Method and system for carrying out clustering analysis on user asset sequence

Abstract

The invention provides a method and system for carrying out clustering analysis on a user asset sequence, equipment and a storage medium. According to the method provided by the invention, characteristics for clustering are not set in advance, and the curve distance of each piece of time sequence data is calculated by directly using the dynamic time warping algorithm after the time sequence data of the user is uniformly processed; and carrying out clustering analysis on the obtained curve distance, and carrying out intra-class curve and inter-class curve comparison in a visual mode to obtain related user behavior performance characteristics. According to the method, the original time sequence data is directly clustered, so that the efficiency problem of extracting features in advance is solved, original information of a curve can be reserved, more valuable information can be found, and game operators can provide more reasonable user experience conveniently.

Description

Technical Field

[0001] The present application relates to the field of information interaction technology, and in particular to a method and system for clustering analysis of user asset sequences, a computing device, and a computer-readable storage medium. Background Art

[0002] In existing online games, it is usually necessary to analyze the behavior of game users and classify users with similar performance or properties into the same category for further analysis and use in game operations. Traditional analysis sequence methods require manual setting of features for clustering, such as the slope, inflection point, and peak value of the curve. The effect of clustering depends largely on the richness of the extracted features. In addition, this method also requires prior verification of the correlation between the features and the target indicators of the analysis to prevent the introduction of irrelevant variables from interfering with the clustering results. Extracting a large number of features is inefficient, while extracting fewer features will result in insufficient information, which greatly affects the efficiency and effectiveness of user behavior analysis. Summary of the invention

[0003] In view of this, embodiments of the present application provide a method and system for clustering analysis of user asset sequences, a computing device, and a computer-readable storage medium to address technical deficiencies in the prior art.

[0004] According to a first aspect of an embodiment of the present application, a method for clustering analysis of a user asset sequence is provided, the method comprising:

[0005] Acquire the user's time series data, and perform unified numerical processing on the time series data;

[0006] Using a dynamic time warping algorithm to calculate the curve distance of the processed time series data, clustering the curve distance to obtain a distance clustering result;

[0007] The correlation between the distance clustering result and the indicator is analyzed to obtain at least one user group having a common behavior performance.

[0008] According to a second aspect of an embodiment of the present application, a system for performing cluster analysis on a user asset sequence is provided, comprising:

[0009] A data acquisition unit, used to acquire the user's time series data and perform unified numerical processing on the time series data;

[0010] A calculation unit, used for calculating the curve distance of the processed time series data using a dynamic time warping algorithm, and clustering the curve distance to obtain a distance clustering result;

[0011] The analyzing unit is used to analyze the correlation between the distance clustering result and the index to obtain at least one user group with common behavior performance.

[0012] According to a third aspect of an embodiment of the present application, a computing device is provided, comprising a memory, a processor, and computer instructions stored in the memory and executable on the processor, wherein when the processor executes the instructions, the steps of the above-mentioned method for clustering analysis of user asset sequences are implemented.

[0013] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, which stores computer instructions, and when the instructions are executed by a processor, the steps of the method for clustering analysis of user asset sequences described above are implemented.

[0014] In the embodiment of the present application, in order to provide a more accurate and efficient user behavior analysis method, the features for clustering are no longer set in advance. Instead, the dynamic time warping algorithm is directly used to calculate the curve distance of each time series data after the user's time series data is uniformly processed; the obtained curve distance is then clustered and analyzed, and the curves within and between classes are compared in a visual way to obtain relevant user behavior performance characteristics. This method not only solves the efficiency problem of extracting features in advance by directly clustering the original time series data, but also helps to retain the original information of the curve, which is conducive to discovering more valuable information and facilitating game operators to provide a more reasonable user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a structural block diagram of a computing device provided in an embodiment of the present application;

[0016] Figure 2 It is a flowchart of a method for clustering analysis of user asset sequences provided in an embodiment of the present application;

[0017] Figure 3 is a schematic diagram of a curve calculated using a dynamic time warping algorithm provided in an embodiment of the present application;

[0018] Figure 4 is another schematic diagram of a curve calculated using a dynamic time warping algorithm provided in an embodiment of the present application;

[0019] Figure 5 It is a structural diagram of a system for performing cluster analysis on user asset sequences provided in an embodiment of the present application. DETAILED DESCRIPTION

[0020] Many specific details are described in the following description to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present application, so the present application is not limited by the specific implementation disclosed below.

[0021] The terms used in one or more embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of the present application. The singular forms of "a", "said" and "the" used in one or more embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in one or more embodiments of the present application refers to and includes any or all possible combinations of one or more associated listed items.

[0022] It should be understood that, although the terms first, second, etc. may be used to describe various information in one or more embodiments of the present application, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of the present application, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "in response to determination".

[0023] In the present application, a method and system for clustering analysis of user asset sequences, a computing device, and a computer-readable storage medium are provided, which are described in detail one by one in the following embodiments.

[0024] Figure 1 The structure block diagram of a computing device 100 according to an embodiment of the present application is shown. The components of the computing device 100 include but are not limited to a memory 110 and a processor 120. The processor 120 is connected to the memory 110 via a bus 130, and the database 150 is used to store data.

[0025] The computing device 100 also includes an access device 140 that enables the computing device 100 to communicate via one or more networks 160. Examples of these networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. The access device 140 may include one or more of any type of network interface (e.g., a network interface card (NIC)) that is wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a World Wide Interoperability for Microwave Access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, a near field communication (NFC) interface, and the like.

[0026] In one embodiment of the present application, the above components of the computing device 100 and Figure 1 Other components not shown in the figure may also be connected to each other, for example, via a bus. It should be understood that Figure 1 The computing device structure block diagram shown is only for the purpose of illustration, and is not intended to limit the scope of the present application. Those skilled in the art may add or replace other components as needed.

[0027] The computing device 100 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, etc.), a mobile phone (e.g., a smart phone), a wearable computing device (e.g., a smart watch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or PC. The computing device 100 may also be a mobile or stationary server.

[0028] In the prior art, a time series is a set of ordered data obtained by observing a certain type of data in chronological order, and these data are usually records of observed events or numerical values ​​of specific patterns. The main research purpose of time series data mining is how to quickly and effectively discover the potential information of time series. The time series distance algorithm represented by the dynamic time warping algorithm is an algorithm for calculating the distance between curves. It can dynamically align two curves, that is, stretch or compress the time of the curve while keeping the chronological order unchanged, and finally obtain an alignment method with the shortest distance, and this shortest distance is the distance between the two curves, and then clustering is performed according to the distance calculated by DTW. But even if the DTW algorithm is used, it is still necessary to verify the correlation between the extracted clustering features and the indicators, and constantly screen the features, which affects the efficiency of the analysis. Therefore, the embodiment of the present application proposes a new method for solving this problem.

[0029] The processor 120 may execute Figure 2 The steps of a method for clustering analysis of user asset sequences shown include steps 202 to 206.

[0030] Step 202: Obtain the user's time series data and perform unified numerical processing on the time series data.

[0031] In online gaming environments, there are a large number of data asset sequences, such as various numerical indicators of users that change over time in the game. For example, the amount of currency, the number of props, the character level, skill points, etc. of the user. These data are arranged in chronological order to form a sequence that can reflect the user's game behavior and progress. This data asset sequence is actually a quantitative representation of user behavior. For example, a user's currency growth may indicate that he is a user who actively participates in game activities; while a slow growth in the number of props may indicate that this is a more cautious user.

[0032] In this step, first obtain the user's time series data. Assume that in a strategy game, users can use in-game currency to speed up the game process or obtain special props. In order to analyze the user's consumption behavior in the game, so as to better understand different types of users and optimize the operation in the game, collect the daily consumption data of multiple users within 60 days, for example. The following is an example of the consumption sequence of three users:

[0033] User A: [0,0,5,0,0,10,0,0,0,15,0,0,20,0,0,…]

[0034] User B: [0,2,0,3,0,0,5,0,8,0,0,10,0,15,0,…]

[0035] User C: [0,0,0,10,0,0,0,0,15,0,0,0,0,25,0,…] ...

[0037] Furthermore, the acquired time series data is uniformly processed numerically to remove junk data that will obviously cause errors in clustering, including:

[0038] Delete abnormally large consumption values. Sometimes, there may be abnormal values ​​in time series data, such as extremely large consumption data caused by system errors.

[0039] Raw data:

[0040] [5,10,0,15,1000000,20,0,25,30,...]

[0041] After processing:

[0042] [5,10,0,15,20,0,25,30,...]

[0043] Specifically, a threshold can be set, for example, data with daily consumption exceeding 5,000 is considered abnormal and deleted.

[0044] Compression or mapping of logarithmic values. In consumption and other behavioral analyses, we often encounter situations where the consumption data of most users may be within a smaller range (such as 0-100), while a few users may have very large consumption data (above 1000). If these raw data are used directly, the following problems will occur, such as: large consumption values ​​will dominate the analysis results, masking the patterns of small consumption values, or clustering algorithms may over-focus on the differences in absolute data and ignore the relative patterns of consumption behavior. Therefore, it is necessary to compress or map the consumption data (logarithmic transformation, square root transformation, etc.) to narrow the gap between large and small consumption values ​​so that the patterns of small consumption can also be well captured. For example:

[0045] Original data: [0,5,10,50,100,500,..]

[0046] After processing (using logarithmic log(x) transformation): [0, 1.79, 2.40, 3.93, 4.62, 6.22]

[0047] It can be seen that the data range gap is greatly narrowed after conversion.

[0048] Select a specific time period. Select the first 30 days of data after the user enters the game. This allows us to focus on analyzing the user's early consumption behavior and ensure that the data of all users has the same time series meaning.

[0049] Original data (60 days): [0,5,10,0,15,20,…,30,35,40]

[0050] After treatment (first 30 days): [0,5,10,0,15,20,…,25,30]

[0051] Unify the sequence length. For example, resampling 30 days of data into 10 data points can improve algorithm performance while preserving the main consumption patterns.

[0052] Raw data (30 days):

[0053] [0,5,10,0,15,20,0,25,30,0,35,40,0,45,50,…,25,30]

[0054] After processing (10 data points):

[0055] [2.5,12.5,22.5,27.5,37.5,47.5,32.5,27.5,22.5,27.5]

[0056] The final processed data is:

[0057] [0,0.2,0.4,0.5,0.6,0.7,0.8,0.9,0.95,1,...]

[0058] Furthermore, the unified numerical processing of the time series data also includes performing Fourier transformation on the asset time series to be processed to obtain the spectrum of the time series data, and analyzing the characteristics of the high-frequency component and the low-frequency component of the user asset series according to the spectrum.

[0059] The processed data sequence is more suitable for the analysis of the DTW algorithm, which can better capture a certain type of behavior pattern and improve the efficiency and accuracy of the algorithm. Those skilled in the art should know that in practical applications, the specific preprocessing steps to be used and how to set the parameters need to be determined according to the characteristics of the data and the purpose of the analysis.

[0060] Step 204: Use the dynamic time warping algorithm to calculate the curve distance of the processed time series data, and cluster the curve distance to obtain a distance clustering result.

[0061] In this step, the dynamic time warping (DTW) algorithm is first used to calculate the curve distance of each time series data.

[0062] For example, here are the 7-day cumulative consumption data of the following 5 users:

[0063] User A: [5,5,15,15,15,15,30]

[0064] User B: [0,8,8,8,20,20,20]

[0065] User C: [3,3,3,10,10,15,15]

[0066] User D: [10,10,20,20,25,25,35]

[0067] User E: [2,2,5,5,8,15,15]

[0068] Directly calculate the DTW distance between these users and get the following results:

[0069]

[0070] Preferably, when performing DTW calculations on large amounts of user data, such as time series data of tens of thousands of people, in order to improve computational efficiency, a small batch calculation method is used to divide the curves into different batches, which can effectively process large-scale data sets.

[0071] For example, analyze the daily active time data of 3,000 users over a 30-day period.

[0072] The data of each user is a sequence of length 30, which indicates the daily active time. For example, the data of user 1 may be:

[0073] [60,45,30,90,120,…,75] ...

[0075] In order to use small batch calculations, the 3,000 users are divided into 100 batches, with 30 users in each batch. Then, DTW calculations are performed within the batch, and DTW distance calculations are performed for the users in each batch.

[0076] For example, for the first batch (users 1-30):

[0077] Create a 30x30 distance matrix and for each cell (i, j) in the matrix, calculate the DTW distance between user i and user j.

[0078] Furthermore, after calculating the DTW distances of users in different batches, the distance matrices are combined to obtain a complete DTW distance matrix. Since the matrices of different batches are symmetrical in the complete distance matrix, a large number of repeated calculation processes are reduced.

[0079] Furthermore, a clustering method is used to calculate the DTW distance matrix to obtain clustering results of multiple different categories.

[0080] For example, for the DTW distance matrix:

[0081]

[0082]

[0083] Using Agglomerative Hierarchical Clustering, starting from the bottom, gradually merge the closest clusters to get the clustering result:

[0084] (((A,B),E),(C,D))

[0085] We can then choose appropriate cut points from this hierarchy. For example, split into two clusters:

[0086] Cluster 1: A, B, E

[0087] Cluster 2: C, D

[0088] Those skilled in the art should know that in practical applications, a variety of clustering algorithms can be used to calculate the distance matrix, including but not limited to k-means clustering, KMedo id clustering, hierarchical clustering, Canopy clustering, etc., which needs to be determined according to the characteristics of the data and the purpose of the analysis.

[0089] Step 206: Analyze the correlation between the distance clustering result and the index to obtain at least one user group with common behavior performance.

[0090] After obtaining the clustering result of step 204, the curves corresponding to the time series values ​​in the same class will show similarity in shape. On this basis, the curves within the class can be further analyzed in a unified manner to obtain the commonalities of the curves in this class and the differences between the curves of different classes; and the differences in indicators in different classes can be analyzed to obtain some behavioral performance characteristics related to the indicators.

[0091] Specifically, the intra-class curves and inter-class curves are compared in a visual way to obtain the commonalities of the intra-class curves and the differences between the inter-class curves, and analyze the differences in indicators in different categories, such as Figure 3 and Figure 4 shown.

[0092] For example, identifying common features of curves within a class.

[0093] Peak: Consumption peaks on the 2nd and 5th days

[0094] Large fluctuation: consumption data fluctuates greatly, with the highest and lowest difference being 120

[0095] Periodicity: Seems to peak every 3 days

[0096] Based on the above analysis, it can be inferred that this type of users may be sensitive to specific activities or content updates (causing periodic peaks).

[0097] To account for the differences between the curves of different categories, the average trend curve of each category is plotted on the same chart to identify the unique characteristics of each category:

[0098] High Value Fluctuation: These are likely to be users who are highly engaged in the game and require regular, high-quality content updates to maintain their interest.

[0099] Stable growth type: These users may gradually deepen their game experience and increase their investment as their game experience improves. These users may pay more attention to long-term game experience and progress.

[0100] Low Value Unstable: May be a casual or less engaged user with limited interest in the game. May require special incentives to increase their engagement and value.

[0101] In the above-mentioned embodiment of the present application, in order to provide a more accurate and efficient user behavior analysis method, the features for clustering are no longer set in advance, but the dynamic time warping algorithm is directly used to calculate the curve distance of each time series data after the user's time series data is uniformly processed; then the obtained curve distance is clustered and analyzed, and the curves within and between classes are compared in a visual way to obtain relevant user behavior performance characteristics. This method not only solves the efficiency problem of extracting features in advance by directly clustering the original time series data, but also helps to retain the original information of the curve, integrate various features into the shape of the curve, and facilitate the discovery of more valuable information.

[0102] Corresponding to the above-mentioned embodiment of the method for analyzing clustering user asset sequences, the present application also provides an embodiment of an analysis system for clustering user asset sequences, such as Figure 5 As shown, the system includes:

[0103] A data acquisition unit, used to acquire the user's time series data and perform unified numerical processing on the time series data;

[0104] A calculation unit, used for calculating the curve distance of the processed time series data using a dynamic time warping algorithm to obtain distance clustering;

[0105] The analysis unit is used to analyze the correlation between the clustering results and the indicators to obtain user groups with certain common behavioral performances.

[0106] The above is a schematic scheme of an analysis system for clustering user asset sequences of this embodiment. It should be noted that the technical scheme of the analysis system for clustering user asset sequences and the technical scheme of the analysis method for clustering user asset sequences belong to the same concept, and the details not described in detail in the technical scheme of the analysis system for clustering user asset sequences can be found in the description of the technical scheme of the analysis method for clustering user asset sequences.

[0107] In one embodiment of the present application, a computing device is also provided, including a memory, a processor, and computer instructions stored in the memory and executable on the processor, wherein the processor implements the steps of the above-described method when executing the instructions.

[0108] The above is a schematic scheme of a computing device of this embodiment. It should be noted that the technical scheme of the computing device and the technical scheme of the above method belong to the same concept, and the details of the technical scheme of the computing device that are not described in detail can be found in the description of the technical scheme of the above method.

[0109] An embodiment of the present application further provides a computer-readable storage medium storing computer instructions, which implement the steps of the method described above when executed by a processor.

[0110] The above is a schematic scheme of a computer-readable storage medium of this embodiment. It should be noted that the technical scheme of the storage medium and the technical scheme of the above-mentioned method for analyzing clustering user asset sequences belong to the same concept, and the details not described in detail in the technical scheme of the storage medium can be referred to the description of the technical scheme of the above-mentioned method for analyzing clustering user asset sequences.

[0111] The above describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0112] The computer instructions include computer program codes, which may be in source code form, object code form, executable files or some intermediate forms, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0113] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described action sequence, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0114] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0115] The preferred embodiments of the present application disclosed above are only used to help explain the present application. The optional embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of the present application. The present application selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present application, so that those skilled in the art can understand and use the present application well. The present application is only limited by the claims and their full scope and equivalents.

Claims

1. A method for clustering analysis of user asset sequences, It is characterized in that include: Acquire the user's time series data, and perform unified numerical processing on the time series data; Using a dynamic time warping algorithm to calculate the curve distance of the processed time series data, clustering the curve distance to obtain a distance clustering result; The correlation between the distance clustering result and the indicator is analyzed to obtain at least one user group having a common behavior performance.

2. The method according to claim 1, in, The unified numerical processing of the time series data includes but is not limited to: Remove outliers, compress or map values, select specific time periods, and unify series lengths.

3. The method according to claim 1, in, The curve distance of the processed time series data is calculated using a dynamic time warping algorithm, and the curve distance is clustered to obtain a distance clustering result including: Dividing the time series data into different batches, calculating the dynamic time warping distances between different batches, and then combining the distance matrices to obtain a complete dynamic time warping distance matrix; The clustering method is used to calculate the complete dynamic time warping distance matrix to obtain the clustering results.

4. The method according to claim 3, in, The clustering methods include but are not limited to: k-means clustering, KMedoi d clustering, hierarchical clustering and Canopy clustering.

5. The method according to claim 1, in, Analyze the correlation between clustering results and indicators, and get at least one user group with common behavior performance, including: By comparing the intra-class curves and inter-class curves in a visual way, the commonalities of the intra-class curves and the differences between the inter-class curves are obtained.

6. The method according to claim 1, in, The user's time series data includes, but is not limited to: the user's asset time series in the game, the user's win-loss time series in the game, and the active number time series.

7. An analysis system for clustering user asset sequences, It is characterized in that include: A data acquisition unit, used to acquire the user's time series data and perform unified numerical processing on the time series data; A calculation unit, used for calculating the curve distance of the processed time series data using a dynamic time warping algorithm, and clustering the curve distance to obtain a distance clustering result; The analyzing unit is used to analyze the correlation between the distance clustering result and the index to obtain at least one user group with common behavior performance.

8. A computing device comprising a memory, a processor, and computer instructions stored in the memory and executable on the processor, It is characterized in that When the processor executes the instructions, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium storing computer instructions. It is characterized in that When the instruction is executed by a processor, the steps of the method described in any one of claims 1 to 6 are implemented.

Images