A game promotion algorithm and system

Abstract

The present invention discloses a game promotion algorithm and system, and specifically relates to the technical field of game promotion, including monitoring the changes of promotion evaluation features with time series, using wavelet to analyze the periodic information of game promotion activities in the game community, determining the expected promotion evaluation features in the game promotion activities through an autoregressive model, determining the continuous heat information in the game community based on the comparison between the data of actual promotion evaluation features and the expected promotion evaluation features, extracting the comments of users on the game promotion after the game promotion, using the sentiment dictionary to assign sentiment scores to each word, obtaining the sentiment score of each comment, and obtaining the comment information of the game promotion according to the number of views and the number of likes of each comment, and monitoring the performance of the game promotion activities in the game community by training a neural network model. The present invention helps to automatically monitor and adjust the game promotion strategy and improve the effectiveness of the promotion activities.

Description

Technical Field

[0001] The present invention relates to the field of game promotion technology, and more specifically, to a game promotion algorithm and system. Background Art

[0002] In the game promotion industry, evaluating the effectiveness of promotion activities is the key to ensuring effective use of resources and improving user engagement. Traditional evaluation methods cannot fully capture the complexity of user behavior and emotions, resulting in insufficient and ineffective adjustments to promotion strategies. Traditional methods fail to effectively identify the cyclical changes in promotion activities over time, limiting in-depth understanding of promotion effectiveness. In addition, many promotion activities fail to fully utilize user comments and feedback, resulting in insufficient understanding of user attitudes and emotions, which in turn affects the formulation of subsequent promotion strategies, ultimately leading to poor accuracy in effectiveness evaluation of promotion activities and timeliness of strategy adjustments.

[0003] In order to solve the above two defects, a technical solution is now provided. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a game promotion algorithm and system to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A game promotion algorithm specifically includes the following steps:

[0007] S1: Based on the periodicity of game promotion activities in the game community, monitor the changes of promotion evaluation characteristics over time series, and use wavelet analysis to analyze the periodicity information of game promotion activities in the game community;

[0008] S2: Determine the expected promotion evaluation features in game promotion activities through the autoregressive model, and determine the continuous popularity information in the game community based on the comparison between the actual promotion evaluation features and the expected promotion evaluation features;

[0009] S3: Quantify the effect of game promotion activities in the community through comprehensive analysis of periodic information and continuous popularity information in the game community;

[0010] S4: Extract users’ comments on game promotion after game promotion, use sentiment dictionary to assign sentiment scores to each word, obtain sentiment scores for each comment, and obtain comment information on game promotion based on the number of views and likes of each comment;

[0011] S5: Quantify the impact of reviews on game promotion and monitor the performance of game promotion activities in the game community by training a neural network model;

[0012] Among them, the periodic information in the game community is represented by the generalized periodicity coefficient;

[0013] The logic for obtaining the promotion periodicity coefficient is: setting a monitoring interval, obtaining the discrete time series data of the number of clicks on the game promotion link in the promotion evaluation feature in the monitoring interval, and marking the discrete time series data of the number of clicks on the game promotion link in the monitoring interval as: DJ n , where n = 1, 2, 3, ..., N, N is a positive integer, n is the number of clicks on the game promotion link collected in the unit interval, usually the interval collected in the unit interval is one day, and Daubechies wavelet is selected to decompose the time series into different scales;

[0014] Calculate the periodic peak coefficient of the game promotion link click volume in the monitoring interval. The calculation formula for the periodic peak coefficient is: Among them, FZ j is the cycle peak coefficient of the jth cycle in the monitoring interval, j = 1, 2, 3, ..., J, J is a positive integer, φ * is the conjugate of the wavelet function, a j is the scale of the jth period, b j is the shift of the wavelet function in the jth period, m is the number of clicks on the game promotion link in the period extracted by wavelet analysis, m = 1, 2, 3, ..., M, M is a positive integer;

[0015] Set the peak threshold and time difference threshold to perform peak detection and time point detection. The peak detection compares the cycle peak coefficient in the cycle extracted after wavelet analysis with the peak threshold. If the cycle peak coefficient is less than the peak threshold, the extracted cycle is filtered.

[0016] If the cycle peak coefficient is greater than the peak threshold, the number position of the peak game promotion link click volume in the cycle is obtained, and the number position of the peak game promotion link click volume in the previous cycle is obtained, and the number position of the peak game promotion link click volume in the obtained cycle is subtracted from the number position of the peak game promotion link click volume in the previous cycle to obtain the time difference;

[0017] The time difference is compared with the time difference threshold. If the weekly time difference is less than the time difference threshold, the extracted period is filtered;

[0018] If the time difference is greater than the time difference threshold, the extracted period is retained, the period peak coefficient of the retained period is obtained, and the period peak coefficient of the retained period is re-marked as: FZ k , where k = 1, 2, 3, ..., K, K is a positive integer;

[0019] Obtain the number of times the monitoring interval is promoted in the game community, and mark the number of times the monitoring interval is promoted in the game community as: c;

[0020] Calculate the promotion periodicity coefficient, the calculation formula is: Among them, TG zq To generalize the periodic coefficient, FZ yz is the peak threshold;

[0021] The continuous popularity information in the game community is expressed by the continuous promotion difference coefficient;

[0022] The logic for obtaining the continuous promotion difference coefficient is as follows: obtaining the discrete time series data of the click volume of the game promotion link in the promotion evaluation feature in the monitoring interval, and using the click volume data of the game promotion link collected in the first R unit intervals in the monitoring interval as the input data of the autoregressive model;

[0023] The expression of the autoregressive model is: t =φ 1 y 1 +φ 2 y 3 +……+φ r y t-r +∈ r ; Among them, y t is the number of clicks on the tth day predicted by the autoregressive model, φ 1 ,φ 2 、……、φ r are the coefficients of the model, ∈ r is the error term;

[0024] Get the predicted value of the number of clicks on the game promotion link collected within the next F unit intervals within the monitoring interval, and mark the predicted value of the number of clicks on the game promotion link collected within the next F unit intervals as: y f , where f = 1, 2, 3, ..., F, F is a positive integer, f is the number of the predicted value in the last F unit intervals, the actual value of the click volume of the collected game promotion link in the last F unit intervals in the monitoring interval is obtained, and the actual value of the click volume of the collected game promotion link in the last F unit intervals in the monitoring interval is marked as: DJ f ;

[0025] Calculate the coefficient of difference of continuous promotion, the calculation formula is: Among them, CY tg To continuously promote the coefficient of difference.

[0026] In a preferred embodiment, quantifying the effect of a game promotion activity in a region includes:

[0027] The effect of game promotion activities in the community is represented by the promotion influence evaluation coefficient. The promotion periodicity coefficient and the continuous promotion difference coefficient are weighted and calculated to build a promotion influence evaluation model and generate the promotion influence evaluation coefficient. The expression of the promotion influence evaluation coefficient is: pg yx =(α 1 +α 2 )ln(α 1 TG zq +α 2 CY tg +1); where pg yx is the promotion impact evaluation coefficient, α 1 is the proportionality coefficient of the generalized periodic coefficient, α 2 is the proportional coefficient of the continuous promotion difference coefficient, α 1 , α 2 Both are greater than 0.

[0028] In a preferred embodiment, obtaining comment information for game promotion includes:

[0029] The comment information of game promotion is expressed by the emotional fermentation coefficient;

[0030] The logic for obtaining the sentiment fermentation coefficient is as follows: extract user comments related to game promotion within the monitoring period from the game community platform, assign sentiment scores to each word through the sentiment dictionary, obtain the sentiment score of each comment, and mark the sentiment score of each comment as: PF q,h , where q = 1, 2, 3, ..., Q, Q is a positive integer, q is the number of each game promotion in the monitoring interval, h = 1, 2, 3, ... H, H is a positive integer, h is the number of the comment in the qth game promotion;

[0031] Get the number of views of each comment after each game promotion within the monitoring interval, and mark the number of views of each comment after each game promotion within the monitoring interval as: LL q,h , get the number of likes for each comment after each game promotion in the monitoring interval, and mark the number of likes for each comment after each game promotion in the monitoring interval as: DZ q,h ;

[0032] Calculate the comment impact coefficient using the following formula: Among them, PF q,h is the comment influence coefficient of the hth comment in the qth game promotion, β 1 is the specific proportional coefficient of pageviews, β 2 is a specific proportional coefficient for the number of likes;

[0033] According to the influence coefficient of each comment, each comment is divided into different categories, and the comment categories are marked as: ω g , where g = 1, 2, 3, ..., G, G is a positive integer, and g is the number of different types of comments;

[0034] According to the comment influence coefficient of each comment, the total comment influence coefficient after different game promotions is obtained through the summation formula, and the total comment influence coefficient after different game promotions is marked as: ZPF q ,in,

[0035] The comment influence coefficient and the proportion of different types of comments after each game promotion are used to build a regression model to generate the emotional fermentation coefficient, which is expressed as: Among them, FJ qg is the emotional fermentation coefficient, BL 1 BL 2 BL 3 ,……,BL q is the proportion of different types of comments after different game promotions, γ 1 , γ 2 , γ 3 ,……,γ q It is the influence weight of the proportion of different types of comments after the promotion of different games.

[0036] In a preferred embodiment, the performance of game promotion activities in the game community is monitored by training a neural network model, including:

[0037] The promotion influence evaluation coefficient, the emotional fermentation coefficient, the number of game promotions and the duration of continuous game promotion are used as the input of the neural network model training, and the promotion warning coefficient is used as the output of the neural network model training, and it is assigned to 0 or 1 to train the model, where 0 means that the promotion warning coefficient is less than the promotion warning coefficient threshold, and 1 means that the promotion warning coefficient is greater than the promotion warning coefficient threshold;

[0038] The game promotion warning coefficient is obtained through the formula. The expression of the promotion warning coefficient is: Among them, YJ tg To promote the early warning coefficient, τ is a specific proportional coefficient.

[0039] In a preferred embodiment, a game promotion system includes a periodicity analysis module, an autoregressive model module, an evaluation module, a sentiment analysis module, and a neural network module, and the modules are signal-connected;

[0040] The periodicity analysis module is used to extract periodicity information by performing wavelet analysis on the time series data of game promotion evaluation features;

[0041] The autoregressive model module is used to predict future promotion evaluation features through the autoregressive model, measure the effectiveness of current promotion activities by comparing the predicted results with actual data, and determine the continuous popularity information;

[0042] Evaluation module, which is used to comprehensively analyze periodic information and continuous popularity information to quantify the promotion effect in the game community;

[0043] Sentiment analysis module, which is used to extract user comments after promotion activities, and use sentiment dictionary to perform sentiment analysis on the comments, assign sentiment scores, and evaluate the emotional impact of game promotion activities on users and community feedback reactions;

[0044] The neural network module is used to integrate the data of each module, conduct comprehensive evaluation through the trained model, and monitor the overall performance of game promotion in real time.

[0045] Technical effects and advantages of the present invention:

[0046] The present invention monitors the changes of promotion evaluation characteristics over time series, predicts the game promotion effect and quantifies the promotion performance of game promotion in the game community by comparing the actual with the expected. It also extracts user comments and quantifies users' evaluation of game promotion. It monitors the performance of game promotion activities in the game community in real time based on a neural network model, which helps to automatically monitor and adjust game promotion strategies and improve the effectiveness of promotion activities. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0048] Figure 1 A flowchart of a game promotion algorithm of the present invention;

[0049] Figure 2 The present invention is a structural schematic diagram of a game promotion system. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] Example 1

[0052] Figure 1 A flowchart of a game promotion algorithm of the present invention is given, which specifically includes the following steps:

[0053] S1: Based on the periodicity of game promotion activities in the game community, monitor the changes of promotion evaluation characteristics over time series, and use wavelet analysis to analyze the periodicity information of game promotion activities in the game community;

[0054] S2: Determine the expected promotion evaluation features in game promotion activities through the autoregressive model, and determine the continuous popularity information in the game community based on the comparison between the actual promotion evaluation features and the expected promotion evaluation features;

[0055] S3: Quantify the effect of game promotion activities in the community through comprehensive analysis of periodic information and continuous popularity information in the game community;

[0056] S4: Extract users’ comments on game promotion after game promotion, use sentiment dictionary to assign sentiment scores to each word, obtain sentiment scores for each comment, and obtain comment information on game promotion based on the number of views and likes of each comment;

[0057] S5: Quantify the impact of reviews on game promotion and monitor the performance of game promotion activities in the game community by training a neural network model.

[0058] Game promotion is usually carried out in the game community, which is an important platform for promoting games and attracting players. Limited-time events within the game, such as holiday events and task challenges, are released in the community to stimulate players' interest and participation. By monitoring promotion evaluation features such as the number of clicks on game promotion links in the game community, it helps to determine the influence of game promotion activities on game community users.

[0059] Among them, the promotion evaluation features include but are not limited to the number of clicks on the game promotion link, the popularity of the topic in the game community, and the download rate of users in the game community. That is, the promotion evaluation features are used as a standard for quantifying the promotion of game activities.

[0060] According to the periodicity of game promotion activities in the game community, the periodicity information in the game community is determined by monitoring the changes in the promotion evaluation characteristics over time series. The continuous heat information in the game community is determined based on the comparison between the actual promotion evaluation characteristic data and the expected promotion evaluation characteristics. The periodicity information and continuous heat information in the game community are represented by the promotion periodicity coefficient and the continuous promotion difference coefficient, respectively.

[0061] It should be noted that the periodicity of game promotion activities in the game community refers to the fact that promotion activities are repeated at specific time intervals to form a regular cycle, that is, promotion activities are repeated in a regular pattern at different time points. The purpose is to continuously stimulate users' interest and participation, and maintain users' attention and activity on the game.

[0062] Analyzing the periodicity of game promotion activities in the game community based on wavelet analysis is helpful to evaluate the performance of game activities in the game community. Game promotion activities may present different user response patterns in different periods. That is, when the activity is just launched, user interaction and click volume may surge, while as time goes by, the interaction volume may weaken. Through wavelet analysis, these dynamic changes can be captured, the promotion effect of each stage can be evaluated, and the time points for promotion activities are identified to help staff optimize promotion time and content.

[0063] When using wavelet analysis to analyze the periodic information of game promotion activities in the game community, the following key points should be noted:

[0064] Select appropriate wavelet basis functions, including Haar wavelet, Daubechies wavelet and Morlet wavelet, etc. Use wavelet basis functions with better smoothness. The specific selection is made by professional staff and will not be elaborated here;

[0065] Choose the appropriate scale and resolution. Wavelet analysis can provide information in both the time domain and the frequency domain, but a balance needs to be made in terms of resolution. A high scale is more suitable for capturing long-term trends, with low time domain resolution (it is impossible to accurately locate the time point of occurrence), but it can discover long-term cycles. A low scale is suitable for capturing short-term high-frequency fluctuations, with high time domain resolution, and can accurately identify the instantaneous response of promotional activities. The specific selection is made by professional staff, so I will not elaborate on it here.

[0066] De-noising processing, during the wavelet transform process, the noise in the signal will also be decomposed into different scales. If not processed, the noise may interfere with the analysis results. Therefore, by selecting an appropriate threshold, the high-frequency components (noise) below the threshold are removed and useful signal information is retained, including setting a frequency peak threshold to screen the frequency peak and remove small fluctuations to avoid misjudging short-term invalid fluctuations as periodic phenomena. A time difference threshold is set to filter out small fluctuations that are not related to actual promotion activities and retain the main cycles related to the promotion activities. The frequency peak threshold and time difference threshold are selected by professional staff and will not be elaborated here.

[0067] The logic for obtaining the promotion periodicity coefficient is: setting a monitoring interval, obtaining the discrete time series data of the number of clicks on the game promotion link in the promotion evaluation feature in the monitoring interval, and marking the discrete time series data of the number of clicks on the game promotion link in the monitoring interval as: DJ n , where n = 1, 2, 3, ..., N, N is a positive integer, n is the number of clicks on the game promotion link collected in the unit interval, usually the interval collected in the unit interval is one day, and Daubechies wavelet is selected to decompose the time series into different scales;

[0068] It should be noted that the monitoring interval is a time series, which includes multiple game promotion activities and is usually a longer period of time, which can reflect the periodic information of the effectiveness of game promotion activities.

[0069] Calculate the periodic peak coefficient of the game promotion link click volume in the monitoring interval. The calculation formula for the periodic peak coefficient is: Among them, FZ j is the cycle peak coefficient of the jth cycle in the monitoring interval, j = 1, 2, 3, ..., J, J is a positive integer, φ * is the conjugate of the wavelet function, a j is the scale of the jth period, b j is the shift of the wavelet function in the jth period, m is the number of clicks on the game promotion link in the period extracted by wavelet analysis, m = 1, 2, 3, ..., M, M is a positive integer;

[0070] Set the peak threshold and time difference threshold to perform peak detection and time point detection. The peak detection compares the cycle peak coefficient in the cycle extracted after wavelet analysis with the peak threshold. If the cycle peak coefficient is less than the peak threshold, the cycle is considered insignificant and ignored, and the extracted cycles are filtered.

[0071] If the cycle peak coefficient is greater than the peak threshold, the number position of the peak game promotion link click volume in the cycle is obtained, and the number position of the peak game promotion link click volume in the previous cycle is obtained, and the number position of the peak game promotion link click volume in the obtained cycle is subtracted from the number position of the peak game promotion link click volume in the previous cycle to obtain the time difference;

[0072] Compare the time difference with the time difference threshold. If the weekly time difference is less than the time difference threshold, the cycle is considered insignificant and ignored, and the extracted cycles are filtered.

[0073] If the time difference is greater than the time difference threshold, the extracted period is retained, the period peak coefficient of the retained period is obtained, and the period peak coefficient of the retained period is re-marked as: FZ k, where k = 1, 2, 3, ..., K, K is a positive integer;

[0074] Obtain the number of times the monitoring interval is promoted in the game community, and mark the number of times the monitoring interval is promoted in the game community as: c;

[0075] Calculate the promotion periodicity coefficient, the calculation formula is: Among them, TG zq To generalize the periodic coefficient, FZ yz is the peak threshold.

[0076] From the formula, we can see that the larger the promotion periodicity coefficient is, the more likely each game promotion is to achieve a higher number of clicks, indicating that each promotion has achieved higher user engagement and response, and that users have a sustained interest in the game or promotion.

[0077] The expected promotion evaluation features in game promotion activities are determined through the autoregressive model, and the promotion activity effect is evaluated by comparing the predicted promotion evaluation features with the actual promotion evaluation features.

[0078] The logic for obtaining the continuous promotion difference coefficient is as follows: obtaining the discrete time series data of the click volume of the game promotion link in the promotion evaluation feature in the monitoring interval, and using the click volume data of the game promotion link collected in the first R unit intervals in the monitoring interval as the input data of the autoregressive model;

[0079] The expression of the autoregressive model is: t =φ 1 y 1 +φ 2 y 3 +……+φ r y t-r +∈ r ; Among them, y t is the number of clicks on the tth day predicted by the autoregressive model, φ 1 ,φ 2 、……、φ r are the coefficients of the model, ∈ r is the error term;

[0080] Get the predicted value of the number of clicks on the game promotion link collected within the next F unit intervals within the monitoring interval, and mark the predicted value of the number of clicks on the game promotion link collected within the next F unit intervals as: y f , where f = 1, 2, 3, ..., F, F is a positive integer, f is the number of the predicted value in the last F unit intervals, the actual value of the click volume of the collected game promotion link in the last F unit intervals in the monitoring interval is obtained, and the actual value of the click volume of the collected game promotion link in the last F unit intervals in the monitoring interval is marked as: DJ f;

[0081] Calculate the coefficient of difference of continuous promotion, the calculation formula is: Among them, CY tg To continuously promote the coefficient of difference.

[0082] From the formula, we can see that the larger the continuous promotion difference coefficient is, the greater the positive impact of the promotion activities on the number of clicks is likely to be. Conversely, the smaller the continuous promotion difference coefficient is, the smaller the continuous impact of the promotion activities on the number of clicks is likely to be, which means that the subsequent promotion activities are not attractive enough and fail to continue to attract users' interest.

[0083] The effect of game promotion activities in the community is represented by the promotion influence evaluation coefficient. The promotion periodicity coefficient and the continuous promotion difference coefficient are weighted to build a promotion influence evaluation model and generate the promotion influence evaluation coefficient. The expression of the promotion influence evaluation coefficient is: pg yx =(α 1 +α 2 )ln(α 1 TG zq +α 2 CY tg +1); where pg yx is the promotion impact evaluation coefficient, α 1 is the proportionality coefficient of the generalized periodic coefficient, α 2 is the proportional coefficient of the continuous promotion difference coefficient, α 1 , α 2 Both are greater than 0.

[0084] Based on each game promotion within the monitoring period, users' comments on the game promotion after the game promotion are extracted, and a sentiment score is assigned to each word using the sentiment dictionary to obtain the sentiment score of each comment. Based on the number of views and likes of each comment, the comment information of the game promotion is obtained, and the comment information of the game promotion is represented by the sentiment fermentation coefficient.

[0085] The comments information promoted through the game can reflect the following information:

[0086] Sentiment analysis can reveal users’ true feelings about the game and its promotions, which is crucial to understanding users’ satisfaction and dissatisfaction;

[0087] Users' emotional feedback on games directly affects their brand image and reputation. Positive comments can enhance brand value, while negative comments may lead to user loss. Through sentiment analysis, developers can identify potential negative public opinion in a timely manner and take corresponding measures to avoid affecting the market performance of the game.

[0088] By analyzing the sentiment of comments after a promotion, developers can evaluate the actual impact of the campaign and determine the user's acceptance of the promotional content.

[0089] The logic for obtaining the sentiment fermentation coefficient is as follows: extract user comments related to game promotion within the monitoring period from the game community platform, assign sentiment scores to each word through the sentiment dictionary, obtain the sentiment score of each comment, and mark the sentiment score of each comment as: PF q,h , where q = 1, 2, 3, ..., Q, Q is a positive integer, q is the number of each game promotion in the monitoring interval, h = 1, 2, 3, ... H, H is a positive integer, h is the number of the comment in the qth game promotion;

[0090] It should be noted that user comments related to each promotion activity are extracted from the game community platform, and the comments should correspond to the time window of the promotion activity to ensure the relevance of the data.

[0091] Get the number of views of each comment after each game promotion within the monitoring interval, and mark the number of views of each comment after each game promotion within the monitoring interval as: LL q,h , get the number of likes for each comment after each game promotion in the monitoring interval, and mark the number of likes for each comment after each game promotion in the monitoring interval as: DZ q,h ;

[0092] Calculate the comment impact coefficient using the following formula: Among them, PF q,h is the comment influence coefficient of the hth comment in the qth game promotion, β 1 is the specific proportional coefficient of pageviews, β 2 is a specific proportional coefficient for the number of likes;

[0093] According to the influence coefficient of each comment, each comment is divided into different categories, and the comment categories are marked as: ω g , where g = 1, 2, 3, ..., G, G is a positive integer, and g is the number of different types of comments;

[0094] According to the comment influence coefficient of each comment, the total comment influence coefficient after different game promotions is obtained through the summation formula, and the total comment influence coefficient after different game promotions is marked as: ZPF q ,in,

[0095] It should be noted that the classification is based on the comment influence coefficient, that is, the basis of classification is the influence of the comment on the overall promotion effect and user sentiment. The specific classification categories and the numerical range of the comment influence coefficient in the category are set by professional staff. For example, according to the numerical size of the comment influence coefficient, the categories are divided into large negative impact, medium negative impact, small negative impact, small positive impact, medium positive impact, and large positive impact.

[0096] The comment influence coefficient and the proportion of different types of comments after each game promotion are used to build a regression model to generate the emotional fermentation coefficient, which is expressed as: Among them, FJ qg is the emotional fermentation coefficient, BL 1 BL 2 BL 3 ,……,BL q is the proportion of different types of comments after different game promotions, γ 1 , γ 2 , γ 3 ,……,γ q It is the influence weight of the proportion of different types of comments after the promotion of different games.

[0097] From the formula, we can see that the larger the emotional fermentation coefficient is, the better the effect of the game promotion is, reflecting that the user's emotional response to the promotion activities increases or expands over time, indicating the long-term influence of the promotion activities in the game community.

[0098] The promotion influence evaluation coefficient, the emotional fermentation coefficient, the number of game promotions and the duration of continuous game promotion are used as the input of the neural network model training, and the promotion warning coefficient is used as the output of the neural network model training, and it is assigned to 0 or 1 to train the model, where 0 means that the promotion warning coefficient is less than the promotion warning coefficient threshold, and 1 means that the promotion warning coefficient is greater than the promotion warning coefficient threshold;

[0099] The game promotion warning coefficient is obtained through the formula. The expression of the promotion warning coefficient is: Among them, YJ tg To promote the early warning coefficient, τ is a specific proportional coefficient.

[0100] It should be noted that the promotion warning coefficient threshold is used to indicate whether the promotion strategy needs to be adjusted. The setting of the promotion warning coefficient threshold should be based on actual business needs and experience data and should be set by professional staff.

[0101] The present invention monitors the changes of promotion evaluation characteristics over time series, predicts the game promotion effect and quantifies the promotion performance of game promotion in the game community by comparing the actual with the expected. It also extracts user comments and quantifies users' evaluation of game promotion. It monitors the performance of game promotion activities in the game community in real time based on a neural network model, which helps to automatically monitor and adjust game promotion strategies and improve the effectiveness of promotion activities.

[0102] Example 2

[0103] Figure 2 A structural schematic diagram of a game promotion system of the present invention is given, which includes a periodic analysis module, an autoregressive model module, an evaluation module, a sentiment analysis module and a neural network module, and signal connections between the modules;

[0104] The periodicity analysis module is used to extract periodicity information by performing wavelet analysis on the time series data of game promotion evaluation features;

[0105] The autoregressive model module is used to predict future promotion evaluation features through the autoregressive model, measure the effectiveness of current promotion activities by comparing the predicted results with actual data, and determine the continuous popularity information;

[0106] Evaluation module, which is used to comprehensively analyze periodic information and continuous popularity information to quantify the promotion effect in the game community;

[0107] Sentiment analysis module, which is used to extract user comments after promotion activities, and use sentiment dictionary to perform sentiment analysis on the comments, assign sentiment scores, and evaluate the emotional impact of game promotion activities on users and community feedback reactions;

[0108] The neural network module is used to integrate the data of each module, conduct comprehensive evaluation through the trained model, and monitor the overall performance of game promotion in real time.

[0109] The above formulas are all calculated by taking their numerical values ​​after removing the dimensions. The formula is a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0110] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0111] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0112] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0113] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0114] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0115] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0116] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A game promotion algorithm, characterized in that: The specific steps include: S1: Based on the periodicity of game promotion activities in the game community, monitor the changes of promotion evaluation characteristics over time series, and use wavelet analysis to analyze the periodicity information of game promotion activities in the game community; S2: Determine the expected promotion evaluation features in game promotion activities through the autoregressive model, and determine the continuous popularity information in the game community based on the comparison between the actual promotion evaluation features and the expected promotion evaluation features; S3: Quantify the effect of game promotion activities in the community through comprehensive analysis of periodic information and continuous popularity information in the game community; S4: Extract users’ comments on game promotion after game promotion, use sentiment dictionary to assign sentiment scores to each word, obtain sentiment scores for each comment, and obtain comment information on game promotion based on the number of views and likes of each comment; S5: Quantify the impact of reviews on game promotion and monitor the performance of game promotion activities in the game community by training a neural network model; Among them, the periodic information in the game community is represented by the generalized periodicity coefficient; The logic for obtaining the promotion periodicity coefficient is: setting a monitoring interval, obtaining the discrete time series data of the number of clicks on the game promotion link in the promotion evaluation feature in the monitoring interval, and marking the discrete time series data of the number of clicks on the game promotion link in the monitoring interval as: DJ n , where n = 1, 2, 3, ..., N, N is a positive integer, n is the number of clicks on the game promotion link collected in the unit interval, usually the interval collected in the unit interval is one day, and Daubechies wavelet is selected to decompose the time series into different scales; Calculate the periodic peak coefficient of the game promotion link click volume in the monitoring interval. The calculation formula for the periodic peak coefficient is: Among them, FZ j is the cycle peak coefficient of the jth cycle in the monitoring interval, j = 1, 2, 3, ..., J, J is a positive integer, φ * is the conjugate of the wavelet function, a j is the scale of the jth period, b j is the shift of the wavelet function in the jth period, m is the number of clicks on the game promotion link in the period extracted by wavelet analysis, m = 1, 2, 3, ..., M, M is a positive integer; Set the peak threshold and time difference threshold to perform peak detection and time point detection. The peak detection compares the cycle peak coefficient in the cycle extracted after wavelet analysis with the peak threshold. If the cycle peak coefficient is less than the peak threshold, the extracted cycle is filtered. If the cycle peak coefficient is greater than the peak threshold, the number position of the peak game promotion link click volume in the cycle is obtained, and the number position of the peak game promotion link click volume in the previous cycle is obtained, and the number position of the peak game promotion link click volume in the obtained cycle is subtracted from the number position of the peak game promotion link click volume in the previous cycle to obtain the time difference; The time difference is compared with the time difference threshold. If the weekly time difference is less than the time difference threshold, the extracted period is filtered; If the time difference is greater than the time difference threshold, the extracted period is retained, the period peak coefficient of the retained period is obtained, and the period peak coefficient of the retained period is re-marked as: FZ k , where k = 1, 2, 3, ..., K, K is a positive integer; Obtain the number of times the monitoring interval is promoted in the game community, and mark the number of times the monitoring interval is promoted in the game community as: c; Calculate the promotion periodicity coefficient, the calculation formula is: Among them, TG zq To generalize the periodic coefficient, FZ yz is the peak threshold; The continuous popularity information in the game community is expressed by the continuous promotion difference coefficient; The logic for obtaining the continuous promotion difference coefficient is as follows: obtaining the discrete time series data of the click volume of the game promotion link in the promotion evaluation feature in the monitoring interval, and using the click volume data of the game promotion link collected in the first R unit intervals in the monitoring interval as the input data of the autoregressive model; The expression of the autoregressive model is: t =φ1y1+φ2y3+……+φ r y t-r +∈ r ; Among them, y t is the number of clicks on the tth day predicted by the autoregressive model, φ1, φ2, …, φ r are the coefficients of the model, ∈ r is the error term; Get the predicted value of the number of clicks on the game promotion link collected within the next F unit intervals within the monitoring interval, and mark the predicted value of the number of clicks on the game promotion link collected within the next F unit intervals as: y f , where f = 1, 2, 3, ..., F, F is a positive integer, f is the number of the predicted value in the last F unit intervals, the actual value of the click volume of the collected game promotion link in the last F unit intervals in the monitoring interval is obtained, and the actual value of the click volume of the collected game promotion link in the last F unit intervals in the monitoring interval is marked as: DJ f ; Calculate the coefficient of difference of continuous promotion, the calculation formula is: Among them, CY tg To continuously promote the coefficient of difference.

2. According to a game promotion algorithm according to claim 1, it is characterized in that: The quantitative game promotion activities in the community include: The effect of game promotion activities in the district is represented by the promotion influence evaluation coefficient. The promotion periodicity coefficient and the continuous promotion difference coefficient are weighted and calculated to build a promotion influence evaluation model and generate the promotion influence evaluation coefficient. The expression of the promotion influence evaluation coefficient is: pg yx =(α1+α2)ln(α1TG zq +α2CY tg +1); where pg yx is the promotion influence evaluation coefficient, α1 is the proportional coefficient of the promotion periodicity coefficient, α2 is the proportional coefficient of the continuous promotion difference coefficient, and α1 and α2 are both greater than 0.

3. A game promotion algorithm according to claim 2, characterized in that: The review information obtained for game promotion includes: The comment information of game promotion is expressed by the emotional fermentation coefficient; The logic for obtaining the sentiment fermentation coefficient is as follows: extract user comments related to game promotion within the monitoring period from the game community platform, assign sentiment scores to each word through the sentiment dictionary, obtain the sentiment score of each comment, and mark the sentiment score of each comment as: PF q,h , where q = 1, 2, 3, ..., Q, Q is a positive integer, q is the number of each game promotion in the monitoring interval, h = 1, 2, 3, ... H, H is a positive integer, h is the number of the comment in the qth game promotion; Get the number of views of each comment after each game promotion within the monitoring interval, and mark the number of views of each comment after each game promotion within the monitoring interval as: LL q,h , get the number of likes for each comment after each game promotion in the monitoring interval, and mark the number of likes for each comment after each game promotion in the monitoring interval as: DZ q,h ; Calculate the comment impact coefficient using the following formula: Among them, PF q,h is the comment influence coefficient of the hth comment in the qth game promotion, β1 is the specific proportional coefficient of the pageviews, and β2 is the specific proportional coefficient of the likes; According to the influence coefficient of each comment, each comment is divided into different categories, and the comment categories are marked as: ω g , where g = 1, 2, 3, ..., G, G is a positive integer, and g is the number of different types of comments; According to the comment influence coefficient of each comment, the total comment influence coefficient after different game promotions is obtained through the summation formula, and the total comment influence coefficient after different game promotions is marked as: ZPF q ,in, The comment influence coefficient and the proportion of different types of comments after each game promotion are used to build a regression model to generate the emotional fermentation coefficient, which is expressed as: Among them, FJ qg is the emotional fermentation coefficient, BL1, BL2, BL3, ..., BL q is the proportion of different types of comments after the promotion of different games, γ1, γ2, γ3, …, γ q It is the influence weight of the proportion of different types of comments after the promotion of different games.

4. A game promotion algorithm according to claim 3, characterized in that: The monitoring of the performance of game promotion activities in the game community by training a neural network model includes: The promotion influence evaluation coefficient, the emotional fermentation coefficient, the number of game promotions and the duration of continuous game promotion are used as the input of the neural network model training, and the promotion warning coefficient is used as the output of the neural network model training, and it is assigned to 0 or 1 to train the model, where 0 means that the promotion warning coefficient is less than the promotion warning coefficient threshold, and 1 means that the promotion warning coefficient is greater than the promotion warning coefficient threshold; The game promotion warning coefficient is obtained through the formula. The expression of the promotion warning coefficient is: Among them, YJ tg To promote the early warning coefficient, τ is a specific proportional coefficient.

5. A game promotion system, used to implement a game promotion algorithm as described in any one of claims 1 to 4, characterized in that: It includes periodic analysis module, autoregressive model module, evaluation module, sentiment analysis module and neural network module, and the signal connections between modules; The periodicity analysis module is used to extract periodicity information by performing wavelet analysis on the time series data of game promotion evaluation features; The autoregressive model module is used to predict future promotion evaluation features through the autoregressive model, measure the effectiveness of current promotion activities by comparing the predicted results with actual data, and determine the continuous popularity information; Evaluation module, which is used to comprehensively analyze periodic information and continuous popularity information to quantify the promotion effect in the game community; Sentiment analysis module, which is used to extract user comments after promotion activities, and use sentiment dictionary to perform sentiment analysis on the comments, assign sentiment scores, and evaluate the emotional impact of game promotion activities on users and community feedback reactions; The neural network module is used to integrate the data of each module, conduct comprehensive evaluation through the trained model, and monitor the overall performance of game promotion in real time.

Images