An index root cause analysis method, device and equipment

Abstract

The application provides a kind of index root cause analysis method, device and equipment, related to communication technical field.The method comprises: taking the subordinate index of the first target index of the data to be analyzed, and the first target index is a root index;According to the contribution degree and / or similarity of the subordinate index, the root cause index in the subordinate index is determined.The scheme of the application solves the problem of low efficiency of manual analysis of data root cause index.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a method, apparatus and equipment for root cause analysis of indicators. Background Technology

[0002] In scenarios such as business management, financial management, and scientific research, root cause analysis is often necessary to thoroughly solve or explain problems.

[0003] In existing root cause analysis, the root causes of changes in the analyzed indicator are determined by manual reasoning based on the trend charts of each subordinate indicator. However, manual reasoning consumes a significant amount of time and effort, greatly reducing the efficiency of the analysis. Summary of the Invention

[0004] The purpose of this invention is to provide a method, apparatus, and equipment for root cause analysis of indicators, which solves the problem of low efficiency in manual analysis of root cause indicators.

[0005] To achieve the above objectives, embodiments of the present invention provide a root cause analysis method, comprising:

[0006] Obtain the sub-indicators of the first target indicator of the data to be analyzed, where the first target indicator is the root indicator;

[0007] The root cause indicators among the lower-level indicators are determined based on their contribution and / or similarity.

[0008] Optionally, the lower-level indicators may include one or more levels of indicators;

[0009] The step of determining the root cause indicators among the lower-level indicators based on their contribution and / or similarity includes:

[0010] Calculate the contribution of the lower-level indicators to the higher-level indicators;

[0011] If at least one of the contributions of the peer indicators is greater than 0, the root cause indicator is selected according to the magnitude of the contribution.

[0012] If the contribution of all peer indicators is less than 0, obtain the similarity between the peer indicator and the next higher level indicator, and determine the indicator with the highest similarity as the root cause indicator.

[0013] Optionally, calculating the contribution of lower-level indicators to higher-level indicators includes:

[0014] Obtain the correlation between the peer-level indicator and the higher-level indicator;

[0015] Based on the aforementioned correlation, determine the summation relationship between the peer indicator and the higher-level indicator;

[0016] Based on the summation relationship, the contribution is calculated using actual and predicted data.

[0017] Optionally, selecting the root cause indicator based on the magnitude of the contribution includes:

[0018] The contribution values ​​are selected sequentially from largest to smallest, and the sum of the selected contribution values ​​is calculated.

[0019] If the sum is greater than or equal to a preset threshold, the indicator corresponding to the selected contribution is determined as the root cause indicator.

[0020] Optionally, obtaining the similarity between the peer indicator and the higher-level indicator includes:

[0021] The similarity between the same-level index and the previous-level index is calculated using the cosine similarity formula.

[0022] Optionally, when the peer indicator is a non-first-level subordinate indicator of the first target indicator, the next-level indicator of the peer indicator is the root cause indicator.

[0023] Optionally, after determining the root cause indicators among the lower-level indicators, the method further includes:

[0024] In determining the root cause index and the superior index of the root cause index, the data source is a second target index of different elements, and the different elements all belong to the target dimension.

[0025] Calculate the contribution of each element's data to the second target indicator;

[0026] If at least one of the contribution values ​​of the data elements is greater than 0, the root cause element is selected according to the magnitude of the contribution.

[0027] If the contribution of each element data is less than 0, the similarity between each element data and the second target indicator is obtained, and the element with the highest similarity is determined as the root cause element.

[0028] Optionally, the target dimension is a geographical region, and the different elements correspond to different regions.

[0029] Optionally, the subordinate indicators of the first target indicator for acquiring the data to be analyzed include:

[0030] If an anomaly is detected in the data of the primary target indicator, obtain the subordinate indicators of the primary target indicator.

[0031] To achieve the above objectives, embodiments of the present invention provide an indicator root cause analysis device, comprising:

[0032] The acquisition module is used to acquire the lower-level indicators of the first target indicator of the data to be analyzed, where the first target indicator is the root indicator.

[0033] The first processing module is used to determine the root cause indicators among the lower-level indicators based on their contribution and / or similarity.

[0034] Optionally, the lower-level indicators may include one or more levels of indicators;

[0035] The first processing module includes:

[0036] The first processing submodule is used to calculate the contribution of the lower-level indicators to the higher-level indicators.

[0037] The second processing submodule is used to select a root cause indicator based on the magnitude of the contribution of at least one of the peer indicators when the contribution of the peer indicators is greater than 0.

[0038] The third processing submodule is used to obtain the similarity between the peer indicator and the previous level indicator when all the contributions of the peer indicators are less than 0, and to determine the indicator with the highest similarity as the root cause indicator.

[0039] Optionally, the first processing submodule includes:

[0040] The acquisition unit is used to acquire the correlation between the peer indicator and the higher-level indicator;

[0041] The first determining unit is used to determine the summation relationship between the peer indicator and the higher-level indicator based on the correlation relationship;

[0042] The first processing unit is used to calculate the contribution based on the summation relationship using actual data and predicted data.

[0043] Optionally, the second processing submodule includes:

[0044] The second processing unit is used to select the contribution values ​​in descending order and calculate the sum of the selected contribution values.

[0045] The second determining unit is used to determine the indicator corresponding to the selected contribution as the root cause indicator when the sum is greater than or equal to a preset threshold.

[0046] Optionally, the third processing submodule is further configured to:

[0047] The similarity between the same-level index and the previous-level index is calculated using the cosine similarity formula.

[0048] Optionally, when the peer indicator is a non-first-level subordinate indicator of the first target indicator, the next-level indicator of the peer indicator is the root cause indicator.

[0049] Optionally, the device further includes:

[0050] The second processing module is used to determine the second target indicators whose data sources are different elements among the root cause indicators and the parent indicators of the root cause indicators, and the different elements all belong to the target dimension.

[0051] The third processing module is used to calculate the contribution of each element data to the second target indicator;

[0052] The fourth processing module is used to select the root cause element based on the magnitude of the contribution of each element data when at least one of the contribution values ​​is greater than 0.

[0053] The fifth processing module is used to obtain the similarity between each element data and the second target indicator when the contribution of each element data is less than 0, and to determine the element with the highest similarity as the root cause element.

[0054] Optionally, the target dimension is a geographical region, and the different elements correspond to different regions.

[0055] Optionally, the acquisition module is further configured to:

[0056] If an anomaly is detected in the data of the primary target indicator, obtain the subordinate indicators of the primary target indicator.

[0057] To achieve the above objectives, embodiments of the present invention provide an indicator root cause analysis device, including a processor, the processor being used for:

[0058] Obtain the sub-indicators of the first target indicator of the data to be analyzed, where the first target indicator is the root indicator;

[0059] The root cause indicators among the lower-level indicators are determined based on their contribution and / or similarity.

[0060] Optionally, the lower-level indicators may include one or more levels of indicators;

[0061] The processor is also used for:

[0062] Calculate the contribution of the lower-level indicators to the higher-level indicators;

[0063] If at least one of the contributions of the peer indicators is greater than 0, the root cause indicator is selected according to the magnitude of the contribution.

[0064] If the contribution of all peer indicators is less than 0, obtain the similarity between the peer indicator and the next higher level indicator, and determine the indicator with the highest similarity as the root cause indicator.

[0065] Optionally, the processor is further configured to:

[0066] Obtain the correlation between the peer-level indicator and the higher-level indicator;

[0067] Based on the aforementioned correlation, determine the summation relationship between the peer indicator and the higher-level indicator;

[0068] Based on the summation relationship, the contribution is calculated using actual and predicted data.

[0069] Optionally, the processor is further configured to:

[0070] The contribution values ​​are selected sequentially from largest to smallest, and the sum of the selected contribution values ​​is calculated.

[0071] If the sum is greater than or equal to a preset threshold, the indicator corresponding to the selected contribution is determined as the root cause indicator.

[0072] Optionally, the processor is further configured to:

[0073] The similarity between the same-level index and the previous-level index is calculated using the cosine similarity formula.

[0074] Optionally, when the peer indicator is a non-first-level subordinate indicator of the first target indicator, the next-level indicator of the peer indicator is the root cause indicator.

[0075] Optionally, the processor is further configured to:

[0076] In determining the root cause index and the superior index of the root cause index, the data source is a second target index of different elements, and the different elements all belong to the target dimension.

[0077] Calculate the contribution of each element's data to the second target indicator;

[0078] If at least one of the contribution values ​​of the data elements is greater than 0, the root cause element is selected according to the magnitude of the contribution.

[0079] If the contribution of each element data is less than 0, the similarity between each element data and the second target indicator is obtained, and the element with the highest similarity is determined as the root cause element.

[0080] Optionally, the target dimension is a geographical region, and the different elements correspond to different regions.

[0081] Optionally, the processor is further configured to:

[0082] If an anomaly is detected in the data of the primary target indicator, obtain the subordinate indicators of the primary target indicator.

[0083] To achieve the above objectives, embodiments of the present invention provide an indicator root cause analysis device, including a transceiver, a processor, a memory, and a program or instructions stored in the memory and executable on the processor; when the processor executes the program or instructions, it implements the indicator root cause analysis method as described above.

[0084] To achieve the above objectives, embodiments of the present invention provide a readable storage medium having a program or instructions stored thereon, which, when executed by a processor, implement the steps in the indicator root cause analysis method as described above.

[0085] The beneficial effects of the above-described technical solution of the present invention are as follows:

[0086] The method of this invention, for the data to be analyzed, obtains the lower-level indicators of its root indicators, and further determines the root cause indicators among the lower-level indicators based on the contribution and / or similarity of the lower-level indicators, thus completing the root cause analysis of the data to be analyzed. Compared with manual analysis, it has a more efficient processing. Attached Figure Description

[0087] Figure 1 This is one of the flowcharts of the root cause analysis method for indicators according to an embodiment of the present invention;

[0088] Figure 2 A diagram illustrating the relationship between telecommunications service revenue indicators;

[0089] Figure 3 This is the second flowchart of the root cause analysis method for indicators according to an embodiment of the present invention;

[0090] Figure 4 This is the third flowchart of the root cause analysis method for indicators in this embodiment of the invention;

[0091] Figure 5 This is a structural diagram of the root cause analysis device according to an embodiment of the present invention;

[0092] Figure 6 This is a structural diagram of the root cause analysis device according to an embodiment of the present invention;

[0093] Figure 7 This is a structural diagram of an indicator root cause analysis device according to another embodiment of the present invention. Detailed Implementation

[0094] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0095] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of the invention. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0096] In various embodiments of the present invention, it should be understood that the sequence number of each process described below does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0097] In addition, the terms "system" and "network" are often used interchangeably in this article.

[0098] In the embodiments provided in this application, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information.

[0099] like Figure 1 As shown, an embodiment of the present invention provides a method for root cause analysis of indicators, comprising:

[0100] Step 101: Obtain the subordinate indicators of the first target indicator of the data to be analyzed, where the first target indicator is the root indicator.

[0101] Step 102: Determine the root cause indicators among the lower-level indicators based on their contribution and / or similarity.

[0102] Here, the root indicator is an indicator with subordinate indicators. Contribution indicates the degree to which a subordinate indicator affects the data of its superior indicator, while similarity indicates the degree to which a subordinate indicator is similar to its superior indicator.

[0103] Thus, the method of this invention, for the data to be analyzed, can obtain the lower-level indicators of its root indicators, and further determine the root cause indicators among the lower-level indicators based on the contribution and / or similarity of the lower-level indicators, thereby completing the root cause analysis of the data to be analyzed. Compared with manual analysis, it has a more efficient processing.

[0104] It should be understood that in this embodiment, since the root indicator may be a lower-level indicator than the top-level indicator, the lower-level indicators obtained in the corresponding data to be analyzed include one or more levels of indicators.

[0105] Taking the analysis of telecommunications service revenue data in financial management as an example, combined with, for instance, Figure 2 The relationships between the indicators shown indicate that both telecommunications revenue and home broadband are root indicators. Therefore, following step 101, the obtained lower-level indicators include: voice call revenue (mobile voice), other voice services, mobile internet access, telecommunications service revenue, home broadband, smart TVs, dedicated corporate lines, cloud computing, big data, internet data centers (IDC), information and communication technology (ICT), other data services, and others. The next-level indicators for home broadband are the basic home broadband average revenue per user (ARPU) and the average number of home broadband users. At this point, the lower-level indicator comprises two levels of indicators.

[0106] Thus, in this embodiment, optionally, as... Figure 3 As shown, the lower-level indicators include one or more levels of indicators, and corresponding to the existence of different levels of indicators, step 102 includes:

[0107] Step 301: Calculate the contribution of the lower-level indicators to the higher-level indicators;

[0108] Step 302: If at least one of the contributions of the peer indicators is greater than 0, select the root cause indicator according to the magnitude of the contribution.

[0109] Step 303: When all the contribution values ​​of the peer indicators are less than 0, obtain the similarity between the peer indicators and the higher-level indicators, and determine the indicator with the highest similarity as the root cause indicator.

[0110] Thus, for the lower-level indicators obtained in step 101, processing is performed on the peer indicators to calculate their contribution to the higher-level indicator, thereby allowing the selection of the root cause indicator based on the magnitude of the contribution. However, considering the possibility of contribution calculation failures, based on the calculated contribution, if at least one peer indicator has a contribution greater than 0, the root cause indicator is selected according to the magnitude of the contribution; if all peer indicators have a contribution less than 0, the similarity between the peer indicator and the higher-level indicator is obtained, and the indicator with the highest similarity is determined as the root cause indicator.

[0111] Here, if at least one of the contributions of the same level indicators is greater than 0, the root cause indicator can be selected simply by calculating the contribution. However, if all the contributions of the same level indicators are less than 0 (contribution calculation fails), the root cause indicator needs to be determined further based on the similarity between the same level indicator and the higher level indicator.

[0112] Continuing the previous example, for the same-level indicators—voice call revenue (mobile voice), other voice services, mobile internet, communication service revenue, home broadband, MP3 players, corporate leased lines, cloud computing, big data, IDC, ICT, data services, and others—calculate the contribution of each indicator. If at least one of the calculated contribution values ​​is greater than 0, the root cause indicator can be selected based on the magnitude of the contribution. If all calculated contribution values ​​are less than 0, the similarity between each indicator and communication service revenue needs to be obtained, and the indicator with the highest similarity is determined as the root cause indicator. Similarly, for the same-level indicators—basic home broadband ARPU and average number of home broadband users—calculate the contribution of these two indicators. If at least one of the contribution values ​​of these two indicators is greater than 0, the root cause indicator can be selected based on the magnitude of the contribution. If both of the contribution values ​​of these two indicators are less than 0, the similarity between these two indicators and home broadband needs to be obtained, and the indicator with the highest similarity is determined as the root cause indicator.

[0113] Of course, in this embodiment, the determination of the root cause index for the lower-level index obtained in step 101 can also be directly achieved by calculating the similarity between the same-level index and the higher-level index in the lower-level index, and the index with the highest similarity is determined as the root cause index, which will not be elaborated here.

[0114] Optionally, when the peer indicator is a non-first-level subordinate indicator of the first target indicator, the next-level indicator of the peer indicator is the root cause indicator.

[0115] In other words, when calculating the contribution of a lower-level indicator to a higher-level indicator, if the lower-level indicator is not a lower-level indicator of the root indicator, then the higher-level indicator of the lower-level indicator must be the root cause indicator. For example... Figure 2 As shown, when performing root cause analysis on telecommunications service revenue, the contributions of "basic household bandwidth ARPU" and "average number of household broadband users" to "household broadband" are only calculated if "household broadband" is a root cause indicator. Thus, in this embodiment, the contribution is calculated for lower-level indicators, which are the next level indicators after the root indicators and the next level indicators after the root cause indicators.

[0116] Optionally, selecting the root cause indicator based on the magnitude of the contribution includes:

[0117] The contribution values ​​are selected sequentially from largest to smallest, and the sum of the selected contribution values ​​is calculated.

[0118] If the sum is greater than or equal to a preset threshold, the indicator corresponding to the selected contribution is determined as the root cause indicator.

[0119] Here, the preset threshold can be defined between 0 and 1, such as setting the preset threshold to 80%. Following the above method, the calculated contribution values ​​are selected sequentially from largest to smallest and summed. The set of indicators {x1, x2, ..., xm} whose sum of contribution values ​​to the root indicator is ≥80% is then determined as the root cause indicator, where m≥1. Of course, if indicators x1, x2, ..., xm are not the lowest-level indicators (i.e., root indicators), the contribution values ​​of their subordinate indicators will be calculated, and these contribution values ​​will be used to determine the root cause indicator.

[0120] Thus, in this embodiment, the root cause indicators determined by the root indicator also include one or more levels. In other words, the root cause indicators are not limited to the next level of indicators of the root indicator.

[0121] Optionally, in this embodiment, calculating the contribution of the lower-level indicators to the higher-level indicators includes:

[0122] Obtain the correlation between the peer-level indicator and the higher-level indicator;

[0123] Based on the aforementioned correlation, determine the summation relationship between the peer indicator and the higher-level indicator;

[0124] Based on the summation relationship, the contribution is calculated using actual and predicted data.

[0125] Thus, calculating the contribution requires knowing the correlation between the lower-level indicators and the higher-level indicators, which is then converted into an additive relationship, and the calculation is completed by combining actual data and predicted data.

[0126] The predicted data is based on historical data over a preset time period.

[0127] For example, in the root cause analysis for March 2020, "Communication Service Revenue" is A. Among its 13 sub-indicators, "Voice Call Revenue (Mobile Voice)" is a1, "Other Voice Revenue" is a2, ..., and "Other" is a13. Therefore:

[0128] A = a1 + a2 + ... + a13, which represents the relationship between the revenue from telecommunications services and its next-level indicators. It can be seen that this relationship is an additive one.

[0129] Based on the historical data of "voice call revenue (mobile voice)" for the past two years (or the past n months, n≥1), the predicted value of "voice call revenue (mobile voice)" in March 2020 is f1, and similarly, the predicted value of "other voice" is f2, ..., and the predicted value of "other" is f13. Therefore, the predicted value of "community service revenue" in March 2020 is F=f1+f2+……+f13;

[0130] Thus, in March 2020, the contribution of "voice call revenue (mobile voice)" to the change in "community service revenue" was: EP1 = (a1 - f1) / (AF). Similarly, the contribution of "other voice calls" to the change in "community service revenue" was EP2 = (a2 - f2) / (AF), ..., and the contribution of "other" to the change in "community service revenue" was EP13 = (a13 - f13) / (AF).

[0131] To further facilitate contribution comparison, this embodiment normalizes the calculated contributions. For example, EP1 to EP13 in the above example are normalized as follows:

[0132] EP1=EP1 / (EP1+EP2+……+EP13)

[0133] EP2=EP2 / (EP1+EP2+……+EP13)

[0134] ...

[0135] EP13=EP13 / (EP1+EP2+...+EP13).

[0136] If the relationship between lower-level indicators and higher-level indicators is not additive, then it needs to be transformed:

[0137] For example, the relationship between "Basic Home Broadband ARPU," "Average Home Broadband Users," and the higher-level indicator "Home Broadband" is X(Home Broadband) = x1(Basic Home Broadband ARPU) × x2(Average Home Broadband Users). Then, a logarithmic transformation is performed on the historical data involved in the forecast and the data from March 2020 for X, x1, and x2. The resulting summation can be expressed as Log(X) = log(x1) + log(x2). Specifically, when x1 or x2 is negative, we need to perform a log transformation based on the sign of X. If X > 0, then: log(X) = log(|x1|) + log(|x2|); if X < 0, then -log(|X|) = (-log(|x1|)) + (-log(|x2|)). Then, referring to the calculation method for the contribution of the next-level indicator of "Communication Service Revenue," we can calculate the contribution of "Basic Home Broadband ARPU" and "Average Home Broadband Users" to the change in "Home Broadband."

[0138] Furthermore, due to the complexity and diversity of data changes, and the fact that contribution calculation requires prediction of indicator values, root cause analysis based solely on contribution can be ineffective. Therefore, as mentioned above, in this embodiment, when all contributions of lower-level indicators are less than 0, the similarity between the lower-level indicator and the higher-level indicator is obtained, and the indicator with the highest similarity is determined as the root cause indicator.

[0139] Optionally, obtaining the similarity between the peer indicator and the higher-level indicator includes:

[0140] The similarity between the same-level index and the previous-level index is calculated using the cosine similarity formula.

[0141] Cosine similarity is a measure of the difference between two vectors in a vector space, using the cosine of the angle between them. The closer the cosine value is to 1, the closer the angle is to 0 degrees, meaning the two vectors are more similar.

[0142] Thus, the cosine similarity formula is: Similarity Among them, P i Q represents the i-th data point of the current indicator within the same level of indicators. i This represents the i-th data point of the previous level indicator of the current indicator, and k represents the total number of data points.

[0143] For example, regarding the relationship between the indicators: Return on Equity = Net Profit Margin × Asset Turnover × Equity Multiplier, the calculation shows that in June 2019, the contributions of "Net Profit Margin", "Asset Turnover", and "Equity Multiplier" to "Return on Equity" were all less than 0, as shown in the table below:

[0144]

[0145] Therefore, the similarity between "net profit margin", "asset turnover", and "equity multiplier" and "return on equity" will be calculated using the cosine similarity formula to determine the root cause indicators. The process is as follows:

[0146] From January 2018 to June 2019, the return on equity (ROE) was B1, B2, ..., B18; the net profit margin was C1, C2, ..., C18; the asset turnover was D1, D2, ..., D18; and the equity multiplier was E1, E2, ..., E18. Therefore, the similarity between the net profit margin and the ROE is:

[0147]

[0148] Similarly, the similarity between asset turnover, equity multiplier, and return on equity can be calculated as similarity2 and similarity3, respectively. Assuming that the three similarities are arranged from largest to smallest as similarity1≥simility2≥simility3, then net profit margin can be considered as the main factor causing changes in return on equity, i.e., "net profit margin" is the root cause indicator.

[0149] Furthermore, considering other elements that affect the data, in this embodiment, optionally, such as Figure 4As shown, after step 102, the following steps are also included:

[0150] Step 401: Determine the second target indicators whose data sources are different elements among the root cause indicators and the parent indicators of the root cause indicators, where each of the different elements belongs to the target dimension.

[0151] Step 402: Calculate the contribution of each element's data to the second target indicator;

[0152] Step 403: If at least one of the contribution values ​​of the element data is greater than 0, select the root cause element according to the magnitude of the contribution value.

[0153] Step 404: When the contribution of each element data is less than 0, obtain the similarity between each element data and the second target indicator, and determine the element with the highest similarity as the root cause element.

[0154] Here, the second target indicator is determined based on the target dimension, and the data for this second target indicator originates from different elements belonging to that target dimension. The root cause element can be selected based on the magnitude of the contribution of each element's data to this second target indicator. However, considering the possibility of contribution calculation failures, the root cause indicator will be selected based on the following conditions: if at least one element's contribution is greater than 0, the root cause indicator will be selected; if all element's contributions are less than 0, the similarity between each element and the second target indicator will be obtained, and the element with the highest similarity will be determined as the root cause element.

[0155] Of course, the way to calculate the contribution of each element's data to the second target indicator is similar to the way to calculate the contribution of the same level indicator to the previous level indicator.

[0156] Optionally, the target dimension is a geographical region, and the different elements correspond to different regions.

[0157] Specifically, different elements are based on provinces, but they can also be based on cities, districts, or defined regional ranges.

[0158] For example, in the root cause analysis of "Communication Service Revenue," where different elements are divided by province, for the root cause indicator "Basic Household Broadband ARPU," since the national basic household broadband ARPU ≠ the sum of the basic household broadband ARPUs of each province, it is clear that "Basic Household Broadband ARPU" is not the second target indicator. Therefore, its superior indicator "Household Broadband" will be used, making "Household Broadband" the second target indicator. Then, the contribution of each province's household broadband revenue to the national household broadband revenue in March 2020 can be calculated. If at least one of the contributions from each province is greater than 0, the contributions are selected sequentially from largest to smallest, and the sum of the selected contributions is calculated. If the sum is greater than or equal to a preset element threshold, the element corresponding to the selected contribution is determined as the root cause element, i.e., the root cause province's household broadband revenue. If all the contributions from each province's household broadband revenue are less than 0, similarity needs to be calculated to determine the root cause element. Here, the preset element threshold can be the same as or different from the preset threshold.

[0159] Of course, the second target indicator, as a superior indicator of the root cause indicator, is not limited to the direct superior indicator of the root cause indicator, but also includes the peer indicator of its direct superior indicator. For example, for the indicator "basic household broadband ARPU" in the example above, the second target indicator, in addition to the direct superior indicator "household broadband", also includes the peer indicator "mobile internet access", calculating the contribution of each province's mobile internet access revenue to the national mobile internet access in March 2020.

[0160] For example, for the root cause indicator set {x1, x2, ..., xm} determined above, after determining the set of second target indicators {y1, y2, ..., yn} for itself or its superior indicators, each indicator will be expanded according to different elements of the target dimension, and the contribution of each element's data to the second target indicator will be calculated respectively.

[0161] Optionally, step 101 includes:

[0162] If an anomaly is detected in the data of the primary target indicator, obtain the subordinate indicators of the primary target indicator.

[0163] Thus, in this embodiment, when a data anomaly occurs in the first target indicator, the lower-level indicators of the first target indicator can be obtained for root cause analysis.

[0164] For example, Figure 2When the "Communication Service Revenue" data shows a sudden change / anomaly at a certain point in time (e.g., March 2020), root cause analysis will be performed according to the above method. Specifically, it automatically identifies the next-level indicator (e.g., voice call revenue, other voice services, etc.) that has a significant impact on the change of "Communication Service Revenue" (e.g., the contribution of the indicator set ≥80%, or any value defined by the user between 0 and 1) according to the hierarchical relationship of the indicators, and obtains the root cause indicators and root cause elements. Assuming that the indicator set whose sum of contribution to "Communication Service Revenue" is ≥80% includes "Home Broadband" and "Mobile Internet", then the contribution of "Basic Home Broadband ARPU" and "Average Number of Home Broadband Users" to "Home Broadband" will be calculated again; while "Mobile Internet" is already the lowest level indicator in the indicator relationship tree, and does not need to be analyzed again according to the indicator dimension. If the indicator set whose sum of contribution to "Home Broadband" is ≥80% is "Basic Home Broadband ARPU". Because this indicator is a comprehensive analytical indicator, the national basic household broadband ARPU is not the sum of the basic household broadband ARPUs of each province. Therefore, this indicator cannot be expanded by province. Thus, it is necessary to conduct root cause analysis again for "household broadband" and other indicators at the same level by province. For example, if the national household broadband indicator value equals the sum of the household broadband indicator values ​​of each province, then the contribution of each province's household broadband revenue to the national household broadband revenue in March 2020 should be calculated separately. Similarly, the contribution of each province's mobile internet revenue to the national mobile internet revenue in March 2020 should be calculated to obtain the root cause elements. Of course, if the contribution calculation fails, then the cosine similarity method should be used to calculate the trend similarity of the indicators.

[0165] In summary, by obtaining the lower-level indicators of the root indicator for the data to be analyzed, the root cause indicators among the lower-level indicators can be further determined by the contribution and / or similarity of the lower-level indicators, thus completing the root cause analysis of the data to be analyzed. Compared with manual analysis, this method is more efficient.

[0166] like Figure 5 As shown, an indicator root cause analysis device according to an embodiment of the present invention includes:

[0167] The acquisition module 510 is used to acquire the lower-level indicators of the first target indicator of the data to be analyzed, wherein the first target indicator is the root indicator.

[0168] The first processing module 520 is used to determine the root cause index among the lower-level indicators based on the contribution and / or similarity of the lower-level indicators.

[0169] Optionally, the lower-level indicators may include one or more levels of indicators;

[0170] The first processing module includes:

[0171] The first processing submodule is used to calculate the contribution of the lower-level indicators to the higher-level indicators.

[0172] The second processing submodule is used to select a root cause indicator based on the magnitude of the contribution of at least one of the peer indicators when the contribution of the peer indicators is greater than 0.

[0173] The third processing submodule is used to obtain the similarity between the peer indicator and the previous level indicator when all the contributions of the peer indicators are less than 0, and to determine the indicator with the highest similarity as the root cause indicator.

[0174] Optionally, the first processing submodule includes:

[0175] The acquisition unit is used to acquire the correlation between the peer indicator and the higher-level indicator;

[0176] The first determining unit is used to determine the summation relationship between the peer indicator and the higher-level indicator based on the correlation relationship;

[0177] The first processing unit is used to calculate the contribution based on the summation relationship using actual data and predicted data.

[0178] Optionally, the second processing submodule includes:

[0179] The second processing unit is used to select the contribution values ​​in descending order and calculate the sum of the selected contribution values.

[0180] The second determining unit is used to determine the indicator corresponding to the selected contribution as the root cause indicator when the sum is greater than or equal to a preset threshold.

[0181] Optionally, the third processing submodule is further configured to:

[0182] The similarity between the same-level index and the previous-level index is calculated using the cosine similarity formula.

[0183] Optionally, when the peer indicator is a non-first-level subordinate indicator of the first target indicator, the next-level indicator of the peer indicator is the root cause indicator.

[0184] Optionally, the device further includes:

[0185] The second processing module is used to determine the second target indicators whose data sources are different elements among the root cause indicators and the superior indicators of the root cause indicators, and the different elements all belong to the target dimension.

[0186] The third processing module is used to calculate the contribution of each element data to the second target indicator;

[0187] The fourth processing module is used to select the root cause element based on the magnitude of the contribution of each element data when at least one of the contribution values ​​is greater than 0.

[0188] The fifth processing module is used to obtain the similarity between each element data and the second target indicator when the contribution of each element data is less than 0, and to determine the element with the highest similarity as the root cause element.

[0189] Optionally, the target dimension is a geographical region, and the different elements correspond to different regions.

[0190] Optionally, the acquisition module is further configured to:

[0191] If an anomaly is detected in the data of the primary target indicator, obtain the subordinate indicators of the primary target indicator.

[0192] This device, for the data to be analyzed, obtains the lower-level indicators of its root indicators, and then, based on the contribution and / or similarity of these lower-level indicators, further determines the root cause indicators among the lower-level indicators, thus completing the root cause analysis of the data to be analyzed. Compared with manual analysis, it has a more efficient processing capability.

[0193] It should be noted that the device uses the aforementioned root cause analysis method, and the implementation of the above method embodiments is applicable to this device and can achieve the same technical effect.

[0194] like Figure 6 As shown, an indicator root cause analysis device 600 according to an embodiment of the present invention includes a processor 610, wherein...

[0195] The processor 610 is used for:

[0196] Obtain the sub-indicators of the first target indicator of the data to be analyzed, where the first target indicator is the root indicator;

[0197] The root cause indicators among the lower-level indicators are determined based on their contribution and / or similarity.

[0198] Optionally, the lower-level indicators may include one or more levels of indicators;

[0199] The processor is also used for:

[0200] Calculate the contribution of the lower-level indicators to the higher-level indicators;

[0201] If at least one of the contributions of the peer indicators is greater than 0, the root cause indicator is selected according to the magnitude of the contribution.

[0202] If the contribution of all peer indicators is less than 0, obtain the similarity between the peer indicator and the next higher level indicator, and determine the indicator with the highest similarity as the root cause indicator.

[0203] Optionally, the processor is further configured to:

[0204] Obtain the correlation between the peer-level indicator and the higher-level indicator;

[0205] Based on the aforementioned correlation, determine the summation relationship between the peer indicator and the higher-level indicator;

[0206] Based on the summation relationship, the contribution is calculated using actual and predicted data.

[0207] Optionally, the processor is further configured to:

[0208] The contribution values ​​are selected sequentially from largest to smallest, and the sum of the selected contribution values ​​is calculated.

[0209] If the sum is greater than or equal to a preset threshold, the indicator corresponding to the selected contribution is determined as the root cause indicator.

[0210] Optionally, the processor is further configured to:

[0211] The similarity between the same-level index and the previous-level index is calculated using the cosine similarity formula.

[0212] Optionally, when the peer indicator is a non-first-level subordinate indicator of the first target indicator, the next-level indicator of the peer indicator is the root cause indicator.

[0213] Optionally, the processor is further configured to:

[0214] In determining the root cause index and the superior index of the root cause index, the data source is a second target index of different elements, and the different elements all belong to the target dimension.

[0215] Calculate the contribution of each element's data to the second target indicator;

[0216] If at least one of the contribution values ​​of the data elements is greater than 0, the root cause element is selected according to the magnitude of the contribution.

[0217] If the contribution of each element data is less than 0, the similarity between each element data and the second target indicator is obtained, and the element with the highest similarity is determined as the root cause element.

[0218] Optionally, the target dimension is a geographical region, and the different elements correspond to different regions.

[0219] Optionally, the processor is further configured to:

[0220] If an anomaly is detected in the data of the primary target indicator, obtain the subordinate indicators of the primary target indicator.

[0221] The aforementioned root cause analysis device also includes a transceiver 620 for receiving and sending data under the control of the processor 610.

[0222] The device in this embodiment, for the data to be analyzed, can obtain the lower-level indicators of its root indicators, and further determine the root cause indicators among the lower-level indicators based on the contribution and / or similarity of the lower-level indicators, thus completing the root cause analysis of the data to be analyzed. Compared with manual analysis, it has a more efficient processing capability.

[0223] Another embodiment of the root cause analysis device of the present invention, such as Figure 7 As shown, it includes a transceiver 710, a processor 700, a memory 720, and a program or instructions stored in the memory 720 and executable on the processor 700; when the processor 700 executes the program or instructions, it implements the above-mentioned root cause analysis method.

[0224] The transceiver 710 is used to receive and send data under the control of the processor 700.

[0225] Among them, Figure 7 In this context, the bus architecture may include any number of interconnected buses and bridges, specifically linking various circuits together, represented by one or more processors (processor 700) and memory (memory 720). The bus architecture may also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 710 may be multiple elements, including transmitters and receivers, providing a unit for communicating with various other devices over a transmission medium. The processor 700 is responsible for managing the bus architecture and general processing, and the memory 720 may store data used by the processor 700 during operation.

[0226] An embodiment of the present invention provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the steps in the root cause analysis method of the indicators as described above and achieve the same technical effect. To avoid repetition, further details are omitted here.

[0227] The processor mentioned above is the processor in the root cause analysis device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0228] It should be further noted that many of the functional components described in this specification are referred to as modules in order to emphasize the independence of their implementation.

[0229] In this embodiment of the invention, the module can be implemented in software so that it can be executed by various types of processors. For example, an identified executable code module may include one or more physical or logical blocks of computer instructions, which may be constructed as objects, procedures, or functions. Nevertheless, the executable code of the identified module does not need to be physically located together, but may include different instructions stored in different bits, which, when logically combined, constitute the module and achieve the module's intended purpose.

[0230] In practice, an executable code module can be a single instruction or many instructions, and can even be distributed across multiple different code segments, different programs, and across multiple memory devices. Similarly, operational data can be identified within the module and can be implemented in any suitable form and organized within any suitable type of data structure. This operational data can be collected as a single dataset or distributed across different locations (including different storage devices), and can exist, at least in part, solely as electronic signals within the system or network.

[0231] When a module can be implemented using software, considering the current level of hardware technology, modules that can be implemented in software can be implemented using hardware circuits by those skilled in the art to achieve the corresponding functions, without considering cost. These hardware circuits include conventional very-large-scale integrated circuits (VLSI) or gate arrays, as well as existing semiconductors such as logic chips and transistors, or other discrete components. Modules can also be implemented using programmable hardware devices, such as field-programmable gate arrays, programmable array logic, and programmable logic devices.

[0232] The exemplary embodiments described above are with reference to the accompanying drawings. Many different forms and embodiments are feasible without departing from the spirit and teachings of the invention. Therefore, the invention should not be construed as limiting the exemplary embodiments set forth herein. Rather, these exemplary embodiments are provided to make the invention complete and convey the scope of the invention to those skilled in the art. In these drawings, component dimensions and relative dimensions may be exaggerated for clarity. The terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. As used herein, unless clearly indicated otherwise, the singular forms “a,” “an,” and “the” are intended to include all such forms. It will be further understood that the terms “comprising” and / or “including”, when used in this specification, indicate the presence of the stated features, integers, steps, operations, components, and / or elements, but do not exclude the presence or addition of one or more other features, integers, steps, operations, components, and / or groups thereof. Unless otherwise indicated, when stated, a range of values ​​includes the upper and lower limits of the range and any subranges in between.

[0233] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method of indicator root cause analysis, characterized by, The method comprises the following steps: acquiring sub-indicators of a first target indicator of data to be analyzed, the first target indicator being a root indicator; determining a root cause indicator among the sub-indicators according to contribution degrees and / or similarities of the sub-indicators; after the determination of the root cause indicator among the sub-indicators, further comprising the steps of: determining a second target indicator among the root cause indicator and a superior indicator of the root cause indicator, the data of which is derived from different elements, the different elements all belonging to a target dimension; calculating contribution degrees of each element data to the second target indicator; in a case where at least one of the contribution degrees of the element data is greater than 0, selecting a root cause element according to the size of the contribution degrees; in a case where all the contribution degrees of the element data are less than 0, acquiring similarities of the element data and the second target indicator, and determining an element with the greatest similarity as the root cause element; wherein the target dimension is a geographical region, the different elements correspond to different regions, the different elements are divided based on provinces, and the second target indicator comprises a household broadband; the sub-indicators comprise multi-level indicators; the determination of the root cause indicator among the sub-indicators according to the contribution degrees and / or similarities of the sub-indicators comprises the steps of: calculating contribution degrees of same-level indicators in the sub-indicators to a superior indicator; in a case where at least one of the contribution degrees of the same-level indicators is greater than 0, selecting a root cause indicator according to the size of the contribution degrees; in a case where all the contribution degrees of the same-level indicators are less than 0, acquiring similarities of the same-level indicators and the superior indicator, and determining an indicator with the greatest similarity as the root cause indicator.

2. The method of claim 1, wherein, the calculation of the contribution degrees of the same-level indicators in the sub-indicators to the superior indicator comprises the steps of: acquiring an association relationship between the same-level indicators and the superior indicator; determining an additive relationship between the same-level indicators and the superior indicator according to the association relationship; calculating the contribution degrees through actual data and predicted data according to the additive relationship.

3. The method of claim 1, wherein, the selection of the root cause indicator according to the size of the contribution degrees comprises the steps of: selecting the contribution degrees in turn from large to small, and calculating a sum of the selected contribution degrees; in a case where the sum is greater than or equal to a preset threshold, determining an indicator corresponding to the selected contribution degrees as the root cause indicator.

4. The method of claim 1, wherein, the acquisition of the similarities of the same-level indicators and the superior indicator comprises the steps of: calculating the similarities of the same-level indicators and the superior indicator through a cosine similarity formula.

5. The method of claim 1, wherein, in a case where the same-level indicators are non-first-level sub-indicators of the first target indicator, the superior indicator of the same-level indicators is the root cause indicator.

6. The method of claim 1, wherein, the acquisition of the sub-indicators of the first target indicator of the data to be analyzed comprises the steps of: in a case where data abnormality of a first target indicator is monitored, acquiring sub-indicators of the first target indicator.

7. An indicator root cause analysis apparatus characterized by comprising: The method comprises the following steps: an acquisition module is configured to acquire sub-indicators of a first target indicator of data to be analyzed, the first target indicator being a root indicator; a first processing module is configured to determine a root cause indicator among the sub-indicators according to contribution degrees and / or similarities of the sub-indicators; The second processing module is configured to determine a second target indicator in which data in the root cause indicator and the superior indicator of the root cause indicator originates from different elements, and the different elements belong to a target dimension; The third processing module is configured to calculate a contribution degree of each element data to the second target indicator; The fourth processing module is configured to, in a case where at least one of the contribution degrees of the element data is greater than 0, select a root cause element according to the size of the contribution degree; The fifth processing module is configured to, in a case where all of the contribution degrees of the element data are less than 0, acquire a similarity between the element data and the second target indicator, and determine an element with the maximum similarity as the root cause element; The target dimension is a geographical region, the different elements correspond to different regions, the different elements are divided based on provinces, and the second target indicator includes a household broadband. The lower-level indicators include multi-level indicators. The first processing module includes: A first processing submodule is configured to calculate a contribution degree of a same-level indicator in the lower-level indicators to a superior indicator; A second processing submodule is configured to, in a case where at least one of the contribution degrees of the same-level indicators is greater than 0, select a root cause indicator according to the size of the contribution degree; A third processing submodule is configured to, in a case where all of the contribution degrees of the same-level indicators are less than 0, acquire a similarity between the same-level indicators and the superior indicator, and determine an indicator with the maximum similarity as the root cause indicator.

8. An indicator root cause analysis device characterized by, The processor is configured to: acquire lower-level indicators of a first target indicator of to-be-analyzed data, the first target indicator being a root indicator; determine a root cause indicator in the lower-level indicators according to a contribution degree and / or a similarity of the lower-level indicators; after the root cause indicator in the lower-level indicators is determined, the method further includes: determining a second target indicator in which data in the root cause indicator and a superior indicator of the root cause indicator originates from different elements, and the different elements belong to a target dimension; calculating a contribution degree of each element data to the second target indicator; in a case where at least one of the contribution degrees of the element data is greater than 0, selecting a root cause element according to the size of the contribution degree; in a case where all of the contribution degrees of the element data are less than 0, acquiring a similarity between the element data and the second target indicator, and determining an element with the maximum similarity as the root cause element; The target dimension is a geographical region, the different elements correspond to different regions, the different elements are divided based on provinces, and the second target indicator includes a household broadband. The lower-level indicators include multi-level indicators. The determination of the root cause indicator in the lower-level indicators according to the contribution degree and / or the similarity of the lower-level indicators includes: calculating a contribution degree of a same-level indicator in the lower-level indicators to a superior indicator; in a case where at least one of the contribution degrees of the same-level indicators is greater than 0, selecting a root cause indicator according to the size of the contribution degree; in a case where all of the contribution degrees of the same-level indicators are less than 0, acquiring a similarity between the same-level indicators and the superior indicator, and determining an indicator with the maximum similarity as the root cause indicator. ​ 9. An indicator root cause analysis device comprising: A transceiver, a processor, a memory, and a program or instructions stored on the memory and executable on the processor; wherein the processor implements the method for index root cause analysis according to any one of claims 1-6 when executing the program or instructions.

10. A readable storage medium, on which a program or instructions are stored, characterized in that, The program or instructions, when executed by the processor, implement the steps in the method for index root cause analysis according to any one of claims 1-6.

Images