Method and device for determining complaint early warning threshold, electronic equipment and storage medium

Abstract

The application provides a complaint early warning threshold determination method and device, electronic equipment and a storage medium, comprising: determining a complaint amount reference value of each dotting according to historical complaint amount data, performing interval division on a complaint amount reference value interval, determining upper and lower limits of an early warning threshold of a reference value subinterval, and generating complaint early warning information according to a comparison result of a real-time complaint amount and the upper and lower limits of the early warning threshold of a target reference value subinterval in which the real-time complaint amount is located. The application fully considers the difference of complaint amount reference values in each period, performs interval division on the complaint amount reference interval in an indefinite section according to historical complaint amount data, and then determines whether the complaint amount of the current dotting is normal according to the upper and lower limits of the early warning threshold of each reference value subinterval. Since the complaint amount reference value interval is analyzed in a more fine-grained manner, the clustering analysis effect of the complaint amount reference value is greatly improved, and the accuracy of complaint early warning is effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to a complaint early warning threshold determination method and device, electronic equipment and a storage medium. BACKGROUND

[0002] Complaint early warning refers to early warning when the complaint volume exceeds the threshold value, so as to discover problems in advance through the complaint volume. The existing complaint early warning method analyzes the historical complaint volume data of different business types, different city dimensions and different time accumulation dimensions as complaint volume reference value data when setting the early warning threshold value, thereby realizing static threshold value setting of different dimensions. However, this static threshold value setting method is not flexible and will inevitably cause certain experience errors.

[0003] In addition, through statistical analysis of complaint data volume in different time periods of a day, it is found that the difference in customer complaint volume in different time periods of a day is huge. The complaint volume average value in different time periods can be calculated based on historical complaint volume data, which is referred to as a complaint volume reference value in the industry. Most of the existing early warning threshold methods do not take into account the difference in complaint volume reference values in different time periods, so the existing early warning method has a coarse-grained fixed segmentation of the complaint volume reference value, which affects the accuracy of complaint early warning. SUMMARY

[0004] The present application provides a complaint early warning threshold determination method and device, electronic equipment and a storage medium, which solves the defect of low complaint early warning accuracy caused by the coarse-grained fixed segmentation of the complaint volume reference value due to the failure to take into account the difference in complaint volume reference values in different time periods, realizes more fine-grained analysis of the complaint volume reference value interval, greatly improves the clustering analysis effect of the complaint volume reference value, and effectively improves the accuracy of complaint early warning.

[0005] In a first aspect, the present application provides a complaint early warning threshold determination method, comprising:

[0006] According to historical complaint volume data, determine the complaint volume reference value of each dot, and the complaint volume reference value corresponding to any dot is determined by the complaint volume average value of the same dot in the previous preset number of days;

[0007] Interval division is performed on the complaint volume reference value interval determined by the maximum value in the complaint volume reference value, and a plurality of continuous reference value subintervals are obtained;

[0008] Determine the upper and lower limits of the early warning threshold value of each reference value subinterval;

[0009] According to the real-time complaint volume reference value of the current dot, determine the target reference value subinterval from the plurality of continuous reference value subintervals;

[0010] According to the comparison of the current real-time complaint amount and the upper and lower limits of the early warning threshold of the target reference value sub-interval, complaint early warning information is generated.

[0011] According to the complaint early warning threshold determination method provided by the application, the complaint amount reference value interval determined by the maximum value in the complaint amount reference value is divided into intervals to obtain a plurality of continuous reference value sub-intervals, which comprises:

[0012] A change trend factor and a complaint amount threshold interval of each threshold model preset are determined;

[0013] The change trend factor is the average value of the difference between adjacent complaint amount reference values of the previous day; the threshold model includes an increment model, an increment amplitude model and an amplitude model; the maximum value of the first complaint amount threshold interval related to the increment model is determined based on the maximum value of all complaint amount reference values of the previous preset number of days before the preset time point; the minimum value of the third complaint amount threshold interval related to the amplitude model is determined based on the maximum value of the complaint amount in the previous preset number of days and a preset complaint amount amplitude threshold; the minimum value of the second complaint amount threshold interval related to the increment amplitude model is the maximum value of the first complaint amount threshold interval, and the maximum value of the second complaint amount threshold interval is the minimum value of the first complaint amount threshold interval;

[0014] The complaint amount reference value interval is divided into intervals based on the change trend factor, the first complaint amount threshold interval, the second complaint amount threshold interval and the third complaint amount threshold interval.

[0015] According to the complaint early warning threshold determination method provided by the application, the complaint amount reference value interval is divided into intervals based on the change trend factor, the first complaint amount threshold interval, the second complaint amount threshold interval and the third complaint amount threshold interval, which comprises:

[0016] The division nodes are determined by increasing the change trend factor from 0 to infinity, and the reference value sub-interval is divided into intervals;

[0017] When any minimum value or any maximum value in the first complaint amount threshold interval, the second complaint amount threshold interval and the third complaint amount threshold interval is contained in any reference value sub-interval, the any minimum value or the any maximum value is increased as a division node of the interval division.

[0018] According to the complaint early warning threshold determination method provided by the application, the determination of the upper and lower limits of the early warning threshold of each reference value sub-interval comprises:

[0019] If any reference value sub-interval is a sub-interval of the first complaint quantity threshold interval, the upper and lower limits of the early warning threshold of the any reference value sub-interval are determined as upper and lower limits of the complaint quantity increment early warning threshold;

[0020] If any reference value sub-interval is a sub-interval of the second complaint quantity threshold interval, the upper and lower limits of the early warning threshold of the any reference value sub-interval include upper and lower limits of the complaint quantity increment early warning threshold and upper and lower limits of the complaint quantity increase early warning threshold;

[0021] If any reference value sub-interval is a sub-interval of the third complaint quantity threshold interval, the upper and lower limits of the early warning threshold of the any reference value sub-interval include upper and lower limits of the complaint quantity increase early warning threshold.

[0022] According to the method for determining the complaint early warning threshold provided by the application, the upper and lower limits of the complaint quantity increment early warning threshold of any reference value sub-interval include upper and lower limits of the complaint quantity increment early warning threshold; and the upper and lower limits of the complaint quantity increase early warning threshold of any reference value sub-interval include upper and lower limits of the complaint quantity increase early warning threshold.

[0023] The upper limit of the complaint quantity increment early warning threshold is determined based on a first preset static threshold corresponding to the any reference value sub-interval; and the upper limit of the complaint quantity increment early warning threshold is determined based on an average value of the complaint quantity increment in the previous preset days related to the any reference value sub-interval.

[0024] The upper limit of the complaint quantity increase early warning threshold is determined based on a second preset static threshold corresponding to the any reference value sub-interval; and the upper limit of the complaint quantity increase early warning threshold is determined based on an average value of the complaint quantity increase in the previous preset days related to the any reference value sub-interval.

[0025] According to the method for determining the complaint early warning threshold provided by the application, after determining the upper and lower limits of the early warning threshold of each reference value sub-interval, the method further includes:

[0026] According to the lower limit of the early warning threshold of each reference value sub-interval, determining early warning lower limits of different early warning levels;

[0027] According to the early warning lower limits of different early warning levels, determining complaint quantity early warning intervals corresponding to different early warning levels.

[0028] According to the method for determining the complaint early warning threshold provided by the application, the early warning levels include a first early warning level, a second early warning level and a third early warning level, and the early warning lower limits of the first early warning level, the second early warning level and the third early warning level increase in turn.

[0029] The method for determining a complaint early warning threshold value according to the present application comprises the following steps: determining a target reference value subinterval from a plurality of continuous reference value subintervals according to a real-time complaint quantity reference value of current dotting, and the target reference value subinterval is determined as a reference value subinterval in which the real-time complaint quantity reference value is located.

[0030] The target reference value subinterval is determined as a reference value subinterval in which the real-time complaint quantity reference value is located.

[0031] The method for determining a complaint early warning threshold value according to the present application comprises the following steps: comparing the real-time complaint quantity of current dotting with upper and lower limits of the early warning threshold value of the target reference value subinterval, and generating complaint early warning information.

[0032] If the target reference value subinterval is determined as a subinterval of the first complaint quantity threshold value interval, then a complaint quantity increment of current dotting is determined according to the real-time complaint quantity.

[0033] The complaint early warning information is generated based on a comparison result between the complaint quantity increment and upper and lower limits of the complaint quantity increment early warning threshold value of the target reference value subinterval.

[0034] If the target reference value subinterval is determined as a subinterval of the second complaint quantity threshold value interval, then a complaint quantity increment and a complaint quantity increment amplitude of current dotting are determined according to the real-time complaint quantity.

[0035] The complaint early warning information is generated by comprehensively combining a comparison result between the complaint quantity increment and upper and lower limits of the complaint quantity increment early warning threshold value of the target reference value subinterval, and a comparison result between the complaint quantity increment amplitude and upper and lower limits of the complaint quantity increment amplitude early warning threshold value of the target reference value subinterval.

[0036] If the target reference value subinterval is determined as a subinterval of the second complaint quantity threshold value interval, then a complaint quantity increment amplitude of current dotting is determined according to the real-time complaint quantity.

[0037] The complaint early warning information is generated based on a comparison result between the complaint quantity increment amplitude and upper and lower limits of the complaint quantity increment amplitude early warning threshold value of the target reference value subinterval.

[0038] The method for determining a complaint early warning threshold value according to the present application comprises the following steps: after determining upper and lower limits of the early warning threshold value of each reference value subinterval, the upper and lower limits of the early warning threshold value of each reference value subinterval are optimized by using detection results of a plurality of continuous predicted work orders, and the optimization specifically comprises the following steps:

[0039] If the detection results are that a first preset number of continuous predicted work orders do not generate early warnings, then the lower limit of the early warning threshold value of each reference value subinterval is reduced.

[0040] If the detection result is that the second preset number of continuous prediction work orders all generate early warning, increase the lower limit of the early warning threshold of each reference value sub-interval;

[0041] Otherwise, keep the upper and lower limits of the early warning threshold of each reference value sub-interval unchanged.

[0042] According to the method for determining a complaint early warning threshold provided by the application, the adjustment formula for reducing the lower limit of the early warning threshold of each reference value sub-interval is:

[0043] α count =β count (1-δ), α rate =β rate -ε;

[0044] The adjustment formula for increasing the lower limit of the early warning threshold of each reference value sub-interval is:

[0045] α count =β count (1+δ), α rate =β rate +ε;

[0046] Wherein, α count is the adjusted complaint amount increment, β count is the complaint amount increment before adjustment, δ is a preset increment adjustment factor, α rate is the adjusted complaint amount increase, β rate is the complaint amount increase before adjustment, and ε is a preset increase adjustment factor.

[0047] In a second aspect, the application further provides a device for determining a complaint early warning threshold, comprising:

[0048] A data preprocessing unit is configured to determine a complaint amount reference value for each dot according to historical complaint amount data, wherein the complaint amount reference value corresponding to any dot is determined by the average value of the complaint amount of the same dot in the previous preset number of days;

[0049] A reference value interval division unit is configured to divide the complaint amount reference value interval determined by the maximum value in the complaint amount reference value into multiple continuous reference value sub-intervals;

[0050] An early warning threshold determination unit is configured to determine the upper and lower limits of the early warning threshold of each reference value sub-interval;

[0051] A dot interval positioning unit is configured to determine a target reference value sub-interval from the multiple continuous reference value sub-intervals according to the real-time complaint amount reference value of the current dot;

[0052] The early warning generation unit generates complaint early warning information according to comparison of the current dotting real-time complaint quantity and the upper and lower limits of the early warning threshold of the target reference value subinterval.

[0053] In a third aspect, the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the complaint early warning threshold determination method according to any one of the above aspects when executing the program.

[0054] In a fourth aspect, the present application further provides a non-transitory computer readable storage medium having a computer program stored thereon, wherein the computer program is executable on a processor to implement the steps of the complaint early warning threshold determination method according to any one of the above aspects.

[0055] The complaint early warning threshold determination method, device, electronic device, and storage medium provided by the present application fully consider the difference of complaint quantity reference values in each period, perform interval division on the complaint quantity reference interval in an indefinite section according to historical complaint quantity data, and then determine whether the complaint quantity of the current dotting is normal according to the upper and lower limits of the early warning threshold of each reference value subinterval. Since the complaint quantity reference value interval is analyzed in a more fine-grained manner, the clustering analysis effect of the complaint quantity reference value is greatly improved, and the accuracy of complaint early warning is effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the present application or prior art, the following will briefly introduce the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0057] Figure 1 is one of the flowcharts of the complaint early warning threshold determination method provided by the present application;

[0058] Figure 2 is another flowchart of the complaint early warning threshold determination method provided by the present application;

[0059] Figure 3 is a flowchart of adjusting the upper and lower limits of the early warning threshold by using predicted work orders provided by the present application;

[0060] Figure 4 is a structural schematic diagram of the complaint early warning threshold determination device provided by the present application;

[0061] Figure 5 is a structural schematic diagram of the electronic device provided by the present application. DETAILED DESCRIPTION

[0062] In order to make the objects, technical solutions and advantages of the present application clearer, the following will clearly and completely describe the technical solutions in the present application with reference to the drawings in the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0063] It should be noted that, in the description of the embodiments of the present application, the terms "comprising", "containing" or any other variants thereof are intended to cover the non-exclusive containing, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitations, the element defined by the statement "including a" does not exclude the presence of another identical element in the process, method, article or device including the element. The above terms in the present application can be understood according to the specific meaning in the present application by a person of ordinary skill in the art.

[0064] The terms "first", "second", and the like in the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than that illustrated or described herein, and the objects distinguished by "first", "second", etc. are generally a category and do not limit the number of objects, for example, the first object can be one or more. In addition, "and / or" means at least one of the connected objects, and the character " / ", generally means that the front and rear associated objects are in a "or" relationship.

[0065] When the prior art communication companies set the complaint warning threshold, they generally analyze the historical warning complaint data of different service types, different city dimensions and different time cumulative dimensions, so as to set static thresholds of different dimensions, but such static threshold setting is not flexible and usually has a certain subjectivity. In particular, when the complaint situation changes, the static threshold needs to be adjusted again to adapt to the new complaint situation, which is extremely low in efficiency.

[0066] Moreover, the existing complaint warning method does not consider the difference of customer complaint data amount in different time periods when setting the static threshold, that is, the customer complaint is in the unit place in some time period, and the customer complaint is in the hundred place or even the thousand place in some time period. The coarse-grained fixed segmentation of the complaint amount reference value leads to the inability to perform similarity clustering analysis on the reference value.

[0067] Therefore, it is urgent to improve the existing complaint early warning threshold setting method to improve the accuracy of complaint early warning. The following will combine Figures 1-5 The application provides a complaint early warning threshold determination method, device, electronic equipment and storage medium.

[0068] Figure 1 The application provides a complaint early warning threshold determination method, device, electronic equipment and storage medium. Figure 1 As shown in the figure, the method comprises the following steps:

[0069] Step 101: According to the historical complaint amount data, the complaint amount reference value of each dot is determined.

[0070] Among them, the complaint amount reference value corresponding to any dot is determined by the average value of the complaint amount of the same dot in each day within the previous preset days.

[0071] The historical complaint amount data can be generated by calling the historical complaint amount of each communication company in different time periods before the current time. Accordingly, the complaint amount reference value is the average value of the preset period complaint amount calculated according to the collected historical complaint amount data.

[0072] As an optional embodiment, the complaint early warning threshold determination method provided by the application can determine 288 dots with 5 minutes as an interval for 24 hours in a day, such as 00:05 representing the first dot, 00:10 representing the second dot, and so on.

[0073] Further, assuming that the preset days are 30 days, then the complaint amount of each dot in 30 days can be obtained according to the historical complaint amount data, and the average value of the complaint amount of each dot is calculated as the complaint amount reference value of each dot.

[0074] After determining the complaint amount reference value of each dot, normally, as long as the complaint amount of each dot is greater than the complaint amount reference value, it is outside the normal range of complaint and needs to be warned.

[0075] Step 102: The complaint amount reference value interval determined by the maximum value in the complaint amount reference value is divided into intervals, and a plurality of continuous reference value subintervals are obtained.

[0076] Specifically, the above-mentioned complaint amount reference value interval can start from 0 to the maximum value of the complaint amount reference value of all dots determined by the historical complaint amount data.

[0077] Suppose the maximum value of the complaint amount reference value of all dots determined by the historical complaint amount data corresponding to the recent 30 days is 100, then the above-mentioned complaint amount reference value interval is [1, 100].

[0078] In the prior art, when performing complaint early warning analysis, the complaint amount reference value is generally segmented by coarse granularity fixed segmentation, for example, the complaint amount reference value interval is artificially divided into [0, 50] and [51, 100].

[0079] The present application considers that the existing complaint amount reference value coarse granularity fixed segmentation cannot effectively perform similarity clustering analysis on the complaint amount reference value, and therefore performs interval division on the complaint amount reference value interval in which the complaint amount reference value is located to obtain a plurality of continuous reference value subintervals.

[0080] Optionally, the reference value subintervals divided by the present application are more fine-grained, and fully consider the difference of complaint amount data at the beginning, middle and end of each month, and the difference of complaint amount data at different time periods each day. For example, compared with the existing complaint amount reference value coarse granularity fixed segmentation, the reference value subintervals obtained by the present application can be [0, 20], (20, 30], (30, 50], (50, 65], (65, 85] and (85, 100], and of course (100, +∞] can also be taken as a reference value subinterval.

[0081] Step 103: determining the upper and lower limits of the early warning threshold of each reference value subinterval.

[0082] Further, the present application performs prediction analysis on the current early warning according to the historical complaint amount of each dot and the complaint amount reference value of the corresponding dimension.

[0083] The complaint amount increment average value and the complaint amount amplitude average value corresponding to each dot can be calculated according to the historical complaint amount data. Then, the upper and lower limits of the early warning threshold of each reference value subinterval are calculated respectively.

[0084] Generally, the upper limit of the early warning threshold of each reference value subinterval can be determined by a static threshold, and the static threshold can also be set in advance by historical complaint amount data. The lower limit of the early warning threshold of each reference value subinterval can be determined according to the complaint amount increment average value and the complaint amount amplitude average value in the past 30 days.

[0085] It should be noted that the upper limits of the early warning thresholds of different reference value subintervals can be the same or different, and the upper limits of the early warning thresholds of adjacent multiple reference value subintervals can also be set to be the same.

[0086] Step 104: determining a target reference value subinterval from the plurality of continuous reference value subintervals according to the real-time complaint amount reference value of the current dot.

[0087] As an optional embodiment, after 288 dots are determined in 24 hours with a time interval of 5 minutes, the current dot can be determined according to the current time. The complaint amount reference value corresponding to the current dot is taken as the real-time complaint amount reference value, and the real-time complaint amount reference value is taken as the target reference value sub-interval according to which interval it falls.

[0088] Suppose that the real-time complaint amount reference value corresponding to the current dot is 25, and the previously determined reference value sub-intervals include [0, 20], (20, 30], (30, 50], (50, 65], (65, 85], (85, 100] and (100, +∞], then the target reference value sub-interval of the current dot can be determined as (20, 30].

[0089] Step 105: generating complaint warning information according to the comparison of the real-time complaint amount of the current dot with the upper and lower limits of the warning threshold of the target reference value sub-interval.

[0090] After the upper and lower limits of the warning threshold of each reference value sub-interval are determined in step 103, the upper and lower limits of the warning threshold of the target reference value sub-interval can be retrieved.

[0091] In this way, the real-time complaint amount of the current dot can be compared with the upper and lower limits of the warning threshold of the target reference value sub-interval, and the warning information can be generated according to the comparison result.

[0092] Suppose that the warning information includes multiple warning levels such as no alarm, yellow alarm, orange alarm and red alarm, and the yellow alarm, orange alarm and red alarm correspond to different intervals in the upper and lower limits of the warning threshold. The alarm information generated when the real-time complaint amount is less than the lower limit of the warning threshold is no alarm, and when the real-time complaint amount is greater than or equal to the lower limit of the warning threshold, the alarm corresponding to the warning level is generated.

[0093] The determination method of the complaint warning threshold provided by the application fully considers the difference of the complaint amount reference value in each period, performs interval division on the complaint amount reference interval in an indefinite section according to the historical complaint amount data, and then judges whether the complaint amount of the current dot is normal according to the upper and lower limits of the warning threshold of each reference value sub-interval. Since the complaint amount reference value interval is analyzed in a more fine-grained manner, the clustering analysis effect of the complaint amount reference value is greatly improved, and the accuracy of the complaint warning is effectively improved.

[0094] Based on the content of the above embodiments, as an optional embodiment, the complaint amount reference value interval determined by the maximum value in the complaint amount reference value is divided into multiple continuous reference value sub-intervals in step 102, which specifically includes but is not limited to the following implementation manners:

[0095] Firstly, a change trend factor is determined, and a complaint quantity threshold interval of each preset threshold model is determined.

[0096] Then, the complaint quantity reference value interval is divided based on the change trend factor, the first complaint quantity threshold interval, the second complaint quantity threshold interval, and the third complaint quantity threshold interval.

[0097] The change trend factor is an average value of differences between adjacent complaint quantity reference values of the previous day.

[0098] The threshold model mainly includes an increment model, an increment-amplitude model, and an amplitude model.

[0099] The maximum value of the first complaint quantity threshold interval related to the increment model is determined based on a maximum value of all complaint quantity reference values of the previous preset number of days before a preset time point.

[0100] The minimum value of the third complaint quantity threshold interval related to the amplitude model is determined based on a maximum value of the complaint quantity in the previous preset number of days and a preset complaint quantity amplitude threshold.

[0101] The minimum value of the second complaint quantity threshold interval related to the increment-amplitude model is the maximum value of the first complaint quantity threshold interval, and the maximum value of the second complaint quantity threshold interval is the minimum value of the first complaint quantity threshold interval.

[0102] As an optional embodiment, in the case of determining 288 points in 24 hours of a day with a time interval of 5 minutes, the change trend factor can be a 5-minute change trend factor, that is, an average value of increases or decreases of adjacent 5-minute complaint quantity reference values of the previous day.

[0103] Further, the division interval for dividing the complaint quantity reference value interval can be referred to according to the change trend factor. The larger the change trend factor, the smaller the corresponding division interval, and the smaller the change trend factor, the larger the corresponding division interval can be set.

[0104] Because the customer complaint quantity is quite different in different time periods of a day, the customer complaint quantity in some time periods is in the unit of ten, and the customer complaint quantity in some time periods is in the unit of hundred or even thousand, the fluctuation of the customer complaint quantity in different time periods is not consistent.

[0105] In addition, the customer influence caused by different situations is different, that is, the fluctuation and influence degree of different complaint quantity reference values and amplitudes are different. Assuming that the complaint quantity reference value is 1000, the increment of the complaint quantity is 100, and the corresponding amplitude is 10%; if the complaint quantity reference value is 100, the increment of the complaint quantity is 100, but the amplitude is 100%; if the complaint quantity reference value is 1, the increment of the complaint quantity is 100, and the amplitude is as high as 10000%.

[0106] Therefore, when the complaint amount reference value is small, only the increment of the complaint amount needs to be considered; when the complaint amount reference value is large, only the increment amplitude needs to be considered; and in other cases, both the increment amplitude and the increment of the complaint amount need to be considered.

[0107] Therefore, the determination method of the complaint early warning threshold provided by the present application can automatically construct a threshold model for different reference value subintervals, adopt three models, namely, a pure increment model, an increment and increment amplitude model, and a pure increment amplitude model, set the complaint amount threshold interval of each threshold model by using the average value of the increment and the increment amplitude of the complaint amount reference value, and can to some extent avoid the influence of the complaint amount reference abnormal value on the threshold setting.

[0108] Taking the complaint amount reference value interval [1, +∞] as an example, a plurality of continuous reference value subintervals are obtained by interval division, as shown in Table 1 below:

[0109] Table 1 Reference value subinterval list

[0110]

[0111] The obtained reference value subintervals specifically include [0, 20), [20, 30), [30, 50), [50, 70), [70, 85), [85, 100), and (100, +∞]. Among them, 100 refers to the maximum value in all complaint amount reference values; the interval division interval 20 is determined according to the change trend factor; the first complaint amount threshold interval, i.e., the complaint amount threshold interval corresponding to the increment model, is [0, 30); the second complaint amount threshold interval, i.e., the complaint amount threshold interval corresponding to the increment and increment amplitude model, is [30, 85]; and the third complaint amount threshold interval, i.e., the complaint amount threshold interval corresponding to the increment amplitude model, is [85, +∞).

[0112] As an optional embodiment, the interval division of the complaint amount reference value interval based on the change trend factor, the first complaint amount threshold interval, the second complaint amount threshold interval, and the third complaint amount threshold interval includes:

[0113] The division nodes are determined by increasing the change trend factor from 0 to infinity for interval division of the reference value subintervals;

[0114] When any minimum value or any maximum value in the first complaint amount threshold interval, the second complaint amount threshold interval, and the third complaint amount threshold interval is included in any divided reference value subinterval, the any minimum value or the any maximum value is increased as a division node of the interval division.

[0115] The following briefly describes how to divide the complaint amount reference value interval according to the above change trend factor, the first complaint amount threshold interval, the second complaint amount threshold interval, and the third complaint amount threshold interval:

[0116] For the first reference value sub-interval, the maximum value of the first reference value sub-interval is determined starting from the minimum value 0, and the interval division interval determined by the change trend factor is considered together with the first complaint amount threshold interval. If the interval division interval is greater than the maximum value of the first complaint amount threshold interval, the maximum value of the complaint amount reference value of the reference value sub-interval is set to the maximum value of the first complaint amount threshold interval. Otherwise, the maximum value of the complaint amount reference value of the reference value sub-interval is set to the complaint amount reference value determined by the interval division interval.

[0117] Taking the reference value sub-interval recorded in the above table 1 as an example, because the interval division interval determined by the change trend factor is 20, and the maximum value of the first complaint amount threshold interval is 30, the first reference value sub-interval is set to [0, 20).

[0118] Further, when dividing the next segment of the reference value sub-interval, the maximum value of the last segment is taken as the minimum value of the current segment, and the interval maximum value is similar to the setting rule of the first reference value sub-interval, that is, the value determined by the interval division interval is compared with the size of the complaint amount reference value defined by each threshold model.

[0119] Taking the reference value sub-interval recorded in the above table 1 as an example, the maximum value 20 of the last segment is set as the minimum value of the current segment, and the value 40 determined by the interval division interval is compared with the maximum value 30 of the increment model. Therefore, the next reference value sub-interval of the first reference value sub-interval is determined to be [20, 30).

[0120] By analogy, other reference value sub-intervals can be divided as follows: [30, 50), [50, 70), [70, 85), [85, 100), and (100, +∞].

[0121] The determination method of the complaint early warning threshold provided in the present application combines the complaint amount threshold intervals set by each threshold model (including the increment model, the increment-amplitude model, and the amplitude model) when dividing the complaint amount reference value interval. If the value of the complaint amount amplitude interval, the complaint amount increment interval, and the complaint amount amplitude increment interval is not the value determined by the interval division interval, the value determined by the interval division interval needs to be set to the maximum value of the complaint amount reference value interval, and the minimum value of the next complaint amount reference value interval starts from the maximum value of the last complaint amount reference value interval, and the maximum value is the minimum value of the complaint amount reference value interval plus the interval division interval.

[0122] The application provides a complaint early warning threshold determination method, which considers that the similarity clustering analysis of the complaint quantity reference value cannot be performed due to the coarse-grained fixed segmentation of the complaint quantity reference value, and therefore adopts a trend change factor to perform the reference value division in an indefinite segment. Meanwhile, the characteristics that the customer complaint quantities of different time periods are greatly different are taken into account, the complaint quantity threshold interval of different threshold models is combined with the trend change factor during the interval division, the more fine-grained division of the complaint quantity reference value interval is realized, the fine increment and increment amplitude are realized, and the complaint quantity reference value clustering analysis effect is greatly improved.

[0123] Based on the content of the above embodiment, as an optional embodiment, the determination of the early warning threshold upper and lower limits of each reference value subinterval specifically includes:

[0124] If any reference value subinterval is a subinterval of the first complaint quantity threshold interval, the early warning threshold upper and lower limits of any reference value subinterval are determined as the complaint quantity increment early warning threshold upper and lower limits;

[0125] If any reference value subinterval is a subinterval of the second complaint quantity threshold interval, the early warning threshold upper and lower limits of any reference value subinterval include the complaint quantity increment early warning threshold upper and lower limits and the complaint quantity increment amplitude early warning threshold upper and lower limits;

[0126] If any reference value subinterval is a subinterval of the third complaint quantity threshold interval, the early warning threshold upper and lower limits of any reference value subinterval include the complaint quantity increment amplitude early warning threshold upper and lower limits.

[0127] In the embodiment, the above table 1 is taken as an example for describing the reference value subinterval.

[0128] The first complaint quantity threshold interval corresponding to the preset increment model is [0, 30), the second complaint quantity threshold interval corresponding to the increment amplitude model is [30, 85], and the third complaint quantity threshold interval corresponding to the amplitude model is [85, +∞).

[0129] Then, the two reference value subintervals with interval segmentation numbers of “1” and “2” are subintervals of the first complaint quantity threshold interval, and therefore the early warning threshold upper and lower limits of the two reference value subintervals only need to consider the influence of the complaint quantity increment, and therefore only include the complaint quantity increment early warning threshold upper and lower limits.

[0130] Correspondingly, the three reference value subintervals with interval segmentation numbers of “3”, “4” and “5” are subintervals of the second complaint quantity threshold interval, and therefore the early warning threshold upper and lower limits of the three reference value subintervals need to consider the influence of the complaint quantity increment and the influence of the complaint quantity increment amplitude, and therefore simultaneously include the complaint quantity increment early warning threshold upper and lower limits and the complaint quantity increment amplitude early warning threshold upper and lower limits.

[0131] Finally, the two reference value sub-intervals with interval segmentation sequence numbers of "6" and "7" are sub-intervals of the third complaint amount threshold interval, and the corresponding early warning threshold upper and lower limits only include the complaint amount increase early warning threshold upper and lower limits.

[0132] The complaint amount increase early warning threshold upper and lower limits of any reference value sub-interval include the complaint amount increase early warning threshold upper limit and the complaint amount increase early warning threshold lower limit.

[0133] The complaint amount increase early warning threshold upper and lower limits of any reference value sub-interval include the complaint amount increase early warning threshold upper limit and the complaint amount increase early warning threshold lower limit.

[0134] As an optional embodiment, the complaint amount increase early warning threshold upper limit can be determined based on the first preset static threshold corresponding to any reference value sub-interval, and the complaint amount increase early warning threshold upper limit can be determined based on the average complaint amount increase in the previous preset number of days related to the reference value sub-interval.

[0135] Correspondingly, the complaint amount increase early warning threshold upper limit can be determined based on the second preset static threshold corresponding to the reference value sub-interval, and the complaint amount increase early warning threshold upper limit can be determined based on the average complaint amount increase in the previous preset number of days related to the reference value sub-interval.

[0136] The first preset static threshold and the second static threshold can be preset based on historical complaint amount data, or can be read from the static threshold set when performing coarse-grained interval division in the existing complaint early warning threshold determination method, and are determined according to the coarse-grained interval to which the reference value sub-interval belongs.

[0137] Suppose that the existing coarse-grained interval division is [0, 50] and [51, 100], and the complaint amount increase early warning threshold upper limit corresponding to the [0, 50] complaint amount reference value interval is 80, and the complaint amount increase early warning threshold upper limit corresponding to the [51, 100] complaint amount reference value interval is 110. Then, the complaint amount increase early warning threshold upper limit of the three reference value sub-intervals with interval segmentation sequence numbers of "1", "2", and "3" shown in Table 1 of the present application, i.e., the first preset static threshold, can be directly set to 80, and the complaint amount increase early warning threshold upper limit of the three reference value sub-intervals with interval segmentation sequence numbers of "4", "5", and "6" can be set to 110.

[0138] Correspondingly, when presetting the complaint amount increase early warning threshold upper limit of each reference value sub-interval, a similar method can also be used, which will not be described here.

[0139] Further, the application also provides a method for setting the lower limit of the complaint volume increment early warning threshold of each reference value sub-interval. Taking the lower limit of the complaint volume increment early warning threshold of the reference value sub-interval [20, 30) with the interval segmentation number of 2 as shown in Table 1 as an example.

[0140] Firstly, according to all the 288 dots, the reference value of the complaint volume of all the dots, hereinafter referred to as target dots, is in the reference value sub-interval [20-30), and it is assumed that there are 18 target dots.

[0141] Then, the complaint volume increments corresponding to the 18 target dots are determined respectively, and the average value of the complaint volume increments corresponding to all the 18 target dots is calculated as the lower limit of the complaint volume increment early warning threshold of the reference value sub-interval [20, 30) with the interval segmentation number of 2.

[0142] Similarly, the lower limit of the complaint volume increment early warning threshold of each reference value sub-interval can be calculated in the same way.

[0143] Further, the lower limit of the complaint volume increment early warning threshold of each reference value sub-interval can also be calculated respectively in the above-mentioned manner.

[0144] The method for determining the complaint early warning threshold provided by the application considers the difference of complaint volume in different time periods, automatically adopts different threshold models for different dots, and divides the intervals according to different complaint volume reference values, so as to realize more fine-grained similarity clustering analysis of the reference value. The scheme discards the traditional early warning control mode according to the complaint same period, the growth rate, the growth amount index and other indicators according to the time sequence, and creatively introduces the same period complaint volume, the increment and the amplitude of the complaint volume reference value for early warning control. At the same time, the upper limit of the threshold is based on the static threshold, and the lower limit of the threshold is based on the average value of the increment and / or amplitude of the historical complaint volume reference value, so as to avoid the influence of the complaint volume reference abnormal value on the threshold setting to a certain extent.

[0145] Based on the content of the above-mentioned embodiment, as an optional embodiment, after determining the upper and lower limits of the early warning threshold of each reference value sub-interval, the method further comprises:

[0146] determining the early warning lower limit of different early warning levels according to the early warning threshold lower limit of each reference value sub-interval;

[0147] determining the complaint volume early warning interval corresponding to different early warning levels according to different early warning lower limits.

[0148] In order to more accurately perform complaint early warning, in the method for determining the complaint early warning threshold provided by the application, different early warning levels are set to intuitively show the emergency degree of early warning.

[0149] Specifically, after the upper and lower limits of the early warning threshold of each reference value sub-interval are determined, the complaint amount early warning intervals of different early warning levels can be set according to the upper and lower limits of the early warning threshold.

[0150] As an optional embodiment, it is assumed that the early warning levels include a first early warning level, a second early warning level and a third early warning level, and the early warning lower limits of the first early warning level, the second early warning level and the third early warning level increase in turn.

[0151] If the first early warning level is marked by red color, it can also be called red early warning level; if the second early warning level is marked by orange color, it can also be called orange early warning level; and if the third early warning level is marked by yellow color, it can also be called yellow early warning level. Among them, from the perspective of emergency degree, red early warning level > orange early warning level > yellow early warning level.

[0152] Generally, the upper limits of the levels of the red early warning level, the orange early warning level and the yellow early warning level can adopt the static threshold corresponding to the reference value sub-interval, and the setting mode of the early warning lower limit can be determined by the following way:

[0153]

[0154]

[0155]

[0156]

[0157]

[0158]

[0159] Wherein, mRate red , mRate orange , mRate yellow respectively represent the increase amplitude early warning lower limit of the red early warning level, the orange early warning level and the yellow early warning level; mCount red , mCount orange , mCount yellow respectively represent the increase amount early warning lower limit of the red early warning level, the orange early warning level and the yellow early warning level; n represents the past n days; λ i represents the complaint amount, μ i represents the complaint amount reference value, respectively represent the threshold rule percentage of the red early warning level, the orange early warning level and the yellow early warning level.

[0160] Suppose the upper and lower limits of the complaint volume increment early warning threshold of any confirmed reference value sub-interval are [100, 200], and the upper and lower limits of the complaint volume growth early warning threshold are [10, 50], then according to the above threshold rule percentage, the complaint volume increment early warning interval of the red early warning level, the orange early warning level and the yellow early warning level corresponding to the reference value sub-interval is: [150, 200], [120, 200] and [100, 200].

[0161] Alternatively, the complaint volume increment early warning intervals of the red early warning level, the orange early warning level and the yellow early warning level can be further optimized to [150, 200], [120, 150) and [100, 120), that is, the complaint volume increment early warning intervals of the red early warning level, the orange early warning level and the yellow early warning level are used to divide the upper and lower limits of the complaint volume increment early warning threshold of the reference value sub-interval.

[0162] Correspondingly, the complaint volume increment early warning intervals of the red early warning level, the orange early warning level and the yellow early warning level are: [15, 50], [12, 50] and [10, 50]. Further optimization can also be made to set them to [15, 20], [12, 15) and [10, 12).

[0163] Based on the content of the above embodiments, as an optional embodiment, the complaint early warning information is generated in step 105 according to the comparison of the current complaint volume and the early warning threshold upper and lower limits of the target reference value sub-interval, specifically including:

[0164] If it is determined that the target reference value sub-interval is a sub-interval of the first complaint volume threshold interval, the complaint volume increment of the current dot is determined according to the real-time complaint volume; and the complaint early warning information is generated based on the comparison result between the complaint volume increment and the complaint volume increment early warning threshold upper and lower limits of the target reference value sub-interval.

[0165] If it is determined that the target reference value sub-interval is a sub-interval of the second complaint volume threshold interval, the complaint volume increment and the complaint volume growth of the current dot are determined according to the real-time complaint volume; and the complaint early warning information is generated by comprehensively considering the comparison result between the complaint volume increment and the complaint volume increment early warning threshold upper and lower limits of the target reference value sub-interval, and the comparison result between the complaint volume growth and the complaint volume growth early warning threshold upper and lower limits of the target reference value sub-interval.

[0166] If it is determined that the target reference value sub-interval is a sub-interval of the second complaint volume threshold interval, the complaint volume growth of the current dot is determined according to the real-time complaint volume; and the complaint early warning information is generated based on the comparison result between the complaint volume growth and the complaint volume growth early warning threshold upper and lower limits of the target reference value sub-interval.

[0167] In the method for determining the complaint early warning threshold provided by the application, if it is needed to determine whether the real-time complaint quantity of the current dotting will generate an alarm or further determine the early warning level, first, it is needed to determine the target reference value sub-interval corresponding to the current dotting.

[0168] This embodiment is described by taking the reference value sub-interval list shown in Table 1 as an example.

[0169] The determination of the target reference value sub-interval according to the real-time complaint quantity collected by the current dotting has been described in the previous embodiments, and will not be repeated here.

[0170] Suppose the determined target reference value sub-interval is the reference value sub-interval with the interval segmentation serial number of “2”, which belongs to the sub-interval of the first complaint quantity threshold interval corresponding to the incremental model, then only the complaint quantity increment between the current dotting and the previous dotting needs to be determined according to the real-time complaint quantity.

[0171] Then, the complaint quantity increment is compared with the complaint quantity increment early warning threshold lower limit corresponding to the target reference value sub-interval, if the complaint quantity increment is greater than the complaint quantity increment early warning threshold lower limit, the generated complaint early warning information is an alarm, and whether it belongs to the red early warning level, the orange early warning level or the yellow early warning level can be determined according to which complaint quantity increment early warning interval the complaint quantity increment belongs to, and will not be repeated here.

[0172] Suppose the determined target reference value sub-interval is the reference value sub-interval with the interval segmentation serial number of “5”, which belongs to the sub-interval of the second complaint quantity threshold interval corresponding to the incremental amplitude model, then not only the complaint quantity increment between the current dotting and the previous dotting needs to be determined according to the real-time complaint quantity, but also the complaint quantity amplitude of the current dotting relative to the previous dotting needs to be further determined.

[0173] It should be noted that the current alarm and the corresponding early warning level can be determined according to the complaint quantity increment and the complaint quantity amplitude respectively, and the final complaint early warning information is determined based on the highest early warning level.

[0174] Suppose that according to the complaint quantity increment, it is determined that there is no need to alarm, and according to the complaint quantity amplitude, it is determined that there is a need to alarm and the early warning level is orange early warning, then the final complaint early warning information is to generate orange early warning.

[0175] Suppose that according to the complaint quantity increment, it is determined that there is a need to generate an alarm and the early warning level is red early warning, and according to the complaint quantity amplitude, it is determined that there is a need to alarm and the early warning level is orange early warning, then the final complaint early warning information is to generate red early warning.

[0176] Similarly, assuming that the determined target reference value sub-interval is the reference value sub-interval with the interval segmentation serial number of 6, the sub-interval belongs to the third complaint quantity threshold interval corresponding to the increasing model, and then only the complaint quantity increment between the current dotting and the previous dotting needs to be determined according to the real-time complaint quantity.

[0177] Then, the calculated complaint quantity increment is compared with the complaint quantity increment early warning threshold lower limit corresponding to the target reference value sub-interval, if the complaint quantity increment is greater than the complaint quantity increment early warning threshold lower limit, the generated complaint early warning information is an alarm, and whether it belongs to the red early warning level, the orange early warning level or the yellow early warning level can be determined according to which complaint quantity increment early warning interval the complaint quantity increment belongs to, which will not be repeated here.

[0178] The complaint early warning threshold determination method provided by the application determines the early warning complaint quantity reference value of each dotting according to the historical complaint quantity data, so as to predict and analyze the current early warning complaint quantity. In order to improve the accuracy of the early warning threshold, the complaint quantity reference value interval is divided into multiple continuous reference value sub-intervals, and the difference of the complaint quantity in different time periods is considered when the interval is divided, different threshold models are determined according to the complaint reference value in different time periods, the clustering analysis effect of the complaint quantity reference value is greatly improved, and the accuracy of the complaint early warning is effectively improved.

[0179] Figure 2 It is a second flowchart of the complaint early warning threshold determination method provided by the application, as an optional embodiment, after determining the upper and lower limits of the early warning threshold of each reference value sub-interval, the detection results of the continuous multiple predicted work orders are used to optimize the upper and lower limits of the early warning threshold of each reference value sub-interval.

[0180] The implementation steps of the complaint early warning threshold determination method provided by the embodiment mainly include:

[0181] Step 201, pre-processing the collected historical complaint quantity data.

[0182] By analyzing the historical complaint quantity data of different statistical dimensions (provincial dimension, city dimension, different business type dimension, time accumulation dimension, etc.), relevant factors of different dimensions are extracted, including day 5-minute change trend factor, one-month day increment average, one-month complaint quantity reference value, one-month day increment average, etc.

[0183] Step 202, using the data processing method provided in the above embodiment, including the change trend factor, the complaint quantity reference value of the last 30 days and the complaint quantity threshold interval of the three threshold models determined in advance, the interval of the complaint quantity reference value is divided into multiple reference value sub-intervals, and the reference value sub-intervals are obtained, as shown in Table 1.

[0184] Step 203, the upper and lower limits of the early warning threshold corresponding to each reference value sub-interval are calculated respectively, which will not be repeated here.

[0185] Step 204, the early warning upper and lower limits currently set are verified by using the prediction work order, and if they are inaccurate, the early warning threshold upper and lower limits corresponding to each reference value sub-interval can be properly adjusted based on a dynamic adjustment factor.

[0186] Figure 3 is a flowchart of adjusting the early warning threshold upper and lower limits by using the prediction work order provided by the application, in combination with Figure 3 , the specific implementation of properly adjusting the early warning threshold upper and lower limits corresponding to each reference value sub-interval based on a dynamic adjustment factor can include but is not limited to:

[0187] If the detection result is that none of the continuous first preset number of prediction work orders generates an early warning, the early warning threshold lower limit of each reference value sub-interval is reduced;

[0188] If the detection result is that the continuous second preset number of prediction work orders all generate an early warning, the early warning threshold lower limit of each reference value sub-interval is increased;

[0189] Otherwise, the early warning threshold upper and lower limits of each reference value sub-interval remain unchanged.

[0190] Optionally, the prediction work order is obtained by manually counting the number of complaints in a preset period, and each prediction work order is provided with a complaint early warning information label.

[0191] The implementation of the whole early warning threshold upper and lower limit optimization can adopt a timing adjustment mode, that is, the early warning threshold upper and lower limit generated in the previous day is adjusted at a certain period of each day, and the specific adjustment rule is as follows:

[0192] First, the prediction work order is counted and generated according to the same dimension, and the complaint early warning information judged by the current early warning threshold upper and lower limit is obtained.

[0193] The generated complaint early warning information is compared with the complaint early warning information label of the prediction work order to obtain a detection result.

[0194] (1) If the detection result meets the early warning generated every day and multiple early warnings generated in continuous days (non-continuous days refer to days that do not generate early warnings), the current early warning threshold upper and lower limit remains unchanged.

[0195] (2) If the detection result meets the early warning generated every day and multiple early warnings generated in continuous days, the current early warning threshold upper and lower limit is adjusted positively based on a preset adjustment factor, and the corresponding positive adjustment is made according to the adjustment factor.

[0196] (3) If the same reference value sub-interval does not generate a pre-warning for consecutive days (X can be adjusted based on the configuration, such as being set to 180 days), the current pre-warning threshold upper and lower limits are correspondingly adjusted in a negative direction according to an adjustment factor.

[0197] The method for determining a complaint pre-warning threshold provided by the application is helpful in dynamically correcting complaint abnormal points by using a dynamic threshold adjustment method, that is, different pre-warning strategies are applied according to the different complaint situations of each province, each business type and each city at different time periods, thereby avoiding dry and wet pre-warning caused by excessively low or high threshold values, and promoting complaint situation rectification at different levels; meanwhile, the scheme provides the ability to predict work orders to verify the accuracy of automatic threshold setting, and according to the accuracy, statistical calculation, probability and other means are integrated to realize the dynamic modification ability of the threshold value.

[0198] Based on the content of the above embodiment, as an optional embodiment, the adjustment formula for reducing the pre-warning threshold lower limit of each reference value sub-interval is:

[0199] α count =β count (1-δ), α rate =β rate -ε;

[0200] The adjustment formula for increasing the pre-warning threshold lower limit of each reference value sub-interval is:

[0201] α count =β count (1+δ), α rate =β rate +ε;

[0202] Wherein, α count is the adjusted complaint amount increment, β count is the complaint amount increment before adjustment, δ is a pre-set increment adjustment factor, α rate is the adjusted complaint amount increase, β rate is the complaint amount increase before adjustment, and ε is a pre-set increase adjustment factor.

[0203] In general, if the same reference value sub-interval does not generate a pre-warning for consecutive days (such as being set to 180 days), the pre-warning threshold upper and lower limits are adjusted in a negative direction according to the percentage of the adjustment factor. The adjustment method of the pre-warning threshold upper and lower limits is to adjust in a descending manner, that is, starting from the pre-warning threshold upper limit, adjusting in stages and at a constant ratio according to the pre-set ratio (the ratio refers to adjusting according to the adjustment factor), and the threshold adjustment range is between the pre-warning threshold lower limit and the pre-warning threshold upper limit. Correspondingly, when it is judged that positive adjustment is needed, the corresponding method can also be used, which is not described here.

[0204] Figure 4 is a structural schematic diagram of the complaint early warning threshold determination device provided by the application, as shown in Figure 4 The application also provides a complaint early warning threshold determination device, mainly comprising:

[0205] The data preprocessing unit 41 is mainly used for determining a complaint amount reference value of each dotting according to historical complaint amount data.

[0206] The reference value interval division unit 42 is mainly used for dividing the complaint amount reference value interval determined by the maximum value in the complaint amount reference value into multiple continuous reference value subintervals.

[0207] The early warning threshold determination unit 43 is mainly used for determining the upper and lower limits of the early warning threshold of each reference value subinterval.

[0208] The dotting interval positioning unit 44 is mainly used for determining a target reference value subinterval from the multiple continuous reference value subintervals according to the real-time complaint amount reference value of the current dotting.

[0209] The early warning generation unit 45 is mainly used for generating complaint early warning information according to the comparison between the real-time complaint amount of the current dotting and the upper and lower limits of the early warning threshold of the target reference value subinterval.

[0210] It should be noted that the complaint early warning threshold determination device provided by the embodiment of the application can execute the complaint early warning threshold determination method described in any of the above embodiments when specifically running, and the embodiment will not be described here.

[0211] The complaint early warning threshold determination device provided by the application fully considers the difference of complaint amount reference values in each period, divides the complaint amount reference interval into indefinite intervals according to historical complaint amount data, and then judges whether the complaint amount of the current dotting is normal according to the upper and lower limits of the early warning threshold of each reference value subinterval. Since the complaint amount reference value interval is analyzed in a more fine-grained manner, the clustering analysis effect of the complaint amount reference value is greatly improved, and the accuracy of complaint early warning is effectively improved.

[0212] Figure 5 is a structural schematic diagram of the electronic device provided by the application, as shown in Figure 5As shown, the electronic device can include a processor 510, a communications interface 520, a memory 530, and a communications bus 540, wherein the processor 510, the communications interface 520, and the memory 530 complete mutual communication through the communications bus 540. The processor 510 can invoke the logic instructions in the memory 530 to execute the determination method of the complaint early warning threshold, which includes: determining a complaint amount reference value of each dot according to historical complaint amount data, wherein the complaint amount reference value corresponding to any dot is determined by the average value of the complaint amount of the same dot in the previous preset number of days; performing interval division on the complaint amount reference value interval determined by the maximum value in the complaint amount reference value, to obtain a plurality of continuous reference value subintervals; determining the upper and lower limits of the early warning threshold of each reference value subinterval; determining a target reference value subinterval from the plurality of continuous reference value subintervals according to the real-time complaint amount reference value of the current dot; and generating complaint early warning information according to the comparison of the real-time complaint amount of the current dot and the upper and lower limits of the early warning threshold of the target reference value subinterval.

[0213] In addition, the logic instructions in the memory 530 described above can be implemented in the form of a software function unit and sold or used as an independent product, and can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various program code storage media.

[0214] In another aspect, the present application also provides a computer program product, which comprises a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions which, when executed by a computer, enable the computer to perform the method for determining a complaint early warning threshold provided by any of the above embodiments, the method comprising: determining, according to historical complaint volume data, a complaint volume reference value for each dot, the complaint volume reference value corresponding to any dot being determined by an average value of the complaint volume of the same dot in the previous preset number of days; performing interval division on a complaint volume reference value interval determined by a maximum value in the complaint volume reference values to obtain a plurality of continuous reference value subintervals; determining upper and lower limits of an early warning threshold for each of the reference value subintervals; determining a target reference value subinterval from the plurality of continuous reference value subintervals according to a real-time complaint volume reference value of a current dot; and generating complaint early warning information according to a comparison between the real-time complaint volume of the current dot and the upper and lower limits of the early warning threshold of the target reference value subinterval.

[0215] In yet another aspect, the present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the method for determining a complaint early warning threshold provided by any of the above embodiments, the method comprising: determining, according to historical complaint volume data, a complaint volume reference value for each dot, the complaint volume reference value corresponding to any dot being determined by an average value of the complaint volume of the same dot in the previous preset number of days; performing interval division on a complaint volume reference value interval determined by a maximum value in the complaint volume reference values to obtain a plurality of continuous reference value subintervals; determining upper and lower limits of an early warning threshold for each of the reference value subintervals; determining a target reference value subinterval from the plurality of continuous reference value subintervals according to a real-time complaint volume reference value of a current dot; and generating complaint early warning information according to a comparison between the real-time complaint volume of the current dot and the upper and lower limits of the early warning threshold of the target reference value subinterval.

[0216] The device embodiments described above are merely illustrative, wherein the units illustrated as separate components can or can not be physically separate, and the components illustrated as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purposes of the present embodiment scheme according to actual needs. Those of ordinary skill in the art can understand and implement without creative labor.

[0217] Those skilled in the art can clearly understand the technical solutions of the various embodiments from the above description of the embodiments, and the various embodiments can be implemented by means of software with the necessary general hardware platforms, and of course, can also be implemented by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part of the prior art that makes a contribution, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0218] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for some technical features therein; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for determining a complaint alert threshold, characterized in that, The method comprises the following steps: According to historical complaint volume data, determine the complaint volume reference value of each dot. The complaint volume reference value corresponding to any dot is determined by the average value of the complaint volume of the same dot in the previous preset number of days; Divide the complaint volume reference value interval determined by the maximum value in the complaint volume reference value into subintervals to obtain a plurality of continuous reference value subintervals; Determine the upper and lower limits of the warning threshold of each reference value subinterval; According to the real-time complaint volume of the current dot, determine the target reference value subinterval from the plurality of continuous reference value subintervals; According to the comparison of the real-time complaint volume of the current dot and the upper and lower limits of the warning threshold of the target reference value subinterval, generate complaint warning information; The method comprises the following steps: Determine the change trend factor and the complaint volume threshold interval of each threshold model preset; The change trend factor is the average value of the difference between the complaint volume reference values of adjacent dots in the previous day. The threshold model includes an increment model, an increment amplitude model, and an amplitude model. The maximum value of the first complaint volume threshold interval related to the increment model is determined based on the maximum value of the complaint volume reference values of all dots before the preset time point in the previous preset number of days. The minimum value of the third complaint volume threshold interval related to the amplitude model is determined based on the maximum value of the complaint volume in the previous preset number of days and the preset complaint volume amplitude threshold. The minimum value of the second complaint volume threshold interval related to the increment amplitude model is the maximum value of the first complaint volume threshold interval, and the maximum value of the second complaint volume threshold interval is the minimum value of the first complaint volume threshold interval; Based on the change trend factor, the first complaint volume threshold interval, the second complaint volume threshold interval, and the third complaint volume threshold interval, divide the complaint volume reference value interval.

2. The method of claim 1, wherein, The method comprises the following steps: From 0 to infinity, determine the division node by increasing the change trend factor to divide the reference value subinterval; When any minimum value or any maximum value in the first complaint volume threshold interval, the second complaint volume threshold interval, and the third complaint volume threshold interval is included in any divided reference value subinterval, increase the any minimum value or any maximum value as a division node of the interval division.

3. The method of determining a complaint alert threshold according to any of claims 1-2, characterized in that, The method comprises the following steps: If any reference value subinterval is a subinterval of the first complaint volume threshold interval, determine the upper and lower limits of the warning threshold of the any reference value subinterval as the complaint volume increment warning threshold upper and lower limits; If any reference value subinterval is a subinterval of the second complaint volume threshold interval, determine the upper and lower limits of the warning threshold of the any reference value subinterval to include the complaint volume increment warning threshold upper and lower limits and the complaint volume amplitude warning threshold upper and lower limits. If any reference value sub-interval is a sub-interval of the third complaint quantity threshold interval, the upper and lower limits of the early warning threshold of the any reference value sub-interval include the complaint quantity increase early warning threshold upper and lower limits.

4. The method of claim 3, wherein, The complaint quantity increase early warning threshold upper and lower limits of any reference value sub-interval include the complaint quantity increase early warning threshold upper limit and the complaint quantity increase early warning threshold lower limit; the complaint quantity increase early warning threshold upper and lower limits of any reference value sub-interval include the complaint quantity increase early warning threshold upper limit and the complaint quantity increase early warning threshold lower limit; The complaint quantity increase early warning threshold upper limit is determined based on the first preset static threshold corresponding to the any reference value sub-interval; the complaint quantity increase early warning threshold upper limit is determined based on the complaint quantity increase average value within the previous preset number of days related to the any reference value sub-interval; The complaint quantity increase early warning threshold upper limit is determined based on the second preset static threshold corresponding to the any reference value sub-interval; the complaint quantity increase early warning threshold upper limit is determined based on the complaint quantity increase average value within the previous preset number of days related to the any reference value sub-interval.

5. The method of claim 3, wherein, After determining the upper and lower limits of the early warning threshold of each reference value sub-interval, it further includes: According to the early warning lower limit of each reference value sub-interval, determine the early warning lower limit of different early warning levels; According to different early warning lower limits, determine the complaint quantity early warning interval corresponding to different early warning levels.

6. The method of claim 5, wherein, The early warning levels include a first early warning level, a second early warning level and a third early warning level, and the early warning lower limits of the first early warning level, the second early warning level and the third early warning level increase in turn.

7. The method of claim 1, wherein, The target reference value sub-interval is determined from a plurality of continuous reference value sub-intervals according to the real-time complaint quantity reference value of the current pitch, including: Determine the reference value sub-interval where the real-time complaint quantity reference value is located as the target reference value sub-interval.

8. The method of claim 3, wherein, The comparison of the real-time complaint quantity of the current pitch and the early warning threshold upper and lower limits of the target reference value sub-interval generates complaint early warning information, including: If it is determined that the target reference value sub-interval is a sub-interval of the first complaint quantity threshold interval, determine the complaint quantity increase of the current pitch according to the real-time complaint quantity; Based on the comparison result between the complaint quantity increase and the complaint quantity increase early warning threshold upper and lower limits of the target reference value sub-interval, generate the complaint early warning information; If it is determined that the target reference value sub-interval is a sub-interval of the second complaint quantity threshold interval, determine the complaint quantity increase and the complaint quantity increase of the current pitch according to the real-time complaint quantity; Integrate the comparison results between the complaint quantity increase and the complaint quantity increase early warning threshold upper and lower limits of the target reference value sub-interval, and the comparison results between the complaint quantity increase and the complaint quantity increase early warning threshold upper and lower limits of the target reference value sub-interval, to generate the complaint early warning information; If it is determined that the target reference value sub-interval is a sub-interval of the second complaint quantity threshold interval, determine the complaint quantity increase of the current pitch according to the real-time complaint quantity; Based on the comparison result between the complaint quantity increase and the complaint quantity increase early warning threshold upper and lower limits of the target reference value sub-interval, generate the complaint early warning information.

9. The method of claim 1, wherein, After determining the upper and lower limits of the early warning threshold of each reference value sub-interval, the upper and lower limits of the early warning threshold of each reference value sub-interval are optimized by using the detection results of a plurality of continuous prediction work orders, and the optimization specifically includes: If the detection result is that no early warning is generated for a continuous first preset number of prediction work orders, the lower limit of the early warning threshold of each reference value sub-interval is reduced; If the detection result is that early warning is generated for a continuous second preset number of prediction work orders, the lower limit of the early warning threshold of each reference value sub-interval is increased; Otherwise, the upper and lower limits of the early warning threshold of each reference value sub-interval remain unchanged.

10. The method of claim 9, wherein, The adjustment formula for reducing the lower limit of the early warning threshold of each reference value sub-interval is: ； ； The adjustment formula for increasing the lower limit of the early warning threshold of each reference value sub-interval is: ； ； wherein, is the adjusted complaint volume delta, is the unadjusted complaint volume delta, is the pre-set delta adjustment factor, is the adjusted complaint volume increment, is the unadjusted complaint volume increment, is the pre-set increment adjustment factor.

11. A device for determining a complaint alert threshold, characterized in that It includes: The data preprocessing unit is configured to determine each complaint volume reference value of the dot according to historical complaint volume data. The complaint volume reference value corresponding to any dot is determined by the average value of the complaint volume of the same dot in each day within a preset number of days. The reference value interval division unit is configured to divide the complaint volume reference value interval determined by the maximum value in the complaint volume reference value into a plurality of continuous reference value sub-intervals, and specifically includes: determining a change trend factor and a preset complaint volume threshold interval of each threshold model; the change trend factor is the average value of the difference between the complaint volume reference values of adjacent dots of the previous day; the threshold model includes an increment model, an increment amplitude model and an amplitude model; the maximum value of the first complaint volume threshold interval related to the increment model is determined based on the maximum value of the complaint volume reference values of all dots before a preset time point within the preset number of days; the minimum value of the third complaint volume threshold interval related to the amplitude model is determined based on the maximum value of the complaint volume within the preset number of days and a preset complaint volume amplitude threshold; the minimum value of the second complaint volume threshold interval related to the increment amplitude model is the maximum value of the first complaint volume threshold interval, and the maximum value of the second complaint volume threshold interval is the minimum value of the first complaint volume threshold interval; the complaint volume reference value interval is divided based on the change trend factor, the first complaint volume threshold interval, the second complaint volume threshold interval and the third complaint volume threshold interval; The early warning threshold determination unit is configured to determine the upper and lower limits of the early warning threshold of each reference value sub-interval. The dot interval positioning unit is configured to determine a target reference value sub-interval from the plurality of continuous reference value sub-intervals according to the real-time complaint volume reference value of the current dot. The early warning generation unit is configured to generate complaint early warning information according to the comparison between the real-time complaint volume of the current dot and the upper and lower limits of the early warning threshold of the target reference value sub-interval.

12. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the steps of the complaint early warning threshold determination method of any one of claims 1 to 10.

13. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the complaint early warning threshold determination method of any one of claims 1 to 10.

14. A computer program product comprising a computer program, characterized in that, The computer program, when executed by a processor, implements the steps of the method for determining a complaint early warning threshold according to any one of claims 1 to 10.

Images