Order business abnormity attribution method and device, electronic equipment and medium

Abstract

The invention provides an order business exception attribution method and device, electronic equipment and a medium, and relates to the technical field of data processing, and the method comprises the steps: determining at least one correlation index corresponding to an exception index; combining the association indexes with preset dimensions to obtain a plurality of sets; determining a first contribution value of the association index in the set to the change of the abnormal index in a preset time period and a second contribution value of the preset dimension in the set to the change of the abnormal index in the preset time period; if the first contribution value of the associated index in the set meets a first preset condition, determining that the associated index in the set is a factor of order business abnormity; and if the second contribution value of the preset dimension in the set satisfies a second preset condition, determining that the preset dimension in the set is a factor of order business abnormity. According to the embodiment of the invention, the association index having the causal relationship with the abnormal index can be found according to the mathematical relationship of the abnormal index, so that the key factor of the business abnormality is found, and the attribution accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular, to a method, device, electronic device and medium for abnormal attribution of order services. Background Art

[0002] In the field of online car-hailing, the operation of online car-hailing is monitored, and order-related metrics are monitored, such as the number of completed orders, the number of pickups, the number of calls, the number of bubbles, etc. When an abnormality in an order-related metric, such as the number of pickups, is monitored, it is determined that the operation situation is abnormal. However, the abnormal indicator of the pickup volume may be caused by other indicators with an indirect causal relationship. If the business abnormality is only attributed to the monitored indicator abnormality, it may lead to inaccurate attribution. Summary of the Invention

[0003] The present invention provides a method, device, electronic device and medium for abnormal attribution of order services, which can find the associated indicators that have a causal relationship with the abnormal indicator according to the mathematical relationship of the abnormal indicators, so as to find the key factors of the business abnormality and improve the accuracy of attribution.

[0004] In a first aspect, an embodiment of the present invention provides a method for abnormal attribution of order services, including:

[0005] Determine the abnormal indicator of the order and multiple preset dimensions;

[0006] According to the mathematical relationship between multiple indicators, determine at least one associated indicator corresponding to the abnormal indicator; and combine at least one associated indicator and multiple preset dimensions to obtain multiple sets;

[0007] According to the current values of the associated indicators and preset dimensions that meet the multiple sets, determine the first contribution value of the associated indicators in at least one set to the change of the abnormal indicator within a preset time period and the second contribution value of the preset dimensions in at least one set to the change of the abnormal indicator within a preset time period;

[0008] Judge whether the first contribution value of the associated indicators in at least one set meets a first preset condition, and judge whether the second contribution value of the preset dimensions in at least one set meets a second preset condition;

[0009] If the first contribution value of the associated indicators in a set meets the first preset condition, determine that the associated indicators in the set are the factors for the abnormal order service; if the second contribution value of the preset dimensions in a set meets the second preset condition, determine that the preset dimensions in the set are the factors for the abnormal order service.

[0010] The above method can determine at least one associated indicator corresponding to an abnormal indicator through the mathematical relationship between multiple indicators, combine the associated indicators with multiple preset dimensions to obtain multiple sets, calculate the contribution value of the associated indicator to the change of the abnormal indicator and the contribution value of the preset dimension to the change of the abnormal indicator in the form of a set, determine whether the contribution value of the associated indicator meets the first preset condition, and whether the contribution value of the preset dimension meets the second preset condition, and use the associated indicators and / or preset dimensions that meet the preset conditions as the factors for order business anomalies, improving the accuracy of attribution.

[0011] In a possible implementation manner, determining whether the first contribution value of the associated indicator in at least one set meets the first preset condition includes:

[0012] According to the hierarchical relationship of the associated indicators, layer by layer, determine whether the first contribution value of the associated indicator in at least one set meets the first preset condition until the first contribution value of the associated indicator in a set meets the first preset condition; wherein, the hierarchical relationship of the associated indicators is determined according to the direct or indirect mathematical relationship between the associated indicators and the abnormal indicator.

[0013] In a possible implementation manner, determining whether the contribution value of the preset dimension in at least one set meets the second preset condition includes:

[0014] According to the number of preset dimensions, one by one, determine whether the contribution value of the preset dimension in at least one set meets the second preset condition until the contribution value of the preset dimension in a set meets the second preset condition.

[0015] In a possible implementation manner:

[0016] If the types of multiple preset dimensions are the same as those of the fixed dimension, each set includes multiple preset dimensions and one associated indicator; wherein, the fixed dimension is a preset dimension specified by the user;

[0017] If the types of multiple preset dimensions are different from those of the fixed dimension, each set includes the fixed dimension, at least one remaining dimension and one associated indicator; wherein, the remaining dimension is a preset dimension excluding the type of the fixed dimension among the multiple preset dimensions.

[0018] In a possible implementation manner:

[0019] If the types of multiple preset dimensions are the same as those of the fixed dimension, the first preset condition is that the ratio of the first contribution value of the associated indicator to the abnormal indicator is less than the first preset value;

[0020] If the types of multiple preset dimensions are different from those of the fixed dimension, and the associated indicators in the set are non-rate indicators, then the first preset condition is that the parameter of the associated indicator in the first set exceeds a second preset value; and the ratio between the target first contribution value and the sum of the first contribution values of the associated indicators corresponding to the target first contribution value exceeds a third preset value; and the ratio of the first contribution value to the abnormal indicator is less than a first preset value; wherein, the first set is a set of preset dimensions of the same type and with the same associated indicator; the parameter of the associated indicator in the first set is a parameter for measuring the gap between the first contribution values of the associated indicators in multiple first sets; the target first contribution value is the maximum value in the largest cluster obtained after clustering the first contribution values of the associated indicators in multiple first sets.

[0021] If the types of multiple preset dimensions are different from those of the fixed dimension, and the associated indicators in the set are rate indicators, then the first preset condition is that the parameter of the associated indicators in multiple first sets exceeds a second preset value; and the ratio of the target first contribution value to the abnormal indicator is less than a fourth preset value.

[0022] In a possible implementation manner:

[0023] If the types of multiple preset dimensions are the same as those of the fixed dimension, then the second preset condition is that the second contribution value is greater than zero;

[0024] If the types of multiple preset dimensions are different from those of the fixed dimension, then the second preset condition is that the parameter of the preset dimension in the second set exceeds a second preset value; and the ratio of the target second contribution value to the abnormal indicator is less than a first preset value; wherein, the second set is a set of preset dimensions of the same type and with the same hierarchical relationship of associated indicators; the preset dimension in the second set is a parameter for measuring the gap between the second contribution values of the preset dimensions in multiple second sets; the target second contribution value is the maximum value in the largest cluster obtained after clustering the second contribution values of the preset dimensions in multiple second sets.

[0025] In a possible implementation manner:

[0026] If the types of multiple preset dimensions are different from those of the fixed dimension, and the associated indicator is a rate indicator, then the first contribution value of the associated indicator is determined by the change amount of the abnormal indicator due to the change of the rate indicator and the change amount of the abnormal indicator due to the change of the non-rate indicator that has a direct mathematical relationship with the rate indicator; wherein, the change amount of the abnormal indicator due to the change of the non-rate indicator is the product of the difference between the past value of the rate indicator and the sum of the first contribution values of the rate indicators in multiple first sets, and the change amount of the non-rate indicator in a preset time period.

[0027] If the types of multiple preset dimensions are different from the type of the fixed dimension, and the associated indicator is a non-rate indicator, the first contribution value of the associated indicator is determined by the change amount of the abnormal indicator due to the change of the non-rate indicator and the change amount of the abnormal indicator due to the change of the rate indicator that has a direct mathematical relationship with the non-rate indicator; wherein, the change amount of the abnormal indicator due to the change of the non-rate indicator is the product of the change amount of the non-rate indicator in the preset time period and the rate indicator.

[0028] The second contribution value of the preset dimension included in the set is determined by the product of the first contribution value of the rate indicator and the sum of the first contribution values of the non-rate indicators in the multiple first sets, the product of the first contribution value of the non-rate indicator and the sum of the first contribution values of the rate indicators in the multiple first sets, and the product of the first contribution value of the rate indicator and the first contribution value of the non-rate indicator.

[0029] If the types of multiple preset dimensions are the same as the type of the fixed dimension, the first contribution value of the rate indicator is determined by the ratio between the current value and the past value of the rate indicator, and the ratio between the current value and the past value of the abnormal indicator; the first contribution value of the non-rate indicator is determined by the ratio between the current value and the past value of the non-rate indicator, and the ratio between the current value and the past value of the abnormal indicator.

[0030] Second, an embodiment of the present invention provides an order business anomaly attribution device, including:

[0031] A determination module, configured to determine the abnormal indicator of the order and multiple preset dimensions;

[0032] A combination module, configured to determine at least one associated indicator corresponding to the abnormal indicator according to the mathematical relationship between multiple indicators; and combine at least one associated indicator and multiple preset dimensions to obtain multiple sets;

[0033] A calculation module, configured to determine the first contribution value of the associated indicator in at least one set to the change of the abnormal indicator in the preset time period and the second contribution value of the preset dimension in at least one set to the change of the abnormal indicator in the preset time period according to the current values of the associated indicator and the preset dimension that meet the multiple sets;

[0034] A judgment module, configured to judge whether the first contribution value of the associated indicator in at least one set meets the first preset condition, and judge whether the second contribution value of the preset dimension in at least one set meets the second preset condition;

[0035] An attribution module, configured to determine, if a first contribution value of an associated metric in a set meets a first preset condition, that the associated metric in the set is a factor for abnormal order operations; and determine, if a second contribution value of a preset dimension in a set meets a second preset condition, that the preset dimension in the set is a factor for abnormal order operations.

[0036] Thirdly, an embodiment of the present invention provides an electronic device, including:

[0037] A processor;

[0038] The processor is configured to execute a computer program or instruction in the memory, so that the order operation abnormality attribution method as described in any one of the first aspects is executed.

[0039] Fourthly, an embodiment of the present invention provides a computer-readable storage medium, when an instruction in the storage medium is executed by a processor, enabling the processor to execute the order operation abnormality attribution method as described in any one of the first aspects.

[0040] Fifthly, an embodiment of the present invention provides a computer program product, the computer program product includes: computer program code, when the computer program code runs on a computer, enabling the computer to execute the order operation abnormality attribution method as described in any one of the first aspects above.

[0041] In addition, for the technical effects brought by any implementation manner in the second aspect to the fifth aspect, reference may be made to the technical effects brought by different implementation manners in the first aspect, which will not be elaborated herein.

[0042] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present invention. Description of the Drawings

[0043] Figure 1 A schematic diagram of user verification attribution provided by an embodiment of the present invention;

[0044] Figure 2 A flowchart of an order operation abnormality attribution method provided by an embodiment of the present invention;

[0045] Figure 3 A schematic diagram of the contribution value of the completed order volume in different cities provided by an embodiment of the present invention;

[0046] Figure 4 A flowchart of the first order operation abnormality attribution method provided by an embodiment of the present invention;

[0047] Figure 5 A flowchart of the second order operation abnormality attribution method provided by an embodiment of the present invention;

[0048] Figure 6 It is a flowchart of the third method for attributing order service anomalies provided by the embodiments of the present invention;

[0049] Figure 7 It is a flowchart of the fourth method for attributing order service anomalies provided by the embodiments of the present invention;

[0050] Figure 8 It is a schematic structural diagram of an order service anomaly attribution device provided by the embodiments of the present invention;

[0051] Figure 9 It is a schematic structural diagram of an electronic device provided by the embodiments of the present invention. Detailed implementation manners

[0052] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0053] Among them, in the description of the embodiments of the present application, unless otherwise specified, in the description of the embodiments of the present application, "a plurality" means two or more than two.

[0054] Hereinafter, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as implying or suggesting relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features.

[0055] In the field of online car-hailing, the operation of online car-hailing is monitored. When abnormal order-related indicators are monitored, it is determined that the operation is abnormal. However, the abnormal indicators may be caused by other indicators with indirect causal relationships. If the service anomaly is only attributed to the monitored abnormal indicators, it may lead to inaccurate attribution.

[0056] Based on this, the embodiments of the present invention provide a solution that can find the mathematical relationships of abnormal indicators, thereby finding associated indicators that have causal relationships with the abnormal indicators, and using the associated indicators and preset dimensions to find the key factors of service anomalies. Compared with directly attributing service anomalies to the monitored abnormal indicators, the accuracy of attribution is improved.

[0057] Glossary:

[0058] The metrics of an order refer to the parameters used to measure the operation of online car-hailing services. For example, the number of completed orders, the number of pick-ups, the number of calls, the number of bubbles, the completion rate, the pick-up rate, and the bubble rate.

[0059] The preset dimensions describe the characteristics or attributes of the business. In the embodiments of the present invention, they refer to the characteristics or attributes of the online car-hailing business. For example, city, mileage, time period, and channel.

[0060] The present invention will be described in detail below with reference to the accompanying drawings:

[0061] The embodiments of the present invention provide an application scenario in the form of Figure 1 as shown. Combining Figure 1 as shown, the embodiments of the present invention provide a monitoring page that displays monitoring metrics and the time intervals of the monitoring metrics. For example, the monitored metric is the number of completed orders, the fixed dimension is Shanghai and the time period in the preset dimensions, and the time intervals are 30 minutes, 1 hour, and 2 hours. The embodiments of the present invention show the number of completed orders in Shanghai at 30-minute intervals, and the time period is from 14:30 to 17:30. The number of completed orders is compared with the preset threshold every 30 minutes during the monitoring. The user can click on the origin corresponding to the time (the black dot in the figure), and the effect of the click is that a ring will appear outside the origin. After clicking, the comparison is made. When the number of completed orders is smaller than the preset threshold, it means that the number of completed orders is abnormal, and the exclamation mark is displayed in red. The user can click the next step for attribution. The attribution method can be to find the mathematical relationship of the abnormal metric, thereby finding the associated metric that has a causal relationship with the abnormal metric, and using the associated metric and the preset dimensions to find the key factors of the business anomaly. For example, it is found that the call volume that has a causal relationship with the number of completed orders is abnormal, and the contribution value of the call volume and the preset dimensions are displayed. The following shows the attribution method.

[0062] Combining Figure 2 as shown, the embodiments of the present invention provide a method for attributing order business anomalies, including:

[0063] S210: Determine the abnormal metrics of the order and multiple preset dimensions;

[0064] Among them, the abnormal metrics of the order can be the monitored abnormal metrics. For example, if it is monitored that the number of completed orders is abnormal, then the abnormal metric is the number of completed orders; if it is monitored that the call volume is abnormal, then the abnormal metric is the call volume.

[0065] The preset dimensions describe the characteristics or attributes of the business. The types of preset dimensions are, for example, city, mileage, time period, and channel. The multiple preset dimensions include cities such as Shanghai, Beijing, and Chengdu, mileage such as 5 km, 10 km, and 15 km, time periods such as 8:00 - 9:00 and 9:00 - 10:00, and channels such as map software A, map software B, and food delivery software A.

[0066] The embodiments of the present invention propose to determine multiple preset dimensions, that is, all dimensions of the order business. The preset dimensions include fixed dimensions, and the fixed dimensions are preset dimensions selected by the user, in the form of Figure 1 As shown, the preset dimension selected by the user, that is, the fixed dimension, is Shanghai and time period.

[0067] S220: According to the mathematical relationships between multiple metrics, determine at least one associated metric corresponding to the abnormal metric; and combine at least one associated metric and multiple preset dimensions to obtain multiple sets;

[0068] Exemplarily, when the metrics include the number of completed orders, the number of answered calls, the number of calls, the number of bubbles, the completion rate, the answer rate, and the burst rate, the mathematical relationships between these metrics are:

[0069] The number of completed orders = the number of answered calls * the completion rate;

[0070] The number of answered calls = the number of calls * the answer rate;

[0071] The number of calls = the number of bubbles * the burst rate.

[0072] When the abnormal metric is the number of completed orders, the associated metrics are the number of answered calls, the number of calls, the number of bubbles, the completion rate, the answer rate, and the burst rate. When the abnormal metric is the number of calls, the associated metrics are the number of bubbles and the burst rate.

[0073] Wherein, if the types of multiple preset dimensions are the same as those of the fixed dimension, each set includes multiple preset dimensions and one associated metric; wherein, the fixed dimension is the preset dimension specified by the user;

[0074] If the types of multiple preset dimensions are different from those of the fixed dimension, each set includes the fixed dimension, at least one remaining dimension, and one associated metric; wherein, the remaining dimension is a preset dimension in the multiple preset dimensions excluding the type of the fixed dimension.

[0075] Exemplarily, when the types of multiple preset dimensions are different from those of the fixed dimension, when combining, any one associated metric, all fixed dimensions, and at least one remaining dimension can be combined to obtain a set, where the remaining dimension is the remaining dimension in the multiple preset dimensions excluding the fixed dimension. When combining, the combination can be performed in ascending order according to the total number of dimensions. For example, first combine any one associated metric, all fixed dimensions, and one remaining dimension, and then combine any one associated metric, all fixed dimensions, and two remaining dimensions, and so on, until all the remaining dimensions are combined.

[0076] For example, if the abnormal indicator is the number of completed orders, the fixed dimensions are Shanghai and the time period from 14:30 to 17:30, then the channel and mileage are the types of the remaining dimensions. The channels include channels 1 to 10, and the mileage includes mileage 1 to mileage 10. So the remaining dimensions are channels 1 to 10, mileage 1 to mileage 10; the sets can be the answer pick-up volume, Shanghai, and the time period from 14:30 to 17:30, channel 1. The sets can be the answer pick-up volume, Shanghai, and the time period from 14:30 to 17:30, channel 2, and so on, to get 10 sets, each set including a different channel. The sets can also be the answer pick-up volume, Shanghai, and the time period from 14:30 to 17:30, mileage 1, and so on, to get 10 sets, each set including a different mileage. The sets can also be the answer pick-up volume, Shanghai, and the time period from 14:30 to 17:30, mileage 1, channel 1, and so on, and 100 sets can be obtained, each set including different mileage and different channels.

[0077] When the type of the fixed dimension is equal to the type of the preset dimension. For example, if the fixed dimensions are Shanghai, mileage 1, the time period from 14:30 to 17:30, and channel 1, then when combining, any one of the associated indicators and all the fixed dimensions are combined to obtain a set. For example, when the abnormal indicator is the number of completed orders, answer pick-up volume - Shanghai - mileage 1 - 14:30 to 17:30 - channel 1 forms a set, call volume - Shanghai - mileage 1 - 14:30 to 17:30 - channel 1 forms a set, bubble volume - Shanghai - mileage 1 - 14:30 to 17:30 - channel 1 forms a set, completion rate - Shanghai - mileage 1 - 14:30 to 17:30 - channel 1 forms a set, answer pick-up rate - Shanghai - mileage 1 - 14:30 to 17:30 - channel 1 forms a set, bubble occurrence rate - Shanghai - mileage 1 - 14:30 to 17:30 - channel 1 forms a set. A total of 6 sets are formed.

[0078] S230: Determine the first contribution value of the associated indicator in at least one set to the change of the abnormal indicator within the preset time period and the second contribution value of the preset dimension in at least one set to the change of the abnormal indicator within the preset time period according to the current values of the associated indicator and the preset dimension that meet the requirements in multiple sets;

[0079] Among them, the current values of the associated indicator and the preset dimension that meet the requirements in the set are calculated in the following way: taking the preset dimension as the attribute of the associated indicator, calculating the value of the associated indicator that meets the preset dimension, and taking this value as the current value of the set.

[0080] For example, when the set is bubble occurrence rate, Shanghai, mileage 1, 14:30 to 17:30, channel 1, taking Shanghai, mileage 1, 14:30 to 17:30, channel 1 as the attributes of the bubble occurrence rate, calculating the bubble occurrence rate that meets the above attributes, and taking the calculated bubble occurrence rate as the current value of the set.

[0081] For associated metrics, they can be divided into rate metrics and non-rate metrics. Briefly speaking, rate metrics are metrics that are rates, such as the order completion rate, the pick-up rate, and the outbreak rate; non-rate metrics are metrics that are not rates, such as the pick-up volume, the call volume, and the bubble volume.

[0082] If the types of multiple preset dimensions are different from the fixed dimension, and the associated metric is a rate metric, then the first contribution value of the associated metric is determined by the change amount of the abnormal metric due to the change of the rate metric and the change amount of the abnormal metric due to the non-rate metric that has a direct mathematical relationship with the rate metric; among them, the change amount of the abnormal metric due to the change of the non-rate metric is the product of the difference between the past value of the rate metric and the sum of the first contribution values of the rate metric in multiple first sets, and the change amount of the non-rate metric in the preset time period;

[0083] For example, calculated by formula 1:

[0084] ΔC = (C1 - C2) * B1 / A1 + (B1 - B2) * (C2 - C) / A1;

[0085] Among them, ΔC is the first contribution value of the rate metric in a set, C1 represents the current value of the rate metric in a set, and C2 represents the past value of the rate metric in a set; B1 is the current value of the non-rate metric that has a direct mathematical relationship with the rate metric, B2 is the past value of the non-rate metric that has a direct mathematical relationship with the rate metric, C is the sum of the first contribution values of the rate metric in multiple first sets, and A1 is the current value of the abnormal metric. The preset time period includes the current period and the past period. The first set is a set that contains the same type of preset dimensions and the same associated metric.

[0086] Among them, (C1 - C2) * B1 / A1 represents the proportion of the change amount (C1 - C2) of the rate metric in the preset time period multiplied by the non-rate metric (B1) that has a direct mathematical relationship with the rate metric to the actual current value (A1) of the abnormal metric, which is the change amount of the abnormal metric due to the change of the rate metric;

[0087] (B1 - B2) * (C2 - C) / A1 represents the proportion of the difference (C2 - C) between the past value of the rate metric and the total value multiplied by the change amount (B1 - B2) of the non-rate metric that has a direct mathematical relationship with the rate metric in the preset time period to the actual current value (A1) of the abnormal metric, which is the change amount of the abnormal metric due to the change of the non-rate metric.

[0088] If the types of multiple preset dimensions are different from the type of the fixed dimension, and the associated indicator is a non-rate indicator, the first contribution value of the associated indicator is determined by the change amount of the abnormal indicator due to the change of the non-rate indicator and the change amount of the abnormal indicator due to the change of the rate indicator that has a direct mathematical relationship with the non-rate indicator; wherein, the change amount of the abnormal indicator due to the change of the non-rate indicator is the product of the change amount of the non-rate indicator in the preset time period and the rate indicator.

[0089] For example, calculated by formula 2:

[0090] ΔB = (B1 - B2) * C1 / A1 + (C1 - C2) * B1 / A1;

[0091] Wherein, ΔB is the first contribution value of the non-rate indicator in a set.

[0092] Among them, (B1 - B2) * C1 / A1 represents the change amount of the non-rate indicator in the preset time period ((B1 - B2)) multiplied by the rate indicator (C1) that has a direct mathematical relationship with the non-rate indicator, which is the change amount of the abnormal indicator due to the change of the non-rate indicator, and the ratio to the actual current value (A1) of the abnormal indicator;

[0093] (C1 - C2) * B1 / A1 represents the change amount of the rate indicator in the preset time period (C1 - C2), multiplied by the non-rate indicator (B1) that has a direct mathematical relationship with the rate indicator, which is the change amount of the abnormal indicator due to the change of the rate indicator, and the ratio to the actual current value (A1) of the abnormal indicator.

[0094] The second contribution value of the preset dimension included in the set is determined by the product of the first contribution value of the rate indicator and the sum of the first contribution values of the non-rate indicators in multiple first sets, the product of the first contribution value of the non-rate indicator and the sum of the first contribution values of the rate indicators in multiple first sets, and the product of the first contribution value of the rate indicator and the first contribution value of the non-rate indicator;

[0095] For example, calculated by formula 3:

[0096] ΔA = ΔB * C + ΔC * B + ΔB * ΔC;

[0097] Wherein, ΔA is the second contribution value of the preset dimension included in the set; B is the sum of the first contribution values of the non-rate indicators in multiple first sets that have a direct mathematical relationship with the rate indicator.

[0098] ΔC * B represents the product of the first contribution value of the rate indicator (ΔC) and the sum of the first contribution values of the non-rate indicators in multiple first sets (B);

[0099] ΔB * C represents the product of the first contribution value of the non-rate indicator (ΔB) and the sum of the first contribution values of the rate indicators in multiple first sets (C);

[0100] ΔB * ΔC represents the product of the first contribution value of the rate index and the first contribution value of the non-rate index.

[0101] Exemplarily, the abnormal index A is the number of completed orders, and B and C are the number of answered calls and the completion rate respectively. The number of answered calls is a non-rate index, i.e., B, and the completion rate is a rate index, i.e., C. Taking the content introduced above as an example, the fixed dimensions are Shanghai and time period, and the remaining dimensions are mileage and channel. There are 20 sets regarding mileage, 20 sets regarding channel, and 200 sets regarding different mileage and different channels. Taking the set of completion rate - Shanghai - 14:30 - 17:30 - channel 1 as an example, calculate the completion rate of Shanghai - 14:30 - 17:30 - channel 1, and take it as the current value C1. Find the previous value of the completion rate of Shanghai - 14:30 - 17:30 - channel 1 as C2. Calculate the number of answered calls of Shanghai - 14:30 - 17:30 - channel 1, and take it as the current value B1. Find the previous value of the number of answered calls of Shanghai - 14:30 - 17:30 - channel 1 as B2. Calculate the completion rate of Shanghai - 14:30 - 17:30 - channel 1 - 10, and take it as C. Calculate the number of answered calls of Shanghai - 14:30 - 17:30 - channel 1 - 10, and take it as B. Take the number of completed orders of Shanghai - 14:30 - 17:30 as A1. According to the above values, calculate the first contribution value of the completion rate in the set of completion rate - Shanghai - 14:30 - 17:30 - channel 1 through formula 1. Similarly, calculate the first contribution value of the number of answered calls in the set of number of answered calls - Shanghai - 14:30 - 17:30 - channel 1 through formula 2. Similarly, calculate the second contribution value of the combination of Shanghai - 14:30 - 17:30 - channel 1 through formula 3.

[0102] If the types of multiple preset dimensions are the same as those of the fixed dimensions, the first contribution value of the associated index is determined by the ratio between the current value and the previous value of the rate index, and the ratio between the current value and the previous value of the abnormal index; the first contribution value of the non-rate index is determined by the ratio between the current value and the previous value of the non-rate index, and the ratio between the current value and the previous value of the abnormal index.

[0103] For example, calculated through formula 4:

[0104] ΔC = Log(C1 / C2) / Log(A1 / A2);

[0105] ΔB = Log(B1 / B2) / Log(A1 / A2);

[0106] Where A2 is the previous value of the abnormal index.

[0107] Among them, there is an index A = B * C. When there is no dimension, the value of ΔB * ΔC is relatively high. Therefore, the above decomposition method is not adopted, and the method of converting the multiplication formula to addition by using Log is adopted.

[0108] Log(C1 / C2) / Log(A1 / A2) represents the ratio of the current value to the previous value of the log function rate index, divided by the ratio of the current value to the previous value of the log function anomaly index;

[0109] Log(B1 / B2) / Log(A1 / A2) represents the ratio of the current value to the previous value of the log function non-rate index, divided by the ratio of the current value to the previous value of the log function anomaly index.

[0110] Taking the above-introduced set as an example, when the fixed dimension is Shanghai, mileage 1, 14:30 - 17:30, and channel 1, that is, the fixed dimension and the preset dimension are of the same type. When calculating the first contribution value of the answer rate in the set of answer rate - Shanghai - mileage 1 - 14:30 - 17:30 - channel 1, ΔB = Log(B1 / B2) / Log(A1 / A2) can be used for calculation. After calculating the first contribution value of the completion rate in the set of completion rate - Shanghai - mileage 1 - 14:30 - 17:30 - channel 1, ΔC = Log(C1 / C2) / Log(A1 / A2) is used for calculation.

[0111] S240: Determine whether the first contribution value of the associated index in at least one set meets the first preset condition, and determine whether the second contribution value of the preset dimension in at least one set meets the second preset condition;

[0112] In some embodiments, determining whether the first contribution value of the associated index in at least one set meets the first preset condition includes:

[0113] According to the hierarchical relationship of the associated indexes in the set, layer by layer, determine whether the first contribution value of the associated index in at least one set meets the first preset condition until the first contribution value of the associated index in a set meets the first preset condition; among them, the hierarchical relationship of the associated indexes is determined according to the direct or indirect mathematical relationship between the associated indexes and the anomaly index.

[0114] Exemplarily, the mathematical metrics of multiple metrics are: completed order volume = answered volume * completion rate; answered volume = call volume * answer rate; call volume = bubble volume * bubble occurrence rate. When the completed order volume is an abnormal metric, the answered volume and the completion rate have a direct mathematical relationship with the abnormal metric, the call volume and the answer rate have an indirect mathematical relationship with the abnormal metric, and the bubble volume and the bubble occurrence rate have an indirect mathematical relationship with the abnormal metric. That is, the answer rate and the completion rate that have a direct mathematical relationship with the abnormal metric are at the first layer, the call volume and the answer rate that have an indirect mathematical relationship separated by one layer from the abnormal metric are at the second layer, and the bubble volume and the bubble occurrence rate that have an indirect mathematical relationship separated by two layers from the abnormal metric are at the third layer.

[0115] During the processing, first determine whether the answered volume and the completion rate meet the first preset condition. If they meet, determine the attribution result and end the judgment. If they do not meet, then determine whether the call volume and the answer rate meet the first preset condition. If they meet, determine the attribution result and end the judgment. If they do not meet, then determine whether the bubble volume and the bubble occurrence rate meet the first preset condition.

[0116] In some embodiments, determining whether the contribution value of a preset dimension in at least one set meets the second preset condition includes:

[0117] According to the number of preset dimensions in the set, successively determine whether the second contribution value of the preset dimension in at least one set meets the second preset condition until the contribution value of the preset dimension in a set meets the second preset condition.

[0118] When the types of preset dimensions include 4 types, for example, the types of preset dimensions are city, mileage, time period, and channel. The number of preset dimensions in the set can be 1, 2, 3, or 4.

[0119] Exemplarily, determine whether the contribution value of one preset dimension in the set meets the second preset condition. If it meets, determine the attribution result and end the judgment. If it does not meet, then determine whether the contribution value of two preset dimensions in the set meets the second preset condition. If it meets, determine the attribution result and end the judgment. If it does not meet, then determine whether the contribution value of three preset dimensions in the set meets the second preset condition. If it meets, determine the attribution result and end the judgment. If it does not meet, then determine whether the contribution value of four preset dimensions in the set meets the second preset condition.

[0120] S250: If the first contribution value of the associated metric in a set meets the first preset condition, determine that the associated metric in the set is a factor for abnormal order business; if the second contribution value of the preset dimension in a set meets the second preset condition, determine that the preset dimension in the set is a factor for abnormal order business.

[0121] In some embodiments, if the types of multiple preset dimensions are the same as the type of fixed dimensions, the first preset condition is that the ratio of the first contribution value of the associated indicator to the abnormal indicator is less than the first preset value; illustratively, the first preset value can be 0.1, i.e. 10%.

[0122] If the types of multiple preset dimensions are different from the types of fixed dimensions, and the associated indicators in the set are non-rate indicators, then the first preset condition is that the parameter of the associated indicators in the first set exceeds the second preset value; and the ratio between the target first contribution value and the sum of the first contribution values ​​of the associated indicators corresponding to the target first contribution value exceeds the third preset value; and the ratio of the first contribution value to the abnormal indicator is less than the first preset value; wherein the first set is a set containing preset dimensions of the same type and the same associated indicators; the parameter of the associated indicators in the first set is a parameter for measuring the gap between the first contribution values ​​of the associated indicators in multiple first sets; the target first contribution value is the maximum value of the first contribution values ​​of the associated indicators in multiple first sets in the largest cluster obtained after clustering;

[0123] The parameter of the correlation indicator may be the Gini coefficient of the correlation indicator.

[0124] Gini coefficient: It is a statistical method to measure unequal distribution. In practical applications, the calculation of Gini coefficient is often used to evaluate the impurity of classification models. The closer the value is to 0, the more evenly distributed the samples are, and the smaller the difference in the proportion of each category is; while the closer the value is to 1, the more unbalanced the samples are, and most of them are concentrated in a few categories.

[0125] Calculation formula:

[0126] Example:

[0127] Combination Figure 3 As shown in the figure, taking the distribution of completed orders on a certain day as an example, the horizontal axis is the cumulative proportion of the first contribution value, and the vertical axis is the cumulative proportion of completed orders. The first contribution value of the completed orders is calculated for all sets of the same type of preset dimension containing the completed orders. For example, the preset dimension type is city, and the first contribution values ​​of the completed orders of multiple cities are composed of Figure 3 .

[0128] Gini coefficient = 1-the sum of the areas of each trapezoid. The Gini coefficient of this figure is 0.57, and the change in Guangzhou is abnormal. When the second preset value is 0.4, the Gini coefficient of this figure is 0.57, which exceeds 0.4, indicating that it is not the entire market that is changing, that is, not everyone is changing, but the samples are extremely unbalanced, and most of them are concentrated in a few categories.

[0129] Cluster the correlation indicators of all the sets for calculating the Gini coefficient to obtain multiple clusters. Find the maximum value in each cluster, compare the maximum values in each cluster, and take the cluster with the maximum value as the largest cluster. For example, obtain clusters A, B, C, and D, with the maximum value a in cluster A, the maximum value b in cluster B, the maximum value c in cluster C, and the maximum value d in cluster D. Compare the maximum values a, b, c, and d. When the value c is the largest, then cluster C is the largest cluster.

[0130] The maximum value c of cluster C is the target first contribution value. The total first contribution value of the correlation indicators corresponding to the target first contribution value is the sum of the first contribution values of the correlation indicators of all the sets for calculating the Gini coefficient.

[0131] Among them, the third preset value can be 0.5, that is, 50%.

[0132] If the types of multiple preset dimensions are different from the type of the fixed dimension, and the correlation indicator in the set is a rate indicator, then the first preset condition is that the parameter of the correlation indicator in multiple first sets exceeds the second preset value; and the proportion of the target first contribution value to the abnormal indicator is less than the fourth preset value.

[0133] Taking the fixed dimension as Shanghai and the time period from 14:30 to 17:30 as an example, for the types of the remaining dimensions of channel and mileage, when the correlation indicator is the order completion rate, the sets of preset dimensions of the same type are the set of preset dimensions including the fixed dimension and the channel, the set of preset dimensions including the fixed dimension and the mileage, and the set of preset dimensions including both the fixed dimension, the channel and the mileage. Use the first contribution value of the order completion rate in all the sets including the fixed dimension and one channel to calculate the Gini coefficient of the order completion rate, and judge whether the Gini coefficient of the order completion rate is greater than 0.4. If it is greater than 0.4, it means there is a problem with the data part in calculating the Gini coefficient of the order completion rate. Cluster the first contribution value of the order completion rate in all the sets including the fixed dimension and one channel to obtain multiple clusters, compare the clusters, find the target first contribution value, calculate the ratio between the target first contribution value and the order completion volume in the time period from 14:30 to 17:30 in Shanghai, and judge whether the ratio is greater than 0.1. If it is greater than 0.1, it means that the order completion rate is not a factor for the abnormal order completion volume in the time period from 14:30 to 17:30 in Shanghai. If it is less than 0.1, it means that the order completion rate is a factor for the abnormal order completion volume in the time period from 14:30 to 17:30 in Shanghai. Similarly, judge whether the order completion rate of the set of mileage is a factor for the abnormal order completion volume in the time period from 14:30 to 17:30 in Shanghai. Similarly, judge whether the order completion rate of the set of mileage and channel is a factor for the abnormal order completion volume in the time period from 14:30 to 17:30 in Shanghai.

[0134] In some embodiments, if the types of multiple preset dimensions are the same as those of the fixed dimension, the second preset condition is that the second contribution value is greater than zero;

[0135] If the types of multiple preset dimensions are different from those of the fixed dimension, the second preset condition is that the parameters of the preset dimensions in the second set exceed the second preset value; and the ratio of the target second contribution value to the abnormal index is less than the first preset value; where the second set is a set of preset dimensions of the same type and associated indicators with the same hierarchical relationship; the preset dimensions in the second set are parameters for measuring the gap between the second contribution values of multiple preset dimensions in the second set; the target second contribution value is the maximum value in the largest cluster obtained after clustering the second contribution values of the preset dimensions in multiple second sets.

[0136] Taking the fixed dimension as Shanghai and the time period from 14:30 to 17:30 as an example, and the channel and mileage as the types of the remaining dimensions, the sets of preset dimensions of the same type are the set of preset dimensions including the fixed dimension and the channel type, the set of preset dimensions including the fixed dimension and the mileage type, and the set of preset dimensions including both the fixed dimension, the channel, and the mileage types. Among them, all sets including the fixed dimension, one channel, and the completion rate are the first set, and all sets including the fixed dimension, one channel, and the answer rate are the first set. Using the first contribution values of the preset dimensions of the type of the fixed dimension and the channel in all sets, calculate the Gini coefficient of the preset dimensions of the types of the fixed dimension and one channel.

[0137] Judge whether the Gini coefficient is greater than 0.4. If it is greater than 0.4, it means there is a problem with the data part in calculating the Gini coefficient. Cluster the first contribution values of all sets including the fixed dimension and one channel, that is, the first contribution values of the completion rate and the answer rate of all sets including the fixed dimension and one channel, to obtain multiple clusters. Compare the clusters to find the target first contribution value, calculate the ratio between the target first contribution value and the completion volume during the period from Shanghai - 14:30 to 17:30, and judge whether the ratio is greater than 0.1. If it is greater than 0.1, it means that the fixed dimension and one channel are not factors for the abnormal completion volume during the period from Shanghai - 14:30 to 17:30. If it is less than 0.1, it means that the fixed dimension and one channel are factors for the abnormal completion volume during the period from Shanghai - 14:30 to 17:30.

[0138] As shown above, combined with Figure 4 shown, an embodiment of the present invention provides a method for abnormal attribution of order services, including:

[0139] S410: Input an abnormal index and a fixed dimension;

[0140] S420: Determine a first - layer associated indicator that has a direct mathematical relationship with the abnormal indicator according to the mathematical relationship between multiple indicators;

[0141] S430: If the type of the fixed dimension is different from the preset dimension, then combine any one associated indicator, the fixed dimension, and a remaining dimension of the same type to obtain multiple sets;

[0142] S440: Determine the first contribution value of the associated indicator in at least one set to the change of the abnormal indicator within the preset time period according to the current values of the associated indicator and the preset dimension that meet the multiple sets;

[0143] S450: If the associated indicator is a non - rate indicator, then determine the Gini coefficient of the associated indicator according to the first contribution values of the associated indicator in all sets, and judge whether the Gini coefficient of the associated indicator is greater than 0.4; if so, execute S460; otherwise, execute S470;

[0144] S460: Cluster the first contribution values of the associated indicator in all sets to obtain multiple clusters, and judge whether the ratio of the first contribution value of the largest cluster to the total first contribution value of the associated indicator corresponding to the target first contribution value is greater than 0.5; if so, execute S480; otherwise, execute S490;

[0145] S470: Judge whether the total first contribution value of the associated indicator corresponding to the target first contribution value is greater than 0.8; if so, execute S490; otherwise, execute S491;

[0146] S490: Determine a second - layer associated indicator that has an indirect mathematical relationship with the abnormal indicator; and continue to execute starting from S420.

[0147] S480: Judge whether the ratio of the first contribution value to the abnormal indicator is greater than 0.1; if so, execute S492, otherwise, execute S493;

[0148] S491: The Gini coefficient of the associated indicator is normal;

[0149] S492: Combine any one associated indicator, the fixed dimension, and another remaining dimension of the same type to obtain multiple sets;

[0150] S493: Determine the factors for the abnormal order business of the associated indicator.

[0151] Combined with Figure 5 As shown, an embodiment of the present invention provides a method for attributing abnormal order business, including:

[0152] S510: Input an abnormal indicator and a fixed dimension;

[0153] S520: Determine the first-level associated indicators that have a direct mathematical relationship with the abnormal indicator based on the mathematical relationships among multiple indicators;

[0154] S530: If the type of the fixed dimension is different from the preset dimension, combine any one of the associated indicators, the fixed dimension, and one remaining dimension of the same type to obtain multiple sets;

[0155] S540: Determine the first contribution value of the associated indicators in at least one set to the change of the abnormal indicator within the preset time period according to the current values of the associated indicators and the preset dimension that meet the multiple sets; Among them, the steps of S500 - S540 are the same as those of S400 - S440;

[0156] S550: If the associated indicator is a rate indicator, determine the Gini coefficient of the associated indicator according to the first contribution values of the associated indicators in all sets, and determine whether the Gini coefficient of the associated indicator is greater than 0.4; If so, execute S560; Otherwise, execute S580;

[0157] S560: Cluster the first contribution values of the associated indicators in all sets to obtain multiple clusters, and determine whether the ratio of the first contribution value of the largest cluster among the multiple clusters to the abnormal indicator is greater than 0.1; If so, execute S570; Otherwise, execute S580;

[0158] S570: Determine the second-level associated indicators that have an indirect mathematical relationship with the abnormal indicator; and continue to execute starting from S530.

[0159] S580: Determine the factors for the abnormal order business of the associated indicator.

[0160] Combined Figure 6 As shown, an embodiment of the present invention provides a method for attributing abnormal order business, including:

[0161] S610: Input the abnormal indicator and the fixed dimension;

[0162] S620: Determine the first-level associated indicators that have a direct mathematical relationship with the abnormal indicator based on the mathematical relationships among multiple indicators;

[0163] S630: If the type of the fixed dimension is different from the preset dimension, combine any one of the associated indicators, the fixed dimension, and one remaining dimension of the same type to obtain multiple sets; Among them, the steps of S610 - S630 are the same as those of S510 - S530;

[0164] S640: Determine the second contribution value of the preset dimension in at least one set to the change of the abnormal indicator within the preset time period according to the current values of the associated indicators and the preset dimension that meet the multiple sets;

[0165] S650: Calculate the Gini coefficient of the preset dimension in multiple second sets, and determine whether the Gini coefficient of the preset dimension in the multiple second sets is greater than 0.4; if so, execute S660; otherwise, execute S670;

[0166] S660: Cluster the second contribution values of the associated indicators in all second sets to obtain multiple clusters, and determine whether the ratio of the second contribution value of the largest cluster among the multiple clusters to the abnormal indicator is greater than 0.1; if so, execute S680; otherwise, execute S690;

[0167] S670: There is no abnormality in the Gini coefficient of the preset dimension in the multiple second sets.

[0168] S680: Combine any one associated indicator, the fixed dimension, and another remaining dimension of the same type to obtain multiple sets; continue to execute S640;

[0169] S690: Determine the factors for which the preset dimension in the second set is an order business anomaly.

[0170] Combined Figure 7 As shown, an embodiment of the present invention provides a method for attributing order business anomalies, including:

[0171] S710: Input an abnormal indicator and a fixed dimension;

[0172] S720: Determine the first-level associated indicators that have a direct mathematical relationship with the abnormal indicator according to the mathematical relationship between multiple indicators;

[0173] Among them, the steps of S710 to S720 are the same as those of S510 to S520;

[0174] S730: If the type of the fixed dimension is the same as the preset dimension, combine any one associated indicator and the fixed dimension to obtain multiple sets;

[0175] S740: Determine the first contribution value of the associated indicator to the change of the abnormal indicator within the preset time period according to the current values of the associated indicators and the preset dimension that meet the multiple sets;

[0176] S750: Determine whether the ratio of the target first contribution value to the abnormal indicator is greater than the first preset value; if so, execute S760; otherwise, execute S770;

[0177] S760: Determine the second-level associated indicators that have an indirect mathematical relationship with the abnormal indicator; and continue to execute starting from S730.

[0178] S770: Determine the factors for which the associated indicator is an order business anomaly.

[0179] Example 1: When the user inputs the metric: the number of completed orders, dimension: Shanghai.

[0180] The algorithm will determine the Gini coefficient based on the user-input metric, the number of completed orders in Shanghai, mainly to determine whether the anomaly is caused by individual cases or by changes in the overall market. That is, if there are anomalies in the entire number of completed orders in Shanghai, if it is caused by changes in the overall market, no further drilling down will be performed, and the conclusion will be output that the anomaly is due to the impact of changes in the overall market. If the anomaly is caused by individual case changes, recursive drilling down will be performed. During the drilling-down process, the metric and dimension will be carried out simultaneously. For example: the Shanghai completed order metric is decomposed into the pick-up volume in Shanghai and the completion rate in Shanghai, and at the same time, all dimensions except 'city' are matched, such as: Shanghai - channel - pick-up volume, Shanghai - channel - completion rate, Shanghai - time period - pick-up volume, Shanghai - time period - completion rate, etc. These conditional 'factors' are brought into the clustering algorithm, and clustering is performed according to the contribution to the overall market to obtain the several clusters with the largest contribution. Suppose it is the completion rate of Shanghai - Gaode and the completion rate of Shanghai - 8 - 9 o'clock. Then the above operations are repeated for these two factors until it converges to less than the contribution threshold.

[0181] Example 2: When the user inputs the metric: the number of bubbles, dimension: Shanghai.

[0182] Suppose the metric cannot be further decomposed under the number of bubbles, but there are anomalies in the bubbles in Shanghai. At this time, the algorithm will only perform drilling down of the dimension. For example, Shanghai - channel - number of bubbles, Shanghai - mileage - number of bubbles, Shanghai - time period - number of bubbles, etc. First, the Gini coefficient is still calculated to determine whether the anomaly is caused by the overall change in the market or by changes in individual dimensions. If there are anomalies in individual dimensions, the contribution will be calculated. If there is a problem with the contribution of the channel, Shanghai - Gaode - number of bubbles, then the above operations will be performed on Shanghai Gaode.

[0183] Example 3: When the user inputs the metric: the number of completed orders, dimension: Shanghai; Gaode; 8 - 9 o'clock; long orders over 5 km.

[0184] Since the dimensions input by the user already cover all dimensions, the algorithm will only perform metric drilling down and no longer perform dimension combination. The anomaly of the completed orders of Shanghai - Gaode - 8 - 9 o'clock - long orders over 5 km will be decomposed into the pick-up number of Shanghai - Gaode - 8 - 9 o'clock - long orders over 5 km and the completion rate of Shanghai - Gaode - 8 - 9 o'clock - long orders over 5 km. At this time, the Gini measurement will no longer be performed, and only the measurement of the contribution to the overall market will be carried out. If the pick-up contribution of Shanghai - Gaode - 8 - 9 o'clock - long orders over 5 km is abnormal, then the call number of Shanghai - Gaode - 8 - 9 o'clock - long orders over 5 km and the pick-up rate problem of Shanghai - Gaode - 8 - 9 o'clock - long orders over 5 km will continue to be decomposed. Until the contribution is less than the threshold.

[0185] Such as Figure 8As shown in the figure, the present invention also provides an order business exception attribution device, including:

[0186] A determination module 810, configured to determine the exception metrics of an order and multiple preset dimensions;

[0187] A combination module 820, configured to determine at least one associated metric corresponding to the exception metric according to the mathematical relationship between multiple metrics; and combine at least one associated metric and multiple preset dimensions to obtain multiple sets;

[0188] A calculation module 830, configured to determine a first contribution value of the associated metric to the change of the exception metric within a preset time period and a second contribution value of the preset dimension to the change of the exception metric within a preset time period in at least one set according to the current values of the associated metric and the preset dimension that meet the multiple sets;

[0189] A judgment module 840, configured to judge whether the first contribution value of the associated metric in at least one set meets a first preset condition, and judge whether the second contribution value of the preset dimension in at least one set meets a second preset condition;

[0190] An attribution module 850, configured to, if the first contribution value of the associated metric in a set meets the first preset condition, determine that the associated metric in the set is a factor for the order business exception; if the second contribution value of the preset dimension in a set meets the second preset condition, determine that the preset dimension in the set is a factor for the order business exception.

[0191] Optionally, the judgment module 840 is specifically configured to:

[0192] Judging layer by layer whether the first contribution value of the associated metric in at least one set meets the first preset condition according to the hierarchical relationship of the associated metrics until the first contribution value of the associated metric in a set meets the first preset condition; wherein, the hierarchical relationship of the associated metrics is determined according to the direct or indirect mathematical relationship between the associated metric and the exception metric.

[0193] Optionally, the judgment module 840 is specifically configured to:

[0194] Judging one by one whether the contribution value of the preset dimension in at least one set meets the second preset condition according to the number of the preset dimensions until the contribution value of the preset dimension in a set meets the second preset condition.

[0195] Optionally:

[0196] If the types of multiple preset dimensions are the same as the type of the fixed dimension, each set includes multiple preset dimensions and one associated metric; wherein, the fixed dimension is a preset dimension specified by the user;

[0197] If the types of multiple preset dimensions are different from the type of the fixed dimension, each set includes the fixed dimension, at least one remaining dimension, and an associated index; wherein, the remaining dimension is a preset dimension among the multiple preset dimensions excluding the type of the fixed dimension.

[0198] Optionally:

[0199] If the types of multiple preset dimensions are the same as the type of the fixed dimension, the first preset condition is that the ratio of the first contribution value of the associated index to the abnormal index is less than the first preset value;

[0200] If the types of multiple preset dimensions are different from the type of the fixed dimension, and the associated index in the set is a non-rate index, the first preset condition is that the parameter of the associated index in the first set exceeds the second preset value; and the ratio between the target first contribution value and the sum of the first contribution values of the associated index corresponding to the target first contribution value exceeds the third preset value; and the ratio of the first contribution value to the abnormal index is less than the first preset value; wherein, the first set is a set of preset dimensions of the same type and with the same associated index; the parameter of the associated index in the first set is a parameter for measuring the gap between the first contribution values of the associated index in multiple first sets; the target first contribution value is the maximum value in the largest cluster obtained after clustering the first contribution values of the associated index in multiple first sets;

[0201] If the types of multiple preset dimensions are different from the type of the fixed dimension, and the associated index in the set is a rate index, the first preset condition is that the parameter of the associated index in multiple first sets exceeds the second preset value; and the ratio of the target first contribution value to the abnormal index is less than the fourth preset value.

[0202] Optionally:

[0203] If the types of multiple preset dimensions are the same as the type of the fixed dimension, the second preset condition is that the second contribution value is greater than zero;

[0204] If the types of multiple preset dimensions are different from the type of the fixed dimension, the second preset condition is that the parameter of the preset dimension in the second set exceeds the second preset value; and the ratio of the target second contribution value to the abnormal index is less than the first preset value; wherein, the second set is a set of preset dimensions of the same type and with the same hierarchical relationship of the associated index; the preset dimension in the second set is a parameter for measuring the gap between the second contribution values of the preset dimension in multiple second sets; the target second contribution value is the maximum value in the largest cluster obtained after clustering the second contribution values of the preset dimension in multiple second sets.

[0205] Optionally:

[0206] If the types of multiple preset dimensions are different from those of the fixed dimension, and the associated indicator is a rate indicator, the first contribution value of the associated indicator is determined by the change amount of the abnormal indicator due to the change of the rate indicator and the change amount of the abnormal indicator due to the change of a non-rate indicator that has a direct mathematical relationship with the rate indicator; wherein, the change amount of the abnormal indicator due to the change of the non-rate indicator is the product of the difference between the past value of the rate indicator and the sum of the first contribution values of the rate indicator in multiple first sets, and the change amount of the non-rate indicator in the preset time period;

[0207] If the types of multiple preset dimensions are different from those of the fixed dimension, and the associated indicator is a non-rate indicator, the first contribution value of the associated indicator is determined by the change amount of the abnormal indicator due to the change of the non-rate indicator and the change amount of the abnormal indicator due to the change of a rate indicator that has a direct mathematical relationship with the non-rate indicator; wherein, the change amount of the abnormal indicator due to the change of the non-rate indicator is the product of the change amount of the non-rate indicator in the preset time period and the rate indicator;

[0208] The second contribution value of the preset dimension included in the set is determined by the product of the first contribution value of the rate indicator and the sum of the first contribution values of the non-rate indicator in multiple first sets, the product of the first contribution value of the non-rate indicator and the sum of the first contribution values of the rate indicator in multiple first sets, and the product of the first contribution value of the rate indicator and the first contribution value of the non-rate indicator;

[0209] If the types of multiple preset dimensions are the same as those of the fixed dimension, the first contribution value of the rate indicator is determined by the ratio between the current value and the past value of the rate indicator, and the ratio between the current value and the past value of the abnormal indicator; the first contribution value of the non-rate indicator is determined by the ratio between the current value and the past value of the non-rate indicator, and the ratio between the current value and the past value of the abnormal indicator.

[0210] In addition, the order service exception attribution method and device according to the embodiments of the present invention described in combination with Figures 1 - 8 can be implemented by an electronic device.

[0211] The electronic device includes: a processor;

[0212] a memory for storing executable instructions of the processor;

[0213] Wherein, the processor is configured to execute the instructions to implement the order service exception attribution method described in any one of the above introductions.

[0214] Based on the above introduction, exemplarily, an electronic device structure is proposed Figure 9 as described above.

[0215] The electronic device may include a processor 901 and a memory 902 storing computer program instructions.

[0216] Specifically, the processor 901 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured as one or more integrated circuits implementing the embodiments of the present invention.

[0217] The memory 902 may include a mass storage for data or instructions. By way of example and not limitation, the memory 902 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In a suitable case, the memory 902 may include a removable or non-removable (or fixed) medium. In a suitable case, the memory 902 may be internal or external to the data processing device. In a specific embodiment, the memory 902 is a non-volatile solid state memory. In a specific embodiment, the memory 902 includes a read only memory (ROM). In a suitable case, the ROM may be a mask programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or a flash memory or a combination of two or more of these.

[0218] The processor 901 reads and executes the computer program instructions stored in the memory 902 to implement any one of the methods of performing tasks in the above embodiments.

[0219] In one example, the electronic device may further include a communication interface 903 and a bus 904. Among them, as Figure 9 shown, the processor 901, the memory 902, and the communication interface 903 are connected through the bus 904 and complete communication with each other.

[0220] The communication interface 903 is mainly used to implement communication between each module, device, unit, and / or device in the embodiments of the present invention.

[0221] Bus 904 includes hardware, software, or both, and couples components of an electronic device to each other. By way of example and not limitation, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a MicroChannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable bus or a combination of two or more of these. In suitable cases, bus 904 may include one or more buses. Although embodiments of the present invention describe and illustrate specific buses, the present invention contemplates any suitable bus or interconnect.

[0222] The electronic device may execute the order service exception attribution method in the embodiments of the present invention based on the received task, so as to implement the combination Figures 1 - 8 of the described order service exception attribution method and apparatus.

[0223] In addition, in combination with the electronic device in the above embodiments, embodiments of the present invention may provide a storage medium, which when the instructions in the storage medium are executed by a processor of the electronic device, enables the electronic device to execute the order service exception attribution method as described in any one of the above.

[0224] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce an apparatus for implementing the specified function in one Figure 1 flow or multiple flows and / or one Figure 1 block or multiple blocks.

[0225] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including an instruction apparatus that implements the specified function in one Figure 1 flow or multiple flows and / or one Figure 1 block or multiple blocks.

[0226] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one process or a plurality of processes and / or blocks Figure 1 one process or a plurality of processes and / or blocks Figure 1 steps for implementing the functions specified in one block or a plurality of blocks.

[0227] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made by those skilled in the art once they learn of the basic inventive concept. Therefore, the appended claims are intended to be construed to cover the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0228] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A method for attributing abnormal order business, characterized in that: include: Determine abnormal indicators and multiple preset dimensions of orders; Determine at least one associated indicator corresponding to the abnormal indicator according to a mathematical relationship between the multiple indicators; and combining at least one associated indicator and a plurality of preset dimensions to obtain a plurality of sets; Determine, according to the current period values ​​that conform to the associated indicators and preset dimensions in the multiple sets, a first contribution value of the associated indicators in at least one set to the change of the abnormal indicator within a preset time period and a second contribution value of the preset dimension in at least one set to the change of the abnormal indicator within the preset time period; Determine whether a first contribution value of an associated indicator in at least one set meets a first preset condition, and determine whether a second contribution value of a preset dimension in at least one set meets a second preset condition; If the first contribution value of an associated indicator in a set satisfies a first preset condition, the associated indicator in the set is determined to be a factor causing abnormal order business; if the second contribution value of a preset dimension in the set satisfies a second preset condition, the preset dimension in the set is determined to be a factor causing abnormal order business.

2. The order business abnormality attribution method according to claim 1 is characterized in that: Determining whether a first contribution value of an associated indicator in at least one set meets a first preset condition includes: According to the hierarchical relationship of the associated indicators, whether the first contribution value of the associated indicators in at least one set meets the first preset condition is determined layer by layer until the first contribution value of the associated indicators in the set meets the first preset condition; wherein the hierarchical relationship of the associated indicators is determined based on the direct or indirect mathematical relationship between the associated indicators and the abnormal indicators.

3. The order business abnormality attribution method according to claim 1 is characterized in that: Determining whether the contribution value of the preset dimension in at least one set meets a second preset condition includes: According to the number of preset dimensions, it is determined one by one whether the contribution value of the preset dimensions in at least one set meets the second preset condition, until the contribution value of the preset dimensions in a set meets the second preset condition.

4. The order business abnormality attribution method according to claim 1 is characterized by: If the multiple preset dimensions are of the same type as the fixed dimension, each set includes the multiple preset dimensions and an associated indicator; wherein the fixed dimension is a preset dimension specified by the user; If the types of the multiple preset dimensions are different from those of the fixed dimension, each set includes the fixed dimension, at least one remaining dimension and an associated indicator; wherein the remaining dimension is a preset dimension of the type of the multiple preset dimensions excluding the fixed dimension.

5. The order business abnormality attribution method according to claim 1 is characterized by: If the types of the multiple preset dimensions and the fixed dimension are the same, the first preset condition is that the ratio of the first contribution value of the associated indicator to the abnormal indicator is less than a first preset value; If the types of multiple preset dimensions and fixed dimensions are different, and the associated indicators in the set are non-rate indicators, then the first preset condition is that the parameter of the associated indicators in the first set exceeds the second preset value; and the ratio between the target first contribution value and the sum of the first contribution values ​​of the associated indicators corresponding to the target first contribution value exceeds the third preset value; and the ratio of the first contribution value to the abnormal indicator is less than the first preset value; wherein, the first set is a set containing preset dimensions of the same type and the same associated indicators; the parameter of the associated indicators in the first set is a parameter for measuring the gap between the first contribution values ​​of the associated indicators in multiple first sets; the target first contribution value is the maximum value of the first contribution values ​​of the associated indicators in multiple first sets in the largest cluster obtained after clustering; If the types of multiple preset dimensions and fixed dimensions are different, and the associated indicators in the set are rate indicators, then the first preset condition is that the parameters of the associated indicators in the multiple first sets exceed the second preset value; and the ratio of the target first contribution value to the abnormal indicator is less than the fourth preset value.

6. The order business abnormality attribution method according to claim 1 is characterized by: If the multiple preset dimensions are of the same type as the fixed dimension, then the second preset condition is that the second contribution value is greater than zero; If the types of multiple preset dimensions are different from those of the fixed dimension, the second preset condition is that the parameter of the preset dimension in the second set exceeds the second preset value; and the ratio of the target second contribution value to the abnormal indicator is less than the first preset value; wherein, the second set is a set of associated indicators containing preset dimensions of the same type and the same hierarchical relationship; the preset dimensions in the second set are parameters for measuring the gap between the second contribution values ​​of the preset dimensions in multiple second sets; the target second contribution value is the maximum value of the second contribution values ​​of the preset dimensions in multiple second sets in the largest cluster obtained after clustering.

7. The order business abnormality attribution method according to claim 1 is characterized by: If the types of the multiple preset dimensions and the fixed dimensions are different, and the associated indicator is a rate indicator, the first contribution value of the associated indicator is determined by the amount of change of the abnormal indicator due to the change of the rate indicator and the amount of change of the abnormal indicator due to the change of a non-rate indicator having a direct mathematical relationship with the rate indicator; wherein the amount of change of the abnormal indicator due to the change of the non-rate indicator is the product of the difference between the previous value of the rate indicator and the sum of the first contribution values ​​of the rate indicators in the multiple first sets, and the amount of change of the non-rate indicator in the preset time period; If the types of the multiple preset dimensions are different from the types of the fixed dimensions, and the associated indicator is a non-rate indicator, the first contribution value of the associated indicator is determined by the amount of change of the abnormal indicator due to the change of the non-rate indicator and the amount of change of the abnormal indicator due to the change of the rate indicator having a direct mathematical relationship with the non-rate indicator; wherein the amount of change of the abnormal indicator due to the change of the non-rate indicator is the product of the amount of change of the non-rate indicator in the preset time period and the rate indicator; The second contribution value of the preset dimension included in the set is determined by multiplying the first contribution value of the rate index by the sum of the first contribution values ​​of the non-rate indexes in the first set, the first contribution value of the non-rate index by the sum of the first contribution values ​​of the rate indexes in the first set, and the first contribution value of the rate index by the first contribution value of the non-rate index; If the types of multiple preset dimensions are the same as the type of fixed dimensions, the first contribution value of the rate indicator is determined by the ratio between the current value and the previous value of the rate indicator, and the ratio between the current value and the previous value of the abnormal indicator; the first contribution value of the non-rate indicator is determined by the ratio between the current value and the previous value of the non-rate indicator, and the ratio between the current value and the previous value of the abnormal indicator.

8. A device for attributing abnormal order business, characterized in that: include: A determination module is used to determine abnormal indicators and multiple preset dimensions of an order; A combination module, used to determine at least one associated indicator corresponding to the abnormal indicator according to the mathematical relationship between the multiple indicators; and combine the at least one associated indicator with the multiple preset dimensions to obtain multiple sets; A calculation module, used to determine, according to current period values ​​that conform to the associated indicators and preset dimensions in the multiple sets, a first contribution value of the associated indicators in at least one set to the change of the abnormal indicator within a preset time period and a second contribution value of the preset dimension in at least one set to the change of the abnormal indicator within a preset time period; A judgment module, used to judge whether a first contribution value of an associated indicator in at least one set meets a first preset condition, and to judge whether a second contribution value of a preset dimension in at least one set meets a second preset condition; The attribution module is used to determine that the associated indicator in the set is a factor causing the abnormal order business if the first contribution value of the associated indicator in the set meets the first preset condition; and to determine that the preset dimension in the set is a factor causing the abnormal order business if the second contribution value of the preset dimension in the set meets the second preset condition.

9. An electronic device, characterized in that: include: Memory, used to store computer programs or instructions; A processor is used to execute the computer program or instructions in the memory so that the order business abnormality attribution method as described in any one of claims 1 to 7 is executed.

10. A computer-readable storage medium, characterized in that: When the instructions in the storage medium are executed by a processor, the processor is enabled to execute the order business abnormality attribution method as described in any one of claims 1-7.

Images