Detailed Description
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present description. This description may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein, as those skilled in the art will be able to make and use the present disclosure without departing from the spirit and scope of the present disclosure.
As mentioned above, after a user encounters a problem, the user will ask for help through a fixed-line telephone or an internet phone within a certain time, for example, the service phone of an organization such as a bank is generally a fixed-line telephone, and the internet mall such as the kyoto is generally an internet phone or a fixed-line telephone. Both fixed telephone service and network telephone service generally comprise two modes of manual service and robot self-service. Summarizing, current IVR presence includes two problems:
1. during the traffic peak period, the user waits for a long time in the IVR without shunting to the robot for self-service, so that the user actively hangs up. The robot self-service is better than the call loss in experience.
2. In the traffic valley period, the seat resources are relatively idle, but the user calls and then is shunted to the robot for self-service, so that the user does not enjoy manual service, thereby causing resource waste and poor service experience.
In view of the current IVR situation, in this specification, a traffic prediction method and apparatus, a traffic scheduling method and apparatus, a computing device, and a storage medium are provided, and details are described in the following embodiments one by one.
Fig. 1 is a block diagram illustrating a configuration of a computing device 100 according to an embodiment of the present specification. The components of the computing device 100 include, but are not limited to, memory 110 and processor 120. The processor 120 is coupled to the memory 110 via a bus 130 and the database 150 is used to store user data.
Computing device 100 also includes access device 140, access device 140 enabling computing device 100 to communicate via one or more networks 160. Examples of such networks include the Public Switched Telephone Network (PSTN), a Local Area Network (LAN), a Wide Area Network (WAN), a Personal Area Network (PAN), or a combination of communication networks such as the internet. Access device 140 may include one or more of any type of network interface (e.g., a Network Interface Card (NIC)) whether wired or wireless, such as an IEEE802.11 Wireless Local Area Network (WLAN) wireless interface, a worldwide interoperability for microwave access (Wi-MAX) interface, an ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a bluetooth interface, a Near Field Communication (NFC) interface, and so forth.
In one embodiment of the present description, the other components of the computing device 100 described above and not shown in FIG. 1 may also be connected to each other, such as by a bus. It should be understood that the block diagram of the computing device architecture shown in FIG. 1 is for purposes of example only and is not limiting as to the scope of the description. Those skilled in the art may add or replace other components as desired.
Computing device 100 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., tablet, personal digital assistant, laptop, notebook, netbook, etc.), a mobile phone (e.g., smartphone), a wearable computing device (e.g., smartwatch, smartglasses, etc.), or other type of mobile device, or a stationary computing device such as a desktop computer or PC. Computing device 100 may also be a mobile or stationary server.
Fig. 5 is a schematic flow chart illustrating a flow prediction model implementation method and a flow prediction method according to an embodiment of the present specification. As shown in fig. 5, the steps of the method in the figure correspond to the embodiments in fig. 2 and 3, and the embodiments in fig. 2 and 3 are described below with reference to fig. 5.
Wherein the processor 120 may perform the steps of the method shown in fig. 2. Fig. 2 is a schematic flow chart diagram illustrating an implementation method of a flow prediction model according to an embodiment of the present description, including step 202, step 204, and step 206.
Step 202: the method comprises the steps of obtaining actual flow of a set time period and actual flow of a plurality of time periods which are continuously spaced for a set time length before the set time period, and obtaining a first predicted flow of the set time period and a first predicted flow of the plurality of time periods by using a first prediction method, wherein the set time period is a time period before a current time period.
In one implementation, the traffic is traffic, for example, actual traffic data before the current time period of today is sampled, and every 30min is used as a time period to obtain the actual traffic before the current time period of today. For example, the real traffic before the present time period of today includes the actual traffic of the (t-1) time period, the actual traffic of the (t-2) time period, and the actual traffic of the (t-3) time period, where (t-1) is the sequence number of the last time period of the present time period, (t-2) is the sequence number of the last time period of the (t-1) time period, and (t-3) is the sequence number of the last time period of the (t-2) time period.
Step 204: the predicted error amount of the set time period is obtained based on the actual flow rate of the set time period and the first predicted flow rate of the set time period, and the predicted error amount sequence of the plurality of time periods is obtained based on the actual flow rates of the plurality of time periods and the first predicted flow rates of the plurality of time periods.
In one implementation, the first prediction method may use an existing data prediction method to perform data prediction to obtain the first predicted flow rate. For example, the first predicted traffic may include the first predicted traffic for a period (t-2), the first predicted traffic for a period (t-1), the first predicted traffic for a period (t + 1); accordingly, taking the actual flow rates of all time periods prior to the current time period today, a sequence of error amounts (f-1, f-2, f-3 …) for all time periods prior to the current time period can be obtained.
Step 206: training a flow prediction model that relates the sequence of prediction error amounts for the plurality of periods to the prediction error amount for the set period.
In one implementation, a deep learning algorithm may be used to generate a prediction model, and the prediction error amount sequences of the multiple time periods and the prediction error amount of the set time period are used as inputs to perform training, so that the prediction error amount sequences of the multiple time periods and the prediction error amount of the set time period are associated with each other, and a model factor of the flow prediction model is obtained through training. Compared with the existing prediction method, the flow prediction model disclosed by the embodiment of the description improves the accuracy of flow prediction by sampling actual flow data and forming an error sequence with the conventional prediction method to perform rolling prediction on the flow error in the next time period.
In one implementation, the trained traffic prediction model is stored in disk for use in subsequent traffic prediction.
Wherein the processor 120 may also perform the steps of the method shown in fig. 3. Fig. 3 is a schematic flow chart diagram illustrating a flow prediction method according to an embodiment of the present description, including step 302, step 304, and step 306.
Step 302: and calling a flow prediction model.
The trained traffic prediction model can be stored in a disk in advance, and can be directly called in the subsequent traffic prediction.
Step 304: and updating the time interval sequence by taking the (t +1) time interval as the set time interval, and obtaining the prediction error quantity of the (t +1) time interval through the flow prediction model, wherein t is the serial number of the current time interval, and (t +1) is the serial number of the next time interval of the current time interval.
The time period sequence can be updated by inputting the current time period, and the model predicts the flow error value, so that the prediction can be updated in a rolling manner.
Step 306: and obtaining a second predicted flow rate of the (t +1) time period according to the predicted error amount of the (t +1) time period and the first predicted flow rate of the (t +1) time period predicted by the first prediction method.
Assuming that the model predicted flow error value is F0, the final output predicted flow F is F0+ FBI,fBIThe predicted flow obtained by the data fitting method is adopted.
The model factors that invoke the traffic prediction model in one implementation are shown in table 1.
In one implementation, the actual flow in the current time period is calculated by calling a flow prediction model to obtain: calling a flow prediction model, updating the time period sequence by taking the current time period as the set time period, and obtaining the prediction error amount of the current time period through the flow prediction model; and obtaining a second predicted flow of the current time period according to the prediction error amount of the current time period and the first predicted flow of the current time period predicted by the first prediction method, and taking the second predicted flow of the current time period as the actual flow of the current time period.
In one implementation, the actual flow rate in the current time period is obtained by fitting the flow rate attribute characteristics in the current time period, for example, the actual flow rate in the current time period may be obtained by calculation according to the average flow rate before the current time in the current time period, or may be obtained by calculation according to the average call duration in the last time period of the current time period, the man-hour utilization rate in the last time period, and the number of shift workers in the current time period.
TABLE 1 model factors for flow prediction models
In the flow prediction model in the embodiment of the description, the actual flow data is sampled and an error sequence is formed with a conventional prediction method to perform rolling prediction on the flow in the next time period, so that compared with the existing prediction method such as a data fitting method, the accuracy of flow prediction is improved.
Wherein the processor 120 may perform the steps of the method shown in fig. 4. Fig. 4 is a schematic flow chart diagram illustrating a traffic scheduling method according to an embodiment of the present specification, including step 402, step 404, and step 406.
Wherein the processor 120 may also perform the steps of the method shown in fig. 4. Fig. 4 is a schematic flow chart diagram illustrating a traffic scheduling method according to an embodiment of the present specification, including step 402, step 404, and step 406.
Step 402: and acquiring a second predicted flow of a (t +1) time period obtained by a pre-established flow prediction model, wherein t is the sequence number of the current time period, and (t +1) is the sequence number of the next time period of the current time period.
Step 404: and determining a grading proportion according to the second predicted flow in the (t +1) time period, the production energy in the (t +1) time period, the proportion of automatic shunting to manual operation, the target call completing rate and the shunting self-service non-acceptance proportion.
In the flow automatic scheduling process, a user dials a customer service telephone to describe a problem, if the resolution of a standard problem corresponding to the problem is greater than a reference value, the problem is output to the robot for self-service, and if not, the problem is output to a manual seat service. The channel scheduling upgrade is to forcibly output the traffic of the self-service traffic scheduled and output by the channel to the manual service according to the upgrade proportion, for example, 30% of the traffic scheduled and output to the robot self-service by the channel is forcibly output to the manual seat service in 30% of the total traffic scheduled and output to the robot self-service by the channel in the time period. The channel scheduling degradation is to schedule and output artificial telephone traffic to the channel, and self-help is forcibly output according to the degradation proportion. For example, 30% of traffic is degraded at 10-11 points, and the robot self-service is forced to be output by the channel scheduling output artificial traffic 30% of the total time in the time period.
In one embodiment, the formula for the hierarchical scheduling is derived as follows:
it is known that: m is the rolling predicted inflow, N is the diversion self-service non-acceptance ratio, K is the productivity, L is the target call-on rate, and T is the ratio of automatic diversion to manual.
The evidence is proved that the flow which is automatically distributed to the manual service and the self-service flow and the carrying gap are respectively set as follows without going through the upgrading and downgrading process: manflow, Botflow, ResponseGap, then
ManFlow=M*(1-T)+M*T*N
BotFlow=M*T*(1-N)
A. When the BotFlow is greater than 0, an upgrade operation may be performed, that is, the traffic automatically shunted to the self-service is partially upgraded to the manual service, assuming that the upgrade ratio is R1,
ResponseGap=K-ManFlow*L
BotFlow*R1=ResponseGap,
B. When BotFlow is less than 0, a destage operation may be performed, i.e., partially destaging to self-service traffic that is automatically shunted to manual service, assuming the proportion of destage is R2,
ResponseGap=ManFlow*L-K
M*(1-T)*(1-N)*R2=ResponseGap
The productivity estimation is that a telephone operator estimated based on the scheduling data can take over telephone traffic, and the productivity estimation can be obtained by calculation according to the number of the scheduling personnel in the next time period and the commitment amount in the next time period.
Step 406: and carrying out the level-adjusting scheduling of the flow according to the level-adjusting proportion.
In one implementation, the step of scheduling the traffic according to the step-rate ratio includes:
step 4062: acquiring problem description information of a user, and determining a corresponding standard problem according to the problem description information.
In an embodiment, when the traffic is traffic, the problem description information may be obtained in a manner of converting speech recognition into text, and then the corresponding standard problem is found according to semantic similarity calculation or keyword search, for example, a pre-established problem recognition model may be called, the similarity between the problem description information and each standard problem is calculated through the problem recognition model, and then the standard problem corresponding to the problem description information is determined according to a magnitude relationship between the similarity and each standard problem.
Step 4064: and determining a skill group, a self-service solution rate and a solution rate reference value corresponding to the standard problem according to a problem and skill mapping table, and distributing the user to self-service or manual service according to the relation between the self-service solution rate and the solution rate reference value, wherein the problem and skill mapping table stores the skill group, the self-service solution rate and the solution rate reference value corresponding to each standard problem.
The self-service solution rate is the solution rate of a standard problem calculated by off-line analysis by adopting the existing method (such as a data fitting method) in a self-service channel. The solution rate reference value is a solution rate threshold value of the problem set according to expert experience in self-service of the robot. The standard problems, skill groups, self-help solution rate and solution rate reference values in the problem and skill mapping table can be updated regularly or irregularly according to the change of the conditions in the operation process.
Step 4066: and forcibly regulating the flow which is distributed to the self-service or manual service according to the regulation proportion.
In an embodiment, a personalized upgrade and downgrade service may be provided for a user according to a user level to improve experience of a high-quality customer, for example, identity information of the user may be identified by a mobile phone number or user name information of the user, and a user database is queried according to the identity information to determine the user level, where the identity information of the user and a corresponding level are stored in the user database. When the upgrade scheduling of the traffic is performed according to the ranking proportion, the upgrade scheduling of the traffic of users whose level is higher than the level threshold (such as VIP level or gold level users) may be performed preferentially. When the traffic is demoted and scheduled according to the grading proportion, the traffic of users (such as non-VIP level or brandy level users) with the level lower than the level threshold value is preferentially demoted and scheduled.
The traffic scheduling scheme of the embodiment of the description performs rolling prediction on the traffic of the next time period, and meanwhile, refers to the service acceptance rate, the capacity and the target call completing rate of the shunting robot to obtain the grading proportion, and then performs traffic ascending and degrading scheduling according to the generated proportion, so that accurate degrading scheduling and drainage are performed before a traffic peak comes.
The user needs to wait for a long time for self-service from the robot, so that the long-time waiting or call loss of the user is avoided, and the user experience is improved; before a traffic valley comes, the traffic of the robot service is distributed to manual service through accurate upgrading scheduling, full utilization of seat resources is guaranteed, and user help seeking experience is improved.
The embodiment of the specification accurately performs the upgrade scheduling by establishing the flow prediction model, avoids the problem of large error caused by manual scheduling by experience, and ensures user experience and reasonable utilization of resources.
Corresponding to the implementation method of the foregoing flow prediction model, this specification further provides an implementation apparatus embodiment of the flow prediction model, where the implementation apparatus embodiment may be implemented by software, or may be implemented by hardware or by a combination of hardware and software. The software implementation is taken as an example, and is formed by reading corresponding computer program instructions in the nonvolatile memory into the memory for operation through the processor of the device where the software implementation is located as a logical means. In terms of hardware, a hardware structure of the apparatus in which the apparatus for generating the event detection model in this specification is located may include a processor, a network interface, a memory, and a nonvolatile memory, and the apparatus in which the apparatus is located in the embodiment may also include other hardware according to an actual function of the event detection, which is not described in detail herein.
Referring to fig. 6, a block diagram of a traffic prediction scheduling system according to an embodiment of the present disclosure is provided. The traffic prediction degree system includes a traffic prediction model implementation device 700, a traffic prediction device 800, a traffic scheduling device 900, and a storage device 1000. The flow prediction model implementation device 700 is configured to generate a flow prediction model according to the training samples, store the generated model in the storage device 1000, and call the model from the storage device 1000 when the flow prediction device 800 performs flow prediction according to the current time period. The flow prediction model implementation device 700, the flow prediction device 800, the flow scheduling device 900, and the storage device 1000 will be described in detail below.
The implementation device 700 of the flow prediction model corresponds to the embodiment shown in fig. 2, and the device 700 includes:
a flow rate acquirer 710 configured to acquire an actual flow rate for a set period and actual flow rates for a plurality of periods that are continuously spaced by a set time period before the set period, and acquire a first predicted flow rate for the set period and first predicted flow rates for the plurality of periods obtained by a first prediction method, wherein the set period is a period before a current period;
an error sequence generator 720 configured to derive the predicted error amount for the set period based on the actual flow rate for the set period and the first predicted flow rate for the set period, and derive the sequence of predicted error amounts for the plurality of periods based on the actual flow rates for the plurality of periods and the first predicted flow rates for the plurality of periods;
a model trainer 730 configured to train a traffic prediction model that associates the sequence of prediction error amounts for the plurality of periods with the prediction error amount for the set period.
The flow prediction apparatus 800 corresponds to the embodiment shown in fig. 3, and the apparatus 800 includes:
a model invoker 810 configured to invoke a traffic prediction model;
an error calculator 820 configured to update the time period sequence with a (t +1) time period as the set time period, and obtain a prediction error amount of the (t +1) time period through the traffic prediction model, wherein t is a serial number of the current time period, and (t +1) is a serial number of a next time period of the current time period;
and a flow rate predictor 830 configured to obtain a second predicted flow rate for the (t +1) period from the prediction error amount for the (t +1) period and the first predicted flow rate for the (t +1) period predicted by the first prediction method.
The traffic scheduling apparatus 900 corresponds to the embodiment shown in fig. 4, and the apparatus 900 includes:
a predicted traffic acquirer 910 configured to acquire a second predicted traffic of a (t +1) time period obtained by a traffic prediction method, where t is a sequence number of a current time period and (t +1) is a sequence number of a next time period of the current time period;
a proportion calculator 920 configured to determine a grading proportion according to the second predicted flow rate in the (t +1) period, the production capacity in the (t +1) period, the proportion of automatic diversion to manual diversion, the target call completing rate and the proportion of self-service non-acceptance diversion;
a scheduler 930 configured to perform the level-adjusting scheduling of the traffic according to the level-adjusting ratio.
In an alternative embodiment, the ratio calculator comprises:
an upgrade calculation unit configured to upgrade the automatically shunted traffic to self-service to the human service at least partially when the automatically shunted traffic to the production capacity of the human service multiplied by the target turn-on rate is less than a (t +1) period, an upgrade ratio R1The calculation formula of (2) is as follows:
m is the second predicted flow rate in the (T +1) time period, N is the diversion self-service non-acceptance proportion, K is the productivity in the (T +1) time period, L is the target call completing rate, and T is the proportion from automatic diversion to manual diversion.
In an alternative embodiment, the ratio calculator comprises:
a degradation calculation unit configured to degrade the traffic automatically shunted to artificial to self-service traffic at least partially when the product of the automatic shunted to artificial service and the target call-in rate is greater than the production capacity for a (t +1) period, a degradation ratio R2The calculation formula of (2) is as follows:
m is the second predicted flow rate in the (T +1) time period, N is the diversion self-service non-acceptance proportion, K is the productivity in the (T +1) time period, L is the target call completing rate, and T is the proportion from automatic diversion to manual diversion.
In an optional embodiment, the scheduler comprises:
the problem determination unit is configured to acquire problem description information of a user and determine a corresponding standard problem according to the problem description information;
the automatic distribution module is configured to determine a skill group, a self-service solution rate and a solution rate reference value corresponding to the standard problem according to a problem and skill mapping table, and pre-distribute the user to self-service or manual service according to the relation between the self-service solution rate and the solution rate reference value, wherein the problem and skill mapping table stores the skill group, the self-service solution rate and the solution rate reference value corresponding to each standard problem;
and the grading module is configured to forcibly grade the flow pre-flowed to self-service or manual service according to the grading proportion.
The devices, modules or units illustrated in the above embodiments may be specifically implemented by computer chips or entities,
or by a product having a certain function. A typical implementation device is a computer, which may take the form of a personal computer, laptop computer, cellular telephone, camera phone, smart phone, personal digital assistant, media player, navigation device, email messaging device, game console, tablet computer, wearable device, or a combination of any of these devices.
The implementation process of the functions and actions of each unit in the above device is specifically described in the implementation process of the corresponding step in the above method, and is not described herein again.
For the device embodiments, since they substantially correspond to the method embodiments, reference may be made to the partial description of the method embodiments for relevant points. The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution in the specification. One of ordinary skill in the art can understand and implement it without inventive effort.
An embodiment of the present specification further provides a computing device, including a memory, a processor, and computer instructions stored on the memory and executable on the processor, where the processor implements the traffic scheduling method when executing the instructions.
An embodiment of the present specification further provides a computer readable storage medium, which stores computer instructions, and when the instructions are executed by a processor, the instructions implement the steps of the traffic scheduling method.
The above is an illustrative scheme of a computer-readable storage medium of the present embodiment. It should be noted that the technical solution of the storage medium belongs to the same concept as the technical solution of the above-mentioned automatic testing method, and details that are not described in detail in the technical solution of the storage medium can be referred to the description of the technical solution of the above-mentioned automatic testing method.
The computer instructions comprise computer program code which may be in the form of source code, object code, an executable file or some intermediate form, or the like. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like. It should be noted that the computer readable medium may contain content that is subject to appropriate increase or decrease as required by legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer readable media does not include electrical carrier signals and telecommunications signals as is required by legislation and patent practice.
It should be noted that, for the sake of simplicity, the foregoing method embodiments are described as a series of acts or combinations, but those skilled in the art should understand that the present disclosure is not limited by the described order of acts, as some steps may be performed in other orders or simultaneously according to the present disclosure. Further, those skilled in the art should also appreciate that the embodiments described in this specification are preferred embodiments and that acts and modules referred to are not necessarily required for this description.
In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.
The preferred embodiments of the present specification disclosed above are intended only to aid in the description of the specification. Alternative embodiments are not exhaustive and do not limit the invention to the precise embodiments described. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the specification and its practical application, to thereby enable others skilled in the art to best understand the specification and its practical application. The specification is limited only by the claims and their full scope and equivalents.