A method for automatically assigning home appliance work orders driven by data

Abstract

The present invention relates to the technical field of big data analysis, and specifically relates to a data-driven automatic dispatching method for household appliance work orders. The method includes obtaining multi-dimensional association information corresponding to each historical work order record; performing big data analysis on the multi-dimensional association information to determine the expected satisfaction, actual satisfaction, and user personalization coefficient corresponding to each historical work order record; determining the user portrait factor corresponding to each historical work order record according to the difference between the expected satisfaction and the actual satisfaction corresponding to each historical work order record and the user personalization coefficient; and determining the target repairman to whom the work order can be dispatched according to the repair ability ratings of the currently schedulable repairmen and the user portrait factors of the historical work order records of the currently reserved users. By performing big data analysis on the multi-dimensional association information corresponding to each historical work order record, the present invention can accurately determine the user portrait factor corresponding to the historical work order record, thereby effectively improving the customer matching degree in work order dispatching.

Description

Technical Field

[0001] The present invention relates to the technical field of big data analysis, and particularly to a method for automatically dispatching household appliance work orders driven by data. Background Art

[0002] In the household appliance repair service industry, the efficient dispatching of work orders is directly related to user satisfaction and operational efficiency. Traditional work order dispatching usually relies on manual scheduling, which is not only cumbersome but also prone to human errors, leading to a series of problems such as repair delays, uneven technician loads, and user dissatisfaction. Existing methods utilize big data analysis and intelligent algorithms to collect and analyze a large amount of data, including user information, fault descriptions, geographical locations, historical repair records, technician skills and availability, etc., to achieve automatic processing and dispatching of work orders, effectively ensuring the scientificity and rationality of work order dispatching.

[0003] Since the core purpose of work order dispatching is to improve the turnover efficiency of maintenance personnel and meet user needs to the greatest extent, in the data-driven automatic work order dispatching method, a work order dispatching strategy of matching maintenance personnel according to user portraits is often adopted. However, for work order data, due to the limited recorded data that cannot cover the work order execution process, directly generating user portraits based on only a small amount of user data may have a large error, thereby reducing the customer matching degree of the work order allocation strategy. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for automatically dispatching household appliance work orders driven by data, which is used to solve the problem that the existing user portrait extraction is inaccurate, resulting in a low customer matching degree in work order dispatching.

[0005] To solve the above technical problems, in a first aspect, the present invention provides a method for automatically dispatching household appliance work orders driven by data, including the following steps:

[0006] Obtain the multi-dimensional correlation information corresponding to each historical work order record, where the multi-dimensional correlation information at least includes: repairman information, user information, and work order information;

[0007] Analyze the repairman information, user information, and work order information in the multi-dimensional correlation information to determine the expected satisfaction, actual satisfaction, and user personalization coefficient corresponding to each historical work order record;

[0008] Determine the user portrait factor corresponding to each historical work order record according to the difference between the expected satisfaction and the actual satisfaction corresponding to each historical work order record and the user personalization coefficient;

[0009] Determine the target repairman to whom the work order can be dispatched according to the repair ability rating of each currently schedulable repairman and the user portrait factors of each historical work order record of the currently reserved user.

[0010] Combined with the first aspect above, in some possible implementation manners, the maintenance worker information at least includes: a maintenance ability rating, a departure address of the maintenance worker, and a time taken for the maintenance worker to reach the maintenance location; the user information at least includes: user feedback information and work order reservation information; the work order reservation information at least includes: the user reservation time, the user order time, and the work order maintenance address of the current work order and its previous work order; the work order information at least includes: the maintenance equipment information of the current work order and the maintenance equipment information of the previous work order of the current work order.

[0011] Combined with the first aspect above, in some possible implementation manners, the steps for determining the expected satisfaction corresponding to each historical work order record include:

[0012] Determine the ratio of the maintenance ability rating to the maintenance duration corresponding to each historical work order record to obtain a first ratio;

[0013] Determine the ratio of the normalized value of the distance from the departure address of the maintenance worker to the work order maintenance address to the time taken for the maintenance worker to reach the maintenance location corresponding to each historical work order record to obtain a second ratio;

[0014] Fuse the first ratio and the second ratio to determine the expected satisfaction corresponding to each historical work order record, and both the first ratio and the second ratio are positively correlated with the expected satisfaction.

[0015] Combined with the first aspect above, in some possible implementation manners, the steps for determining the actual satisfaction corresponding to each historical work order record include:

[0016] Determine the user feedback factor corresponding to each historical work order record according to the user feedback information corresponding to each historical work order record;

[0017] Analyze the work order reservation information corresponding to each historical work order record, the time difference between the user reservation time and the order time of the current work order, the maintenance equipment information of the current work order and the maintenance equipment information of the previous work order of the current work order, and the interval duration between the user reservation times of the current work order and its previous work order to determine the information difference index corresponding to each historical work order record;

[0018] Use the information difference index to correct the user feedback factor to obtain the actual satisfaction corresponding to each historical work order record.

[0019] Combined with the first aspect above, in some possible implementation manners, determining the information difference index corresponding to each historical work order record includes:

[0020] Determine the completeness of the reservation information corresponding to each historical work order record according to the work order reservation information corresponding to each historical work order record;

[0021] Determine the ratio of the time difference between the user appointment time and the order placement time of this work order corresponding to each historical work order record to the completeness of the appointment information, to obtain a third ratio;

[0022] Conduct a similarity analysis on the maintenance equipment information of this work order corresponding to each historical work order record and the maintenance equipment information of the previous work order of this work order, to determine the similarity of equipment failure types corresponding to each historical work order record;

[0023] Determine the ratio of the negative correlation normalization result of the interval duration between the user appointment times of this work order and its previous work order to the similarity of the equipment failure types, to obtain a fourth ratio;

[0024] Fuse the third ratio and the fourth ratio, and determine the information difference index corresponding to each historical work order record. Both the third ratio and the fourth ratio are positively correlated with the information difference index.

[0025] Combined with the first aspect above, in some possible implementation manners, the steps for determining the user personalization coefficient corresponding to each historical work order record include:

[0026] Determine at least two non-correlated element information in the multi-dimensional association information;

[0027] Cluster all historical work order records according to the same type of element information corresponding to all historical work order records, to obtain each first work order record sub-cluster corresponding to each type of element information;

[0028] Determine the proportion of the number of historical work order records corresponding to the customer's non-feedback in the first work order record sub-cluster where each historical work order record is located in the total number of the first work order record sub-cluster;

[0029] According to the proportion corresponding to each historical work order record and the difference between the user feedback factor corresponding to each historical work order record and the user feedback factors corresponding to other historical work order records in the first work order record sub-cluster where this historical work order record is located, determine the personalization coefficient of each historical work order record corresponding to each type of element information;

[0030] Comprehensively determine the user personalization coefficient corresponding to each historical work order record based on the personalization coefficients of each historical work order record corresponding to various types of element information.

[0031] Combined with the first aspect above, in some possible implementation manners, the maintenance equipment information at least includes the type of faulty equipment, and the element information at least includes: work order maintenance address and type of faulty equipment.

[0032] Combined with the first aspect above, in some possible implementation manners, determining the user portrait factor corresponding to each historical work order record includes:

[0033] Determine the absolute value of the difference between the expected satisfaction and the actual satisfaction corresponding to each historical work order record to obtain the satisfaction deviation;

[0034] Based on the user personalization coefficient corresponding to each historical work order record and the satisfaction deviation, determine the user portrait factor corresponding to each historical work order record, where both the user personalization coefficient and the satisfaction deviation are positively correlated with the user portrait factor.

[0035] Combined with the first aspect above, in some possible implementation manners, determining the target repairman who can be dispatched includes:

[0036] Determine the average value of the user portrait factors of the historical work order records of the current reserved user to obtain the average user portrait factor;

[0037] Normalize the average user portrait factor to obtain the first normalized value;

[0038] Normalize the repair ability ratings of each dispatchable repairman to obtain the second normalized value;

[0039] Determine the dispatchable repairman corresponding to the minimum difference between the second normalized value and the first normalized value as the target repairman who can be dispatched.

[0040] Combined with the first aspect above, in some possible implementation manners, the method further includes:

[0041] If there is no historical work order record for the current reserved user, cluster all historical work order records according to the user information and work order information corresponding to each historical work order record to obtain each second work order record sub-cluster;

[0042] Match the user information and work order information of the current reserved user with the user information and work order information of the historical work order records corresponding to the second work order record sub-cluster to obtain the matching second work order record sub-cluster of the current reserved user;

[0043] Determine the average value of the user portrait factors of the historical work order records in the matching second work order record sub-cluster to obtain the user portrait factor of the reserved work order of the current reserved user;

[0044] Based on the repair ability ratings of the current dispatchable repairmen and the user portrait factor of the reserved work order of the current reserved user, determine the target repairman who can be dispatched.

[0045] To solve the above technical problems, in a second aspect, the present invention further provides a data-driven automatic assignment system for household appliance work orders, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, so that the device executes the method in the above first aspect or any possible implementation manner of the first aspect.

[0046] To solve the above technical problems, in a third aspect, the present invention further provides a computer program product, which includes: computer program code, when the computer program code runs on a computer, the computer is caused to execute the method in the above first aspect or any possible implementation manner of the first aspect.

[0047] To solve the above technical problems, in a fourth aspect, the present invention further provides a computer-readable storage medium, which stores computer program code, and when the computer program code runs on a computer, the computer is caused to execute the method in the above first aspect or any possible implementation manner of the first aspect.

[0048] The present invention has the following beneficial effects: By obtaining the multi-dimensional association information corresponding to each historical work order record of the current reservation user, and performing big data analysis and mining on the multi-dimensional association information including repairman information, user information, and work order information, the expected satisfaction, actual satisfaction, and user personalization coefficient corresponding to each historical work order record are determined. And according to the difference between the expected satisfaction and the actual satisfaction corresponding to each historical work order record and the user personalization coefficient, the user portrait factor corresponding to each historical work order record is determined. By comprehensively considering the user portrait factors of each historical work order record of the current reservation user and combining the repair ability ratings of the current schedulable repairmen, the target repairman to whom the work order can be assigned is finally determined. The present invention performs big data analysis on the multi-dimensional association information of each historical work order record of the current reservation user, so that the user portrait factor of the current reservation user can be accurately determined, and finally the customer matching degree in work order assignment is effectively improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0050] Figure 1 It is a flowchart of the steps of a data-driven automatic assignment method for household appliance work orders according to an embodiment of the present invention;

[0051] Figure 2 Flow chart of the determination steps for the expected satisfaction degree of the embodiments of the present invention;

[0052] Figure 3 Flow chart of the determination steps for the actual satisfaction degree of the embodiments of the present invention;

[0053] Figure 4 Flow chart of the determination steps for the user personalization coefficient of the embodiments of the present invention;

[0054] Figure 5 Flow chart of the determination of the user portrait factors corresponding to each historical work order record in the embodiments of the present invention;

[0055] Figure 6 Flow chart of the determination of the target repairmen to whom work orders can be assigned in the embodiments of the present invention;

[0056] Figure 7 Schematic structural diagram of a data-driven automatic assignment system for household appliance work orders according to the embodiments of the present invention. Detailed implementation manners

[0057] To clearly illustrate the technical features of this solution, the present invention will be elaborated in detail below through specific implementation manners in combination with the accompanying drawings.

[0058] Embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not used to limit the protection scope of the present invention.

[0059] It should be understood that the various steps recited in the method embodiments of the present invention can be executed in a different order and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this regard.

[0060] The term "including" and its variants used herein are open-ended, that is, "including but not limited to". The term "based on" is "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0061] It should be noted that concepts such as "first" and "second" mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order or mutual dependence of the functions performed by these devices, modules or units.

[0062] In the embodiments of the present invention, although operations or steps are described in a specific order in the drawings, it should not be understood that these operations or steps are required to be performed in the specific order shown or in a serial order, or that all the shown operations or steps are required to be performed to obtain the desired result. In the embodiments of the present invention, these operations or steps can be performed serially; they can also be performed in parallel; or a part of these operations or steps can be performed.

[0063] At the same time, it can be understood that the data involved in the technical solution of the present invention (including but not limited to the data itself, the acquisition or use of the data) should comply with the requirements of corresponding laws, regulations and related provisions. Unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs, and all parameters or indicators in the formulas involved in the present invention are normalized values that eliminate the influence of dimensions.

[0064] To solve the problem that the existing user portraits are extracted inaccurately, resulting in a low customer matching degree in work order dispatching, the embodiments of the present invention provide a data-driven automatic home appliance work order dispatching method. This method calculates the expected satisfaction of the user by the dispatching situation of the repairmen corresponding to each historical work order record of the current reserved user, and then estimates the information difference between the repairman and the user in each historical work order, obtains the actual satisfaction of the user according to the information difference, and then calculates the personalized coefficient of the user. For users with a larger personalized coefficient, their feedback results may be more difficult to predict, and the bias weight of work order dispatching needs to be increased. Further, according to the difference between the expected satisfaction and the actual satisfaction of the user, and the user's personalized coefficient, a user portrait factor is obtained, and the user portrait factor is matched with the repair ability ratings of the current schedulable repairmen to accurately dispatch real-time work orders, effectively improving the user satisfaction and the work order operation efficiency.

[0065] Next, a data-driven automatic home appliance work order dispatching method provided by the embodiments of the present invention will be introduced in detail with reference to the drawings.

[0066] Figure 1 The basic flow schematic diagram of a data-driven automatic home appliance work order dispatching method provided by the embodiments of the present invention is shown, as Figure 1 shown, and the method specifically includes the following steps:

[0067] Step S100: Obtain the multi-dimensional association information corresponding to each historical work order record. The multi-dimensional association information at least includes: repairman information, user information, and work order information.

[0068] In the work order dispatching system, the detailed information of each work order is tracked and updated in real time, and real-time data from different sources can be integrated and analyzed. Therefore, all historical work order data processed in the past year is read from the system. Each historical work order data is a complete home appliance failure repair record, including but not limited to the following element data: 1. User data: user contact information, address and geographical location, historical service records, etc.; 2. Work order data: fault description and type, urgency, service time requirement, etc.; 3. Repairman data: repairman's work years, skills and qualifications, current work status and availability, geographical location and working area, etc.; 4. Equipment data: equipment type and model, warranty status, repair history, etc.; 5. Time data: repair time, appointment time window, order placement time, etc.; 6. Geographical and traffic data: distance and traffic conditions, regional service coverage, etc.; 7. User feedback data: user satisfaction score, user comments and suggestions, etc.

[0069] Take a complete home appliance failure repair record processed in the past year as a historical work order record. By extracting the above-mentioned element data corresponding to the historical work order record, the multi-dimensional association information corresponding to each historical work order record can be obtained. The multi-dimensional association information at least includes: repairman information, user information, and work order information. Among them, the repairman information includes: repair ability rating, repairman's departure address, time consumed for the repairman to reach the repair location, repair time consumed, etc.; the user information includes: user feedback information, work order appointment information, etc.; the work order appointment information includes: user appointment time of this work order and its previous work order, user order placement time, work order repair address, etc.; the work order information includes: repair equipment information of this work order, repair equipment information of the previous work order of this work order, etc.

[0070] Specifically, the maintenance ability rating is determined based on the length of service, skills, and qualifications of the maintenance worker. The system can quantify the ability rating of the maintenance worker through programming, with the value constrained between 0 and 1. The closer the value is to 1, the stronger the maintenance ability of the corresponding maintenance worker. The departure address of the maintenance worker refers to the original departure address from which the maintenance worker goes to the maintenance address of the corresponding historical work order. The time taken for the maintenance worker to reach the maintenance location is the time consumed by the maintenance worker from the departure address to the maintenance address of the corresponding historical work order, that is, the time required for the maintenance worker to travel from the initial position to the target maintenance address, which can be measured in minutes. The time taken for maintenance is the time interval between the time when the maintenance worker arrives at the maintenance address recorded in the corresponding historical work order and the time when the final work order is completed, which can be measured in minutes. The user feedback information includes the customer's star rating, phone number, and evaluation comments, etc. The work order reservation information includes: the user reservation time, user order time, work order maintenance address, fault description and type, urgency, service time requirements, etc. of this work order and its previous work order. The user reservation time and user order time refer to the time point of the reserved home maintenance and the actual order time point respectively. The user reservation time of the previous work order of this work order refers to the user reservation time of the previous work order of this work order corresponding to the maintenance worker before this work order. The maintenance equipment information of this work order refers to information such as the type of faulty equipment, fault frequency, repair time, required tools, etc. in the corresponding historical work order. The maintenance equipment information of the previous work order of this work order includes information such as the type of faulty equipment, fault frequency, repair time, required tools, etc. of the previous work order when the maintenance worker received the historical work order.

[0071] Step S200: Analyze the maintenance worker information, user information, and work order information in the multi-dimensional association information to determine the expected satisfaction, actual satisfaction, and user personalization coefficient corresponding to each historical work order record.

[0072] For the work order dispatching platform, it is necessary to coordinate the allocation relationship between maintenance personnel and users, meet the user needs to the greatest extent, and provide convenience for maintenance personnel as much as possible, so that maintenance personnel can have a better turnover rate in the case of a large number of work orders. Therefore, by analyzing the multi-dimensions corresponding to each historical work order record, the expected satisfaction corresponding to each historical work order record is determined.

[0073] Since the most intuitive efficiency performance of each work order record is the time length between the user order placement node and the maintenance completion node, being able to dispatch orders in a timely manner and repair home appliance faults to meet user needs is the basic requirement of maintenance order dispatching. However, there are several force majeure factors during the process of the work order, such as the user is not at home when the maintenance worker arrives, traffic congestion, the actual home appliance fault does not match the user's description of the fault, etc. These factors will all lead to more time spent, and even the order dispatching strategy itself may be problematic. Therefore, the expected satisfaction is calculated for each work order record.

[0074] In the embodiments of the present invention, as Figure 2 shown, the steps for determining the expected satisfaction corresponding to each historical work order record include:

[0075] Step S201: Determine the ratio of the repair ability rating corresponding to each historical work order record to the repair duration, and obtain a first ratio;

[0076] Step S202: Determine the ratio of the normalized value of the distance from the repairman's departure address to the work order repair address corresponding to each historical work order record to the duration consumed by the repairman to reach the repair location, and obtain a second ratio;

[0077] Step S203: Integrate the first ratio and the second ratio to determine the expected satisfaction corresponding to each historical work order record, and both the first ratio and the second ratio are positively correlated with the expected satisfaction.

[0078] For the above steps, as an example, for any historical work order record, taking the i-th historical work order record as an example, obtain the distance from the repairman's departure address to the repair address of this work order. Since the repairman's jurisdiction is planned by area, there is generally a maximum value within the area for this distance value. Use this maximum value to perform maximum normalization on the distance value corresponding to the i-th historical work order record to obtain the normalized distance value. At the same time, obtain the duration consumed by the repairman corresponding to the i-th historical work order record to reach the repair location, the repair duration, and the repair ability rating. The duration consumed by the repairman to reach the repair location refers to the time required for the repairman to travel from the initial position to the target repair address. The repair duration refers to the time interval between the time when the repairman reaches the target address of the i-th historical work order record and the time when the work order is finally completed. The repair ability rating reflects the repair ability of the repairman.

[0079] On this basis, based on the normalized distance value, the duration consumed by the repairman to reach the repair location, the repair duration, and the repair ability rating corresponding to the i-th work order record, determine the expected satisfaction corresponding to the i-th historical work order record:

[0080]

[0081] Among them, A i represents the expected satisfaction corresponding to the i-th historical work order record; J i represents the repair ability rating corresponding to the i-th historical work order record; represents the repair duration corresponding to the i-th historical work order record; D i represents the normalized value of the distance from the repairman's departure address to the historical work order repair address corresponding to the i-th historical work order record, that is, the normalized distance value; It represents the time taken for the repairman corresponding to the i-th historical work order record to reach the repair location; norm represents the normalization function.

[0082] In the above calculation formula, the repair ability rating J i and the repair time-consuming The ratio is the first ratio The larger it is, the higher the rating of the repairman assigned in the i-th historical work order record, and the shorter the repair time; the normalized value D of the distance from the repairman's departure address to the historical work order repair address i and the time taken for the repairman to reach the repair location The ratio is the second ratio The larger it is, the faster the repairman reaches the target repair address in the i-th historical work order record; therefore, the larger the product of the two, the higher the expected satisfaction A of the repairman assigned by the system to the user in the i-th historical work order record i The higher.

[0083] In the above manner, the expected satisfaction corresponding to each historical work order record can be determined. The expected satisfaction refers to the expected user satisfaction determined based on the repairman assignment situation, but there may still be a certain difference from the actual user satisfaction. Therefore, in order to extract the user demand portrait that is an important reference basis for formulating the work order dispatching strategy, it is also necessary to determine the actual satisfaction corresponding to each historical work order record.

[0084] In the embodiment of the present invention, as Figure 3 shown, the steps for determining the actual satisfaction corresponding to each historical work order record include:

[0085] Step S211: Determine the user feedback factor corresponding to each historical work order record according to the user feedback information corresponding to each historical work order record;

[0086] Step S212: Analyze the work order reservation information corresponding to each historical work order record, the time difference between the user reservation time and the order placement time of this work order, the repair equipment information of this work order and the repair equipment information of the previous work order of this work order, and the interval duration between the user reservation times of this work order and its previous work order, and determine the information difference index corresponding to each historical work order record;

[0087] Step S213: Use the information difference index to correct the user feedback factor to obtain the actual satisfaction corresponding to each historical work order record.

[0088] For the above steps, as an example, taking the i-th historical work order record as an example, for the user feedback information corresponding to the i-th historical work order record, that is, the user star rating, phone number, and evaluation message, use the trained natural language model or neural network evaluation model to quantify the level of evaluation. That is, input the user feedback information into the trained natural language model or neural network evaluation model, and the model directly outputs the user feedback factor. The value range of the user feedback factor is between 0 and 1. The larger the user feedback factor, the higher the evaluation of the user for this historical work order. It should be understood that if there is no corresponding user feedback information for the historical work order record, it can be considered that the user has no objection to this repair process at this time, and the corresponding user feedback factor can be directly set to 1.

[0089] Considering that it is impossible to record all details in detail during the execution of the work order, and the user needs vary, the user's feedback result may be inconsistent with the expected satisfaction. This may be due to a certain degree of information gap between the repairman and the user during the docking process, resulting in user dissatisfaction. Therefore, it is necessary to analyze the multi-dimensional associated information corresponding to the historical work order record, such as the overall work order reservation information, the time difference between the user's reservation time and the order placement time, the repair equipment information of this work order and the repair equipment information of the previous work order of this work order, and the interval duration between the user's reservation times of this work order and its previous work order by the repairman, etc. Analyze the information to obtain the information difference index corresponding to the historical work order record, and use this information difference index to correct the user feedback factor corresponding to the historical work order record, so as to accurately obtain the actual satisfaction corresponding to the historical work order record.

[0090] In the embodiment of the present invention, determining the information difference index corresponding to each historical work order record, the implementation steps include:

[0091] According to the work order reservation information corresponding to each historical work order record, determine the integrity of the reservation information corresponding to each historical work order record;

[0092] Determine the ratio of the time difference between the user's reservation time and the order placement time of this work order corresponding to each historical work order record to the integrity of the reservation information, and obtain the third ratio;

[0093] Conduct a similarity analysis on the repair equipment information of this work order and the repair equipment information of the previous work order corresponding to each historical work order record, and determine the similarity of the equipment failure types corresponding to each historical work order record;

[0094] Determine the ratio of the negative correlation normalization result of the interval duration between the user's reservation times of this work order and its previous work order to the similarity of the equipment failure types, and obtain the fourth ratio;

[0095] Fuse the third ratio and the fourth ratio to determine the information difference index corresponding to each historical work order record. Both the third ratio and the fourth ratio are positively correlated with the information difference index.

[0096] For the above steps, taking the i-th historical work order record as an example, obtain the work order reservation information corresponding to the i-th historical work order record, and determine the reservation information completeness corresponding to each historical work order record according to this work order reservation information. In implementation, the repairman can directly score the work order reservation information corresponding to the i-th historical work order record to determine the corresponding reservation information completeness. The value range of this reservation information completeness is 0-1. The larger the value of the reservation information completeness, the higher the completeness of the corresponding work order reservation information. Of course, the trained neural network evaluation model can also be used to quantify the completeness of the work order reservation information to obtain the corresponding reservation information completeness.

[0097] According to the repair equipment information of this work order and the repair equipment information of the previous work order corresponding to the i-th historical work order record, that is, the type of faulty equipment, fault frequency, repair time, required tools, etc., after vectorizing these repair equipment information into features, calculate the cosine similarity between the corresponding two feature vectors, and use this cosine similarity as the equipment fault type similarity corresponding to the i-th historical work order record.

[0098] Based on the reservation information completeness corresponding to the i-th historical work order record, the time difference between the user reservation time and the order placement time of this work order, the equipment fault type similarity, and the interval duration between the user reservation times of this work order and its previous work order, determine the information difference index corresponding to the i-th historical work order record:

[0099]

[0100] Among them, M i represents the information difference index corresponding to the i-th historical work order record; represents the time difference between the user reservation time and the order placement time of this work order corresponding to the i-th historical work order record; S i represents the reservation information completeness corresponding to the i-th historical work order record; represents the interval duration between the user reservation times of this work order and its previous work order corresponding to the i-th historical work order record; cos i represents the equipment fault type similarity corresponding to the i-th historical work order record; exp represents the exponential function with the natural constant e as the base.

[0101] In the above calculation formula, represents the time difference between the user reservation time and the order placement time corresponding to the historical work order record The ratio with the appointment information completeness S i , that is, the third ratio, is the time difference between the user's appointment time and the order placement time . The longer it is, the more time - stressed the user is recently. A longer time difference may generate more uncertainties, and at the same time, the completeness of the appointment information is relatively low, which may result in more consumption of time and energy; represents the time interval between the user's appointment time of this work order and its previous work order . The negative - correlation normalization result of and the similarity cos of the equipment failure types between the i - th historical work order and its previous work order i . That is, the fourth ratio. The larger the numerator of the fourth ratio and the smaller the denominator, it means that the i - th historical work order is greatly affected by its previous work order and the failure types are quite different, and there may be problems such as un - timely equipped accessories and tools; therefore, the larger the product of the two, the greater the information difference corresponding to the i - th historical work order record, and the larger the value of the corresponding information - difference index.

[0102] After determining the information - difference index corresponding to the i - th historical work order record in the above - mentioned manner, use this information - difference index to correct the user feedback factor corresponding to this i - th historical work order record, so as to accurately obtain the actual satisfaction corresponding to the i - th historical work order record:

[0103]

[0104] Among them, U i represents the actual satisfaction corresponding to the i - th historical work order record; M i represents the information - difference index corresponding to the i - th historical work order record; e represents the natural constant; γ i represents the user feedback factor corresponding to the i - th historical work order record.

[0105] In the above - mentioned calculation formula, the larger the information - difference index, the more distorted the user's feedback factor may be. On the contrary, the smaller the information - difference index, the higher the credibility of the user's feedback factor. Therefore, after performing a negative - correlation mapping on this information - difference index and multiplying it by the user feedback factor, the actual satisfaction of the user in the i - th historical work order record is obtained.

[0106] In the above - mentioned manner, the actual satisfaction corresponding to each historical work order record can be determined.

[0107] Considering that in addition to the information difference, there is a large correlation between the user feedback result and the user's own personality. Some users have high requirements for service quality and pay attention to feedback, while some users are tolerant of the service process and do not care about the feedback process; therefore, it is not accurate to directly determine the user portrait factor using the user feedback factor, and it is also necessary to extract the user's personalized characteristics.

[0108] In the embodiments of the present invention, for exampleFigure 4 As shown, the steps for determining the user personalization coefficient corresponding to each historical work order record include:

[0109] Step S221: Determine at least two mutually independent element information in the multi-dimensional association information;

[0110] Step S222: Cluster all historical work order records according to the same type of element information corresponding to all historical work order records to obtain each first work order record sub-cluster corresponding to each type of element information;

[0111] Step S223: Determine the proportion of the number of historical work order records corresponding to the customer's non-feedback in each first work order record sub-cluster in the total number of records in the entire first work order record sub-cluster;

[0112] Step S224: Determine the personalization coefficient of each historical work order record corresponding to each type of element information according to the proportion corresponding to each historical work order record and the difference between the user feedback factor corresponding to each historical work order record and the user feedback factors corresponding to other historical work order records in the first work order record sub-cluster where the historical work order record is located;

[0113] Step S225: Synthesize the personalization coefficients of each historical work order record corresponding to various types of element information to determine the user personalization coefficient corresponding to each historical work order record.

[0114] For the above steps, as an example, first determine at least two types of element information according to the multi-dimensional association information. For example, the Pearson correlation coefficient of any two observed elements can be calculated, and the observed element with the lowest correlation coefficient is selected as the element information. In the embodiment of the present invention, the failure equipment type in the maintenance equipment information of the work order and the work order maintenance address are selected as the two types of element information because both are completely independent elements that are not related to each other. That is, the occurrence of a certain type of failure and where the failure occurs can be regarded as random. The reason for selecting unrelated element information is that due to the limitation of the integrity of the work order record information, there is a lack of logical basis for directly inferring the source of the user feedback result. Therefore, the personalized characteristics of the user feedback information can be observed through the information of users in multiple groups of users with similar unrelated random elements.

[0115] Vectorize the same type of element information recorded in all historical work orders separately, and calculate the difference between the cosine similarity of the value 1 and any two vectors as the distance between the corresponding two historical work order records. Based on this distance, use DBSCAN (Density-Based Spatial Clustering of Applications with Noise) to classify all historical work order records according to the same type of element information, and obtain each first work order record sub-cluster corresponding to each type of element information. Based on the amount of historical work order data within the cluster of each first work order record sub-cluster corresponding to the i-th historical work order record for each type of element information, the number of historical work order records with a user feedback factor of 1, and the difference between the user feedback factor of the i-th historical work order record and the user feedback factors of other work orders within the cluster, determine the user personalization coefficient corresponding to the i-th historical work order record:

[0116]

[0117] where P i,a represents the personalization coefficient of the i-th historical work order record corresponding to the a-th type of element information; γ i represents the user feedback factor of the i-th historical work order record; μ(γ i,a ) represents the average value of the user feedback factors of all historical work order records in the first work order record sub-cluster corresponding to the i-th historical work order record for the a-th type of element information; F i,a represents the number of historical work order records for which the customer has not provided feedback in the first work order record sub-cluster corresponding to the i-th historical work order record for the a-th type of element information, and can be approximated by the number of historical work order records with a user feedback value of 1 in the sub-cluster; G i,a represents the total number of all historical work order records in the first work order record sub-cluster corresponding to the i-th historical work order record for the a-th type of element information.

[0118] In the above calculation formula, the larger |γ i - μ(γ i,a )| is, the greater the difference between the user feedback factor of the user of the i-th historical work order record and that of other similar users in the first work order record sub-cluster corresponding to the a-th type of element information. The larger the numerator in the middle is, it means that there are more similar users who have not given user feedback in the first work order record sub-cluster corresponding to the i-th historical work order record for the a-th type of element information, and its proportion in the first work order record sub-cluster is relatively large. Therefore, the product of the two represents that the users of the i-th historical work order record have not received feedback from other similar users in the first work order record sub-cluster corresponding to the a-th type of element information, and the user feedback factor of the i-th historical work order record is quite different from that of other similar users, indicating that the personalization coefficient of the users of the i-th historical work order record in the group of the first work order record sub-cluster corresponding to the a-th type of element information is higher.

[0119] In the above manner, the personalization coefficient corresponding to each historical work order record for each type of element information can be determined. Taking the i-th historical work order record as an example, calculate the product of the personalization coefficients corresponding to all element information of the i-th historical work order record, and perform normalization processing on this product, so as to obtain the user personalization coefficient corresponding to the i-th historical work order record:

[0120]

[0121] Among them, K i represents the user personalization coefficient corresponding to the i-th historical work order record; P i,a represents the personalization coefficient corresponding to the i-th historical work order record for the a-th type of element information; Q represents the number of types of element information, and in the embodiment of the present invention, Q = 2; norm represents the normalization function.

[0122] In the above calculation formula, the product of the personalization coefficients of all element information The larger it is, it means that the users of the i-th historical work order record show a high degree of personalization in the first work order record sub-clusters of each irrelevant random element information. Then, the user feedback result of this historical work order record is more likely to stem from personalization factors. On the contrary, it means that the user feedback result stems from general factors. After normalizing this product, the user personalization coefficient K of the i-th historical work order record is obtained i . The user personalization coefficient K i The larger it is, the more difficult it is to predict the feedback result of the users of the i-th historical work order record, and it requires a very high quality of maintenance personnel. Therefore, for such users, it is necessary to increase the bias weight of work order assignment.

[0123] Through the above steps, the user personalization coefficient corresponding to each historical work order record can be determined.

[0124] Step S300: Determine the user portrait factor corresponding to each historical work order record according to the difference between the expected satisfaction and the actual satisfaction corresponding to each historical work order record and the user personalization coefficient.

[0125] After determining the expected satisfaction, actual satisfaction, and user personalization coefficient corresponding to each historical work order record through the above steps, based on the difference between the expected satisfaction and actual satisfaction corresponding to each historical work order record, and in combination with the user personalization coefficient, determine the user portrait factor corresponding to each historical work order record.

[0126] In the embodiment of the present invention, as Figure 5 shown, the steps for determining the user portrait factor corresponding to each historical work order record include:

[0127] Step S301: Determine the absolute value of the difference between the expected satisfaction and actual satisfaction corresponding to each historical work order record to obtain the satisfaction deviation;

[0128] Step S302: Based on the user personalization coefficient and the satisfaction deviation corresponding to each historical work order record, determine the user portrait factor corresponding to each historical work order record. Both the user personalization coefficient and the satisfaction deviation are positively correlated with the user portrait factor.

[0129] For the above steps, as an example, to determine the user portrait factor corresponding to each historical work order record:

[0130] Q i =(1 + K i ) × |A i - U i |;

[0131] Among them, Q i represents the user portrait factor corresponding to the i-th historical work order record; K i represents the user personalization coefficient corresponding to the i-th historical work order record; A i represents the expected satisfaction corresponding to the i-th historical work order record; U i represents the actual satisfaction corresponding to the i-th historical work order record.

[0132] In the above calculation formula, |A i - U i | represents the absolute value of the difference between the expected satisfaction and actual satisfaction based on the i-th historical work order, that is, the satisfaction deviation. The larger this absolute value of the difference, the more difficult it is to meet the corresponding user's needs, and a greater bias weight is required. The greater the user personalization coefficient K i , the more the bias weight needs to be increased for the user. Therefore, by combining the two, the user portrait factor corresponding to the i-th historical work order record is finally obtained.

[0133] Through the above steps, the user portrait factor corresponding to each historical work order record can be determined.

[0134] Step S400: Determine the target repairman to whom an order can be assigned based on the current repair ability ratings of each schedulable repairman and the user portrait factors of each historical work order record of the current reserved user.

[0135] Based on the current repair ability ratings of each schedulable repairman and the user portrait factors of each historical work order record of the current reserved user, determine the target repairman to whom an order can be assigned.

[0136] In an embodiment of the present invention, as Figure 6 shown, the steps for determining the target repairman to whom an order can be assigned include:

[0137] Step S401: Determine the average value of the user portrait factors of each historical work order record of the current reserved user to obtain an average user portrait factor;

[0138] Step S402: Normalize the average user portrait factor to obtain a first normalized value;

[0139] Step S403: Normalize the repair ability rating of each schedulable repairman to obtain a second normalized value;

[0140] Step S404: Determine the schedulable repairman corresponding to the minimum difference between the second normalized value and the first normalized value as the target repairman to whom an order can be assigned.

[0141] For the above steps, as an example, if a user has multiple historical work order records, then calculate the mean value of the user portrait factors corresponding to each of its historical work order records to obtain an average user portrait factor. For any real-time reserved user, that is, the current reserved user, normalize the average user portrait factor of the current reserved user using the maximum value of the average user portrait factors of all users to obtain a first normalized value; at the same time, normalize the repair ability rating of each current schedulable repairman using the maximum value of the repair ability ratings to obtain a second normalized value, and determine the difference between the first normalized value and the second normalized value:

[0142]

[0143] Among them, E represents the minimum value of the difference between the first normalized value corresponding to the current reserved user and the second normalized values corresponding to each current schedulable repairman; Q represents the average user portrait factor corresponding to the current reserved user; Q max represents the maximum value of the average user portrait factors of all users; J represents the repair ability rating of each current schedulable repairman; J max represents the maximum value of the repair ability ratings; min represents the minimum value function.

[0144] In the above calculation formula, Represents the maximum normalization of the average user portrait factor of real-time reservation users, Represents the maximum normalization of the repair ability rating of any current schedulable repairman, When it represents the minimum value of the absolute value of the difference between the two, the matching degree of dispatching repairmen for the work orders of real-time reservation users is the highest.

[0145] After determining the minimum value E of the difference between the first normalization value corresponding to the current reservation user and the second normalization values corresponding to each current schedulable repairman in the above manner, the schedulable repairman corresponding to the minimum value E is determined as the target repairman to be dispatched, and a work order is dispatched to the target repairman.

[0146] It should be understood that if there is no historical work order record for the current reservation user, the target repairman to be dispatched cannot be determined according to the above method at this time. At this time, other methods need to be used to determine the target repairman to be dispatched, that is:

[0147] If there is no historical work order record for the current reservation user, then according to the user information and work order information corresponding to each historical work order record, all historical work order records are clustered to obtain each second work order record sub-cluster;

[0148] Match the user information and work order information of the current reservation user with the user information and work order information corresponding to the historical work order records in the second work order record sub-cluster to obtain the matching second work order record sub-cluster of the current reservation user;

[0149] Determine the average value of the user portrait factors of the historical work order records in the matching second work order record sub-cluster to obtain the user portrait factors of the reservation work order of the current reservation user;

[0150] According to the repair ability ratings of each current schedulable repairman and the user portrait factors of the reservation work order of the current reservation user, determine the target repairman to be dispatched.

[0151] For the above steps, as an example, if there is no historical work order record for the current reservation user, at this time, the user information and work order information corresponding to each historical work order record are vectorized, and based on the vectors of each obtained historical work order record, all historical work order records are clustered in the same manner as obtaining each first work order record sub-cluster above to obtain each second work order record sub-cluster.

[0152] Vectorize the user information and work order information of the current reservation user to obtain the vector to be matched. At the same time, vectorize the user information and work order information of the same type in each historical work order record in each second work order record sub-cluster to obtain the target matching vector. Match the vector to be matched with the target matching vectors of each historical work order record in each second work order record sub-cluster, that is, calculate the average value of the cosine similarity between the vector to be matched and the target matching vectors of each historical work order record in each second work order record sub-cluster, and use the second work order record sub-cluster corresponding to the maximum value of the average cosine similarity as the matching second work order record sub-cluster. Calculate the average value of the user portrait factors of each historical work order record in the matching second work order record sub-cluster, and use this average value as the user portrait factor of the reservation work order of the current reservation user.

[0153] At this time, according to the repair ability ratings of the current schedulable repairmen and the user portrait factor of the reservation work order of the current reservation user, the target repairman to be dispatched can be determined according to the above steps S401 - S404.

[0154] Based on the same inventive concept, an embodiment of the present invention further provides a data-driven automatic assignment system for home appliance work orders, as Figure 7 shown. The system includes: a memory 701, a processor 702, and a computer program 703 stored in the memory 701 and running on the processor 702. When the processor 702 executes the computer program 703, the system can execute any of the above-described data-driven automatic assignment methods for home appliance work orders.

[0155] Embodiments of the present invention can divide the functions of the system according to the above method examples. For example, each function module can be corresponding, or two or more functions can be integrated into one processing module. The above integrated modules can be implemented in the form of hardware. It should be noted that the division of modules in this embodiment is illustrative, only a logical function division, and there can be other division methods in actual implementation.

[0156] Based on the same inventive concept, an embodiment of the present invention further provides a computer program product, which includes: computer program code. When the computer program code runs on a computer, the computer is caused to execute any of the above-described data-driven automatic assignment methods for home appliance work orders.

[0157] Based on the same inventive concept, an embodiment of the present invention further provides a computer-readable storage medium, which stores computer program code. When the computer program code runs on a computer, the computer is caused to execute any of the above-described data-driven automatic assignment methods for home appliance work orders.

[0158] It should be noted that the above-described embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the various embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A data-driven automatic assignment method for household appliance work orders, characterized in that, Including the following steps: Obtain the multi-dimensional correlation information corresponding to each historical work order record, where the multi-dimensional correlation information at least includes: maintenance worker information, user information, and work order information; Analyze the maintenance worker information, user information, and work order information in the multi-dimensional correlation information to determine the expected satisfaction, actual satisfaction, and user personalization coefficient corresponding to each historical work order record; Determine the user portrait factor corresponding to each historical work order record according to the difference between the expected satisfaction and the actual satisfaction corresponding to each historical work order record and the user personalization coefficient; Determine the target maintenance worker to be dispatched according to the maintenance ability rating of each currently schedulable maintenance worker and the user portrait factors of each historical work order record of the currently reserved user; The maintenance worker information at least includes: maintenance ability rating, departure address of the maintenance worker, time consumed for the maintenance worker to reach the maintenance location, and maintenance time consumed; the user information at least includes: user feedback information, work order reservation information; the work order reservation information at least includes: user reservation time, user order time, and work order maintenance address of this work order and its previous work order; the work order information at least includes: maintenance equipment information of this work order, maintenance equipment information of the previous work order of this work order; The steps for determining the actual satisfaction corresponding to each historical work order record include: Determine the user feedback factor corresponding to each historical work order record according to the user feedback information corresponding to each historical work order record; Determine the completeness of the reservation information corresponding to each historical work order record according to the work order reservation information corresponding to each historical work order record; Determine the ratio of the time difference between the user reservation time and the order time of this work order corresponding to each historical work order record to the completeness of the reservation information to obtain a third ratio; Conduct a similarity analysis on the maintenance equipment information of this work order and the maintenance equipment information of the previous work order of this work order corresponding to each historical work order record to determine the equipment failure type similarity corresponding to each historical work order record; Determine the ratio of the negative correlation normalization result of the interval duration between the user reservation times of this work order and its previous work order to the equipment failure type similarity to obtain a fourth ratio; Fuse the third ratio and the fourth ratio to determine the information difference index corresponding to each historical work order record, and both the third ratio and the fourth ratio are positively correlated with the information difference index; Use the information difference index to correct the user feedback factor to obtain the actual satisfaction corresponding to each historical work order record.

2. The automatic assignment method of home appliance work orders driven by data according to claim 1, characterized in that The steps for determining the expected satisfaction corresponding to each historical work order record include: Determine the ratio of the maintenance ability rating to the maintenance time consumed corresponding to each historical work order record to obtain a first ratio; Determine the ratio of the normalized value of the distance from the departure address of the maintenance worker to the work order maintenance address to the time consumed for the maintenance worker to reach the maintenance location corresponding to each historical work order record to obtain a second ratio; Fuse the first ratio and the second ratio to determine the expected satisfaction corresponding to each historical work order record, and both the first ratio and the second ratio are positively correlated with the expected satisfaction.

3. The automatic assignment method of home appliance work orders driven by data according to claim 1, characterized in that The steps for determining the user personalization coefficient corresponding to each historical work order record include: Determine at least two mutually independent element information in the multi-dimensional association information; Cluster all historical work order records according to the same type of element information corresponding to all historical work order records, and obtain each first work order record sub-cluster corresponding to each type of element information; Determine the proportion of the number of historical work order records not feedback by the customer in each first work order record sub-cluster where each historical work order record is located in the total number of the first work order record sub-cluster; Determine the personalized coefficient of each historical work order record corresponding to each type of element information according to the proportion of the number corresponding to each historical work order record and the difference between the user feedback factor corresponding to each historical work order record and the user feedback factors corresponding to other historical work order records in the first work order record sub-cluster where the historical work order record is located; Comprehensively determine the user personalized coefficient corresponding to each historical work order record according to the personalized coefficients of each historical work order record corresponding to various types of element information.

4. A method for automatically dispatching home appliance work orders driven by data according to claim 3, characterized in that, The maintenance equipment information at least includes the type of faulty equipment, and the element information at least includes: the work order repair address and the type of faulty equipment.

5. A method for automatically dispatching household appliance work orders driven by data according to claim 1, characterized in that Determine the user portrait factor corresponding to each historical work order record, including: Determine the absolute value of the difference between the expected satisfaction and the actual satisfaction corresponding to each historical work order record to obtain the satisfaction deviation; Comprehensively determine the user portrait factor corresponding to each historical work order record according to the user personalized coefficient corresponding to each historical work order record and the satisfaction deviation, and both the user personalized coefficient and the satisfaction deviation are positively correlated with the user portrait factor.

6. The automatic assignment method of home appliance work orders driven by data according to claim 1, wherein Determine the target repairman to whom the work order can be assigned, including: Determine the average value of the user portrait factors of the historical work order records of the current reservation user to obtain the average user portrait factor; Normalize the average user portrait factor to obtain the first normalized value; Normalize the maintenance ability ratings of each schedulable repairman to obtain the second normalized value; Determine the schedulable repairman corresponding to the minimum difference between the second normalized value and the first normalized value as the target repairman to whom the work order can be assigned.

7. A method for automatically dispatching household appliance work orders using data-driven according to claim 1, characterized in that, This method further includes: If there are no historical work order records for the current reservation user, cluster all historical work order records according to the user information and work order information corresponding to each historical work order record to obtain each second work order record sub-cluster; Match the user information and work order information of the current reservation user with the user information and work order information corresponding to the historical work order records in the second work order record sub-cluster to obtain the matching second work order record sub-cluster of the current reservation user; Determine the average value of the user portrait factors of the historical work order records in the matching second work order record sub-cluster to obtain the user portrait factor of the reservation work order of the current reservation user; Determine the target repairman to whom the work order can be assigned according to the maintenance ability ratings of each current schedulable repairman and the user portrait factor of the reservation work order of the current reservation user.

Images