A user behavior analysis system driven by big data

Abstract

The present invention relates to the field of behavior analysis technology, specifically to a user behavior analysis system driven by big data, the system comprising an active user identification module, an activity change analysis module, a satisfaction analysis module and an update impact assessment module. The present invention, by analyzing user interaction data, can evaluate the user's activity in real time, and by dynamically adjusting the interaction activity threshold, accurately identify the user group, improve the accuracy of active user identification, and by statistically analyzing the user's activity data before and after the software update, reveal the actual impact of the update on user behavior, so that decision makers can evaluate the effect of software updates based on detailed data, and combine user feedback with activity data to evaluate software satisfaction, providing a scientific basis for optimizing user experience and improving satisfaction, and by integrating activity and satisfaction change data, comprehensively analyzing user behavior data, and quantifying the overall impact of software product updates.

Description

Technical Field

[0001] The present invention relates to the technical field of behavior analysis, and in particular to a user behavior analysis system driven by big data. Background Art

[0002] The core of the behavioral analysis technology field is data mining and statistical analysis, which aims to identify behavioral trends, preferences and abnormal behaviors through historical behavioral data. Behavioral analysis is widely used in many application areas, such as marketing, network security, health monitoring and social science research. The technology supports decision making by processing and analyzing large amounts of data and providing behavioral insights.

[0003] Among them, the user behavior analysis system driven by big data is a system that analyzes and predicts user behavior, which is mainly achieved by collecting and analyzing large-scale user data. The system uses huge data sets obtained from multiple channels to reveal user behavior patterns and preferences through data analysis techniques such as pattern recognition and predictive modeling. The system is often used in e-commerce, financial services, and social media to optimize user experience, improve customer satisfaction, and drive personalized marketing strategies.

[0004] When processing user behavior data, traditional analysis systems lack full utilization of data dynamics and fail to adjust data processing thresholds in real time, resulting in insufficient recognition accuracy and response speed when user behavior changes rapidly. When analyzing the impact of software updates, traditional systems only rely on simple data comparisons and lack evaluation of changes in user satisfaction, making it difficult to fully understand the actual impact of updates on user experience and satisfaction. This limitation makes it difficult for decision makers to make optimal decisions when facing rapidly changing markets and user needs, resulting in untimely or inaccurate adjustments to product strategies, affecting user retention and brand competitiveness. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a big data driven user behavior analysis system.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solution: a user behavior analysis system driven by big data, the system comprising:

[0007] The active user identification module evaluates the user's interactive activity based on the user's behavioral data on the software and the number of interactions and interaction time per unit time, and identifies active users by dynamically adjusting the activity threshold to obtain active user identification results;

[0008] The activity change analysis module counts the number of active users before and after the software update based on the active user identification results, analyzes the degree of change in the number of active users, and calculates the average activity of users before and after the update, analyzes the degree of change in the average activity of users, and obtains activity change information;

[0009] The satisfaction analysis module evaluates the satisfaction score of the software based on the active user identification result, the user's feedback score on the software, and the user's activity level, and analyzes the degree of change in the satisfaction score based on the satisfaction score before and after the software update to obtain software satisfaction information;

[0010] The update impact assessment module is based on the activity change information and software satisfaction information, and utilizes the change degree of the number of active users, the change degree of the average activity of users, and the change degree of the satisfaction score to quantitatively assess the impact of the update on the software, determine whether the software update meets expectations, and obtain the update impact analysis results.

[0011] The present invention is improved in that the method for evaluating the user's interactive activity is:

[0012] Based on the user's behavior data on the software, collect the number of user interactions and total interaction time per unit time to obtain interaction data information;

[0013] Based on the interaction data information, calculate the average interaction duration of each user interaction to obtain average interaction duration information;

[0014] Based on the average interaction duration information, the formula is:

[0015]

[0016] Calculate the activity index A, and evaluate the user's interactive activity according to the size of the activity index A to obtain user activity information, where α A is the reinforcement factor of the number of interactions, β A is the attenuation coefficient of interaction duration, γ A is the time accumulation adjustment parameter, e is the base of the natural logarithm, T is the total interaction time of the user, and A represents the user's activity.

[0017] The present invention is improved in that the step of obtaining the active user identification result is:

[0018] Based on the user activity information, by summarizing the acquired user activity indexes, calculating the average of the user activity indexes, and obtaining average activity information;

[0019] Based on the average activity information, the formula:

[0020]

[0021] Calculate the activity threshold T A , where λ T is a dynamic adjustment parameter, σ T is the standard deviation adjustment factor, A i represents the activity index of the i-th user, is the average value of the user activity index, and N is the total number of users;

[0022] Based on the activity threshold T A , the activity index exceeds the activity threshold T A The users are marked as active users, and the active user identification results are obtained.

[0023] The present invention is improved in that the method for analyzing the degree of change in the number of active people is:

[0024] Based on the active user identification result, collect the number of active users before and after the software update to obtain active user information;

[0025] Based on the active number information, the formula is:

[0026]

[0027] Calculate the change degree of active users ΔH to obtain the change information of active users, where H pre is the number of active users before the software update, H post is the number of active users after the software update, ∈ H is the stability constant, δ H is the fluctuation adjustment coefficient, and ΔH is the degree of change in the number of active users.

[0028] The present invention is improved in that the step of obtaining the activity change information is:

[0029] Based on the active user identification result, collect the overall user activity index before and after the software update, and calculate the average user activity before and after the update to obtain the average user activity information;

[0030] Based on the average activity information of the users, the formula is:

[0031]

[0032] Calculate the change in average activity ΔA, where and are the average values ​​of user activity index before and after the update, k B is the adjustment parameter, ω B is the smoothing coefficient, σ B is the standard deviation of the user activity index before the update, and ΔA represents the degree of change in the average activity;

[0033] Based on the change degree ΔA of the average activity, the change degree ΔH of the number of active people is integrated to obtain activity change information.

[0034] The present invention is improved in that the method for evaluating the satisfaction rating of the software is:

[0035] Based on the user feedback information, extract each user's feedback score on the software to obtain feedback score information;

[0036] Based on the active user identification result, according to the user activity index, the formula is:

[0037]

[0038] Calculate the user's rating weight factor w i , get the user rating weight information, where w i is the scoring weight factor of the i-th user, N is the total number of users, j is the user index in the summation process, and A i is the activity index of the i-th user;

[0039] Based on the user rating weight information and feedback rating information, the formula:

[0040]

[0041] Calculate the software satisfaction score S and get the satisfaction evaluation result, where R i Score the feedback of the i-th user, w i is the rating weight factor of the i-th user, N is the total number of users, and S is the satisfaction rating of the software.

[0042] The present invention is improved in that the steps of obtaining the software satisfaction information are:

[0043] Based on the satisfaction evaluation results, extract the satisfaction scores before and after the software update to obtain satisfaction change correlation data;

[0044] Based on the satisfaction change correlation data, the formula is:

[0045]

[0046] Calculate the change degree of satisfaction score ΔS to obtain software satisfaction information, where S pre and S post Represents the satisfaction scores before and after the software update, α S is the adjustment coefficient, β S is the smoothing coefficient, γ S is the offset, and ΔS represents the degree of change in the satisfaction score.

[0047] The present invention is improved in that the step of obtaining the update impact analysis result is:

[0048] Based on the activity change information and software satisfaction information, extract the change degree of the number of active users, the change degree of the average activity of users and the change degree of the satisfaction score to obtain update impact association data;

[0049] Based on the update affecting the associated data, by the formula:

[0050]

[0051] Calculate the impact I, where ΔH is the change in the number of active users, ΔA is the change in average activity, and ΔS is the change in satisfaction score; λ I , μ I and ν I They are the weight coefficients of the change in the number of active users, the change in activity level, and the change in satisfaction;

[0052] Based on the impact degree I, it is compared with a preset impact threshold to determine whether the software update meets expectations and obtain an update impact analysis result.

[0053] Compared with the prior art, the advantages and positive effects of the present invention are:

[0054] In the present invention, by analyzing user interaction data, it is possible to evaluate user activity in real time, and by dynamically adjusting the interaction activity threshold, it is possible to accurately identify user groups, thereby improving the accuracy of active user identification. By counting and analyzing user activity data before and after software updates, the actual impact of updates on user behavior is revealed, enabling decision makers to evaluate the effects of software updates based on detailed data, and to evaluate software satisfaction in combination with user feedback and activity data, thereby providing a scientific basis for optimizing user experience and improving satisfaction. By integrating activity and satisfaction change data, user behavior data is comprehensively analyzed, and the overall impact of software product updates is quantified, thereby improving the depth of understanding of the impact of software updates and the accuracy of management, thereby helping to refine product strategy adjustments and achieve continuous optimization and iteration. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 is a system flow chart of the present invention;

[0056] Figure 2 A flow chart of evaluating the interactive activity of a user in the present invention;

[0057] Figure 3 A flow chart of obtaining active user identification results for the present invention;

[0058] Figure 4A flow chart for analyzing the degree of change in the number of active people for the present invention;

[0059] Figure 5 A flowchart of the present invention for obtaining activity change information;

[0060] Figure 6 A flowchart for scoring satisfaction of the software for evaluating the present invention;

[0061] Figure 7 A flowchart for obtaining software satisfaction information for the present invention;

[0062] Figure 8 A flow chart of obtaining update impact analysis results for the present invention. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0064] 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", etc., indicating positions or positional relationships, are based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0065] See also Figure 1 The present invention provides a technical solution: a user behavior analysis system driven by big data, the system comprising:

[0066] The active user identification module evaluates the user's interactive activity based on the user's behavioral data on the software and the number of interactions and interaction time per unit time, and identifies active users by dynamically adjusting the activity threshold to obtain active user identification results;

[0067] The activity change analysis module counts the number of active users before and after the software update based on the active user identification results, analyzes the degree of change in the number of active users, and calculates the average activity of users before and after the update, analyzes the degree of change in the average activity of users, and obtains the activity change information;

[0068] The satisfaction analysis module evaluates the software satisfaction score based on the active user identification results, the user feedback score on the software, and the user activity. It also analyzes the degree of change in the satisfaction score before and after the software update to obtain software satisfaction information.

[0069] The update impact assessment module is based on activity change information and software satisfaction information. It uses the change in the number of active users, the change in the average user activity, and the change in the satisfaction score to quantitatively assess the impact of the update on the software, determine whether the software update meets expectations, and obtain the update impact analysis results.

[0070] The active user identification results include activity scores and user classification labels. The activity change information includes the active user change ratio, average activity increase or decrease value, and activity distribution information. The software satisfaction information includes the satisfaction score change rate and score range distribution information. The update impact analysis results include update benefit rating and update effectiveness information.

[0071] See also Figure 2 ,The method for evaluating the user's interactive activity is:

[0072] Based on the user's behavior data on the software, collect the number of user interactions and total interaction time per unit time to obtain interaction data information;

[0073] Based on the interactive data information, through the formula:

[0074]

[0075] Calculate the average interaction time L of each user interaction to obtain the average interaction time information, where T is the total interaction time of the user, n A is the number of interactions, L is the average interaction duration of each user interaction;

[0076] Based on the average interaction time information, the formula is:

[0077]

[0078] Calculate the activity index A, and evaluate the user's interactive activity according to the size of the activity index A to obtain user activity information, where α A is the reinforcement factor of the number of interactions, β A is the attenuation coefficient of interaction duration, γ A is the time accumulation adjustment parameter, e is the base of the natural logarithm, T is the total interaction time of the user, and A represents the user's activity.

[0079] formula:

[0080]

[0081] and

[0082]

[0083] Parameter meaning and acquisition method:

[0084] n A :The number of user interactions per unit time. The value is obtained by analyzing the user's operation log, for example, counting all clicks or input times of a user in a day. The specific method of obtaining is to extract user activity records within a certain period of time from the software backend database.

[0085] T: The total interaction time of the user in a unit of time. The software tracks and records the start and end time of each user interaction, and finally adds up the duration of all interactions to get the total interaction time.

[0086] α A , β A and γ A : It is a tuning parameter used to adjust the impact of the number of interactions and interaction duration on the activity score. Adjust according to different application scenarios to ensure the accuracy and adaptability of the evaluation model. A common setting method is to determine the optimal parameters through regression analysis so that the activity score can most effectively predict user behavior.

[0087] Calculation example:

[0088] Set the number of interactions a user has in a day n A = 100 times, total interaction time T = 3600 seconds, and the average interaction time L is calculated as follows:

[0089]

[0090] The adjustment parameters are set as: α A =0.5,β A =0.05,γ A =0.1.

[0091] The activity index A is calculated as follows:

[0092]

[0093] The calculation result 61.319 reflects the user's interactive activity within a certain period of time. A higher A value indicates that the user's interactive activity is very high.

[0094] See also Figure 3 , the steps to obtain the active user identification results are:

[0095] Based on the user activity information, the activity indexes of the acquired users are summarized, and the average of the activity indexes of the users is calculated to obtain the average activity information;

[0096] Based on the average activity information, the formula is:

[0097]

[0098] Calculate the activity threshold T A , where λ T is a dynamic adjustment parameter, σ T is the standard deviation adjustment factor, A i represents the activity index of the i-th user, is the average value of the user activity index, and N is the total number of users;

[0099] Based on the activity threshold T A , the activity index exceeds the activity threshold T A The users are marked as active users, and the active user identification results are obtained.

[0100] formula:

[0101]

[0102] Parameter meaning and acquisition method:

[0103] The average value of user activity index is calculated by calculating the activity index of a single user A i It is calculated by dividing the sum of by the total number of users N. It reflects the average interactive activity of all users.

[0104] λ T : Dynamically adjust parameters to adjust the activity threshold based on the variability analysis of user behavior to adapt to the fluctuation of user activity in different time periods or activities. Parameters are usually derived from historical data through statistical methods such as regression analysis to ensure sensitive response to user activity.

[0105] σ T : Standard deviation adjustment factor, used to adjust the threshold to more accurately reflect the fluctuations in user group behavior. Obtained by weighting the importance of the standard deviation of user activity.

[0106] N: The total number of users, which is the number of users directly counted from the dataset.

[0107] Calculation example:

[0108] Suppose there are three users, whose activity index is A i They are 100, 150 and 200 respectively.

[0109] Calculate average activity

[0110]

[0111] Set the dynamic adjustment parameter λ T is 0.8, the standard deviation adjustment coefficient σ T is 0.5.

[0112] Calculate part of the standard deviation:

[0113]

[0114] Calculate the standard deviation:

[0115]

[0116] Calculate the activity threshold:

[0117] T A =0.8×150+0.5×40.82

[0118] ≈120+20.41

[0119] =140.41

[0120] Calculation result, activity threshold T A It is 140.41. It means that the activity index A i Users with a score greater than 140.41 will be considered active users.

[0121] See also Figure 4 , the method to analyze the change in the number of active users is:

[0122] Based on the active user identification results, collect the number of active users before and after the software update to obtain the active user information;

[0123] Based on the active number of people information, through the formula:

[0124]

[0125] Calculate the change degree of active users ΔH to obtain the change information of active users, where H pre is the number of active users before the software update, H post is the number of active users after the software update, ∈ H is the stability constant, δ H is the fluctuation adjustment coefficient, and ΔH is the degree of change in the number of active users.

[0126] formula:

[0127]

[0128] Parameter meaning and acquisition method:

[0129] H pre and H post : The parameters represent the number of active users before and after the software update. The data is obtained by analyzing user logs, and the number of users marked as active before and after the update is counted.

[0130] ∈ H : A stability constant used to avoid the denominator becoming zero during the calculation process and to help reduce statistical errors caused by small samples. This parameter is usually set to a small positive value, such as 0.01, to ensure numerical stability.

[0131] δ H : is the fluctuation adjustment coefficient, which is used to smooth the impact of large changes in the number of users and make the results more stable. The value is set according to the historical volatility of the data, for example, it is set to 0.001 to balance the fluctuations in the calculation.

[0132] Calculation example:

[0133] Set the number of active users before and after the software update H pre =200 and H post =250. Parameter ∈ H =0.01 and δ H =0.001.

[0134] Calculate the numerator and adjusted denominator of the active users change ratio:

[0135] Numerator = H post -H pre =250-200=50

[0136] Denominator = H pre +∈ H =200+0.01=200.01

[0137] Calculate the fractional part:

[0138]

[0139] Then, calculate the adjustment factor in the denominator, which includes the volatility adjustment:

[0140] 1+δ H (H post -H pre ) 2

[0141] =1+0.001×50 2

[0142] =1+2.5

[0143] =3.5

[0144] Combine the above two parts to calculate the degree of change:

[0145]

[0146] The calculated result is ΔH=7.1425, which is a large positive number. This indicates that the software update has a positive impact on increasing active users. The result can help the development team understand the actual effect of the update and further adjust the product strategy to optimize the user experience.

[0147] See also Figure 5 , the steps to obtain activity change information are:

[0148] Based on the active user identification results, collect the overall user activity index before and after the software update, and calculate the average user activity before and after the update to obtain the average user activity information;

[0149] Based on the average user activity information, the formula is:

[0150]

[0151] Calculate the change in average activity ΔA, where and are the average values ​​of user activity index before and after the update, k B is the adjustment parameter, ω B is the smoothing coefficient, σ B is the standard deviation of the user activity index before the update, and ΔA represents the degree of change in the average activity;

[0152] Based on the change in average activity ΔA, the change in the number of active people ΔH is integrated to obtain activity change information.

[0153] formula:

[0154]

[0155] Parameter meaning and acquisition method:

[0156] and Represents the average of all user activity indexes before and after the update. The calculation method is to divide the sum of all users' activity indexes by the total number of users to get the average activity in each period.

[0157] k B: is a tuning parameter used to adjust the weight of the activity index before the update to reflect the impact of the update on user behavior. This parameter is adjusted based on historical data and pre-set assumptions to reconcile the comparison of the activity index between different versions.

[0158] ω B : Smoothing coefficient, used to adjust the statistical deviation caused by the small number of users in the calculation. It helps to stabilize the activity evaluation under small samples.

[0159] σ B : represents the standard deviation of the user activity index before the update, which is obtained by calculating the sum of the squares of the differences between each user activity index and the average value, taking the average and then taking the square root.

[0160] Calculation example:

[0161] The following data and parameters are set:

[0162] Total number of users before and after the update n = 1000, average activity index before the update Updated average activity index Adjustment parameter k B =0.95, smoothing coefficient ω B =0.05, standard deviation σ B =10.

[0163] Calculation process:

[0164] Calculate the numerator and denominator of the activity change before and after the update:

[0165] molecular:

[0166] Denominator:

[0167] Calculate the average activity change:

[0168]

[0169]

[0170] The results show that the change in the average user activity after the update is 29.88, which is a high value, indicating a large change in activity.

[0171] See also Figure 6 , the method to evaluate the satisfaction score of the software is:

[0172] Based on the user feedback information, extract each user's feedback score on the software to obtain feedback score information;

[0173] Based on the active user identification results and the user activity index, the formula is:

[0174]

[0175] Calculate the user's rating weight factor w i , get the user rating weight information, where w i is the scoring weight factor of the i-th user, N is the total number of users, j is the user index in the summation process, and A i is the activity index of the i-th user;

[0176] Based on the user rating weight information and feedback rating information, the formula is:

[0177]

[0178] Calculate the software satisfaction score S and get the satisfaction evaluation result, where R i Score the feedback of the i-th user, w i is the rating weight factor of the i-th user, N is the total number of users, and S is the satisfaction rating of the software.

[0179] formula:

[0180]

[0181] and

[0182]

[0183] Parameter meaning and acquisition method:

[0184] R i : Feedback rating of the ith user. Obtained through ratings submitted after a satisfaction survey in the software or on related platforms. Ratings are usually extracted directly from user feedback systems.

[0185] w i : The scoring weight factor of the i-th user, calculated based on the user's activity.

[0186] A i : The activity index of the ith user, calculated using the previous method.

[0187] N: Total number of users, the number of users directly counted from the dataset.

[0188] Calculation example:

[0189] Assume that there are three users, whose activity indexes are 10, 15, and 25 respectively. Their feedback scores are 3, 4, and 5 respectively. The total number of users N = 3.

[0190] Calculate the weight factor w for each user i :

[0191] Total activity:

[0192]

[0193] Calculate the satisfaction score S:

[0194]

[0195] The calculation result S = 4.3 shows that the average satisfaction score is 4.3 based on the user activity. This weighting method ensures that more active users have a greater impact on the final satisfaction score, thus providing a satisfaction score that more accurately reflects the true feelings of the user group.

[0196] See also Figure 7 ,The steps to obtain software satisfaction information are:

[0197] Based on the satisfaction evaluation results, the satisfaction scores before and after the software update are extracted to obtain the satisfaction change correlation data;

[0198] Based on the satisfaction change correlation data, the formula is:

[0199]

[0200] Calculate the change degree of satisfaction score ΔS to obtain software satisfaction information, where S pre and S post Represents the satisfaction scores before and after the software update, α S is the adjustment coefficient, β S is the smoothing coefficient, γ S is the offset, and ΔS represents the degree of change in the satisfaction score.

[0201] formula:

[0202]

[0203] Parameter meaning and acquisition method:

[0204] S pre and S post : The parameter represents the user satisfaction score before and after the software update, which is calculated using the previous method.

[0205] α S : is the adjustment factor, which is used to adjust the impact of the score before the update. The weight of the old score in the new calculation can be adjusted to reflect the importance of the data before the update, which can be preset through regression analysis or expert evaluation.

[0206] β S: Smoothing coefficient, used to smooth the denominator to prevent fluctuations or extreme values ​​in the results due to a too small denominator. It helps to improve the stability and reliability of the calculation and is set by regression analysis.

[0207] γ S : Offset, used to ensure that the denominator is not zero to prevent calculation errors. For example, it can be set to 0.01.

[0208] Calculation example:

[0209] Set the satisfaction scores before and after the software update to S pre =8.0 and S post =8.5.

[0210] Set the adjustment coefficient α S =0.95, smoothing coefficient β S = 0.95, and the offset γ S =0.01.

[0211] Calculate the change in satisfaction ratings:

[0212]

[0213]

[0214] The calculation results show that the change in user satisfaction after the software update is 11.84, which is a large and positive value, indicating that the update has a significant positive impact on the score.

[0215] See also Figure 8 , the steps to obtain the update impact analysis results are:

[0216] Based on the activity change information and software satisfaction information, the change degree of the number of active users, the change degree of the average user activity and the change degree of the satisfaction score are extracted to obtain the update impact correlation data;

[0217] Based on the update affecting the associated data, the formula is:

[0218]

[0219] Calculate the impact I, where ΔH is the change in the number of active users, ΔA is the change in average activity, and ΔS is the change in satisfaction score; λ I , μ I and ν I They are the weight coefficients of the change in the number of active users, the change in activity level, and the change in satisfaction;

[0220] Based on the impact degree I, it is compared with the preset impact threshold to determine whether the software update meets expectations and obtain the update impact analysis result.

[0221] formula:

[0222]

[0223] Parameter details and how to obtain them:

[0224] ΔH: The degree of change in the number of active users, calculated using the previous method.

[0225] ΔA: The change in average activity, calculated using the previous method.

[0226] ΔS: The degree of change in satisfaction score, calculated using the previous method.

[0227] λ I , μ I and ν I : Weight coefficient, obtained through historical data analysis, business impact assessment and expert opinions, is used to balance the impact of different changing factors on the overall assessment results.

[0228] Calculation example:

[0229] Set the weight coefficient λ I =0.2, μ I =0.5, ν I =0.3, the coefficient is set based on past experience and expected influence. Change data ΔH=7, ΔA=30, ΔS=12.

[0230] Calculate the contribution of each part:

[0231] Contribution of changes in active users:

[0232] λ I ΔH=0.2×7=1.4

[0233] Activity change contribution:

[0234] μ I ·ln(1+ΔA)=0.5×ln(1+30)≈0.5×3.43≈1.72

[0235] Contribution to change in satisfaction:

[0236]

[0237] Calculate the total impact score I:

[0238]

[0239] The calculated comprehensive impact score I = 4.1596, which quantifies the comprehensive positive impact of software updates. A higher I value indicates that the update has achieved significant results in increasing active users, improving user activity, and improving user satisfaction.

[0240] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A user behavior analysis system driven by big data, characterized in that: The system comprises: The active user identification module evaluates the user's interactive activity based on the user's behavioral data on the software and the number of interactions and interaction time per unit time, and identifies active users by dynamically adjusting the activity threshold to obtain active user identification results; The activity change analysis module counts the number of active users before and after the software update based on the active user identification results, analyzes the degree of change in the number of active users, and calculates the average activity of users before and after the update, analyzes the degree of change in the average activity of users, and obtains activity change information; The satisfaction analysis module evaluates the satisfaction score of the software based on the active user identification result, the user's feedback score on the software, and the user's activity level, and analyzes the degree of change in the satisfaction score based on the satisfaction score before and after the software update to obtain software satisfaction information; The method for evaluating the satisfaction score of the software is: Based on the user feedback information, extract each user's feedback score on the software to obtain feedback score information; Based on the active user identification result, according to the user activity index, the formula is: ; Calculate the user's rating weight factor , get the user rating weight information, where, For the The rating weight factor of each user, is the total number of users, is the user index in the summation process, For the Activity index of each user; Based on the user rating weight information and feedback rating information, the formula: , calculate the software satisfaction score , and obtain the satisfaction evaluation results, where For the Feedback ratings from users, For the The rating weight factor of each user, is the total number of users, Rate your satisfaction with the software; The steps for obtaining the software satisfaction information are as follows: Based on the satisfaction evaluation results, extract the satisfaction scores before and after the software update to obtain satisfaction change correlation data; Based on the satisfaction change correlation data, the formula is: ; Calculate the change in satisfaction ratings , get software satisfaction information, where and Represents the satisfaction scores before and after the software update, is the adjustment coefficient, is the smoothing coefficient, is the offset, Indicates the degree of change in satisfaction ratings; The update impact assessment module is based on the activity change information and software satisfaction information, and utilizes the change degree of the number of active users, the change degree of the average activity of users, and the change degree of the satisfaction score to quantitatively assess the impact of the update on the software, determine whether the software update meets expectations, and obtain the update impact analysis result; The steps for obtaining the update impact analysis result are: Based on the activity change information and software satisfaction information, extract the change degree of the number of active users, the change degree of the average activity of users and the change degree of the satisfaction score to obtain update impact association data; Based on the update affecting the associated data, by the formula: ; Calculate the impact ,in, is the change in the number of active users, Indicates the degree of change in average activity. Indicates the degree of change in satisfaction ratings, , and They are the weight coefficients of the change in the number of active users, the change in activity level, and the change in satisfaction; Based on the degree of impact , and compare it with the preset impact threshold to determine whether the software update meets expectations and obtain the update impact analysis results.

2. The user behavior analysis system driven by big data according to claim 1, characterized in that: The method for evaluating the user's interactive activity is: Based on the user's behavior data on the software, collect the number of user interactions and total interaction time per unit time to obtain interaction data information; Based on the interaction data information, calculate the average interaction duration of each user interaction to obtain average interaction duration information; Based on the average interaction duration information, the formula is: ; Calculating activity index , according to the activity index The size of the user's interactive activity is evaluated to obtain user activity information, where is the reinforcement factor of the number of interactions, is the attenuation coefficient of the interaction duration, is the time accumulation adjustment parameter, is the base of natural logarithms, is the total user interaction time, Indicates the user's activity level.

3. The user behavior analysis system driven by big data according to claim 2, characterized in that: The steps for obtaining the active user identification result are: Based on the user activity information, the average activity index of the users is calculated by summarizing the acquired activity indexes of the users to obtain the average activity information; Based on the average activity information, the formula: ; Calculating activity threshold ,in, is a dynamically adjusted parameter. is the standard deviation adjustment factor, Indicates The activity index of a user, is the average value of the user activity index, is the total number of users; Based on the activity threshold , the activity index exceeds the activity threshold The users are marked as active users, and the active user identification results are obtained.

4. The user behavior analysis system driven by big data according to claim 1, characterized in that: The method for analyzing the change in the number of active users is: Based on the active user identification result, collect the number of active users before and after the software update to obtain active user information; Based on the active number information, the formula is: ； Calculate the change in the number of active users , get the active number change information, where, is the number of active users before the software update, is the number of active users after the software update, is the stability constant, is the fluctuation adjustment coefficient, The change in the number of active users.

5. The user behavior analysis system driven by big data according to claim 4, characterized in that: The steps for obtaining the activity change information are as follows: Based on the active user identification result, collect the overall user activity index before and after the software update, and calculate the average user activity before and after the update to obtain the average user activity information; Based on the average activity information of the users, the formula is: ; Calculate the change in average activity ,in, and are the average values ​​of user activity index before and after the update, To adjust the parameters, is the smoothing coefficient, is the standard deviation of the user activity index before the update, Indicates the degree of change in average activity; Based on the change in the average activity , integrating the degree of change in the number of active people , get the activity change information.

Images