Game advertisement putting effect monitoring and analyzing system

Abstract

The invention relates to the technical field of advertisement putting monitoring, in particular to a game advertisement putting effect monitoring and analyzing system which comprises a task rhythm disturbance recognition module, an active starting offset analysis module, an advertisement display period overlap ratio analysis module, a behavior disturbance event matching module and an advertisement interference effect confirmation module. According to the method, the advertisement playing frequency in the task period is compared with the task time change, the performance interfered by the advertisement in the game process rhythm is captured, the login time continuous offset is analyzed in combination with the login time sequence, the advertisement display period and the offset behavior are subjected to intersection matching, and the dynamic display period is divided, so that the dynamic display efficiency is improved. The advertisement overlapping analysis precision is improved, a cross-behavior type marking combination and overlapping rate two-dimensional screening mode is adopted, interference event nodes are recognized, behavior nodes are matched based on the putting section and frequency information, the coverage and guide direction of the advertisement to game behaviors is defined, the advertisement interference recognition range is expanded, and the advertisement putting effect analysis process is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of advertisement delivery monitoring, and in particular to a game advertisement delivery effect monitoring and analysis system. Background Art

[0002] The field of advertising monitoring technology encompasses various technical solutions for collecting, recording, analyzing, and evaluating data from the entire ad display, click, and conversion process. Its core approach is to utilize information processing techniques to meticulously track and compile statistics on advertising activity across various media platforms. By establishing linkages between ad reach, user behavior data, and advertising resources, this technology enables continuous monitoring and data feedback of advertising performance throughout the advertising lifecycle. This technology is widely used in scenarios such as internet advertising, e-commerce promotion, mobile app advertising, and interactive media advertising.

[0003] Among them, the game advertising delivery effect monitoring and analysis system refers to the advertising data monitoring and evaluation system applied to the game scene. It collects data, attributes events and analyzes the delivery of advertising materials in embedded advertising content, incentivized video ads, trial ads or built-in promotional positions in the game. It mainly tracks and organizes the user behavior chain after the advertisement is delivered through recording the user's advertising exposure behavior during the game, analyzing the logs of user click actions and subsequent conversion paths, and modeling the correlation between advertising content and user interaction behavior. Usually, the monitoring and attribution analysis of the game advertising effect are completed by reading log buried data, extracting advertising event identification information, comparing the correlation between user identification and behavior time series, and statistical matching analysis between advertising touchpoints and conversion events.

[0004] Existing analysis of game advertising effectiveness primarily focuses on the reach and conversion results of the ads themselves, often overlooking the dynamic relationship between player behavior rhythms and ad display timing. This lacks temporal trend tracking and the ability to identify rhythmic disturbances in behavioral pattern analysis, making it difficult to identify sources of interference in task execution anomalies or changes in user activity. This limits in-depth understanding of advertising effectiveness within gaming scenarios. Static analysis approaches to ad data ignore the persistent nature of behavioral shifts, leading to analytical gaps within continuous behavioral chains. For example, when a player gradually adjusts their login time over multiple days, it's difficult to determine whether a single behavior point is responsible for the shift, thereby masking the interference signal. The lack of a modeling mechanism for ad display cycles prevents accurate matching of overlapping behavior segments, leading to a fuzzy understanding of the interactive relationship between user behavior changes and ad exposure. Due to the lack of systematic integrated analysis of task flows, login behavior, and ad cycles, existing solutions struggle to accurately identify behavioral disturbance events or provide a detailed breakdown of ad-influenced behavior categories, hindering the effective optimization of advertising strategies. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a game advertising effect monitoring and analysis system.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solution: A system for monitoring and analyzing the effect of game advertising includes:

[0007] The task rhythm disturbance identification module obtains the current game task data and historical task data of the same period, calculates the difference between the task duration and the number of ad playbacks in the two cycles, determines whether the current task duration has been extended and the number of ad playbacks has increased, and generates a game task rhythm disturbance mark record;

[0008] The active launch offset analysis module obtains the sequence of players' first login time points over multiple consecutive days, compares the login time differences between adjacent dates, determines whether there is a continuous backward time offset, compares it with the ad display period, and generates a login behavior offset marker sequence;

[0009] The advertising display cycle overlap analysis module organizes the advertising display time points, divides the advertising exposure time periods, combines the task rhythm interference mark record and the login behavior offset mark sequence time period information, records the start and end range of the overlap section, and generates a detailed list of advertising cycle overlaps;

[0010] The behavior disturbance event matching module integrates the task rhythm interference mark record, the login behavior offset mark sequence and the advertising cycle overlap detailed list, filters the node combinations with simultaneous marking traces and different behavior types, determines whether there is an intersection in the time segment, and generates an advertising behavior disturbance event node set.

[0011] As a further solution of the present invention, the game task rhythm interference mark record includes the interference time period, time offset amplitude, and advertising playback frequency change; the login behavior offset mark sequence includes the offset days, offset amplitude, and offset trend; the advertising cycle overlap details list includes the overlapping segment start and end time, overlapping advertising sequence, and display cycle overlap degree; the advertising behavior disturbance event node set includes event time period, behavior type combination, and overlap rate index.

[0012] As a further solution of the present invention, the task rhythm disturbance identification module includes:

[0013] The task time acquisition submodule extracts the current task start time, end time and task cycle identifier based on the system record data of the current game task, calculates the difference between the task execution time period length and the cycle number within the cycle, and generates the task time difference;

[0014] The ad frequency calculation submodule counts the number of ad plays based on the ad insertion records within the current game task cycle, and calculates the difference between the number of ad plays in the current cycle and the number of ad plays in the same historical tasks, generating the ad play count difference.

[0015] The rhythm interference mark submodule determines whether there is a rhythm abnormality during the execution of the current task based on the task time difference and the advertisement playback number difference, and whether the two conditions of the task time difference exceeding the period average time benchmark value and the advertisement playback number difference exceeding the advertisement frequency benchmark value are simultaneously met, and generates a game rhythm interference mark record.

[0016] As a further solution of the present invention, the active startup offset analysis module includes:

[0017] The login time extraction submodule obtains the time sequence of players' first logins for multiple consecutive days, arranges each player's first login time in chronological order, removes missing data items and performs cleaning processing to generate a continuous login time sequence;

[0018] The time difference determination submodule compares the login time differences between adjacent dates based on the continuous login time sequence, and determines whether there is a continuous time backward shift trend, extracts all player data that meet the continuous shift condition, and generates a login time shift trend record;

[0019] The behavior deviation marking submodule obtains the daily ad display start and end time interval based on the login time deviation trend record, and compares it with the player's daily login time using the formula:

[0020]

[0021] Calculate the coupling degree of player deviation behavior on day i Compare the deviation behavior coupling degree with the deviation intervention judgment threshold to obtain the daily login behavior deviation status and mark it, and generate the login behavior deviation mark sequence, where: Represents the time when the player first logs in on day i, and Respectively represent the start and end time points of the ad display on day i, Represents the total duration of the ad display interval in the past n days. The login time on day i-1.

[0022] As a further solution of the present invention, the advertisement display period overlap analysis module includes:

[0023] The period division submodule obtains all ad display time points, collects the time intervals between adjacent display times, calculates the natural interval sequence of ad display based on the moving average method, divides the sequence into continuous display period segments based on the relationship between the time point intervals and the preset time window threshold, and generates an ad display period segment set;

[0024] The interference matching submodule obtains all time period boundary values in the task rhythm interference mark record and the login behavior offset mark sequence based on the advertisement display cycle segment set, compares each of them to see if they fall within the start and end time range of any advertisement display cycle segment, identifies all overlapping segments, and generates a time period overlap identification record;

[0025] The overlap record submodule obtains the numbers and boundaries of all overlapping display segments according to the overlap identification records of the time period, extracts the display length and interference duration of each overlapping segment, and uses the formula:

[0026]

[0027] Calculate the advertising cycle overlap R, combine the timeline sorting of each segment, establish a complete index of the number and start and end range, and generate a detailed list of advertising cycle overlaps, where G j represents the length of the ad exposure segment in the jth display cycle, D j represents the length of the jth interference or offset segment, and J represents the number of overlapping time periods.

[0028] As a further solution of the present invention, the behavior disturbance event matching module includes:

[0029] The tag screening submodule obtains the task rhythm interference tag record, the login behavior offset tag sequence, and the advertising cycle overlap detailed list, screens node combinations with simultaneous tag features and different behavior types on the unified timeline, establishes behavior interleaving mapping relationships according to tag types, and generates a behavior node combination list;

[0030] The time overlap calculation submodule, based on the behavior node combination list, performs intersection judgment on the task interference period, login offset period, and advertising cycle segment in each group, obtains the overlapping duration, and extracts the total duration of the advertising display segment and the corresponding player behavior offset days using the formula:

[0031]

[0032] Calculate the time overlap index O N , compared with the industry evaluation standards, screened the combined time periods that meet the valid conditions, and generated a valid overlapping segment index set, where Z k is the length of overlapping time in the kth segment, H kis the advertisement display duration in the kth segment, B N is the number of days of the behavior shift of the Nth group, and K is the number of advertising cycle segments involved in the current group;

[0033] The event node extraction submodule extracts the corresponding interference type, login offset type and advertising cycle number according to the valid overlapping segment index set, constructs a ternary behavior intervention mapping relationship, and sorts the time period numbers where the intervention occurs to obtain the advertising behavior disturbance event node set.

[0034] As a further embodiment of the present invention, the system further comprises:

[0035] The advertising interference effect confirmation module matches the delivery plan segment and display frequency range indicated in the advertising delivery record one by one according to the node time period of the node concentration of the advertising behavior disturbance event node, determines whether the corresponding node covers the delivery segment and analyzes whether it affects the player's gaming behavior. If the conditions are met, the corresponding advertisement number is extracted and the content of the affected behavior is marked to obtain the game advertising delivery effect analysis results;

[0036] The game advertisement delivery effect analysis results include advertisement number, impact behavior type, and impact confirmation status.

[0037] As a further solution of the present invention, the advertising interference effect confirmation module includes:

[0038] The plan matching submodule obtains the node time period information in the advertising behavior disturbance event node set, compares the delivery plan segment and display frequency range in the advertising delivery record one by one, screens whether there is a node time period that is completely covered by the plan segment, records the corresponding advertisement number and hit count, and generates a delivery plan coverage match list;

[0039] The behavior impact determination submodule extracts the player behavior characteristic indicators associated with each hit node based on the coverage match list of the delivery plan, and compares them with the behavior of uncovered nodes in the same period to determine whether there are behavioral change characteristics. It combines the ad display frequency with the player login behavior for joint analysis to determine the existence of interference behavior and obtain the ad behavior interference determination label set;

[0040] The result generation submodule extracts the advertisement numbers with identifiable impact and the corresponding behavior type information based on the advertisement behavior interference determination tag set, performs one-to-one matching between the advertisement numbers and the behavior items, and obtains the analysis results of the game advertisement delivery effect.

[0041] Compared with the prior art, the advantages and positive effects of the present invention are:

[0042] In the present invention, by comparing the changes in the frequency of advertisement playback and the task time within the task cycle, the performance of advertisement interference in the rhythm of the game process is captured, and abnormal marks of task behavior are formed. The continuous offset of login time is analyzed in combination with the multi-day login time series, and the intersection matching of the advertisement display period and the offset behavior is performed to construct a behavior rhythm interference identification mechanism. The natural interval threshold of advertisements is generated by moving average statistics, the dynamic display cycle is divided, and the accuracy of advertisement overlap analysis is improved. A two-dimensional screening method of cross-behavior type tag combination and overlap rate is adopted to accurately identify interference event nodes, match behavior nodes based on delivery segment and frequency information, and clarify the coverage and guidance direction of advertisements on behavior. The overall processing logic realizes the interactive fusion of three types of data: advertisement rhythm, task execution and behavior offset, enhances the recognition granularity, attribution accuracy and behavior chain integrity, forms a systematic analysis structure for the path of advertising behavior impact, expands the scope of advertising interference identification, and improves the behavioral level verification capability of delivery effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is a system flow chart of the present invention;

[0044] Figure 2 This is a flow chart of the task rhythm disturbance identification module of the present invention;

[0045] Figure 3 This is a flow chart of the active start-up offset analysis module of the present invention;

[0046] Figure 4 This is a flow chart of the advertisement display cycle overlap analysis module of the present invention;

[0047] Figure 5 This is a flow chart of the behavior disturbance event matching module of the present invention;

[0048] Figure 6 This is a flow chart of the advertising interference effect confirmation module of the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to 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.

[0050] 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, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0051] See also Figure 1 , a game advertising delivery effect monitoring and analysis system includes:

[0052] The task rhythm disturbance identification module obtains the current game task start time, end time and the number of ad playbacks within the task cycle. Combined with historical task data for the same period, it calculates the difference between the task duration and the number of ad playbacks between the two cycles, and determines whether the current task duration has been extended and the number of ad playbacks has increased. If the conditions are met, a game task rhythm disturbance mark record is generated.

[0053] The active launch offset analysis module obtains a sequence of players' first login times over multiple consecutive days, compares the login time differences between adjacent dates, determines whether time offsets continue to occur, extracts ad display time periods, and compares them with daily login times to generate a sequence of login behavior offset markers.

[0054] The ad display cycle overlap analysis module organizes ad display time points and divides them into multiple display cycle segments based on a preset time window threshold (this threshold uses the moving average method to calculate the natural interval distribution of ads, defined as the mean of the time intervals between adjacent ad displays plus 1.5 times the standard deviation, obtained through training with historical delivery data). The module then combines the time period information in the task rhythm interference mark record and the login behavior offset mark sequence to analyze whether the corresponding time period information falls within any ad exposure period set, records the start and end ranges of all overlapping segments, and generates a detailed list of ad cycle overlaps.

[0055] The behavior disturbance event matching module integrates task rhythm interference mark records, login behavior offset mark sequences, and a detailed list of advertising cycle overlaps. It screens for node combinations with simultaneous mark traces and different behavior types on the timeline to determine whether there is an intersection in the time segment. If the time overlap index meets the threshold standard (using the general standard for advertising effectiveness evaluation in the gaming industry, when the time overlap index is ≥30% and the cumulative number of behavior offset days is ≥3), it is considered a valid overlap. This standard refers to the recommended parameters in the "Mobile Game Advertising Effectiveness Evaluation White Paper" v2.1), the corresponding time and behavior type are extracted to generate an advertising behavior disturbance event node set.

[0056] The advertising interference effect confirmation module matches the delivery plan segment and display frequency range indicated in the advertising delivery record one by one according to the node time period where the advertising behavior disturbance event node is concentrated, determines whether the corresponding node covers the delivery segment and analyzes whether it affects the player's gaming behavior. If the conditions are met, the corresponding advertising number is extracted and the content affecting the behavior is marked to obtain the analysis results of the game advertising delivery effect.

[0057] The game task rhythm interference mark record includes the interference time period, time offset amplitude, and advertising playback frequency change. The login behavior offset mark sequence includes the offset days, offset amplitude, and offset trend. The advertising cycle overlap details list includes the overlapping segment start and end time, overlapping advertising sequence, and display cycle overlap degree. The advertising behavior disturbance event node set includes event time period, behavior type combination, and overlap rate indicator. The game advertising delivery effect analysis results include the advertisement number, affected behavior type, and impact confirmation status.

[0058] See also Figure 2 ,The task rhythm disturbance identification module includes:

[0059] The task time acquisition submodule extracts the current task start time, end time and task cycle identifier based on the system record data of the current game task, calculates the difference between the task execution time period length and the cycle number within the cycle, and generates the task time difference;

[0060] Based on the system record data of the current game task, it is necessary to first call the task scheduling log in the task management platform and extract the "task start time", "task end time" and "task cycle number" of each task to form the basic time dimension data. In the specific implementation, it can be queried through the task ID "TSK001" that its start time is 8:00 on March 12, 2024, and the end time is 20:00 on March 12, 2024. The cycle number is "Cycle 5". Next, it is necessary to calculate the task time, that is, the time difference between the end time and the start time, which can be converted into a timestamp difference for calculation. For example: the end time corresponds to the time The time stamp is 1678603200, the start time is 1678560000, the time difference is 43200 seconds, and the corresponding application time is 12 hours. Then, according to the task cycle number, the average time of the task with the same number is extracted from the historical records. Assuming that the time taken for the first four tasks in the fifth cycle is 11 hours, 12 hours, 13 hours and 11.5 hours respectively, the average time is (11+12+13+11.5) / 4=11.875 hours. The difference between the current task time and the historical average is calculated, that is, 12-11.875=0.125 hours, which is converted to 7.5 minutes to generate the task time difference.

[0061] The ad frequency calculation submodule counts the number of ad plays based on the ad insertion records within the current game task cycle, and calculates the difference between the number of ad plays in the current cycle and the number of ad plays in the same historical tasks, generating the ad play count difference.

[0062] According to the advertisement insertion records in the current game task cycle, it is necessary to retrieve the advertisement playback data in the current cycle from the advertisement delivery log to obtain the cumulative number of advertisements inserted in each time period. Assuming that advertisements are delivered at 9:00, 11:00, 13:00, 15:00 and 18:00 in the current cycle, and the records are 2 times, 3 times, 1 time, 2 times, and 3 times respectively, the cumulative number of plays is 2+3+1+2+3=11 times. Then, based on the advertisement playback records of the 5th cycle task in the same period in history, it is assumed that the first 4 advertisement playback times in the historical records are 9 times, 10 times, 12 times and 10 times, and the average is (9+10+12+10) / 4=10.25 times. The difference in the number of advertisement plays is 11-10.25=0.75 times. For the convenience of storage, one decimal place can be retained to generate the difference result 0.8 times, and the difference in the number of advertisement plays is obtained.

[0063] The rhythm interference marking submodule determines whether there is a rhythm abnormality during the current task execution based on the difference between the task time difference and the number of advertisement playbacks, and whether the two conditions of the task time difference exceeding the period average time benchmark and the number of advertisement playbacks exceeding the advertisement frequency benchmark are met at the same time, and generates a game rhythm interference marking record;

[0064] According to the task time difference and the ad play count difference, it is necessary to judge whether the two indicators exceed the corresponding benchmark values. The task time difference benchmark is set to 10 minutes. This value is based on the stable interval formed by the task execution time variance and the maximum offset in the past 5 consecutive cycles. It is specifically defined by the fluctuation range of the task time in the statistical cycle within the range of ±1σ standard deviation. 10 minutes is the upper limit of the difference covered by 95% of the data in the cycle group. If the scale of subsequent tasks changes, its corresponding benchmark value needs to be adjusted synchronously. The shortening of the total task duration will tighten the threshold, and vice versa. The ad play count difference benchmark is set to 1.0 times. It comes from the premise that the average frequency of advertisement delivery (average number of times per hour) in the current task cycle is about 2 times. The deviation exceeds ±10 % threshold response setting, build a reasonable judgment threshold, this value will be adjusted with the density of advertisement insertion period and the duration of task, if the task time period is concentrated and advertisements are frequently delivered, the corresponding threshold will also rise synchronously, among which the task time difference is 7.5 minutes, which does not exceed the 10-minute threshold; the advertisement playback number difference is 0.8 times, which does not exceed the 1.0 threshold. Both conditions are not met, so it does not constitute a rhythm interference situation. If any condition is met, interference needs to be marked; judgment logic needs to be introduced here, and the difference value is input and compared with the threshold item by item. If all are False, the interference status is generated as "no interference". If all or any one is True, the status is "interference exists". The current status is "no interference", and finally a game rhythm interference mark record is generated.

[0065] Table 1 Task and advertising data record table

[0066]

[0067] Table 1 lists the start and end time of the current task, the duration and number of advertisements in the current cycle, and the corresponding historical average data. This table allows for intuitive comparison to determine whether each difference is below the threshold judgment standard.

[0068] See also Figure 3 , the active start offset analysis module includes:

[0069] The login time extraction submodule obtains the time sequence of players' first logins for multiple consecutive days, arranges each player's first login time in chronological order, removes missing data items and performs cleaning processing to generate a continuous login time sequence;

[0070] To obtain the time sequence of players' first logins for multiple consecutive days, first create a unique identifier for each player and extract the first login time field from the daily server log. Use the player ID and date as indexes for permutations and combinations to construct a continuous login time point dataset. Taking the user login records of a platform from March 1 to March 7, 2024 as an example, the login time sequence of player A is: March 1 08:12, March 2 08:09, March 3 08:16, March 4 08:19, March 5 08:25, March 6 08:31, March 7 08: 38. All data is read in timestamp format and converted to minute level for standardization. After eliminating invalid records caused by system failures or account switching, valid continuous data is obtained. During the data cleaning process, the minimum login interval threshold is set to 5 minutes. That is, when the interval between consecutive logins of the same account is less than 5 minutes, only the first record is retained. For example, if player B logs in at 07:59 and 08:01 on March 2, it is considered a single login behavior and 07:59 is taken as the valid record. The time series constructed by the above process serves as the basic data for subsequent offset judgment, generating a continuous login time series.

[0071] The time difference determination submodule compares the login time differences between adjacent dates based on the continuous login time series, and determines whether there is a continuous time backward shift trend. It extracts all player data that meet the continuous shift conditions and generates a login time shift trend record.

[0072] According to the continuous login time series, the time field in the daily record is called to calculate the time difference between the login times of any two adjacent days. The daily time points are converted into absolute values in minutes, and then a pairwise subtraction operation is performed to obtain the time offset vector. Taking player A as an example, the minute values corresponding to his daily login time are 492, 489, 496, 499, 505, 511, and 518, respectively. The time differences between two adjacent days are -3, 7, 3, 6, 6, and 7, respectively. The judgment logic for determining whether it is a continuous backward offset trend is: if the time difference for three or more consecutive days is positive and the difference is not less than the set offset base The offset benchmark value is 3 minutes, then the player is considered to have a continuous backward deviation behavior. The setting of the offset benchmark value here is based on the historical platform average offset value ± standard deviation results. The platform average in February 2024 is 1.2 minutes, and the standard deviation is 0.9 minutes. The offset benchmark value is 1.2 + 2 × 0.9 = 3 minutes. According to the above data, player A's offset values were positive for five consecutive days from March 3 to March 7, and were not less than 3 minutes, which meets the judgment conditions. Therefore, it is confirmed that there is a continuous deviation behavior. The player's login time offset trend value can be expressed as a trend vector [1, 1, 1, 1, 1], where 1 represents a backward deviation.

[0073] The behavior deviation marking submodule obtains the daily ad display start and end time interval based on the login time deviation trend record, and compares it with the player's daily login time using the formula:

[0074]

[0075] Calculate the coupling degree of player deviation behavior on day i Compare the deviation behavior coupling degree with the deviation intervention judgment threshold to obtain the daily login behavior deviation status and mark it, and generate the login behavior deviation mark sequence, where: Represents the time when the player first logs in on day i, and Respectively represent the start and end time points of the ad display on day i, Represents the total duration of the ad display interval in the past n days. The login time on day i-1;

[0076] Based on the login time offset trend record, the daily ad display start and end time periods are obtained. The first login time of each player is compared with the ad interval to determine whether there is any overlapping interval. Taking the platform delivery strategy as an example, the ad display time is set from 08:00 to 08:20 every day. If a player's first login time on the same day is later than 08:00 and earlier than 08:20, their login behavior is determined to be interfered with by advertising. The following formula is used for coupling calculation.

[0077] in, is the login time on day i (minutes), and is the start and end time of the advertisement, and the login time on the i-1th day is recorded as The total advertising time in the past 7 days is That is 140 minutes; taking player A's login time on the 6th and 7th days as an example, they are 08:31 and 08:38 respectively, that is The advertising period is 480-500 minutes, i.e. Substituting into the formula we get:

[0078]

[0079] The deviation intervention threshold of 200 is set based on aggregating the deviation behavior data of hundreds of consecutively logged-in users on the platform, extracting the corresponding coupling value distribution, and selecting the coupling value corresponding to the 75th percentile in this distribution as the intervention threshold for deviation identification to improve the sensitivity of identification to typical deviation behaviors. This value is closely related to the length of the ad display interval and the fluctuation range of login time. The shorter the ad duration and the smaller the login time fluctuation, the lower the threshold should be, and vice versa, it should be increased to dynamically match the interaction intensity between deviation behavior and ad influence at different stages. The results show that player A's deviation behavior coupling on the 7th day was 250.94. If the deviation intervention threshold is set to 200, this value is greater than the threshold, indicating that his login behavior highly overlaps with the ad display, and the system records it as 1; otherwise, it is recorded as 0. Therefore, the generated login behavior deviation mark sequence for the 7-day period is [0, 0, 0, 0, 1, 1, 1], where 1 indicates that the behavior deviation is successfully identified.

[0080] Table 2 Calculation table of advertising display and login overlap

[0081] date Login time (minutes) Last day's login time Ad start time Ad end time Coupling value Day 6 511 505 480 500 182.74 Day 7 518 511 480 500 250.94

[0082] As shown in Table 2, the player's login behavior from the 6th to the 7th day was in a continuous shift state and was all within the advertising interval. The corresponding coupling value was higher than the shift threshold of 200, and was thus recorded in the login behavior shift mark sequence.

[0083] The operational logic of this formula is to comprehensively measure the degree of overlap between login time offset behavior and ad display interval and its changing trend. The multiplication operation of the numerator is It represents the combination of the degree of deviation between the player's login time and the ad start time and the total duration of historical ad display. This product reflects the interactive relationship between the depth of the login time falling into the ad interval and the intensity of the ad. The closer the player's login time is to the end of the ad display and the longer the historical ad duration is, the larger the value of this part is, indicating a greater possibility of being disturbed by the ad. The denominator is the square root of the sum of the difference between the login time change and the square of the ad duration on that day, that is, The two types of variable factors are aggregated by addition, and the square root processing is used to reflect their harmonizing and compressing effects on the overall interference coupling degree, so that the formula achieves a balance between the numerical amplification factor and the normalization factor. Finally, the absolute value operation is used to ensure that the result is non-negative, which is convenient for judgment and subsequent threshold comparison.

[0084] See also Figure 4 , the advertising display cycle overlap analysis module includes:

[0085] The period division submodule obtains all ad display time points, collects the time intervals between adjacent display times, calculates the natural interval sequence of ad display based on the moving average method, divides the sequence into continuous display period segments based on the relationship between the time point intervals and the preset time window threshold, and generates an ad display period segment set;

[0086] To obtain all ad display time points, first filter the timestamps of ad exposure triggers in the daily record log. During the screening process, the complete log should be used as the basis, and abnormal timestamps (such as less than 1 second or greater than 24 hours) should be excluded to ensure the accuracy of the sampling time. Then, the ad display time points on the same day are sorted in chronological order. Suppose the ad display time points on a certain day are 13:00, 13:05, 13:11, 14:00, 14:50, 15:20, and 16:00, then the time intervals between adjacent display points are calculated as 5 minutes, 6 minutes, 49 minutes, 50 minutes, 30 minutes, and 40 minutes respectively. All intervals are collected as an interval sequence, and the rolling calculation method is used to obtain the moving average and standard deviation of the sequence. Specifically, the sequence mean is 30 minutes, and the standard deviation is 6 minutes. If the time window threshold is 15 minutes, then the time window threshold is set to 30 + 1.5 × 15 = 52.5 minutes. After rounding, 53 minutes is taken as the time window threshold. The time intervals are traversed again. If the time intervals between adjacent time points are less than or equal to 53 minutes, they are considered to be the same display cycle segment. Otherwise, they are divided into a new display cycle. For example, in the above example, the intervals from 13:00 to 13:11 are both less than the threshold and are classified as the first display cycle. The intervals from 13:11 to 14:00 are 49 minutes, which is less than the threshold and is still classified as the first display cycle. The intervals from 14:50 to 15:20 and 16:00 do not exceed the threshold, respectively, so they form the second display cycle. Finally, the display cycle segment can be divided into two period segments: 13:00-14:00 and 14:50-16:00. The resulting display cycle segment set is shown in the following table:

[0087] Table 3 Advertisement display period segment set table

[0088]

[0089]

[0090] As shown in Table 3, the display cycle division relies on the dynamic thresholds set by the mean and standard deviation of the display time interval. The time window threshold is set based on the statistical data of rolling sampling, effectively avoiding the deviation of artificial threshold setting. The generated advertising display cycle segment set is composed of the above-mentioned display time periods and serves as a reference set for the subsequent matching process.

[0091] The interference matching submodule obtains all time period boundary values in the task rhythm interference mark record and login behavior offset mark sequence based on the ad display cycle segment set, compares each value to see if it falls within the start and end time range of any ad display cycle segment, identifies all overlapping segments, and generates time period overlap identification records;

[0092] According to the set of advertising display cycle segments, the task rhythm interference mark records and the login behavior offset mark sequence are extracted separately. The extracted content is limited to the mark records with clear time interval boundaries. Assuming that there are two mark time periods in the rhythm interference mark sequence, namely 13:30-13:50 and 15:00-15:30, and the login behavior offset sequence has a time period of 14:55-15:10, then the three interference time periods are respectively constructed into a time interval set, and the intervals are compared with the segments in the previous advertising display cycle segment set one by one to determine whether the start time and end time of each interference time period are both within the start and end time of a certain advertising display segment, and the time coverage logic is used for comparison. Condition matching: For example, the interference time period 13:30-13:50 completely overlaps with display period 1 (13:00-14:00), so the match is established. 35:00-15:30 and 14:50-16:00 also belong to display period 2, so the match is also established. 14:55-15:10 also falls within display period 2, so the match is established. Therefore, all three interference time periods are marked as overlapping. The indexes of the successfully matched time periods are constructed as 1, 2, and 2, and the time period overlap status is "yes, yes, yes". The corresponding interference segment index table is established as the input condition for the subsequent overlap calculation. The generated time period overlap identification record is used to quantify whether the interference segment is nested or overlapped in the display segment.

[0093] The overlap record submodule obtains the numbers and boundaries of all overlapping display segments based on the overlap identification records of the time period, and extracts the display length and interference duration of each overlapping segment using the formula:

[0094]

[0095] Calculate the advertising cycle overlap R, combine the timeline sorting of each segment, establish a complete index of the number and start and end range, and generate a detailed list of advertising cycle overlaps, where G j represents the length of the ad exposure segment in the jth display cycle, D j represents the length of the jth interference or offset segment, and J represents the number of overlapping time periods;

[0096] According to the time period overlap identification records, extract all interference time periods marked as "yes" and their corresponding display cycles, and input the advertising segment length and interference segment length in the matching results as parameters. Assuming that the display cycle segment length is G1 = 60 minutes and G2 = 70 minutes, the corresponding lengths of the interference time periods are D1 = 20 minutes, D2 = 30 minutes, and D3 = 15 minutes, the matching sequences are (G1, D1), (G2, D2), and (G2, D3), a total of 3 groups of matches, which are substituted into the formula for calculation.

[0097] First calculate in sequence:

[0098] molecular:

[0099] G1·D1+G2·D2+G2·D3=60·20+70·30+70·15=4350;

[0100] Denominator:

[0101]

[0102] Substituting into the calculation, we get:

[0103]

[0104] A nonlinear overlap value is constructed by the ratio of the product and the sum of squares between the interference segment and the advertising display segment, so that long-term or short-term but high-frequency interference situations are included in the quantitative system, effectively balancing the comprehensive impact between display intensity and interference density; the result shows that the advertising cycle overlap in the current sample is 35.59. Combined with the set overlap benchmark value (for example, 40), it can be judged that the overlap is in the medium to low range and does not constitute a strong overlap. Based on this, a detailed list of advertising cycle overlap can be established, whose structure is the corresponding display cycle number, start and end time and the corresponding interference segment number and start and end time. The overlap value is listed in the last column as a reference indicator, which is used to identify concentrated overlapping risk segments in multiple samples.

[0105] The operational logic of this formula is to quantitatively express the degree of overlap between the advertising display segment and the interference time segment by constructing a relative ratio relationship between the numerator and the denominator. The numerator of the formula uses the display segment length G j and the interference segment length D jThe product of the two is summed to represent the joint effect area formed by the two in each overlapping pair, reflecting the overall intensity of the distribution of interference time within the advertising time. The denominator is constructed by summing the squares of the lengths of all display segments and the squares of the interference segment lengths and taking the square root. This constructs a normalized measurement structure similar to the Euclidean norm, which is used to eliminate the magnitude effect caused by differences in the length of the original data, thereby avoiding the dominance of a single dimension. The multiplication operation reflects the joint effect strength formed by the simultaneous existence of the two, and the addition and squaring operations are used to accumulate statistics over the entire interval. The square root of the square sum introduces the concept of normalization, making the overlap values comparable and quantifiable. The overall structure combines the two mechanisms of area weighting and norm normalization to form a nonlinear sensitive response to the overlap intensity.

[0106] See also Figure 5 , the behavior disturbance event matching module includes:

[0107] The tag filtering submodule obtains task rhythm interference tag records, login behavior offset tag sequences, and a detailed list of advertising cycle overlaps. It then filters node combinations with simultaneous tag features and different behavior types on a unified timeline, establishes behavior interleaving mapping relationships by tag type, and generates a list of behavior node combinations.

[0108] To obtain the task rhythm interference mark record, login behavior offset mark sequence and advertising cycle overlap details list, you first need to extract the time boundary value, mark type and behavior category of each record from the three types of records. For the task rhythm interference mark record, you can use the execution timestamp field in the daily task data to extract the start and end time period of the task, and record the date number when it is marked as interference. For example, April 1 to April 3 is the interference period. The login behavior offset mark comes from the user's continuous login record. The offset day set is extracted based on the offset judgment logic. For example, there is an obvious delayed login behavior on the 2nd, 3rd and 5th day. The advertising cycle overlap details record the advertising display time period, such as 08:30-09:00, 12:00-12:30, etc. These records standardize the expression of all time periods and unify them into the start and end timestamp format for comparison. Then, a time axis index is established based on the day, and the above three types of records are projected onto the time axis respectively. A behavior type comparison table is constructed, and labeling is performed according to whether the time period of each day contains the three types of records. When a certain date appears in rhythm interference and also appears in login offset, it is judged as a "combination mark" node. If any two of the three types of behaviors have records in the same time period, they can be preliminarily screened as possible interference nodes. For example, 08:30-09:00 on April 3, 2024 overlaps in task interference and advertising display, then this time period is stored in the list as a combination mark time period, and finally a behavior node combination list is obtained.

[0109] The time overlap calculation submodule, based on the behavior node combination list, determines the intersection of the task interference period, login offset period, and advertising cycle segment in each group, obtains the overlapping duration, and extracts the total duration of the advertising display segment and the corresponding player behavior offset days using the formula:

[0110]

[0111] Calculate the time overlap index O N , compared with the industry evaluation standards, screened the combined time periods that meet the valid conditions, and generated a valid overlapping segment index set, where Z k is the length of overlapping time in the kth segment, H k is the advertisement display duration in the kth segment, B N is the number of days of the behavior shift of the Nth group, and K is the number of advertising cycle segments involved in the current group;

[0112] Based on the list of behavioral node combinations, it is necessary to compare each group of nodes one by one to see if there is any intersection in each time period. Specifically, the task interference period, login offset period, and advertising display period are set as time period triplets respectively, and the line segment intersection judgment method is used to determine whether there is an intersection interval. For example, if the interference time period is 09:00-09:30 and the advertising time period is 09:15-09:45, then the intersection is 09:15-09:30, a total of 15 minutes. The cumulative number of days of login offset can be obtained by counting the number of consecutive offset dates in the behavior offset sequence. If the offset mark exists on April 1, April 2, and April 3, 2024, the cumulative number of days of offset is 3. If the advertising display time is 30 minutes and the overlap time is 15 minutes, the overlap rate is 50%, which meets the set 30% threshold. Among them, B N =3, then:

[0113]

[0114] The time overlap index is 1.08475. If the industry benchmark is set to 0.8, this group of records meets the validity conditions. The corresponding index and behavior type number are identified and recorded to form a valid overlapping segment index set.

[0115] Table 4 Advertising interference time overlap sample table

[0116]

[0117] Table 4 lists the overlapping sample data of advertisements, task interference, and offset logins, showing the temporal combination characteristics and participation quantification data of the overlapping segments. The above data provides the parameter source for formula calculation. The time of the overlapping segments and the duration of advertisements are extracted from actual user interaction logs and converted to minutes by timestamp conversion. The number of days of login offset is based on behavioral sequence statistics.

[0118] The operational logic of the formula is to form a quantitative assessment of the intensity of the disturbance event by jointly measuring the ad overlap rate and the persistence of login offset. Part of it represents the ratio of all overlapping durations to the total duration of the corresponding ad display, which is used to reflect the overlap density of behavior and ads on the timeline. The higher the ratio, the more likely the behavior disturbance is directly related to the ad cycle. Secondly, The square root of the cumulative number of days of behavioral deviation is introduced, and the square root operation is used to reduce the linear impact of the increase in the number of days of deviation, retaining its trend enhancement but not over-magnifying the value, reflecting the nonlinear accumulation characteristics of the continuity of behavioral intervention; the two are added to form the total impact measure, that is, the degree of spatial overlap and the continuity in time are considered in parallel, and then uniformly multiplied by The purpose of the coefficient is to keep the composite indicator value within a comparable range and avoid inflating the overall value due to the merging of multiple dimensions. The final output comprehensive result can be compared with the preset interference intensity threshold to determine whether it constitutes a valid disturbance event. The entire structure uses a combination of additive fusion, multiplicative scaling, and root dimension reduction to ensure effective superposition of multi-source interference information while controlling its scale consistency within the evaluation system.

[0119] The event node extraction submodule extracts the corresponding interference type, login offset type, and advertising cycle number based on the valid overlapping segment index set, constructs a ternary behavior intervention mapping relationship, and sorts the time period numbers of the intervention to obtain the advertising behavior disturbance event node set;

[0120] According to the valid overlapping segment index set, the task rhythm interference, login offset and advertising cycle information corresponding to each record are read respectively, and the time boundary and event number of each type of behavior are extracted from the record. For example, the interference segment corresponding to index 1 is task number T05, the offset sequence is the continuous offset record D04 under user ID U21, and the advertising number is A1023. The classification to which it belongs is retrieved in the third-party data source, such as the task is a limited-time challenge, the login behavior is the first login offset, and the advertisement is a splash screen recommendation. Then, the time periods of similar events are merged in chronological order, and the interference event level is marked. The event type combination is output in the form of triples, and finally the advertising behavior disturbance event node set is constructed.

[0121] See also Figure 6 , the advertising interference effect confirmation module includes:

[0122] The plan matching submodule obtains the node time period information in the advertising behavior disturbance event node set, compares the delivery plan segment and display frequency range in the advertising delivery record one by one, screens whether there is a node time period that is completely covered by the plan segment, records the corresponding ad number and hit count, and generates a delivery plan coverage match list;

[0123] To obtain the node time period in the advertising behavior disturbance event node set, you first need to read the timestamp data in the set one by one and parse its corresponding start time and end time. For example, if a node time period is from 14:30 to 15:00 on May 10, 2024, then the total length of its time period is 30 minutes. Then match the time period with the advertising delivery records one by one. Each advertising delivery record contains the delivery start time, end time and corresponding display frequency range. If the advertising plan records that the delivery period of an advertisement is from 14:00 to 15:30 on May 10, 2024, and the display frequency is 3 times per hour, it is necessary to determine whether the start and end times of the disturbance node are completely included in the advertising plan segment. In this case, the node is completely covered, so it is considered a successful match. Further determine whether the display frequency meets the matching requirements, that is, the time period should cover no less than 1 display frequency. Here, since the frequency is 3 times per hour, it is displayed once every 20 minutes on average. The theoretical number of displays within 30 minutes is 1.5 times, which is rounded up to 2 times, meeting the minimum display requirements. Therefore, it is considered that the node meets the coverage conditions, and the advertisement number (such as ADV-1032) and the number of hits are recorded. All time periods in the node set are traversed in turn and compared with the delivery records. Under a large amount of data, a double-layer loop structure can be constructed to complete the traversal, in which the inner layer is the traversal of the delivery plan, and the outer layer is the cyclic matching of the time period of the disturbance node. Finally, a set of corresponding relationships between advertisements and node matches is output to support subsequent behavior impact judgment operations. This process finally generates a delivery plan coverage match list.

[0124] The behavior impact determination submodule extracts the player behavior characteristic indicators associated with each hit node based on the delivery plan coverage match list, and compares them with the behavior of uncovered nodes in the same period to determine whether there are behavioral change characteristics. It combines the ad display frequency with the player login behavior for joint analysis to determine the existence of interference behavior and obtain the ad behavior interference determination label set;

[0125] Based on the coverage matching list of the delivery plan, a behavioral analysis is performed on each combination of hit advertising nodes. Specifically, the player game behavior records during the matching node coverage period, such as login time, number of active times, task completion rate, etc., are extracted and used as the interference group behavior data. At the same time, the player behavior records of the uncovered nodes in the same time period are collected as the control group data. The two sets of data correspond one to one with the player ID as the dimension. For example, if the time period of a certain advertisement hit is from 14:30 to 15:00 on May 10, 2024, the players involved are P101 and P102, whose average login time during this period is 12 minutes and the number of active times is 3 times, while the players in the control group are as follows: The average login time of P103 and P104 is 17 minutes, and the number of active times is 4. In terms of login time, there is a decrease of 5 minutes and the number of active times decreases by 1. To further confirm the interference effect, the difference in behavioral indicators is calculated based on the behavioral change dimension to determine whether the interference group is generally lower than the control group. If the lower ratio exceeds 50%, the behavior item is marked as interference behavior. Further reference is made to historical industry data to set the benchmark threshold. For example, the previous judgment standard was set as the decrease in activity must be greater than 15%, the decrease in login time must be greater than 10%, and the sample size must be no less than 10 people. In actual operation, the sampled behavior data is statistically summarized, as shown in Table 5.

[0126] Table 5 Advertising node behavior comparison data table

[0127] Player Number Grouping Type Login duration (minutes) Active times (times) Task completion rate (%) P101 Interference group 12 3 62 P102 Interference group 12 2 58 P103 control group 17 4 69 P104 control group 16 4 70

[0128] As shown in Table 5, the players in the interference group have a certain degree of decline in various behavioral indicators. Combined with the number of samples in the behavior group, the time coverage and the industry benchmark value, a comprehensive judgment is made as to whether it constitutes the influence of advertising behavior. If all the judgment conditions are met, the group of behaviors and the advertisement number are marked as interference labels together, and finally the advertising behavior interference judgment label set is obtained.

[0129] The result generation submodule extracts the ad numbers with identifiable impact and the corresponding behavior type information based on the ad behavior interference determination label set, performs a one-to-one matching between the ad numbers and the behavior items, and obtains the results of the game ad delivery effect analysis;

[0130] According to the advertising behavior interference judgment label set, the advertising numbers that meet the interference judgment criteria are extracted one by one, and the interference types of their corresponding behavior items are obtained, such as decreased login time, decreased active frequency, fluctuations in task completion rate, etc., and then combined with the corresponding time period range and disturbance node index, the corresponding relationship structure data between the advertising number and the interference behavior is constructed. During the operation, multiple interference behavior items need to be classified and deduplicated to ensure that the behavior items under a single advertising number are not recorded repeatedly. For example, if ADV-1032 simultaneously marks a decrease in active times and a decrease in task completion rate, the behavior item needs to be archived in a collective manner. Finally, the structural data of all advertising numbers and their affected behavior items are summarized to form a mapping table that can be used for systematic analysis, and output as a structured result data set, which is the analysis result of the game advertising effect.

[0131] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A system for monitoring and analyzing the effect of game advertising, characterized in that: The system comprises: The task rhythm disturbance identification module obtains the current game task data and historical task data of the same period, calculates the difference between the task duration and the number of ad playbacks in the two cycles, determines whether the current task duration has been extended and the number of ad playbacks has increased, and generates a game task rhythm disturbance mark record; The active launch offset analysis module obtains the sequence of players' first login time points over multiple consecutive days, compares the login time differences between adjacent dates, determines whether there is a continuous backward time offset, compares it with the ad display period, and generates a login behavior offset marker sequence; The advertising display cycle overlap analysis module organizes the advertising display time points, divides the advertising exposure time periods, combines the task rhythm interference mark record and the login behavior offset mark sequence time period information, records the start and end range of the overlap section, and generates a detailed list of advertising cycle overlaps; The behavior disturbance event matching module integrates the task rhythm interference mark record, the login behavior offset mark sequence and the advertising cycle overlap detailed list, filters the node combinations with simultaneous marking traces and different behavior types, determines whether there is an intersection in the time segment, and generates an advertising behavior disturbance event node set.

2. The game advertisement delivery effect monitoring and analysis system according to claim 1, characterized in that: The game task rhythm interference mark record includes the interference time period, time offset amplitude, and advertising playback frequency change; the login behavior offset mark sequence includes the offset days, offset amplitude, and offset trend; the advertising cycle overlap details list includes the overlapping segment start and end time, overlapping advertising sequence, and display cycle overlap degree; the advertising behavior disturbance event node set includes event time period, behavior type combination, and overlap rate index.

3. The game advertisement delivery effect monitoring and analysis system according to claim 1, characterized in that: The task rhythm disturbance identification module includes: The task time acquisition submodule extracts the current task start time, end time and task cycle identifier based on the system record data of the current game task, calculates the difference between the task execution time period length and the cycle number within the cycle, and generates the task time difference; The ad frequency calculation submodule counts the number of ad plays based on the ad insertion records within the current game task cycle, and calculates the difference between the number of ad plays in the current cycle and the number of ad plays in the same historical tasks, generating the ad play count difference. The rhythm interference mark submodule determines whether there is a rhythm abnormality during the execution of the current task based on the task time difference and the advertisement playback number difference, and whether the two conditions of the task time difference exceeding the period average time benchmark value and the advertisement playback number difference exceeding the advertisement frequency benchmark value are simultaneously met, and generates a game rhythm interference mark record.

4. The game advertisement delivery effect monitoring and analysis system according to claim 1, characterized in that: The active startup offset analysis module includes: The login time extraction submodule obtains the time sequence of players' first logins for multiple consecutive days, arranges each player's first login time in chronological order, removes missing data items and performs cleaning processing to generate a continuous login time sequence; The time difference determination submodule compares the login time differences between adjacent dates based on the continuous login time sequence, and determines whether there is a continuous time backward shift trend, extracts all player data that meet the continuous shift condition, and generates a login time shift trend record; The behavior deviation marking submodule obtains the daily ad display start and end time interval based on the login time deviation trend record, and compares it with the player's daily login time using the formula: Calculate the coupling degree of player deviation behavior on day i Compare the deviation behavior coupling degree with the deviation intervention judgment threshold to obtain the daily login behavior deviation status and mark it, and generate the login behavior deviation mark sequence, where: Represents the time when the player first logs in on day i, and Respectively represent the start and end time points of the ad display on day i, Represents the total duration of the ad display interval in the past n days. The login time on day i-1.

5. The game advertisement delivery effect monitoring and analysis system according to claim 1, characterized in that: The advertisement display cycle overlap analysis module includes: The period division submodule obtains all ad display time points, collects the time intervals between adjacent display times, calculates the natural interval sequence of ad display based on the moving average method, divides the sequence into continuous display period segments based on the relationship between the time point intervals and the preset time window threshold, and generates an ad display period segment set; The interference matching submodule obtains all time period boundary values in the task rhythm interference mark record and the login behavior offset mark sequence based on the advertisement display cycle segment set, compares each of them to see if they fall within the start and end time range of any advertisement display cycle segment, identifies all overlapping segments, and generates a time period overlap identification record; The overlap record submodule obtains the numbers and boundaries of all overlapping display segments according to the overlap identification records of the time period, extracts the display length and interference duration of each overlapping segment, and uses the formula: Calculate the advertising cycle overlap R, combine the timeline sorting of each segment, establish a complete index of the number and start and end range, and generate a detailed list of advertising cycle overlaps, where G j represents the length of the ad exposure segment in the jth display cycle, D j represents the length of the jth interference or offset segment, and J represents the number of overlapping time periods.

6. The game advertisement delivery effect monitoring and analysis system according to claim 1, characterized in that: The behavior disturbance event matching module includes: The tag screening submodule obtains the task rhythm interference tag record, the login behavior offset tag sequence, and the advertising cycle overlap detailed list, screens node combinations with simultaneous tag features and different behavior types on the unified timeline, establishes behavior interleaving mapping relationships according to tag types, and generates a behavior node combination list; The time overlap calculation submodule, based on the behavior node combination list, performs intersection judgment on the task interference period, login offset period, and advertising cycle segment in each group, obtains the overlapping duration, and extracts the total duration of the advertising display segment and the corresponding player behavior offset days using the formula: Calculate the time overlap index O N , compared with the industry evaluation standards, screened the combined time periods that meet the valid conditions, and generated a valid overlapping segment index set, where Z k is the length of overlapping time in the kth segment, H k is the advertisement display duration in the kth segment, B N is the number of days of the behavior shift of the Nth group, and K is the number of advertising cycle segments involved in the current group; The event node extraction submodule extracts the corresponding interference type, login offset type and advertising cycle number according to the valid overlapping segment index set, constructs a ternary behavior intervention mapping relationship, and sorts the time period numbers where the intervention occurs to obtain the advertising behavior disturbance event node set.

7. The game advertisement delivery effect monitoring and analysis system according to claim 1, characterized in that: The system further comprises: The advertising interference effect confirmation module matches the delivery plan segment and display frequency range indicated in the advertising delivery record one by one according to the node time period of the node concentration of the advertising behavior disturbance event node, determines whether the corresponding node covers the delivery segment and analyzes whether it affects the player's gaming behavior. If the conditions are met, the corresponding advertisement number is extracted and the content of the affected behavior is marked to obtain the game advertising delivery effect analysis results; The game advertisement delivery effect analysis results include advertisement number, impact behavior type, and impact confirmation status.

8. The game advertisement delivery effect monitoring and analysis system according to claim 7, characterized in that: The advertising interference effect confirmation module includes: The plan matching submodule obtains the node time period information in the advertising behavior disturbance event node set, compares the delivery plan segment and display frequency range in the advertising delivery record one by one, screens whether there is a node time period that is completely covered by the plan segment, records the corresponding advertisement number and hit count, and generates a delivery plan coverage match list; The behavior impact determination submodule extracts the player behavior characteristic indicators associated with each hit node based on the coverage match list of the delivery plan, and compares them with the behavior of uncovered nodes in the same period to determine whether there are behavioral change characteristics. It combines the ad display frequency with the player login behavior for joint analysis to determine the existence of interference behavior and obtain the ad behavior interference determination label set; The result generation submodule extracts the advertisement numbers with identifiable impact and the corresponding behavior type information based on the advertisement behavior interference determination tag set, performs one-to-one matching between the advertisement numbers and the behavior items, and obtains the analysis results of the game advertisement delivery effect.

Images