CN114663107A - Customer complaint risk prediction method, apparatus, computer device and storage medium - Google Patents

Customer complaint risk prediction method, apparatus, computer device and storage medium Download PDF

Info

Publication number
CN114663107A
CN114663107A CN202011542126.3A CN202011542126A CN114663107A CN 114663107 A CN114663107 A CN 114663107A CN 202011542126 A CN202011542126 A CN 202011542126A CN 114663107 A CN114663107 A CN 114663107A
Authority
CN
China
Prior art keywords
waybill
characteristic
time
customer complaint
data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202011542126.3A
Other languages
Chinese (zh)
Inventor
张策
江洋
潘舒静
聂仁桐
董珊
陈志文
黎碧君
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
SF Technology Co Ltd
Original Assignee
SF Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by SF Technology Co Ltd filed Critical SF Technology Co Ltd
Priority to CN202011542126.3A priority Critical patent/CN114663107A/en
Publication of CN114663107A publication Critical patent/CN114663107A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/01—Customer relationship services
    • G06Q30/015—Providing customer assistance, e.g. assisting a customer within a business location or via helpdesk
    • G06Q30/016—After-sales
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06311—Scheduling, planning or task assignment for a person or group
    • G06Q10/063114—Status monitoring or status determination for a person or group
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/083—Shipping

Landscapes

  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Physics & Mathematics (AREA)
  • Development Economics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Operations Research (AREA)
  • Tourism & Hospitality (AREA)
  • Quality & Reliability (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The application relates to a customer complaint risk prediction method, a customer complaint risk prediction device, a computer device and a storage medium. The method comprises the following steps: acquiring waybill information of a waybill to be predicted, wherein the waybill information comprises a waybill identifier; acquiring first waybill circulation characteristics of the waybill to be predicted according to waybill circulation data associated with the waybill identification at the current time; and obtaining a customer complaint risk prediction result of the waybill to be predicted according to the first waybill characteristic. By adopting the method, the possibility of the occurrence of the customer complaint of the waybill can be predicted in real time in the waybill circulation process, so that in-service monitoring is realized, and corresponding rescue measures can be taken in time to reduce the loss caused by the customer complaint.

Description

Customer complaint risk prediction method, apparatus, computer device, and storage medium
Technical Field
The present application relates to the field of logistics technology, and in particular, to a method, an apparatus, a computer device, and a storage medium for predicting a customer complaint risk.
Background
With the development of the logistics industry, express delivery service plays an increasingly important role in the society. Express delivery aging is an important concept in express delivery services, and represents the time interval between the receipt of a good by an express company and the sign-in of the good by a recipient. Express delivery ageing often influences the evaluation of customer to express delivery service, but because the whole circulation process of express delivery is very complicated, can lead to express delivery ageing to receive the influence because some ineffectiveness factors, for example exceed commitment ageing, and these influences can directly bring the visitor and complain the risk, increase cost.
At present, the scheme for rescuing the express mail with the risk of the effective complaint of the customer mainly aims at carrying out optimal transportation capacity recommendation on an overtime effective freight bill, belongs to post monitoring and has the problem of untimely rescue.
Disclosure of Invention
In view of the above, it is necessary to provide a customer complaint risk prediction method, apparatus, computer device, and storage medium that improve timeliness of rescue in response to the above technical problems.
A customer complaint risk prediction method, the method comprising:
acquiring waybill information of a waybill to be predicted, wherein the waybill information comprises a waybill identifier;
acquiring first waybill transfer characteristics of the waybill to be predicted according to waybill transfer data associated with the waybill identification until the current time;
and obtaining the customer complaint risk prediction result of the waybill to be predicted according to the first waybill characteristic.
A customer complaint risk prediction device, the device comprising:
the acquiring module is used for acquiring waybill information of the waybill to be predicted, wherein the waybill information comprises a waybill identifier;
the processing module is used for obtaining a first waybill characteristic of the waybill to be predicted according to waybill circulation data associated with the waybill identification until the current time;
and the prediction module is used for obtaining the customer complaint risk prediction result of the waybill to be predicted according to the first waybill characteristic.
A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:
acquiring waybill information of a waybill to be predicted, wherein the waybill information comprises a waybill identifier;
acquiring first waybill transfer characteristics of the waybill to be predicted according to waybill transfer data associated with the waybill identification until the current time;
and obtaining the customer complaint risk prediction result of the waybill to be predicted according to the first waybill characteristic.
A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:
acquiring waybill information of a waybill to be predicted, wherein the waybill information comprises a waybill identifier;
acquiring a first waybill characteristic of the waybill to be predicted according to waybill circulation data associated with the waybill identification at the current time;
and obtaining the customer complaint risk prediction result of the waybill to be predicted according to the first waybill characteristic.
The method, the device, the computer equipment and the storage medium for forecasting the risk of the customer complaint acquire the waybill information of the waybill to be forecasted, wherein the waybill information comprises a waybill identifier; acquiring first waybill circulation data associated with the waybill identification at the current time to obtain first waybill characteristics of the waybill to be predicted; and obtaining a customer complaint risk prediction result of the freight bill to be predicted according to the first freight bill characteristic. Therefore, the possibility of the occurrence of the customer complaint of the waybill can be predicted in real time in the waybill circulation process, so that in-service monitoring is realized, corresponding rescue measures can be taken in time, and the loss caused by the customer complaint is reduced.
Drawings
FIG. 1 is a schematic flow chart illustrating a method for predicting a complaint risk according to an embodiment;
FIG. 2 is a flowchart illustrating the steps of obtaining a complaint risk prediction result of a waybill to be predicted according to a first waybill characteristic in one embodiment;
FIG. 3 is a flowchart illustrating a training method of the customer complaint risk prediction model according to an embodiment;
FIG. 4 is a block diagram showing the structure of a customer complaint risk prediction apparatus according to an embodiment;
FIG. 5 is a diagram illustrating an internal structure of a computer device according to an embodiment;
FIG. 6 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
In an embodiment, as shown in fig. 1, a customer complaint risk prediction method is provided, and this embodiment is illustrated by applying the method to a server, it is to be understood that the method may also be applied to a terminal, and may also be applied to a system including the terminal and the server, and is implemented by interaction between the terminal and the server. In this embodiment, taking the age complaint type as an example, the method includes the following steps S102 to S106.
S102, acquiring waybill information of the waybill to be predicted, wherein the waybill information comprises a waybill identifier.
The freight note to be predicted is a freight note in circulation, and specifically can be a freight note which arrives at a transit station or a network point but has not finished dispatching. The waybill identifier corresponds to the waybill one to one and is used for distinguishing different waybill, and specifically can be a waybill number.
And S104, obtaining a first waybill characteristic of the waybill to be predicted according to the waybill circulation data associated with the waybill identification at the current time.
The current time is the current time for forecasting the customer complaint risk of the freight note, and a plurality of moments (for example, every hour between 10 and 22) can be taken as the time for forecasting the customer complaint risk of the freight note in each day. The waybill circulation data associated with the waybill identification at the current time can be obtained through the real-time routing information and the real-time operation data of the waybill, and the real-time state of the waybill in the circulation process can be reflected. The circulation data includes data which may cause an aging complaint, and may include, for example, but not limited to, mail related data, aging related data, retention data, check piece data, dispatch data, and the like.
And S106, obtaining a customer complaint risk prediction result of the waybill to be predicted according to the first waybill characteristic.
The first waybill characteristic can reflect the possibility that the waybill will suffer from the aged customer complaint from the real-time dimension, so that the aged customer complaint risk of the waybill is predicted according to the first waybill characteristic, and the fact that in-situ monitoring is facilitated. Specifically, the first waybill characteristic can be input into the customer complaint risk prediction model, and a customer complaint risk prediction result of the waybill to be predicted is obtained. The customer complaint risk prediction model is obtained by training a flow waybill which is known whether customer complaints occur before a preset time point as a training sample, wherein the input of the customer complaint risk prediction model is a first waybill characteristic of the waybill, the output of the customer complaint risk prediction model is a result of the customer complaint risk prediction of the waybill.
In one embodiment, the customer complaint risk prediction result comprises the prediction probability that the freight note will have an aged customer complaint, when the prediction probability exceeds a threshold value, the aged customer complaint risk is judged to be high, and at the moment, a worker in the current site can be reminded to take rescue measures in time, for example, the worker communicates with related personnel in advance through an intelligent outbound mode, the possibility that the related personnel initiate the aged customer complaint is reduced, and therefore cost reduction and efficiency improvement are achieved.
In the method for predicting the risk of the customer complaint, waybill information of a waybill to be predicted is obtained, wherein the waybill information comprises a waybill identifier; acquiring first waybill circulation characteristics of the waybill to be predicted according to waybill circulation data associated with the waybill identification at the current time; and obtaining a customer complaint risk prediction result of the waybill to be predicted according to the first waybill characteristic. Therefore, the possibility of the occurrence of the customer complaint of the waybill can be predicted in real time in the waybill circulation process, so that in-service monitoring is realized, corresponding rescue measures can be taken in time, and the loss caused by the customer complaint is reduced.
In one embodiment, the first waybill characteristic of the waybill to be predicted can be directly obtained from the waybill circulation data, and can also be obtained by carrying out statistical calculation on related data in the waybill circulation data.
Waybill flow data may include, but is not limited to, the following: sending time, promised delivery time, first arrival time at the current site, second arrival time at the current city, retention data in a time period from the sending time to the current time, checking data in a time period from the sending time to the current time, and hastening data in a time period from the sending time to the current time. The first waybill feature may include, but is not limited to, the following features: retention characteristics, current site retention time, current city retention time, piece checking characteristics, dispatching characteristics, circulation duration, time difference between the last piece checking time and the piece sending time, and time difference between the first piece checking time and the promised delivery time.
The mail sending time can indicate the time for receiving the mail. The committed delivery time may be the latest delivery time, for example, if an invoice corresponds to a mail time of 2020/12/612: 00 and the corresponding committed or expected time period is 48 hours, then the committed delivery time for the invoice is 2020/12/812: 00. The current site is an operation site where the waybill is currently located, such as a website or a transit site. The time interval from the mail sending time to the current time corresponds to the real-time circulation process of the waybill.
The retention data comprises information used for representing whether the waybill is retained in the real-time circulation process and corresponding retention times. Retention characteristics are determined from the retention data, including a "whether retained" characteristic and a "number of retains" characteristic. For example, if a waybill stays for 1 time in the real-time circulation process, the "whether the waybill stays" is characterized as "yes", and the "staying times" is characterized as "1 time".
The first arrival time is the time when the waybill arrives at the current site, and the residence time of the current site is determined according to the current time and the first arrival time, specifically, the time difference between the current time and the first arrival time. For example, if a waybill corresponds to a first arrival time of 2020/12/88: 00 and a current time of 2020/12/810: 00, the "current site dwell time" characteristic of the waybill is "2 hours".
The second arrival time is the time when the waybill arrives at the current city, the current city is the city where the current site is located, and the residence time of the current city is determined according to the current time and the second arrival time, specifically, the time difference between the current time and the second arrival time. For example, if the second arrival time corresponding to a waybill is 2020/12/87: 00 and the current time is 2020/12/810: 00, the "current city stay time" of the waybill is characterized as "3 hours".
The piece checking data comprise information for representing whether the waybill is checked in the real-time circulation process and corresponding piece checking times. The piece checking characteristics are determined according to piece checking data and comprise a characteristic of 'whether to check pieces' and a characteristic of 'number of times to check pieces'. For example, if a waybill is checked for 1 time in the real-time circulation process, the "whether to check for a piece" feature of the waybill is yes, and the "number of times to check for a piece" feature of the waybill is 1.
The dispatching data comprises information for representing whether the freight note is dispatched in the real-time circulation process and corresponding dispatching times. The forcing characteristics are determined according to the forcing data and comprise a 'whether forcing' characteristic and a 'forcing frequency' characteristic. For example, if a delivery order is dispatched 1 time in the real-time circulation process, the "whether to dispatch" feature of the delivery order is yes, and the "dispatch times" feature is 1 time.
The circulation duration is determined according to the current time and the sending time, specifically, the time difference between the current time and the sending time. For example, if the consignment time corresponding to a waybill is 2020/12/610: 00 and the current time is 2020/12/810: 00, the "circulation duration" of the waybill is "48 hours".
And determining the time difference between the last piece checking time and the piece sending time according to the piece checking data and the piece sending time. The time difference between the first piece checking time and the promised delivery time is determined according to the piece checking data and the promised delivery time. It can be understood that the waybill may be checked for a plurality of times in the real-time circulation process, and the piece checking data may further include the time of each piece checked by the waybill. For example, the parcel sending time corresponding to one waybill is 2020/12/610: 00, 2 times of parcel checking are carried out in the real-time circulation process, the corresponding parcel checking times are 2020/12/710: 00 and 2020/12/722: 00 respectively, namely the first parcel checking time corresponding to the waybill is 2020/12/710: 00, the last parcel checking time is 2020/12/722: 00, and the characteristic of the time difference between the last parcel checking time and the parcel sending time of the waybill is 36 hours; if the committed delivery time corresponding to the waybill is 2020/12/812: 00, the "time difference between the first piece-checking time and the committed delivery time" of the waybill is characterized as "-12 hours".
In the embodiment, the real-time data generated by the waybill circulation at the current time is combined with the associated operation data such as checking, dispatching and the like to form a specific scene, and the first waybill characteristic of the real-time dimension is generated according to the data in the specific scene. It should be noted that the first waybill feature is not limited to the feature mentioned in the above embodiment, and other scenarios may be formed according to actual requirements, so as to obtain real-time dimensional features in other scenarios.
In one embodiment, the waybill information further includes: first parameter data for identifying a sender and second parameter data for identifying a receiver; obtaining a second waybill characteristic of the waybill to be predicted according to first historical waybill data associated with the first parameter data and second historical waybill data associated with the second parameter data; the second waybill feature includes: and the consignment rate is determined according to the first historical waybill data, and the consignment rate is determined according to the second historical waybill data.
The historical waybill data refers to waybill data in a historical period (for example, the last six months, three months and the like), each waybill data corresponds to one historical waybill, and each waybill data is associated with one sender identification data and one receiver identification data. First historical waybill data associated with the first parameter data and second historical waybill data associated with the second parameter data can be obtained from the historical waybill data. The second waybill characteristic of the waybill to be predicted can be obtained by carrying out statistical calculation on related data in historical waybill data, wherein the related data comprises a consignment rate and a consignment rate.
In one embodiment, the outgoing customer rate and the incoming customer rate may be obtained by: obtaining a sending volume associated with the first parameter data and a sending customer volume for sending a customer in the sending volume according to first historical waybill data associated with the first parameter data, and obtaining a sending customer rate according to the ratio of the sending customer volume to the sending volume; according to the second historical waybill data related to the first parameter data, the receiving amount related to the first parameter data and the receiving and customer complaint amount of the customer complaints in the receiving amount are obtained, and according to the ratio of the receiving and customer complaint amount to the receiving amount, the receiving and customer complaint rate is obtained.
It should be noted that the identification data of the sender (or the receiver) in this document only relates to the contact telephone of the sender (or the receiver) (encryption processing), and does not include other sensitive data relating to the privacy information of the client. Furthermore, historical waybill data, outgoing customer rates, incoming customer rates, and the like do not relate to customer privacy information as well.
In one embodiment, the mail data, the mail finding data, the call urging data and the complaint data associated with the first parameter data may be obtained according to the first historical waybill data, and specifically, the following data may be obtained: the total sending quantity (J) in the historical period, the total sending customer quantity (K) in which the customer occurs in the total sending quantity, and the first sending quantity (J) which is checked for a times in the total sending quantity1) The first sending customer quantity (K) in which the customer has occurred1) The second sending quantity (J) is sent for b times in the total sending quantity2) And a second sending customer quantity (K) in which a customer has occurred in the second sending quantity2) And the third sending quantity (J) of c times of checked pieces after reaching the destination network point in the total sending quantity3) And a third delivery customer quantity (K) in which a customer has occurred in the third delivery quantity3) And the fourth sending quantity (J) which is urged to be sent d times after the destination network point is reached in the total sending quantity4) A fourth sending customer quantity in which a customer has occurred in the fourth sending quantity(K4) And the fifth sending quantity (J) which reaches the destination network point and is checked for e times after sending is selected from the total sending quantity5) And a fifth sending volume (K) in which a customer complaint occurs among the fifth sending volumes5) The sixth sending quantity (J) which reaches the destination network point in the total sending quantity and is urged to be sent for f times after sending6) A sixth mail volume (K) in which a customer has occurred6). The values a, b, c, d, e, and f are positive integers, and the specific values may be set in combination with actual requirements, for example, the values may be 1 and 2, which is not limited herein.
Based on the first historical waybill data, the second waybill characteristic of the waybill to be predicted may include the following characteristics: total sending customer complaint rate (P) in historical period, sending customer complaint rate (P) of a times of checking1) Dispatch of delivery Parsley Rate (P) for b orders2) Pick-up customer complaint rate (P) of c times of check-up after arrival at destination site3) Delivery of delivery customer complaints d times after arrival at the destination site (P)4) Delivery customer complaint rate (P) of arrival at destination site and check for delivery e times5) Delivery customer complaint rate (P) of arriving at destination network site and urging delivery f times after delivery6). Each of the sending customer rate can be determined by the ratio of the sending customer volume to the corresponding sending volume, and the calculation method is as follows: p ═ K/J, P1＝K1/J1，P2＝K2/J2，P3＝K3/J3，P4＝K4/J4，P5＝K5/J5，P6＝K6/J6。
According to the second historical waybill data, the receiving data, the piece checking data, the dispatching data and the complaint data related to the second parameter data can be obtained, and the following data can be obtained specifically: total receiving amount (S) in historical time period, total receiving and customer calling amount (T) of customer calling in total receiving amount, and first receiving amount (S) of a times of checked piece in total receiving amount1) A first customer acceptance amount (T) at which a customer acceptance occurs in the first customer acceptance amount1) The second receiving amount is dispatched for b times in the total receiving amount (S)2) And a second customer acceptance amount (T) at which customer acceptance occurs among the second customer acceptance amounts2) In the total amount of receivedThird receiving amount of c times checked pieces after reaching destination network point (S)3) And a third customer acceptance amount (T) at which customer acceptance occurs in the third acceptance amount3) And the fourth receiving amount is dispatched d times after the destination network point is reached in the total receiving amount (S)4) And a fourth customer acceptance amount (T) at which customer acceptance occurs in the fourth acceptance amount4) The first receiving and sending quantity which reaches the destination network point and is checked for e times after sending is selected from the total receiving and sending quantity (S)5) Fifth receiving amount (T) at which customer complaints occur among the fifth receiving amounts5) The sixth receiving amount which reaches the destination network point and is dispatched f times after dispatching is obtained from the total receiving amount (S)6) And a sixth delivery amount (T) at which the customer complaint occurs in the sixth delivery amount6). The values a, b, c, d, e, and f are positive integers, and the specific values may be set in combination with actual requirements, for example, the values may be 1 and 2, which is not limited herein.
Based on the second historical waybill data, the second waybill characteristic of the waybill to be predicted may include the following characteristics: total incoming call rate (Q) in historical period, and incoming call rate (Q) of a times of checking1) Urging the customer to ask for a complaint b times (Q)2) Customer acceptance rate (Q) of c times of checking up after arrival at destination network site3) And the customer-receiving complaint rate (Q) is sent d times after the customer arrives at the destination network4) Customer acceptance rate (Q) of arrival at destination network site and check piece e times after delivery5) The customer receiving and telling rate (Q) of arriving at the destination network site and urging delivery for f times after delivery6). The customer receiving rate can be determined by the ratio of the customer receiving quantity to the corresponding customer receiving quantity, and the calculation mode is as follows: q ═ T/S, Q1＝T1/S1，Q2＝T2/S2，Q3＝T3/S3，Q4＝T4/S4，Q5＝T5/S5，Q6＝T6/S6。
In the above embodiment, the historical waybill data associated with the sender and the receiver is calculated separately, mutual interference between the sender and the receiver can be eliminated, the customer complaint data and the associated operation data such as checking and dispatching are combined to form a specific scene, and the second waybill feature of the dimension of the target object (including the sender and the receiver) is generated according to the data in the specific scene. It should be noted that the second waybill feature is not limited to the feature mentioned in the above embodiment, and other scenes may be formed according to actual requirements to obtain dimensional features of the target object in other scenes.
In one embodiment, the waybill information further includes: a third waybill feature. The third waybill feature is a feature related to consignment, age or service recorded on the waybill, and may include, but is not limited to, the following features: the type of the consignment, the aging type, the product code, whether to guarantee the price or not and whether to add value to the service or not.
In an embodiment, as shown in fig. 2, the step of obtaining a customer complaint risk prediction result of the waybill to be predicted according to the first waybill characteristic may specifically include the following steps S202 to S204.
And S202, determining target characteristics of the waybill to be predicted according to the first waybill characteristic, the second waybill characteristic and the third waybill characteristic.
The target feature refers to a final feature for prediction, and after the first waybill feature, the second waybill feature and the third waybill feature are obtained, the first waybill feature, the second waybill feature and the third waybill feature can be integrated to form the target feature. It will be appreciated that the first, second and third waybill features may each include a number of features, for example, the first, second and third waybill features include a number of features N1, N2 and N3 respectively, and the number of target features may be N1+ N2+ N3. In other embodiments, the characteristics included in the first waybill characteristic, the second waybill characteristic and the third waybill characteristic can be further screened, and a characteristic with relatively stronger prediction capability is screened out from the characteristics as the target characteristic.
And S204, predicting based on the target characteristics to obtain the customer complaint risk prediction result of the waybill to be predicted.
The first waybill characteristic, the second waybill characteristic and the third waybill characteristic respectively describe the possibility that the waybill will generate the timeliness customer complaint in the future from different dimensions, and the target characteristics obtained by combining the three dimensional characteristics can be used for more comprehensively and accurately predicting the waybill with the timeliness customer complaint risk in the later period, so that the prediction effect is improved.
In an embodiment, the step of performing prediction based on the target features to obtain a prediction result of the customer complaint risk of the waybill to be predicted may specifically be: and inputting the target characteristics into the trained customer complaint risk prediction model to obtain the customer complaint risk prediction result of the waybill to be predicted. As shown in fig. 3, the training method of the customer complaint risk prediction model includes the following steps S302 to S308.
S302, acquiring waybill information of the waybill sample and a corresponding sample label, wherein the waybill information comprises: the waybill label is used for indicating whether the waybill sample generates a customer complaint before a preset time point.
The waybill sample is a circulating waybill that is known whether a customer complaint occurred before a preset time point. In one embodiment, waybill data arriving at a network point (or a transit point) within a preset time period (e.g., a week) can be obtained, such as waybill data arriving at network points 2020/12/1-2020/12/7, multiple time points (e.g., 10:00, 16:00, 22:00) are selected each day as preset time points, and whether a customer complaint occurs before each preset time point in a waybill in a flow is queried, so that a waybill sample is obtained.
It is understood that the customer appeal status of the same waybill may be the same or different at different points in time. For example, assume that the sign-in time is 2020/12/118: 00 for waybill A, i.e., waybill A is the in-circulation waybill before 2020/12/118: 00. Assuming customer complaints occurred on waybill A at 2020/12/18: 00, then both waybill A occurred before 2020/12/110: 00 and 2020/12/116: 00; assuming that the waybill A has taken a customer complaint at 2020/12/112: 00, waybill A did not take a customer complaint before 2020/12/110: 00 and took a customer complaint before 2020/12/116: 00. Thus, from the manifest data for manifest A prior to 2020/12/110: 00 and 2020/12/116: 00, two manifest samples can be obtained.
In view of the above, a plurality of real-time dimension characteristics of the waybill can be obtained by setting a plurality of preset time points, richer information is provided for model prediction, and since the occurrence of customer complaint is generally a small probability event, a plurality of positive samples (i.e. the samples of the customer complaint) can be generated based on waybill data of the customer complaint by setting a plurality of preset time points, so that the number of the positive samples of the model is increased, and the improvement of the model prediction effect is facilitated.
S304, obtaining a first waybill characteristic of the waybill sample according to the waybill circulation data associated with the waybill identification of the waybill sample at the preset time point, and obtaining a second waybill characteristic of the waybill sample according to historical waybill data associated with the parameter data of the waybill sample. For the detailed description of this step, reference may be made to the foregoing embodiments, which are not repeated herein.
And S306, obtaining target characteristics of the waybill samples according to the first waybill characteristics, the second waybill characteristics and the third waybill characteristics of the waybill samples, inputting the target characteristics into a to-be-trained customer complaint risk prediction model, and obtaining prediction results corresponding to the waybill samples.
After the three-dimensional feature data of the first waybill feature, the second waybill feature and the third waybill feature are obtained, feature processing can be performed on the three-dimensional feature data to obtain a target feature which is used as training data of the model. The process of converting the original data into the training data of the model is called feature engineering, and the feature engineering is beneficial to obtaining better training data features and improving the performance of the model.
In an embodiment, the step of obtaining the target characteristic of each waybill sample according to the first waybill characteristic, the second waybill characteristic, and the third waybill characteristic of each waybill sample may specifically include the following steps: determining the information quantity of each characteristic in the first waybill characteristic, the second waybill characteristic and the third waybill characteristic; and screening out the features of which the information quantity meets the preset conditions from all the features according to the information quantity of each feature, and taking the features as the target features of each waybill sample.
The information amount of the feature is used to characterize the predictive ability of the feature, and it can be understood that the larger the information amount of the feature, the stronger the predictive ability of the feature, and the smaller the information amount of the feature, the weaker the predictive ability of the feature. The information amount meeting the preset condition may mean that the information amount exceeds a preset value, or the information amount exceeds a preset proportion of all the information amounts. The preset value and the preset proportion can be set according to actual requirements, and are not limited here. In other embodiments, the features with the information quantity meeting the preset conditions are screened out from all the features to be used as the features to be selected, the influence degree of each feature to be selected on the time-dependent customer complaints can be further analyzed, and the features with higher influence degree are selected out to be used as the target features.
In one embodiment, the step of determining the information content of each of the first waybill characteristic, the second waybill characteristic, and the third waybill characteristic may include the steps of: for any one of the first waybill characteristic, the second waybill characteristic and the third waybill characteristic, the following processing is carried out: acquiring each grouping feature corresponding to the feature, and acquiring the weight of each grouping feature according to the quantity proportion of the corresponding positive sample to all the positive samples and the quantity proportion of the corresponding negative sample to all the negative samples, wherein the positive samples and the negative samples respectively represent waybill samples with and without customer complaints; and determining the information quantity of the features according to the weights of all the grouped features.
The characteristics can be grouped through data analysis, for example, for the characteristics of the 'aging type', if the aging type comprises three types of express, ordinary and slow pieces, the aging type can be divided into three grouping characteristics of 'express', 'ordinary' and 'slow piece'; for the feature of "check if" it can be divided into two grouping features of "yes" and "no"; for the feature of "total incoming complaint rate", it can be divided into [0, X1)、[X1,X2) And [ X ]2,1]Three grouping features, wherein X1And X2The specific value of (b) can be set by combining with the actual situation, and is not limited here. Other features may be grouped in a similar manner and will not be described in further detail herein.
For any grouping feature, the corresponding positive sample is a sample in which the customer complaint occurs in the samples with the grouping feature, that is, all the positive samples contain the positive sample of the grouping feature; the corresponding negative sample is a sample in which no customer complaint occurs in the samples with the grouping feature, that is, all negative samples contain the negative sample of the grouping feature. The weight of a group feature is used to characterize the likelihood that the group feature is responsive to an aging complaint, it being understood that the greater the weight of a group feature, the greater the likelihood that the group feature is responsive to an aging complaint, and the lesser the weight of a group feature, the lesser the likelihood that the group feature is responsive to an aging complaint.
For any feature, the weight (represented by woe) of any of the grouped features in the feature can be calculated by the following formula:
Figure BDA0002854999800000111
where i represents the ith grouping feature of the features, woeiWeight of the ith packet feature, # yiIndicates the number of positive samples corresponding to the ith packet signature, # niIndicates the number of negative samples, # y, corresponding to the ith grouping featureTIndicates the number of all positive samples, # nTDenotes the number of all negative examples, pyiRepresenting the number ratio of positive samples to all positive samples, pn, for i grouped signaturesiRepresenting the ratio of the negative examples corresponding to the i grouping features to the number of all negative examples.
The amount of information (denoted by iv) for any of the grouped features in the feature can be calculated by the following formula:
Figure BDA0002854999800000112
Figure BDA0002854999800000121
wherein iv isiThe amount of information characterizing the ith packet. The information amount of the feature can be obtained by adding the information amounts of all the grouping features thereof.
After the target features are obtained, the target features can be coded before being input into the customer complaint risk prediction model to be trained. For the discrete features, one-hot (one-hot) or label-encoding (label-encoding) can be adopted for encoding. In this embodiment, a label-encoding mode is selected to encode the discrete type features, for example, if the aging type includes three types, namely, express, ordinary and slow pieces, the aging type features may be divided into three grouping features, namely, express, ordinary and slow pieces, and the encoding is 1, 2 and 3 respectively.
And S308, adjusting parameters of the customer complaint risk prediction model to be trained based on the prediction result and the sample label until a training end condition is met, and obtaining the trained customer complaint risk prediction model.
In one embodiment, the customer complaint risk prediction model is an Xgboost model, which has good characteristics, such as missing value processing, parallel computation, and the like, and the embodiment optimizes the adjustment parameters for the unbalanced data set in the Xgboost model, and obtains the optimal parameters of the adjustment parameters through a large number of tests.
For the training set, as described in the foregoing embodiment, the real-time performance is realized by means of the preset time point, and when calculating the dimensional features of each waybill in the training set, the feature data of each waybill before the preset time point is considered. In order to verify the validity and stability of the model, three days of waybill data separated by one week from the training set can be selected as a test set, 13 time points (10 points, 11 points, 12 points, 13 points, 14 points, 15 points, 16 points, 17 points, 18 points, 19 points, 20 points, 21 points and 22 points) are selected each day, each time point corresponds to one test set, and a test result is obtained, so that the model is verified to have strong robustness at each time point. The evaluation indexes can comprise the AUC value, the accuracy and the recall rate of the model, and the hyper-parameters are continuously adjusted based on the indexes to reach the optimal state of the model. The training end condition may be that each index satisfies a preset condition, and the preset condition may be set in combination with an actual requirement, which is not limited here.
The customer complaint risk prediction model obtained by the embodiment has the accuracy rate of 20% and the recall rate of 5%. The specific application can be as follows: taking the network points as an example, for the waybill arriving at the destination network point, accessing all dimensional data of the waybill in real time through kafka, calculating all dimensional characteristics corresponding to the waybill, generating target characteristics with the same format as that of the model during training, taking the target characteristics as input to the model, and outputting a prediction result of whether the waybill can generate the time-effect customer complaint risk by the model. After the prediction result is obtained by the relevant site, the waybill with the time-efficient customer complaint risk label can be subjected to special flow processing, so that time-efficient customer complaint risk rescue can be performed more specifically, for example, workers are reminded through instructions, and the related personnel are communicated in advance through an intelligent outbound mode, so that the possibility that the related personnel initiate customer complaints is reduced, and cost reduction and efficiency improvement are realized.
It should be noted that the above embodiments can be applied to risk prediction of an aged customer complaint type, and can also provide a technical template for risk prediction of other customer complaint types.
It should be understood that, although the steps in the flowcharts related to the above embodiments are shown in sequence as indicated by the arrows, the steps are not necessarily executed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a part of the steps in each flowchart related to the above embodiments may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of performing the steps or stages is not necessarily sequential, but may be performed alternately or alternately with other steps or at least a part of the steps or stages in other steps.
In one embodiment, as shown in fig. 4, there is provided a customer complaint risk prediction apparatus 400 including: an acquisition module 410, a processing module 420, and a prediction module 430, wherein:
the obtaining module 410 is configured to obtain waybill information of a waybill to be predicted, where the waybill information includes a waybill identifier.
And the processing module 420 is configured to obtain a first waybill characteristic of the waybill to be predicted according to the waybill circulation data associated with the waybill identifier up to the current time.
And the prediction module 430 is configured to obtain a customer complaint risk prediction result of the waybill to be predicted according to the first waybill characteristic.
In one embodiment, the waybill flow data includes at least one of: sending time, promised delivery time, first arrival time of arrival at the current site, second arrival time of arrival at the current city, retention data in a time period from the sending time to the current time, checking data in a time period from the sending time to the current time, and hastening data in a time period from the sending time to the current time; the first waybill feature includes at least one of: retention characteristics, current site retention time, current city retention time, piece checking characteristics, dispatching characteristics, circulation duration, time difference between the last piece checking time and the piece sending time, and time difference between the first piece checking time and the promised delivery time.
In one embodiment, the waybill information further includes: first parameter data for identifying a sender and second parameter data for identifying a receiver; the processing module 420 is further configured to: obtaining a second waybill characteristic of the waybill to be predicted according to first historical waybill data associated with the first parameter data and second historical waybill data associated with the second parameter data; the second waybill feature includes: and the consignment rate is determined according to the first historical waybill data, and the consignment rate is determined according to the second historical waybill data.
In an embodiment, when obtaining the second waybill characteristic of the waybill to be predicted according to the first historical waybill data associated with the first parameter data and the second historical waybill data associated with the second parameter data, the processing module 420 is specifically configured to: obtaining a sending volume associated with the first parameter data and a sending customer volume for sending a customer in the sending volume according to first historical waybill data associated with the first parameter data, and obtaining a sending customer rate according to the ratio of the sending customer volume to the sending volume; according to the second historical waybill data related to the second parameter data, the receiving amount related to the second parameter data and the receiving and customer complaint amount of the customer complaints in the receiving amount are obtained, and according to the ratio of the receiving and customer complaint amount to the receiving amount, the receiving and customer complaint rate is obtained.
In one embodiment, the waybill information further includes: a third waybill feature, the third waybill feature comprising at least one of: the type of the consignment, the type of the time effect, the product code, whether to guarantee the price and whether to add value to the service.
In one embodiment, the prediction module 430, when obtaining the customer complaint risk prediction result of the waybill to be predicted according to the first waybill characteristic, is specifically configured to: determining target characteristics of the waybill to be predicted according to the first waybill characteristics, the second waybill characteristics and the third waybill characteristics; and predicting based on the target characteristics to obtain the customer complaint risk prediction result of the freight bill to be predicted.
In an embodiment, the prediction module 430 is specifically configured to, when performing prediction based on the target features and obtaining a prediction result of the risk of customer complaint of the waybill to be predicted: and inputting the target characteristics into the trained customer complaint risk prediction model to obtain the customer complaint risk prediction result of the waybill to be predicted.
In one embodiment, the apparatus further comprises: and the training module is used for training to obtain a customer complaint risk prediction model. The training module comprises: an acquisition unit, a processing unit, a prediction unit and an adjustment unit, wherein:
the acquisition unit is used for acquiring waybill information of the waybill sample and a corresponding sample label, wherein the waybill information comprises: the waybill label is used for indicating whether the waybill sample generates a customer complaint before the target time point.
And the processing unit is used for obtaining a first waybill characteristic of the waybill sample according to the waybill circulation data associated with the waybill identification of the waybill sample at the preset time point, obtaining a second waybill characteristic of the waybill sample according to the historical waybill data associated with the parameter data of the waybill sample, and obtaining a target characteristic of each waybill sample according to the first waybill characteristic, the second waybill characteristic and the third waybill characteristic of each waybill sample.
And the prediction unit is used for inputting the target characteristics into the customer complaint risk prediction model to be trained to obtain the prediction results corresponding to the waybill samples.
And the adjusting unit is used for adjusting the parameters of the customer complaint risk prediction model to be trained based on the prediction result and the sample label until the training end condition is met, so as to obtain the trained customer complaint risk prediction model.
In one embodiment, when the processing unit obtains the target feature of each waybill sample according to the first waybill feature, the second waybill feature, and the third waybill feature of each waybill sample, the processing unit is specifically configured to: determining the information quantity of each characteristic in the first waybill characteristic, the second waybill characteristic and the third waybill characteristic; and screening out the features of which the information quantity meets the preset conditions from all the features according to the information quantity of each feature, and taking the features as the target features of each waybill sample.
In one embodiment, when determining the information amount of each of the first waybill characteristic, the second waybill characteristic, and the third waybill characteristic, the processing unit is specifically configured to: for any one of the first waybill characteristic, the second waybill characteristic and the third waybill characteristic, the following processing is carried out: acquiring each grouping feature corresponding to the feature, and acquiring the weight of each grouping feature according to the quantity proportion of the corresponding positive sample to all the positive samples and the quantity proportion of the corresponding negative sample to all the negative samples, wherein the positive samples and the negative samples respectively represent waybill samples with and without customer complaints; and determining the information quantity of the sample characteristics according to the weights of all the grouping characteristics.
For specific limitations of the customer complaint risk prediction device, reference may be made to the above limitations of the customer complaint risk prediction method, which are not described herein again. All or part of each module in the customer complaint risk prediction device can be realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent of a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, and the internal structure thereof may be as shown in fig. 5. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operating system and the computer program to run on the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a customer complaint risk prediction method.
In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 6. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a customer complaint risk prediction method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those skilled in the art that the configurations shown in fig. 5 or fig. 6 are only block diagrams of some configurations relevant to the present application, and do not constitute a limitation on the computer apparatus to which the present application is applied, and a particular computer apparatus may include more or less components than those shown in the drawings, or may combine some components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the steps of the above-described method embodiments when executing the computer program.
In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of the respective method embodiment as described above.
In one embodiment, a computer program product or computer program is provided that includes computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the steps in the various method embodiments described above.
It should be understood that the terms "first", "second", etc. in the above-described embodiments are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implying any number of technical features indicated. Further, in the description of the present application, the meaning of "a plurality" means at least two unless otherwise specified.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), for example.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims (12)

1. A customer complaint risk prediction method, characterized in that the method comprises:
acquiring waybill information of a waybill to be predicted, wherein the waybill information comprises a waybill identifier;
acquiring first waybill transfer characteristics of the waybill to be predicted according to waybill transfer data associated with the waybill identification until the current time;
and obtaining the customer complaint risk prediction result of the waybill to be predicted according to the first waybill characteristic.
2. The method of claim 1, wherein the waybill flow data comprises at least one of: sending time, promised delivery time, first arrival time of the current site, second arrival time of the current city, retention data in the time period from the sending time to the current time, checking data in the time period from the sending time to the current time, and expediting data in the time period from the sending time to the current time;
the first waybill feature comprises at least one of: the system comprises a retention characteristic, a current site retention time, a current city retention time, a piece checking characteristic, a piece dispatching prompting characteristic, a circulation time length, a time difference between the last piece checking time and the piece dispatching time, and a time difference between the first piece checking time and the promised delivery time.
3. The method of claim 1, wherein the waybill information further comprises: first parameter data for identifying a sender and second parameter data for identifying a receiver; the method further comprises the following steps:
obtaining a second waybill characteristic of the waybill to be predicted according to first historical waybill data related to the first parameter data and second historical waybill data related to the second parameter data; the second waybill feature comprises: and the sending customer complaint rate is determined according to the first historical waybill data, and the receiving customer complaint rate is determined according to the second historical waybill data.
4. The method of claim 3, wherein obtaining a second waybill characteristic of the waybill to be predicted according to first historical waybill data associated with the first parameter data and second historical waybill data associated with the second parameter data comprises:
obtaining a sending volume associated with the first parameter data and a sending customer volume of sending a customer complaint in the sending volume according to first historical waybill data associated with the first parameter data, and obtaining a sending customer complaint rate according to the ratio of the sending customer volume to the sending volume;
according to second historical waybill data related to the second parameter data, obtaining the receiving amount related to the second parameter data and the receiving customer complaint amount of the customer complaints in the receiving amount, and according to the ratio of the receiving customer complaint amount to the receiving amount, obtaining the receiving customer complaint rate.
5. The method of claim 3, wherein the waybill information further comprises: a third waybill feature that includes at least one of: the type of the consignment, the type of the time effect, the product code, whether to guarantee the price and whether to add value to the service.
6. The method of claim 5, wherein obtaining a complaint risk prediction result for the waybill to be predicted according to the first waybill characteristic comprises:
determining the target characteristics of the waybill to be predicted according to the first waybill characteristics, the second waybill characteristics and the third waybill characteristics;
and predicting based on the target characteristics to obtain the customer complaint risk prediction result of the freight bill to be predicted.
7. The method of claim 6, wherein predicting based on the target characteristics to obtain a prediction result of the risk of customer complaint of the waybill to be predicted comprises:
inputting the target characteristics into a trained customer complaint risk prediction model to obtain a customer complaint risk prediction result of the waybill to be predicted;
the training method of the customer complaint risk prediction model comprises the following steps:
acquiring waybill information of a waybill sample and a corresponding sample label, wherein the waybill information comprises: the waybill label is used for indicating whether the waybill sample generates a customer complaint before a preset time point;
acquiring a first waybill characteristic of the waybill sample according to waybill circulation data associated with the waybill identification of the waybill sample at the preset time point, and acquiring a second waybill characteristic of the waybill sample according to historical waybill data associated with parameter data of the waybill sample;
obtaining target characteristics of the waybill samples according to the first waybill characteristics, the second waybill characteristics and the third waybill characteristics of the waybill samples, inputting the target characteristics into a to-be-trained complaint risk prediction model, and obtaining prediction results corresponding to the waybill samples;
and adjusting parameters of the customer complaint risk prediction model to be trained based on the prediction result and the sample label until a training end condition is met, and obtaining the trained customer complaint risk prediction model.
8. The method of claim 7, wherein obtaining the target characteristics of each waybill sample from the first waybill characteristic, the second waybill characteristic, and the third waybill characteristic of each waybill sample comprises:
determining the information content of each characteristic in the first waybill characteristic, the second waybill characteristic and the third waybill characteristic;
and screening out the features of which the information quantity meets the preset conditions from all the features according to the information quantity of each feature, and taking the features as the target features of each waybill sample.
9. The method of claim 8, wherein determining the amount of information for each of the first waybill characteristic, the second waybill characteristic, and the third waybill characteristic comprises:
for any one of the first waybill characteristic, the second waybill characteristic and the third waybill characteristic, performing the following processing:
acquiring each grouping feature corresponding to the feature, and acquiring the weight of each grouping feature according to the quantity proportion of the corresponding positive sample to all the positive samples and the quantity proportion of the corresponding negative sample to all the negative samples, wherein the positive samples and the negative samples respectively represent waybill samples with and without customer complaints;
and determining the information quantity of the sample characteristics according to the weights of all the grouping characteristics.
10. A customer complaint risk prediction device, characterized in that the device comprises:
the acquiring module is used for acquiring waybill information of the waybill to be predicted, wherein the waybill information comprises a waybill identifier;
the processing module is used for obtaining a first waybill characteristic of the waybill to be predicted according to waybill circulation data associated with the waybill identification until the current time;
and the prediction module is used for obtaining the customer complaint risk prediction result of the waybill to be predicted according to the first waybill characteristic.
11. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor realizes the steps of the method of any one of claims 1 to 9 when executing the computer program.
12. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 9.
CN202011542126.3A 2020-12-23 2020-12-23 Customer complaint risk prediction method, apparatus, computer device and storage medium Pending CN114663107A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011542126.3A CN114663107A (en) 2020-12-23 2020-12-23 Customer complaint risk prediction method, apparatus, computer device and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011542126.3A CN114663107A (en) 2020-12-23 2020-12-23 Customer complaint risk prediction method, apparatus, computer device and storage medium

Publications (1)

Publication Number Publication Date
CN114663107A true CN114663107A (en) 2022-06-24

Family

ID=82024543

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011542126.3A Pending CN114663107A (en) 2020-12-23 2020-12-23 Customer complaint risk prediction method, apparatus, computer device and storage medium

Country Status (1)

Country Link
CN (1) CN114663107A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116402236A (en) * 2023-05-31 2023-07-07 北京京东乾石科技有限公司 Information generation method and device

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116402236A (en) * 2023-05-31 2023-07-07 北京京东乾石科技有限公司 Information generation method and device

Similar Documents

Publication Publication Date Title
CN107679674B (en) OTA platform overseas hotel room type service defect prediction method and system
CN109961248B (en) Method, device, equipment and storage medium for predicting waybill complaints
CN109905882B (en) Network capacity expansion method and device
CN111294730B (en) Method and device for processing network problem complaint information
CN114638391A (en) Waybill risk scene identification processing method and device, computer equipment and medium
WO2019157779A1 (en) Incoming call processing method, electronic device and computer-readable storage medium
CN111047264B (en) Logistics task distribution method and device
CN109598519B (en) Vehicle auditing method, device, computer equipment and storage medium
CN111582771A (en) Risk assessment method, device, equipment and computer readable storage medium
CN110348472A (en) Data Detection rule generating method, device, computer equipment and storage medium
CN114663107A (en) Customer complaint risk prediction method, apparatus, computer device and storage medium
CN110765351A (en) Target user identification method and device, computer equipment and storage medium
CN113095647B (en) Vehicle inspection system
CN115115157A (en) Overdue risk prediction method, overdue risk prediction device, computer equipment and storage medium
CN117291558A (en) Personnel matching management method and system based on multiple attendance rules
CN111860893A (en) Data processing method and device
CN111798151A (en) Enterprise fraud risk assessment method, device, equipment and readable storage medium
CN116934246A (en) Method, device, equipment and readable storage medium for auditing declaration project data
CN108075918B (en) Internet service change detection method and system
CN111311150B (en) Distribution task grouping method, platform, electronic equipment and storage medium
CN111105284B (en) Order processing method and device, M-layer order processing model, electronic equipment and storage medium
CN115127856A (en) Method and device for sampling and identifying concrete test block compression test robot
CN114581130A (en) Bank website number assigning method and device based on customer portrait and storage medium
CN114971446A (en) Method and device for constructing loss risk prediction model and computer equipment
CN111445157A (en) Service data management method, device, equipment and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination