change rate is greater than the set threshold, the work and rest cycle shift trend is significant, and the dynamic prediction result of the work and rest cycle shift is finally generated, see Table 4.

[0043] Table 4 Work and rest cycle deviation calculation data table As shown in Table 4, time points with higher calculated offset values ​​indicate larger fluctuations, which can be used to adjust the work and rest cycle prediction later.

[0044] See also Figure 6 , the user portrait generation module includes: The work and rest activity variability analysis submodule calls the work and rest cycle deviation dynamic prediction results to calculate the activity deviation value of the user in the differentiated time period using the formula: ; Obtain the daily activity variability index through calculation and obtain the distribution parameters of work-rest deviation; in, represents the index of daytime activity variability, Representative Activity records at a certain point in time, Represents the total number of time points measured, Represents the average activity time at all time points, represents the standard deviation of all time points, and u represents the time point index; formula: ; in, represents the index of daytime activity variability, Indicates Activity records at a certain point in time, is the total number of time points, is the average activity time at all time points, is the standard deviation of all time points. This process involves multiple calculation steps. First, the user's activity data within 24 hours of a day is collected, and the activity status (such as walking, stillness, exercise intensity, etc.) is recorded every 15 minutes, for a total of 96 time points. For example, the activity data of a user at some time points are as follows: Table 5 User activity data records As shown in Table 5, the user's activity data at different time points is recorded and stored in an array . Then, calculate the mean of these data: ; Next, calculate the standard deviation : ; ; Then, calculate the exponential offset value for each time point: ; ; Similarly, calculate the offset values ​​of all time points and find the average to get : ; This value represents the variability of the user's daily activity. This indicates that the daily routine is highly unstable. This indicates that the work and rest schedule is relatively regular. Combined with practical applications, this data can be used to predict the user's work and rest schedule deviation trend and provide parameter support for subsequent work and rest pattern identification.

[0045] The holiday pattern recognition submodule compares the differences in users' work and rest schedules on weekdays and holidays based on the work and rest schedule deviation distribution parameters, counts the behavioral change trends in differentiated time periods, calculates the work and rest schedule deviations in multiple time periods, and selects holiday behavior feature values ​​based on the mean and standard deviation of the work and rest schedule deviations, selects the time period with the most obvious changes, and obtains holiday pattern labels; First, select the data for 7 consecutive days and calculate the : ; Compare the weekdays (first four days) and holidays (last three days) to calculate the mean: ; ; Calculate the standard deviation: ; ; by and As the screening criterion, , it is determined that the user has significant differences in his / her daily routine during holidays and is marked as "holiday mode".

[0046] The spatiotemporal behavior statistics submodule calls the holiday mode label, collects users' high-frequency path nodes and core dwelling locations, counts the dwelling time of multi-path nodes, calculates the dwelling time ratio of differentiated areas, calculates the spatiotemporal activity distribution according to the discrete degree of dwelling time, generates user behavior portraits, and obtains a dynamic behavior portrait parameter set.

[0047] First, select the residence time data of a user in different locations: Table 6 Statistics of user spatiotemporal behavior As shown in Table 6, the proportion of stay time in each area is calculated, and the distribution dispersion is evaluated based on the standard deviation to calculate the spatiotemporal activity index: ; ; ; when When , it means that the user's spatiotemporal behavior is relatively dispersed. The user's activity area is relatively concentrated. This parameter can be used to generate a dynamic behavior portrait of the user, extract high-frequency paths and core stop locations, and combine it with changes in holiday patterns to ultimately form a set of spatiotemporal behavior feature parameters.

[0048] A digital marketing management method for telecommunication services, the digital marketing management method for telecommunication services is implemented based on the digital marketing management system for telecommunication services, and includes the following steps: S1: Obtain cellular communication records, collect event type, signaling strength, communication frequency, and dwell time, sort them according to timestamps, calculate the mean, maximum value, and fluctuation rate by partition, and obtain the event intensity change rate value within the user's continuous segment; S2: Based on the event intensity change rate value within the user's continuous segment, determine the trajectory status, collect the base station number access status timestamp, sort and generate a node sequence, count the switching frequency, and obtain a high-frequency path node sequence; S3: Call the high-frequency path node sequence, detect the stable residence time, collect the residence data of multiple regions in the same period, count the frequency of downlink requests and uplink events, merge the frequency values, filter the stable interval according to the residence time, filter the repeated occurrence time period, sort out the intra-day distribution, accumulate the number of occurrence days, and establish the distribution value of the core work and rest time period; S4: According to the distribution value of the core work and rest period, call the work and rest time of consecutive days, detect the period offset, filter the offset period, sort the work and rest nodes, and obtain the dynamic prediction result of the work and rest cycle offset; S5: Call the dynamic prediction results of the work and rest cycle offset, combine them with the telecommunications business push data, filter the active time periods of the target users, match the adaptive marketing push time, and obtain the optimized value of the user marketing reach time period.

[0049] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A digital marketing management system for telecommunication services, characterized in that: The system comprises: The user behavior monitoring module obtains cellular communication records, collects event types, signaling strength, communication frequency, and dwell time, sorts them according to timestamps, and calculates the mean maximum fluctuation rate by partition to obtain the event intensity change rate value within the user's continuous segment; The path trajectory extraction module determines the trajectory state based on the event intensity change rate value, collects the base station number access state timestamp, sorts and generates a node sequence, counts the switching frequency, and obtains a high-frequency path node sequence; The dynamic work and rest modeling module detects the stable residence time based on the high-frequency path node sequence, collects the residence data of multiple regions in the same period, counts the frequency of downlink requests and uplink events, merges the frequency values, determines the stable interval, screens the repeated time periods, sorts out the intra-day distribution, accumulates the number of days, builds the work and rest model, and obtains the distribution value of the core work and rest period; The behavior cycle analysis module calls the work and rest time of consecutive days according to the distribution value of the core work and rest period, detects the period offset, filters the offset period, sorts the work and rest nodes, and obtains the dynamic prediction result of the work and rest cycle offset.

2. The digital marketing management system for telecommunication services according to claim 1, characterized in that: The event intensity change rate value within the user's continuous segment includes dynamic indicators, extreme value indicators, and fluctuation indicators. The high-frequency path node sequence includes core nodes, timing associations, and switching identifiers. The core work and rest period distribution value includes main period identifiers, distribution characteristics, and concentration indicators. The work and rest cycle offset dynamic prediction results include period offsets, offset amplitudes, and dynamic trends.

3. The digital marketing management system for telecommunication services according to claim 2, characterized in that: The user behavior monitoring module includes: The cellular communication data collection submodule obtains cellular communication record data, collects event type, signaling strength, communication frequency and dwell time, and sorts all data according to timestamps to obtain time-series cellular communication data; The event signal analysis submodule calculates the mean and maximum values ​​of the signaling strength, communication frequency and dwell time in the same time interval based on the time-series cellular communication data, divides the multiple data into time series, identifies the change trend in the differentiated intervals, and obtains the interval signal statistical characteristics; The continuous segment intensity change calculation submodule calls the interval signal statistical features, divides the data according to the timestamp, and uses the formula: ； Calculate the signal strength change rate in the continuous segment to obtain the user's continuous segment signal change rate value; in, Represents the rate of change of signal strength, Representative The signal strength at each time point, represents the total number of data points in the time window of the selected segment, Represents the total duration of the time window, Represents the difference in signal strength at adjacent time points, Represents a time point index.

4. The digital marketing management system for telecommunication services according to claim 3, characterized in that: The path trajectory extraction module comprises: The trajectory state determination submodule calculates the rate change difference between adjacent time points based on the event intensity change rate value, determines whether the trajectory state has changed, and if the rate change exceeds the set threshold, marks it as a trajectory switching point, calls the switching point set, matches the corresponding timestamp, and establishes a trajectory switching time sequence; The node sequence generation submodule calls the trajectory switching time sequence, extracts the base station number and access status in chronological order, matches the timestamps and sorts them to generate the initial node sequence, filters the base station numbers of continuous access, removes duplicate and abnormal data, and constructs the trajectory node sequence; The high-frequency path analysis submodule calls the trajectory node sequence, counts the base station switching frequency, and calculates the connection strength between nodes based on the frequency, using the formula: ； Obtain the influence weights between nodes through calculation, filter out high-frequency nodes above the threshold, and establish a high-frequency path node sequence; in, Represents the node influence weight, represents the number of switching times, p , q represents the node index, and r represents the other node indexes.

5. The digital marketing management system for telecommunication services according to claim 4, characterized in that: The dynamic work and rest modeling module includes: The residence period screening submodule detects the stable residence time of multiple regions based on the high-frequency path node sequence, collects the residence data of multiple regions in the same period, calculates the residence time proportion values ​​of multiple regions, and screens the period whose residence time proportion exceeds the set threshold to obtain the period with high residence proportion; The stable interval judgment submodule counts the downlink request frequency and the uplink event frequency of the corresponding period based on the high residence ratio period, calculates the combined frequency value, judges whether the period belongs to the stable interval, and selects the period that appears repeatedly in multiple differentiated days to obtain the stable residence interval distribution; The core work and rest construction submodule sorts out the distribution of multiple time periods within a day based on the stable residence interval distribution, accumulates the number of days with the same time period, and uses the formula: ； The core work and rest period distribution values ​​are obtained by calculation and sorted into the time intervals within the day to obtain the core work and rest period distribution; in, Represents the distribution value of the core work and rest period, represents the cumulative duration of the stable residence period on the sth day, represents the combined frequency value of the s-th day period, Represents the total duration of stay on day s. represents the number of repetitions in the s-th day period, Represents the statistical day, and s represents the day index.

6. The digital marketing management system for telecommunication services according to claim 5, characterized in that: The behavior cycle analysis module includes: The time period offset screening submodule calls the work and rest time of consecutive days based on the core work and rest time period distribution value, calculates the time period intervals of multiple time points, obtains the work and rest frequency change values ​​of differentiated time periods, screens the data points with significant time period changes, extracts the offset amplitude and direction of multiple work and rest time points, and calculates the standardized value and cumulative offset rate of the offset period to obtain the offset period data; The work and rest node arrangement submodule calls the offset period data, detects the work and rest time points that appear frequently, calculates the offset trend of adjacent time points, determines the stability of the time nodes, selects the time points with a lower fluctuation range, re-divides the work and rest cycle nodes, and obtains the work and rest cycle stable nodes; The cycle shift prediction submodule calls the work and rest cycle stability node, calculates the shift trend and fluctuation range of multiple time points, and establishes a cycle shift calculation model using the formula: ； Calculate the cycle shift composite index, obtain the shift trend vector, analyze the shift change rate in combination with past data, and generate a dynamic prediction result for the work and rest cycle shift; in, represents the periodic shift composite index, represents the offset of day t, represents the offset direction coefficient on the tth day, represents the stability coefficient on day t, Represents the total number of days in the statistical period. represents the expected value of the offset, represents the regularization factor, and t represents the cycle day index.

7. The digital marketing management system for telecommunication services according to claim 6, characterized in that: The system further comprises: The user profile generation module calls the dynamic prediction result of the work and rest cycle offset, identifies the holiday mode label, collects high-frequency path nodes and core dwell features, counts the spatiotemporal distribution, generates user behavior profiles, and obtains a dynamic behavior profile parameter set; The dynamic behavior profile parameter set includes pattern labels, path characteristics, and time and space parameters.

8. The digital marketing management system for telecommunication services according to claim 7, characterized in that: The user portrait generation module includes: The work and rest activity variability analysis submodule calls the work and rest cycle deviation dynamic prediction result to calculate the activity deviation value of the user in the differentiated time period using the formula: ； Obtain the daily activity variability index through calculation and obtain the distribution parameters of work-rest deviation; in, represents the index of daytime activity variability, Representative Activity records at a certain point in time, Represents the total number of time points measured, Represents the average activity time at all time points, represents the standard deviation of all time points, and u represents the time point index; The holiday pattern recognition submodule compares the difference between the user's work and rest schedule on weekdays and holidays based on the work and rest schedule deviation distribution parameters, counts the behavior change trends of differentiated time periods, calculates the work and rest schedule deviations of multiple time periods, selects holiday behavior feature values ​​according to the work and rest schedule deviation mean and standard deviation, selects the time period with the most obvious changes, and obtains holiday pattern labels; The spatiotemporal behavior statistics submodule calls the holiday mode label, collects users' high-frequency path nodes and core dwelling locations, counts the dwelling time of multi-path nodes, calculates the dwelling time ratio of differentiated areas, calculates the spatiotemporal activity distribution according to the discrete degree of dwelling time, generates user behavior portraits, and obtains a dynamic behavior portrait parameter set.

9. A digital marketing management method for telecommunication services, characterized in that: The digital marketing management system for telecommunication services according to any one of claims 1 to 8 comprises the following steps: S1: Obtain cellular communication records, collect event type, signaling strength, communication frequency, and dwell time, sort them according to timestamps, calculate the mean, maximum value, and fluctuation rate by partition, and obtain the event intensity change rate value within the user's continuous segment; S2: Based on the event intensity change rate value in the continuous segment of the user, determine the trajectory state, collect the base station number access state timestamp, sort and generate a node sequence, count the switching frequency, and obtain a high-frequency path node sequence; S3: Call the high-frequency path node sequence, detect the stable residence time, collect the residence data of multiple regions in the same period, count the frequency of downlink requests and uplink events, merge the frequency values, filter the stable interval according to the residence time, filter the repeated time period, sort out the intra-day distribution, accumulate the number of days, and establish the distribution value of the core work and rest period; S4: according to the core work and rest period distribution value, call the work and rest time of consecutive days, detect the period offset, filter the offset period, sort the work and rest nodes, and obtain the dynamic prediction result of the work and rest cycle offset; S5: Call the dynamic prediction result of the work and rest cycle offset, combine it with the telecommunications service push data, filter the active time period of the target user, match the adaptive marketing push time, and obtain the optimized value of the user marketing reach time period.

Images