Playing data anomaly detection method and device, storage medium and electronic equipment
By comparing the abnormal threshold value of the playback indicator with the real-time value and introducing the comparison of related indicators, the problem of inaccurate abnormal monitoring of playback indicators in the prior art is solved, and the accuracy of abnormal detection is improved.
Patent Information
- Application Number
- CN202410974372.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-19
- Publication Date
- 2024-11-22
AI Technical Summary
In the prior art, the abnormal monitoring results of the playback indicator are inaccurate, resulting in the inability to accurately determine whether there is an abnormality in the playback.
By obtaining the abnormal threshold of the playback indicator and the real-time value to compare the associated indicator, determine the associated indicator, and compare the real-time value of the playback indicator with the real-time value of the associated indicator to improve the accuracy of abnormal detection.
Improve the accuracy of abnormal detection of playback indicators and ensure that the abnormal detection results of playback indicators are more reliable.
Smart Images

Figure CN119011954A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of playback data monitoring, and in particular to a playback data anomaly detection method, device, storage medium and electronic device. Background Art
[0002] In the prior art, users can watch live broadcasts or videos through a client or a web page. During the process of watching the program, it is necessary to monitor the playback indicators, such as whether there are abnormalities in the playback data, whether the playback is stuck, etc.
[0003] In the prior art, whether the playback index is abnormal is usually determined by estimating the real-time value based on historical data, thereby determining whether the real-time value of the playback index is abnormal. However, the abnormal monitoring result caused by this method is not accurate. Summary of the invention
[0004] The present application provides a method, device, storage medium and electronic device for detecting anomalies in playback data to solve the technical problem of inaccurate abnormal monitoring results of playback indicators.
[0005] In a first aspect, the present application provides a method for detecting anomalies in playback data, comprising: obtaining an anomaly threshold of a playback indicator to be monitored and a real-time value obtained by monitoring the playback indicator; when a comparison result between the real-time value and the anomaly threshold indicates that an anomaly exists in the real-time value, determining an associated indicator of the playback indicator; obtaining the real-time value of the associated indicator; when a comparison result between the real-time value of the playback indicator and the real-time value of the associated indicator indicates that an anomaly exists in the playback indicator, determining that an anomaly exists in the playback indicator; when a comparison result between the real-time value of the playback indicator and the real-time value of the associated indicator indicates that an anomaly exists in the playback indicator, determining that the playback indicator is normal.
[0006] In a second aspect, the present application provides an anomaly detection device for playback data, comprising: a first acquisition module, used to acquire an anomaly threshold value of a playback indicator to be monitored and a real-time value obtained by monitoring the above playback indicator; a first determination module, used to determine an associated indicator of the above playback indicator when a comparison result between the above real-time value and the above anomaly threshold value indicates that the above real-time value is abnormal; a second acquisition module, used to acquire the real-time value of the above associated indicator; a second determination module, used to determine that an abnormality exists in the above playback indicator when a comparison result between the real-time value of the above playback indicator and the real-time value of the above associated indicator indicates that an abnormality exists in the above playback indicator; and determine that the above playback indicator is normal when a comparison result between the real-time value of the above playback indicator and the real-time value of the above associated indicator indicates that the above playback indicator is normal.
[0007] As an optional example, the first acquisition module includes: an estimation unit for acquiring historical data of the playback indicator; estimating an estimated value of the playback indicator at any time point based on the historical data; and determining the product of the estimated value and a predetermined ratio as the abnormal threshold, wherein the predetermined ratio is determined based on any one of the time points.
[0008] As an optional example, the first acquisition module further includes: a setting unit, for, after estimating the estimated value of the playback indicator at any time point based on the historical data, setting the predetermined ratio to a first value greater than 1 when the any time point is in a popular time period, and setting the predetermined ratio to a second value greater than 1 when the any time point is in an unpopular time period, wherein the closer any time point is to the center of the popular time period, the larger the first value is, and the closer any time point is to the center of the unpopular time period, the smaller the second value is, and the minimum value of the first value is greater than the maximum value of the second value.
[0009] As an optional example, the above-mentioned first determination module includes: a determination unit, used to determine one or more indicators as the above-mentioned associated indicators from multiple second indicators related to the above-mentioned playback indicators in the indicator relationship network, wherein the above-mentioned indicator relationship network includes the first indicator and the above-mentioned second indicators, each of the above-mentioned first indicators is related to one or more of the above-mentioned second indicators, each of the above-mentioned second indicators is related to one or more of the above-mentioned first indicators, and the above-mentioned playback indicator is one of the above-mentioned first indicators.
[0010] As an optional example, the determination unit includes: a determination subunit, configured to determine the number of the associated indicators according to the abnormality level, wherein the higher the abnormality level, the greater the number of the associated indicators.
[0011] As an optional example, the above-mentioned device also includes: a first comparison module, which is used to compare the real-time value of the above-mentioned associated indicator with the real-time value of the above-mentioned playback indicator after obtaining the real-time value of the above-mentioned associated indicator to obtain a first ratio; determine a ratio interval based on the ratio of the historical value of the above-mentioned associated indicator to the historical data of the above-mentioned playback indicator; when the above-mentioned first ratio is within the above-mentioned ratio interval, determine that the above-mentioned comparison result indicates that the above-mentioned playback indicator is normal, and when the above-mentioned first ratio is outside the above-mentioned ratio interval, determine that the above-mentioned comparison result indicates that the above-mentioned playback indicator is abnormal.
[0012] As an optional example, the above-mentioned device also includes: a second comparison module, which is used to obtain the historical data of the above-mentioned playback indicator and the above-mentioned associated indicator after obtaining the real-time value of the above-mentioned associated indicator; draw a playback curve according to the historical data and the real-time value of the above-mentioned playback indicator, and draw an associated curve according to the historical data and the real-time value of the above-mentioned associated indicator; when the trajectory similarity between the above-mentioned playback curve and the above-mentioned associated curve is greater than a preset similarity threshold, determine that the above-mentioned comparison result indicates that the above-mentioned playback indicator is normal, and when the above-mentioned trajectory similarity is less than the above-mentioned preset similarity threshold, determine that the above-mentioned comparison result indicates that the above-mentioned playback indicator is abnormal.
[0013] As an optional example, the above-mentioned device also includes: a third comparison module, which is used to, after obtaining the real-time value of the above-mentioned related indicator, determine that the above-mentioned comparison result indicates that the above-mentioned playback indicator is normal when the real-time value of the above-mentioned playback indicator increases or decreases in the same proportion as the real-time value of the above-mentioned related indicator, and determine that the above-mentioned comparison result indicates that the above-mentioned playback indicator is abnormal when the real-time value of the above-mentioned playback indicator and the real-time value of the above-mentioned related indicator do not change in the same proportion.
[0014] In a third aspect, the present application provides an electronic device comprising: at least one communication interface; at least one bus connected to the at least one communication interface; at least one processor connected to the at least one bus; and at least one memory connected to the at least one bus, wherein the memory stores a computer program, and the processor is configured to implement any one of the above-mentioned methods for detecting anomalies in playback data when executing the computer program.
[0015] In a fourth aspect, the present application further provides a computer storage medium storing computer executable instructions, wherein the computer executable instructions are used to execute any of the above-mentioned methods for detecting anomalies in playback data of the present application.
[0016] The above-mentioned technical scheme provided by the embodiment of the present application has the following advantages over the prior art: the scheme provided by the embodiment of the present application, by obtaining the abnormal threshold value of the playback indicator to be monitored and the real-time value obtained by monitoring the above-mentioned playback indicator; when the comparison result of the above-mentioned real-time value and the above-mentioned abnormal threshold value indicates that the above-mentioned real-time value is abnormal, determining the associated indicator of the above-mentioned playback indicator; obtaining the real-time value of the above-mentioned associated indicator; when the comparison result of the real-time value of the above-mentioned playback indicator and the real-time value of the above-mentioned associated indicator indicates that the above-mentioned playback indicator is abnormal, determining that the above-mentioned playback indicator is abnormal; when the comparison result of the real-time value of the above-mentioned playback indicator and the real-time value of the above-mentioned associated indicator indicates that the above-mentioned playback indicator is normal, determining that the above-mentioned playback indicator is normal, so that when performing abnormality detection on the playback indicator, the associated indicator of the playback indicator can be introduced, and the real-time value of the playback indicator is compared with the real-time value of the associated indicator to check whether the playback indicator is really abnormal, thereby improving the accuracy of abnormality detection on the playback indicator. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.
[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0019] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.
[0020] Figure 1 A flowchart of a method for detecting abnormalities in playback data provided in an embodiment of the present application;
[0021] Figure 2 A relationship diagram between a playback index and a correlation index of a playback data anomaly detection method provided in an embodiment of the present application;
[0022] Figure 3 A diagram of determining the number of associated indicators for a method for detecting anomalies in playback data provided in an embodiment of the present application;
[0023] Figure 4 A fluctuation diagram of real-time values of playback indicators and related indicators of a playback data anomaly detection method provided in an embodiment of the present application;
[0024] Figure 5 A flowchart of another method for detecting abnormalities in playback data provided in an embodiment of the present application;
[0025] Figure 6 A flowchart of another method for detecting abnormalities in playback data provided in an embodiment of the present application;
[0026] Figure 7 A schematic diagram of the structure of a device for detecting abnormalities in playback data provided in an embodiment of the present application;
[0027] Figure 8 A schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0028] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.
[0029] The disclosure below provides many different embodiments or examples to implement different structures of the present invention. In order to simplify the disclosure of the present invention, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present invention. In addition, the present invention can repeat reference numbers and / or letters in different examples. This repetition is for the purpose of simplification and clarity, and does not itself indicate the relationship between the various embodiments and / or settings discussed.
[0030] In order to solve the technical problem of inaccurate abnormal monitoring results of playback indicators in the prior art, the present application provides a method for detecting abnormalities in playback data, which can improve the accuracy of abnormal detection of playback indicators.
[0031] Figure 1 A flowchart of a method for detecting abnormalities in playback data provided by an embodiment of the present application. Figure 1 As shown, the above-mentioned method for detecting abnormalities in playback data includes:
[0032] S102, obtaining an abnormal threshold value of a playback indicator to be monitored and a real-time value obtained by monitoring the playback indicator;
[0033] S104, when the comparison result between the real-time value and the abnormal threshold indicates that the real-time value is abnormal, determining the associated index of the playback index;
[0034] S106, obtaining the real-time value of the above-mentioned related indicators;
[0035] S108, when the comparison result of the real-time value of the above-mentioned playback indicator with the real-time value of the above-mentioned associated indicator indicates that there is an abnormality in the above-mentioned playback indicator, it is determined that there is an abnormality in the above-mentioned playback indicator; when the comparison result of the real-time value of the above-mentioned playback indicator with the real-time value of the above-mentioned associated indicator indicates that the above-mentioned playback indicator is normal, it is determined that the above-mentioned playback indicator is normal.
[0036] The above-mentioned abnormality detection method of playback data can be applied in the field of live broadcasting. In the field of live broadcasting, a user can start a live broadcast and other users can watch the live broadcast of the user. The above-mentioned playback indicators can be the playback data of a single live broadcast room, or the playback data of some or all live broadcast rooms of a platform. The playback indicators may include the number of playback users, playback traffic, likes, dislikes, forwarding, comments, interactions, virtual gifts, the number of abnormal network users or collections, etc. The above are only examples and do not constitute limitations.
[0037] In this embodiment, each playback indicator corresponds to one or more abnormal thresholds. The abnormal threshold is a theoretical peak value determined based on historical data of the playback indicator. In addition, in this embodiment, abnormal thresholds at different time points or time periods can also be determined for each playback indicator.
[0038] For example, taking the playback indicator as the number of playback users in a live broadcast room, three time periods within a day can be set, such as 3 am to 12 pm, 12 pm to 6 pm, and 6 pm to 3 am the next day. Each time period corresponds to a different abnormal threshold.
[0039] The abnormal threshold is determined based on historical data. For example, if the number of users playing the video in the past few days is within the first interval from 3 am to 12 pm, the number of users playing the video from 12 pm to 6 pm is within the second interval, and the number of users playing the video from 6 pm to 3 am the next day is within the third interval, then the theoretical number of users in each time period is determined based on the first to third intervals, and the theoretical number of users is the abnormal threshold for each time period. Alternatively, after calculating the theoretical number of users, the value can be adjusted to obtain the abnormal threshold.
[0040] The above real-time value is the real-time value obtained by monitoring the playback index. For example, at a certain time point on a certain day, the number of playback users in a certain live broadcast room is 10,000 users. This data is the real-time value.
[0041] If the real-time value is greater than the abnormal threshold, it means that the playback indicator may be abnormal. In this case, the associated indicator of the playback indicator can be obtained.
[0042] The associated indicator of the playback indicator is an indicator associated with the playback indicator, which can be obtained according to a preset association relationship. In addition, the system can also determine the indicator whose fluctuation is consistent with the playback indicator based on historical data as its associated indicator. For example, the playback indicator has certain numerical fluctuations over time, and the fluctuation ratio of another indicator is very similar to the fluctuation ratio of the playback indicator, then the other indicator can be used as the associated indicator of the playback indicator. The judgment of whether it is similar can be determined by calculating the curve similarity. It should be noted that at this time, when determining the associated indicator, it is necessary to use data without abnormalities for determination.
[0043] For the associated indicators, the real-time values of the associated indicators are obtained, and then the real-time values of the playback indicators are compared with the real-time values of the associated indicators. If the comparison result indicates that the playback indicator is abnormal, it is truly determined that the playback indicator is abnormal. If the comparison result indicates that the playback indicator is normal, it is normal that the previous real-time value of the playback indicator exceeds the abnormal threshold.
[0044] The solution provided by the embodiment of the present application obtains the abnormal threshold value of the playback indicator to be monitored and the real-time value obtained by monitoring the above playback indicator; when the comparison result of the above real-time value and the above abnormal threshold value indicates that the above real-time value is abnormal, determines the associated indicator of the above playback indicator; obtains the real-time value of the above associated indicator; when the comparison result of the real-time value of the above playback indicator and the real-time value of the above associated indicator indicates that the above playback indicator is abnormal, determines that the above playback indicator is abnormal; when the comparison result of the real-time value of the above playback indicator and the real-time value of the above associated indicator indicates that the above playback indicator is normal, determines that the above playback indicator is normal, so that when abnormality detection is performed on the playback indicator, the associated indicator of the playback indicator can be introduced, and the real-time value of the playback indicator is compared with the real-time value of the associated indicator to check whether the playback indicator is really abnormal, thereby improving the accuracy of abnormality detection on the playback indicator.
[0045] As an optional example, the above-mentioned acquisition of the abnormal threshold of the playback indicator to be monitored includes: acquiring the historical data of the above-mentioned playback indicator; estimating the estimated value of the above-mentioned playback indicator at any time point based on the above-mentioned historical data; and determining the product of the above-mentioned estimated value and a predetermined ratio as the above-mentioned abnormal threshold, wherein the above-mentioned predetermined ratio is determined based on any of the above-mentioned time points.
[0046] In this embodiment, when determining the abnormal threshold of any playback indicator, it is necessary to obtain the historical data of the playback indicator and estimate the abnormal threshold of the playback indicator based on the historical data. After the historical data is obtained, a year-on-year and month-on-month comparison can be performed, that is, the change in the data of the playback indicator in each time period and the change in the data of the playback indicator between different periods can be determined. By collecting and comparing data from multiple periods, the approximate value or approximate range of the data of the playback indicator in a certain period in the future can be determined. The determined approximate value or approximate range is the above-mentioned estimated value.
[0047] Then, the estimated value may be multiplied by a predetermined ratio, which is a value greater than 1, so as to increase the size of the estimated value and avoid the situation where a playback indicator is considered abnormal if it slightly exceeds the daily value.
[0048] As an optional example, after estimating the estimated value of the above-mentioned playback indicator at any time point based on the above-mentioned historical data, the above-mentioned method also includes: when any of the above-mentioned time points is in a popular time period, setting the above-mentioned predetermined ratio to a first value greater than 1, and when any of the above-mentioned time points is in an unpopular time period, setting the above-mentioned predetermined ratio to a second value greater than 1, wherein, the closer any of the above-mentioned time points is to the center of the above-mentioned popular time period, the larger the above-mentioned first value is, and the closer any of the above-mentioned time points is to the center of the above-mentioned unpopular time period, the smaller the above-mentioned second value is, and the minimum value of the above-mentioned first value is greater than the maximum value of the above-mentioned second value.
[0049] In this embodiment, the predetermined ratio is a value greater than 1, and the predetermined ratio is related to any time point. For example, the abnormal threshold value estimated at 12 noon and the abnormal threshold value estimated at 6 pm, even if the estimated values are the same, the abnormal threshold values determined are different due to the different predetermined ratios. The predetermined ratio can be specifically determined based on whether any time point is a popular time point. If there are many people watching the live broadcast at 12 noon, then 12 noon is a popular time point. The closer to the popular time point, the larger the predetermined ratio, and the closer to the unpopular time point, the smaller the predetermined ratio. The minimum predetermined ratio is also a value greater than 1. The maximum value of the predetermined ratio for the unpopular time period is smaller than the minimum value of the predetermined ratio for the popular time period.
[0050] As an optional example, when the comparison result between the real-time value and the abnormal threshold indicates that there is an abnormality in the real-time value, determining the associated indicator of the playback indicator includes: determining one or more indicators as the associated indicators from the indicator relationship network, from multiple second indicators related to the playback indicator, wherein the indicator relationship network includes a first indicator and the second indicator, each of the first indicators is related to one or more of the second indicators, each of the second indicators is related to one or more of the first indicators, and the playback indicator is one of the first indicators.
[0051] In this embodiment, when determining the associated index of the playback index, the associated index can be determined from the index relationship network. Figure 2 As shown, Figure 2 In the example, playback index 1 is associated with associated index 1 and associated index 2, playback index 2 is associated with associated index 2, playback index 3 is associated with associated index 2, and playback index 4 is associated with associated indexes 3 and 4. If the associated index of playback index 2 is obtained, associated index 2 is obtained, and if the associated index of playback index 4 is obtained, associated index 3 and associated index 4 are obtained.
[0052] As an optional example, when the comparison result between the real-time value and the abnormal threshold indicates that there is an abnormality in the real-time value, determining the associated indicators of the playback indicators also includes: determining the number of the associated indicators according to the degree of abnormality, wherein the higher the degree of abnormality, the greater the number of the associated indicators.
[0053] In this embodiment, when the real-time value of the above-mentioned playback index exceeds the abnormal threshold, the number of related indicators can be determined according to the exceeded range. The more the abnormal threshold is exceeded, the more the corresponding related indicators are, so as to make a judgment from the perspective of multiple related indicators.
[0054] For example, Figure 3 As shown, when the real-time value of the playback indicator minus the abnormal threshold value, the remaining value is less than 10 (can be a negative value), it corresponds to a related indicator. The larger the remaining value, the more related indicators there are. The upper limit of the related indicator can be set. Figure 3 The numerical values and interval ranges are only examples and do not constitute limitations.
[0055] As an optional example, after obtaining the real-time value of the above-mentioned associated indicator, the above-mentioned method also includes: comparing the real-time value of the above-mentioned associated indicator with the real-time value of the above-mentioned playback indicator to obtain a first ratio; determining a ratio range based on the ratio of the historical value of the above-mentioned associated indicator to the historical data of the above-mentioned playback indicator; when the above-mentioned first ratio is within the above-mentioned ratio range, determining that the above-mentioned comparison result indicates that the above-mentioned playback indicator is normal, and when the above-mentioned first ratio is outside the above-mentioned ratio range, determining that the above-mentioned comparison result indicates that the above-mentioned playback indicator is abnormal.
[0056] In this embodiment, after obtaining the real-time value of the associated indicator, the ratio between the real-time value of the playback indicator and the real-time value of the associated indicator can be checked, and the real-time value of the associated indicator is divided by the real-time value of the playback indicator to obtain a first ratio. Then, based on the historical values of the associated indicator and the historical data of the playback indicator, the ratio interval is determined. If the first ratio is within the ratio interval, it can be determined that the playback indicator is normal. If the first ratio is outside the ratio interval, it can be determined that the playback indicator is abnormal.
[0057] For example, in the historical data of the playback index and the associated index, the ratio range is 1 / 3-2 / 3, and the first ratio obtained is 1 / 2, which is within the interval, so it can be considered that the playback index is normal.
[0058] As an optional example, after obtaining the real-time values of the above-mentioned associated indicators, the above-mentioned method also includes: obtaining historical data of the above-mentioned playback indicators and the above-mentioned associated indicators; drawing a playback curve based on the historical data and real-time values of the above-mentioned playback indicators, and drawing an associated curve based on the historical data and real-time values of the above-mentioned associated indicators; when the trajectory similarity between the above-mentioned playback curve and the above-mentioned associated curve is greater than a preset similarity threshold, determining that the above-mentioned comparison result indicates that the above-mentioned playback indicator is normal, and when the above-mentioned trajectory similarity is less than the above-mentioned preset similarity threshold, determining that the above-mentioned comparison result indicates that the above-mentioned playback indicator is abnormal.
[0059] In this embodiment, after obtaining the real-time value of the associated index, it is possible to check whether the curve fluctuations of the playback curve formed by the real-time value of the playback index and the real-time value of the associated index are consistent with the curve fluctuations of the associated curve, that is, to compare the trajectory similarity of the two. Figure 4 As shown, Figure 4 The left side shows the real-time value of the playback indicator changing over time. Figure 4 The right side of the chart shows how the real-time values of the related indicators change over time. Although the specific values are different, the changes in their directions are basically the same, and the trajectory similarity is high. Therefore, it can be determined that the playback indicators are normal.
[0060] As an optional example, after obtaining the real-time value of the above-mentioned related indicator, the above-mentioned method also includes: when the real-time value of the above-mentioned playback indicator increases or decreases in the same proportion as the real-time value of the above-mentioned related indicator, determining that the above-mentioned comparison result indicates that the above-mentioned playback indicator is normal; when the real-time value of the above-mentioned playback indicator and the real-time value of the above-mentioned related indicator do not change in the same proportion, determining that the above-mentioned comparison result indicates that there is an abnormality in the above-mentioned playback indicator.
[0061] In this embodiment, after obtaining the real-time value of the associated indicator, it can be checked whether the real-time value of the associated indicator increases or decreases synchronously with the real-time value of the playback indicator. If they increase or decrease synchronously, the playback indicator is considered normal. If the two do not increase or decrease synchronously, the playback indicator is considered abnormal. If there are multiple associated indicators, each associated indicator must change synchronously with the playback indicator before the playback indicator can be considered normal. If any associated indicator does not change synchronously with the synchronization indicator, the playback indicator is considered abnormal. In addition, this embodiment can also detect the playback indicator and the associated indicator through a model, so as to obtain a comparison result and determine whether the playback indicator is abnormal.
[0062] Figure 5 and Figure 6 FIG. 1 is a flow chart of this embodiment, and is illustrated by taking a scenario as an example.
[0063] In the fields of business analysis, data monitoring, and report production, year-on-year and month-on-month changes in data are important indicators for measuring data changes. Through these indicators, trends, cyclical fluctuations, and seasonal changes in business performance can be tracked. Taking data monitoring as an example, real-time monitoring systems can use these indicators to detect potential anomalies, such as year-on-year and month-on-month declines in sales, which represent some kind of anomaly.
[0064] The anomaly detection method proposed in this embodiment introduces multiple reference systems, adaptively switches between timestamps and associated indicators, prioritizes timestamps as the reference system, and detects changes in playback indicators at the same time. When a large anomaly occurs in the playback indicator, the coordinate system is switched to use the associated indicator as the reference system, and then the problem can be effectively located by performing anomaly detection through the associated indicator.
[0065] Take the online video playback crash number as an example. Under normal circumstances, the number of playback crashes at this time point in the early morning is relatively small. A simple year-on-year analysis of a single indicator can determine whether the number of online playback crashes is normal. However, when things like the Spring Festival and New Year's Eve occur, the number of online crashes will inevitably increase with the increase in the number of users. The year-on-year analysis will definitely produce anomalies. However, due to the large base number of users, the increase in the number of crashes may not be an abnormal phenomenon.
[0066] However, simply using the crash rate (number of crashes / number of online users) as an indicator still cannot effectively detect anomalies through simple year-on-year and month-on-month characteristics. Since there are some devices that have been crashing for a long time, the crash rate indicator will be significantly higher when the number of online users is small.
[0067] Based on this situation, multiple reference systems are introduced to adaptively switch between timestamps and related indicators, with time being used as the reference system first. At the same time, changes in playback indicators are detected. When playback indicators produce large anomalies, the coordinate system is switched to use related indicators as the reference system, which can effectively locate the problem.
[0068] Still taking the number of crashes as an example, the number of online users is introduced as a related indicator, and anomaly detection is performed first through the same period last year and month-on-month features. When an anomaly is detected, the same period last year and month-on-month anomaly detection is performed on its related indicators. If the number of online users, which is its related indicator, is normal, it is determined that the number of crashes is abnormal. If the related indicator is abnormal, the abnormal judgment is corrected: specifically, the fitting value of the current number of crashes is fitted by training the number of online users with historical data, and the fitting value is used as the same period last year and month-on-month feature for secondary anomaly detection to determine whether an abnormality occurs.
[0069] The specific process is to select the main indicators (such as the number of playback crashes) and related indicators (such as the number of online users). First, perform data preprocessing, including cleaning, missing value processing, normalization, etc. Then establish an anomaly detection framework: establish a time series-based year-on-year anomaly detection model for the first level of screening. Define the threshold of the outlier, which can be based on business knowledge, historical data analysis, or automated methods such as the 3-sigma principle. Train a fitting model through the relationship between the related indicators and the playback indicators in the historical data, such as training with a multilayer perceptron (MLP) model. Perform multi-reference system anomaly detection: enable the first anomaly detection model, perform year-on-year detection on the playback indicators, and identify potential anomalies. When the first round of detection marks an anomaly, check the related indicators. Use the year-on-year method to detect whether the related indicators also show anomalies. If the related indicators do not show anomalies, keep the original anomaly state. If the related indicators show anomalies, take the following steps: Use the related indicators to obtain the fitted value of the current main indicator through the trained fitting model. Use the fitted crash value as the year-on-year reference object to compare with the actual observed value for the second round of anomaly detection. If the playback indicators are detected as abnormal by both rounds of abnormality detection, it is judged as abnormal; otherwise, the playback indicators are judged to be normal.
[0070] Figure 7 This is a schematic diagram of the structure of a device for detecting abnormalities in playback data provided by an embodiment of the present application. Figure 7 As shown, the above-mentioned abnormality detection device for playing data includes:
[0071] The first acquisition module 702 is used to acquire an abnormal threshold value of a playback indicator to be monitored and a real-time value obtained by monitoring the playback indicator;
[0072] The first determination module 704 is used to determine the associated index of the playback index when the comparison result between the real-time value and the abnormal threshold indicates that the real-time value is abnormal;
[0073] The second acquisition module 706 is used to obtain the real-time value of the associated indicator;
[0074] The second determination module 708 is used to determine that there is an abnormality in the playback indicator when the comparison result of the real-time value of the playback indicator with the real-time value of the associated indicator indicates that there is an abnormality in the playback indicator; when the comparison result of the real-time value of the playback indicator with the real-time value of the associated indicator indicates that the playback indicator is normal, determine that the playback indicator is normal.
[0075] The above-mentioned abnormality detection method of playback data can be applied in the field of live broadcasting. In the field of live broadcasting, a user can start a live broadcast and other users can watch the live broadcast of the user. The above-mentioned playback indicators can be the playback data of a single live broadcast room, or the playback data of some or all live broadcast rooms of a platform. The playback indicators may include the number of playback users, playback traffic, likes, dislikes, forwarding, comments, interactions, virtual gifts, the number of abnormal network users or collections, etc. The above are only examples and do not constitute limitations.
[0076] In this embodiment, each playback indicator corresponds to one or more abnormal thresholds. The abnormal threshold is a theoretical peak value determined based on historical data of the playback indicator. In addition, in this embodiment, abnormal thresholds at different time points or time periods can also be determined for each playback indicator.
[0077] For example, taking the playback indicator as the number of playback users in a live broadcast room, three time periods within a day can be set, such as 3 am to 12 pm, 12 pm to 6 pm, and 6 pm to 3 am the next day. Each time period corresponds to a different abnormal threshold.
[0078] The abnormal threshold is determined based on historical data. For example, if the number of users playing the video in the past few days is within the first interval from 3 am to 12 pm, the number of users playing the video from 12 pm to 6 pm is within the second interval, and the number of users playing the video from 6 pm to 3 am the next day is within the third interval, then the theoretical number of users in each time period is determined based on the first to third intervals, and the theoretical number of users is the abnormal threshold for each time period. Alternatively, after calculating the theoretical number of users, the value can be adjusted to obtain the abnormal threshold.
[0079] The above real-time value is the real-time value obtained by monitoring the playback index. For example, at a certain time point on a certain day, the number of playback users in a certain live broadcast room is 10,000 users. This data is the real-time value.
[0080] If the real-time value is greater than the abnormal threshold, it means that the playback indicator may be abnormal. In this case, the associated indicator of the playback indicator can be obtained.
[0081] The associated indicator of the playback indicator is an indicator associated with the playback indicator, which can be obtained according to a preset association relationship. In addition, the system can also determine the indicator whose fluctuation is consistent with the playback indicator based on historical data as its associated indicator. For example, the playback indicator has certain numerical fluctuations over time, and the fluctuation ratio of another indicator is very similar to the fluctuation ratio of the playback indicator, then the other indicator can be used as the associated indicator of the playback indicator. The judgment of whether it is similar can be determined by calculating the curve similarity. It should be noted that at this time, when determining the associated indicator, it is necessary to use data without abnormalities for determination.
[0082] For the associated indicators, the real-time values of the associated indicators are obtained, and then the real-time values of the playback indicators are compared with the real-time values of the associated indicators. If the comparison result indicates that the playback indicator is abnormal, it is truly determined that the playback indicator is abnormal. If the comparison result indicates that the playback indicator is normal, it is normal that the previous real-time value of the playback indicator exceeds the abnormal threshold.
[0083] The solution provided by the embodiment of the present application obtains the abnormal threshold value of the playback indicator to be monitored and the real-time value obtained by monitoring the above playback indicator; when the comparison result of the above real-time value and the above abnormal threshold value indicates that the above real-time value is abnormal, determines the associated indicator of the above playback indicator; obtains the real-time value of the above associated indicator; when the comparison result of the real-time value of the above playback indicator and the real-time value of the above associated indicator indicates that the above playback indicator is abnormal, determines that the above playback indicator is abnormal; when the comparison result of the real-time value of the above playback indicator and the real-time value of the above associated indicator indicates that the above playback indicator is normal, determines that the above playback indicator is normal, so that when abnormality detection is performed on the playback indicator, the associated indicator of the playback indicator can be introduced, and the real-time value of the playback indicator is compared with the real-time value of the associated indicator to check whether the playback indicator is really abnormal, thereby improving the accuracy of abnormality detection on the playback indicator.
[0084] As an optional example, the first acquisition module includes: an estimation unit for acquiring historical data of playback indicators; estimating an estimated value of the playback indicators at any time point based on the historical data; and determining the product of the estimated value and a predetermined ratio as an abnormal threshold, wherein the predetermined ratio is determined based on any time point.
[0085] In this embodiment, when determining the abnormal threshold of any playback indicator, it is necessary to obtain the historical data of the playback indicator and estimate the abnormal threshold of the playback indicator based on the historical data. After the historical data is obtained, a year-on-year and month-on-month comparison can be performed, that is, the change in the data of the playback indicator in each time period and the change in the data of the playback indicator between different periods can be determined. By collecting and comparing data from multiple periods, the approximate value or approximate range of the data of the playback indicator in a certain period in the future can be determined. The determined approximate value or approximate range is the above-mentioned estimated value.
[0086] Then, the estimated value may be multiplied by a predetermined ratio, which is a value greater than 1, so as to increase the size of the estimated value and avoid the situation where a playback indicator is considered abnormal if it slightly exceeds the daily value.
[0087] As an optional example, the first acquisition module also includes: a setting unit, which is used to set the predetermined ratio to a first value greater than 1 when any time point is in a popular time period after estimating the estimated value of the playback indicator at any time point based on historical data, and to set the predetermined ratio to a second value greater than 1 when any time point is in an unpopular time period, wherein the closer any time point is to the center of the popular time period, the larger the first value, and the closer any time point is to the center of the unpopular time period, the smaller the second value, and the minimum value of the first value is greater than the maximum value of the second value.
[0088] In this embodiment, the predetermined ratio is a value greater than 1, and the predetermined ratio is related to any time point. For example, the abnormal threshold value estimated at 12 noon and the abnormal threshold value estimated at 6 pm, even if the estimated values are the same, the abnormal threshold values determined are different due to the different predetermined ratios. The predetermined ratio can be specifically determined based on whether any time point is a popular time point. If there are many people watching the live broadcast at 12 noon, then 12 noon is a popular time point. The closer to the popular time point, the larger the predetermined ratio, and the closer to the unpopular time point, the smaller the predetermined ratio. The minimum predetermined ratio is also a value greater than 1. The maximum value of the predetermined ratio for the unpopular time period is smaller than the minimum value of the predetermined ratio for the popular time period.
[0089] As an optional example, the first determination module includes: a determination unit, used to determine one or more indicators as associated indicators from multiple second indicators related to the playback indicator in the indicator relationship network, wherein the indicator relationship network includes first indicators and second indicators, each first indicator is related to one or more second indicators, each second indicator is related to one or more first indicators, and the playback indicator is one of the first indicators.
[0090] In this embodiment, when determining the associated index of the playback index, the associated index can be determined from the index relationship network. Figure 2 As shown, Figure 2 In the example, playback index 1 is associated with associated index 1 and associated index 2, playback index 2 is associated with associated index 2, playback index 3 is associated with associated index 2, and playback index 4 is associated with associated indexes 3 and 4. If the associated index of playback index 2 is obtained, associated index 2 is obtained, and if the associated index of playback index 4 is obtained, associated index 3 and associated index 4 are obtained.
[0091] As an optional example, the determination unit includes: a determination subunit, which is used to determine the number of associated indicators according to the degree of abnormality, wherein the higher the degree of abnormality, the greater the number of associated indicators.
[0092] In this embodiment, when the real-time value of the above-mentioned playback index exceeds the abnormal threshold, the number of related indicators can be determined according to the exceeded range. The more the abnormal threshold is exceeded, the more the corresponding related indicators are, so as to make a judgment from the perspective of multiple related indicators.
[0093] For example, Figure 3 As shown, when the real-time value of the playback indicator minus the abnormal threshold value, the remaining value is less than 10 (can be a negative value), it corresponds to a related indicator. The larger the remaining value, the more related indicators there are. The upper limit of the related indicator can be set. Figure 3 The numerical values and interval ranges are only examples and do not constitute limitations.
[0094] As an optional example, the above-mentioned device also includes: a first comparison module, which is used to compare the real-time value of the associated indicator with the real-time value of the playback indicator after obtaining the real-time value of the associated indicator to obtain a first ratio; determine a ratio interval based on the ratio of the historical value of the associated indicator to the historical data of the playback indicator; when the first ratio is within the ratio interval, determine that the comparison result indicates that the playback indicator is normal; when the first ratio is outside the ratio interval, determine that the comparison result indicates that the playback indicator is abnormal.
[0095] In this embodiment, after obtaining the real-time value of the associated indicator, the ratio between the real-time value of the playback indicator and the real-time value of the associated indicator can be checked, and the real-time value of the associated indicator is divided by the real-time value of the playback indicator to obtain a first ratio. Then, based on the historical values of the associated indicator and the historical data of the playback indicator, the ratio interval is determined. If the first ratio is within the ratio interval, it can be determined that the playback indicator is normal. If the first ratio is outside the ratio interval, it can be determined that the playback indicator is abnormal.
[0096] For example, in the historical data of the playback index and the associated index, the ratio range is 1 / 3-2 / 3, and the first ratio obtained is 1 / 2, which is within the interval, so it can be considered that the playback index is normal.
[0097] As an optional example, the above-mentioned device also includes: a second comparison module, which is used to obtain the historical data of the playback indicator and the associated indicator after obtaining the real-time value of the associated indicator; draw a playback curve according to the historical data and the real-time value of the playback indicator, and draw an associated curve according to the historical data and the real-time value of the associated indicator; when the trajectory similarity between the playback curve and the associated curve is greater than a preset similarity threshold, determine that the comparison result indicates that the playback indicator is normal, and when the trajectory similarity is less than the preset similarity threshold, determine that the comparison result indicates that the playback indicator is abnormal.
[0098] In this embodiment, after obtaining the real-time value of the associated index, it is possible to check whether the curve fluctuations of the playback curve formed by the real-time value of the playback index and the real-time value of the associated index are consistent with the curve fluctuations of the associated curve, that is, to compare the trajectory similarity of the two. Figure 4 As shown, Figure 4 The left side shows the real-time value of the playback indicator changing over time. Figure 4 The right side of the chart shows how the real-time values of the related indicators change over time. Although the specific values are different, the changes in their directions are basically the same, and the trajectory similarity is high. Therefore, it can be determined that the playback indicators are normal.
[0099] As an optional example, the above-mentioned device also includes: a third comparison module, which is used to, after obtaining the real-time value of the associated indicator, determine that the comparison result indicates that the playback indicator is normal when the real-time value of the playback indicator and the real-time value of the associated indicator increase or decrease in the same proportion; and determine that the comparison result indicates that there is an abnormality in the playback indicator when the real-time value of the playback indicator and the real-time value of the associated indicator do not change in the same proportion.
[0100] In this embodiment, after obtaining the real-time value of the associated indicator, it can be checked whether the real-time value of the associated indicator increases or decreases synchronously with the real-time value of the playback indicator. If they increase or decrease synchronously, the playback indicator is considered normal. If the two do not increase or decrease synchronously, the playback indicator is considered abnormal. If there are multiple associated indicators, each associated indicator must change synchronously with the playback indicator before the playback indicator can be considered normal. If any associated indicator does not change synchronously with the synchronization indicator, the playback indicator is considered abnormal. In addition, this embodiment can also detect the playback indicator and the associated indicator through a model, so as to obtain a comparison result and determine whether the playback indicator is abnormal.
[0101] For other examples of this embodiment, please refer to the above examples and will not be described in detail here.
[0102] like Figure 8 As shown, an embodiment of the present application provides an electronic device, including a processor 111, a communication interface 112, a memory 113 and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114.
[0103] Memory 113, used for storing computer programs;
[0104] In one embodiment of the present application, the processor 111 is used to implement the abnormality detection method for playback data provided by any one of the aforementioned method embodiments when executing the program stored in the memory 113.
[0105] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for detecting anomalies in playback data provided by any one of the aforementioned method embodiments is implemented.
[0106] The device embodiments described above are merely illustrative, wherein the units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0107] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the above methods of each embodiment or some parts of the embodiments.
[0108] It should be understood that the terms used herein are only for the purpose of describing specific example embodiments and are not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms "one", "an" and "above" as used herein may also be meant to include plural forms. The terms "include", "comprise", "contain", and "have" are inclusive, and therefore specify the existence of stated features, steps, operations, elements and / or parts, but do not exclude the existence or addition of one or more other features, steps, operations, elements, parts, and / or combinations thereof. The method steps, processes, and operations described herein are not interpreted as necessarily requiring them to be performed in the specific order described or illustrated, unless the execution order is clearly indicated. It should also be understood that additional or alternative steps may be used.
[0109] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features applied herein.
Claims
1. A method for detecting abnormalities in playback data, It is characterized in that include: Obtaining an abnormal threshold of a playback indicator to be monitored and a real-time value obtained by monitoring the playback indicator; When the comparison result between the real-time value and the abnormal threshold indicates that the real-time value is abnormal, determining the associated index of the playback index; Obtaining real-time values of the associated indicators; When the comparison result of the real-time value of the playback indicator with the real-time value of the associated indicator indicates that the playback indicator is abnormal, it is determined that the playback indicator is abnormal; when the comparison result of the real-time value of the playback indicator with the real-time value of the associated indicator indicates that the playback indicator is normal, it is determined that the playback indicator is normal.
2. The method according to claim 1, It is characterized in that The abnormal threshold of the playback indicator to be monitored is obtained by: Obtaining historical data of the playback indicator; estimating an estimated value of the playback indicator at any time point according to the historical data; The product of the estimated value and a predetermined ratio is determined as the abnormal threshold, wherein the predetermined ratio is determined according to the arbitrary time point.
3. The method according to claim 2, It is characterized in that After estimating the estimated value of the playback indicator at any time point according to the historical data, the method further includes: When any one of the time points is in a popular time period, the predetermined ratio is set to a first value greater than 1; and when any one of the time points is in an unpopular time period, the predetermined ratio is set to a second value greater than 1, wherein the closer any one of the time points is to the center of the popular time period, the larger the first value is, and the closer any one of the time points is to the center of the unpopular time period, the smaller the second value is, and the minimum value of the first value is greater than the maximum value of the second value.
4. The method according to claim 1, It is characterized in that When the comparison result between the real-time value and the abnormal threshold indicates that the real-time value is abnormal, determining the associated index of the playback index further includes: From the indicator relationship network, one or more indicators are determined as the associated indicators from multiple second indicators related to the playback indicator, wherein the indicator relationship network includes first indicators and second indicators, each of the first indicators is related to one or more second indicators, each of the second indicators is related to one or more first indicators, and the playback indicator is one of the first indicators.
5. The method according to claim 4, It is characterized in that When the comparison result between the real-time value and the abnormal threshold indicates that the real-time value is abnormal, determining the associated index of the playback index further includes: The number of the associated indicators is determined according to the degree of abnormality, wherein the higher the degree of abnormality is, the greater the number of the associated indicators is.
6. The method according to claim 1, It is characterized in that After obtaining the real-time value of the associated indicator, the method further includes: Comparing the real-time value of the associated indicator with the real-time value of the playback indicator to obtain a first ratio; Determining a ratio interval according to a ratio of a historical value of the associated indicator to historical data of the playback indicator; When the first ratio is within the ratio interval, it is determined that the comparison result indicates that the playback indicator is normal, and when the first ratio is outside the ratio interval, it is determined that the comparison result indicates that the playback indicator is abnormal.
7. The method according to claim 1, It is characterized in that After obtaining the real-time value of the associated indicator, the method further includes: Acquire historical data of the playback indicator and the associated indicator; Draw a playback curve according to the historical data and real-time values of the playback index, and draw a correlation curve according to the historical data and real-time values of the correlation index; When the trajectory similarity between the playback curve and the association curve is greater than a preset similarity threshold, it is determined that the comparison result indicates that the playback indicator is normal; when the trajectory similarity is less than the preset similarity threshold, it is determined that the comparison result indicates that the playback indicator is abnormal.
8. The method according to claim 1, It is characterized in that After obtaining the real-time value of the associated indicator, the method further includes: When the real-time value of the playback indicator and the real-time value of the associated indicator increase or decrease in the same proportion, it is determined that the comparison result indicates that the playback indicator is normal. When the real-time value of the playback indicator and the real-time value of the associated indicator do not change in the same proportion, it is determined that the comparison result indicates that there is an abnormality in the playback indicator.
9. A device for detecting abnormality in playback data, It is characterized in that include: A first acquisition module is used to acquire an abnormal threshold value of a playback indicator to be monitored and a real-time value obtained by monitoring the playback indicator; A first determination module, configured to determine an associated index of the playback index when a comparison result between the real-time value and the abnormal threshold indicates that the real-time value is abnormal; A second acquisition module is used to obtain the real-time value of the associated indicator; The second determination module is used to determine that there is an abnormality in the playback indicator when the comparison result of the real-time value of the playback indicator with the real-time value of the associated indicator indicates that there is an abnormality in the playback indicator; and to determine that the playback indicator is normal when the comparison result of the real-time value of the playback indicator with the real-time value of the associated indicator indicates that the playback indicator is normal.
10. An electronic device, It is characterized in that include: at least one communication interface; at least one bus connected to the at least one communication interface; at least one processor coupled to the at least one bus; At least one memory connected to the at least one bus, wherein the memory stores a computer program, and when the processor executes the computer program, the method for detecting anomalies in playback data described in any one of claims 1 to 8 is implemented.
11. A computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions are used to execute the method for detecting anomalies in playback data according to any one of claims 1 to 8 of the present application.
