CN116993188A

CN116993188A - Maintenance manufacturer evaluation method, device, equipment and storage medium

Info

Publication number: CN116993188A
Application number: CN202211009106.9A
Authority: CN
Inventors: 黄蕴思; 江恒; 吴宇明
Original assignee: China Mobile Communications Group Co Ltd; China Mobile Group Guangdong Co Ltd
Current assignee: China Mobile Communications Group Co Ltd; China Mobile Group Guangdong Co Ltd
Priority date: 2022-08-22
Filing date: 2022-08-22
Publication date: 2023-11-03

Abstract

The application relates to the technical field of operation and maintenance services, and provides a maintenance manufacturer evaluation method, a maintenance manufacturer evaluation device, maintenance manufacturer evaluation equipment and a storage medium. The method comprises the following steps: acquiring evaluation data of a maintenance manufacturer and constructing a feature vector set of the maintenance manufacturer; the feature vector set comprises a plurality of first feature vectors; any one of the first feature vectors is composed of one or more evaluation index values of a maintenance service provided by a maintenance manufacturer; performing principal component analysis on the feature vector set to obtain a key index data set of a maintenance manufacturer; determining the type of a maintenance manufacturer according to the key index data set; the service score of the maintenance vendor is determined in conjunction with the vendor type. The key index data set is extracted through principal component analysis, so that maintenance manufacturers are classified, and the outstanding service items of the maintenance manufacturers can be determined; the service score of the maintenance manufacturer is determined by combining the type of the maintenance manufacturer and the key index data set, so that comprehensive evaluation can be performed on each maintenance service of the maintenance manufacturer, and the accuracy of an evaluation result is improved.

Description

Maintenance manufacturer evaluation method, device, equipment and storage medium

Technical Field

The application relates to the technical field of operation and maintenance services, in particular to a maintenance manufacturer evaluation method, a maintenance manufacturer evaluation device, maintenance manufacturer evaluation equipment and a storage medium.

Background

Along with the development of information technology and popularization of paperless office concepts, more and more informationized systems are built by enterprises in various industries, and the enterprises can improve collaborative office or work efficiency by means of informationizing through building the informationized systems. Different informationized systems have different functions, and professional technicians are required to maintain the systems so as to ensure the normal operation of the systems. Because of the wide use of information systems, the system maintenance field is wide, and there are many maintenance manufacturers that provide maintenance services, and as the system functions tend to be complicated, the maintenance manufacturers can generally provide various maintenance service items.

The enterprise evaluates and examines the service items of the maintenance manufacturer, which is the basis for penalty or surfacing of the maintenance manufacturer, and is also the reference basis for whether the maintenance manufacturer is continuously selected in the follow-up process. At present, most of the enterprises are responsible authorities responsible for managing maintenance manufacturers, service items of the maintenance manufacturers are evaluated according to assessment indexes formulated by each service item, the assessment standards of different responsible authorities are different, personal subjectivity is high, errors are easy to occur in manual assessment, and accuracy of assessment results cannot be guaranteed.

Disclosure of Invention

The embodiment of the application provides a maintenance manufacturer evaluation method, a device, equipment and a storage medium, which are used for solving the technical problems of inaccurate evaluation results caused by different evaluation standards and strong subjectivity in the existing manual evaluation mode.

In a first aspect, an embodiment of the present application provides a maintenance manufacturer evaluation method, including:

acquiring evaluation data of a maintenance manufacturer, and constructing a feature vector set of the maintenance manufacturer based on the evaluation data; the feature vector set comprises a plurality of first feature vectors; any one of the first feature vectors is composed of one or more evaluation index values of a maintenance service provided by the maintenance manufacturer;

performing principal component analysis on the feature vector set to obtain a key index data set of the maintenance manufacturer;

determining the manufacturer type of the maintenance manufacturer according to the key index data set;

and determining service scores of the maintenance manufacturer according to the manufacturer type of the maintenance manufacturer and the key index data set.

In one embodiment, the step of determining the service score of the maintenance vendor according to the vendor type of the maintenance vendor and the key index data set includes:

Acquiring an optimal service data set and a weight value corresponding to the manufacturer type; the optimal service data set comprises optimal scores of all evaluation indexes corresponding to the manufacturer types;

extracting an evaluation index value corresponding to the manufacturer type from the key index data set to obtain a target index data set of the maintenance manufacturer;

determining the similarity between the target index data set and the optimal service data set, and determining the initial score of the maintenance manufacturer according to the similarity;

and determining the service score of the maintenance manufacturer according to the weight value and the initial score.

In one embodiment, the step of performing principal component analysis on the feature vector set to obtain a key index data set of the maintenance manufacturer includes:

calculating the average value of each first feature vector in the feature vector set to obtain a second feature vector;

based on the second feature vector, performing decentration processing on any one of the first feature vectors to obtain a first index data set;

calculating a covariance matrix of any third eigenvector in the first index data set, and performing eigenvalue decomposition on the covariance matrix to obtain a fourth eigenvector of the covariance matrix; the third feature vector is any feature vector in the first index data set;

Sorting all the characteristic values in the fourth characteristic vector, selecting a preset number of characteristic values according to the sorting order, and generating a target characteristic vector of the covariance matrix;

and determining a key index data set of the maintenance manufacturer according to each target feature vector.

In one embodiment, the step of determining the vendor type of the maintenance vendor from the key index data set includes:

inputting the key index data set into a pre-trained manufacturer classification detection model; the vendor classification detection model comprises a plurality of classifiers;

based on the service types of the maintenance services corresponding to the key index data sets, classifying and deciding the key index data sets by utilizing the classifiers to obtain the prediction results of the service types of the key index data sets by the classifiers; the prediction result of the jth classifier is the accumulated value of the prediction value of the jth classifier and the prediction value of the previous j-1 classifiers, and j is a positive integer greater than 1;

calculating residual values of the prediction result of any target classifier and the prediction result of the adjacent classifier; the target classifier is any classifier in the manufacturer classification detection model;

And determining an optimal decision result of the vendor classification detection model on the service type of the key index data set based on the gradient descent principle of the residual error value, and determining the vendor type of the maintenance vendor according to the optimal decision result.

In one embodiment, the step of constructing the feature vector set of the maintenance vendor based on the evaluation data includes:

classifying each evaluation index in the evaluation data according to the evaluation party of the evaluation data to obtain a second index data set corresponding to each evaluation party;

classifying each evaluation index in the second index data set according to the service type of each maintenance service provided by the maintenance manufacturer to obtain a third index data set corresponding to each service type; and constructing a first feature vector based on the third index data set;

and constructing a feature vector set of the maintenance manufacturer according to each first feature vector.

In one embodiment, the step of obtaining the evaluation data of the maintenance manufacturer includes:

acquiring evaluation records of each evaluation party on maintenance manufacturers; the evaluation record comprises a manufacturer code, a manufacturer name and a plurality of evaluation index data;

Identifying an abnormal evaluation record from the evaluation records; the abnormality evaluation record includes abnormality data; the abnormal data comprise index missing data and index repeated data;

and cleaning the abnormal evaluation record to obtain the evaluation data of the maintenance manufacturer.

In one embodiment, after determining the service score of the maintenance manufacturer according to the manufacturer type of the maintenance manufacturer and the key index data set, the method further includes:

acquiring historical scores of the maintenance manufacturer;

calculating Euclidean distance between the service score and the historical score, and verifying the validity of the service score according to the Euclidean distance;

and if the verification is passed, outputting and displaying the service score to an evaluator of the maintenance manufacturer to acquire feedback information of the evaluator.

In a second aspect, an embodiment of the present application provides a maintenance manufacturer evaluation apparatus, including:

the feature construction module is used for acquiring evaluation data of a maintenance manufacturer and constructing a feature vector set of the maintenance manufacturer based on the evaluation data; the feature vector set comprises a plurality of first feature vectors; any one of the first feature vectors is composed of one or more evaluation index values of a maintenance service provided by the maintenance manufacturer;

The feature dimension reduction module is used for carrying out principal component analysis on the feature vector set to obtain a key index data set of the maintenance manufacturer;

the classification detection module is used for determining the manufacturer type of the maintenance manufacturer according to the key index data set;

and the comprehensive evaluation module is used for determining the service score of the maintenance manufacturer according to the manufacturer type of the maintenance manufacturer and the key index data set.

In a third aspect, an embodiment of the present application provides an electronic device, including a processor and a memory storing a computer program, where the processor implements the steps of the maintenance manufacturer evaluation method described in the first aspect when executing the program.

In a fourth aspect, an embodiment of the present application provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the maintenance vendor evaluation method of the first aspect.

According to the maintenance manufacturer evaluation method, the device, the equipment and the storage medium, the key index data set is obtained by carrying out principal component analysis on the feature vector set formed by the evaluation index values of the maintenance manufacturer, and the manufacturer type of the maintenance manufacturer is determined according to the key index data set, so that the service type of better maintenance service provided by the maintenance manufacturer can be determined, the subjective influence of evaluation data is reduced, and the accuracy of the evaluation data is improved; according to the key index data set and the manufacturer type, service scores of maintenance manufacturers are determined, various maintenance services provided by the maintenance manufacturers can be comprehensively evaluated, and accuracy of evaluation results is improved.

Drawings

In order to more clearly illustrate the application or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the application, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic flow chart of a maintenance manufacturer evaluation method according to an embodiment of the present application;

FIG. 2 is a second flow chart of a maintenance manufacturer evaluation method according to an embodiment of the present application;

FIG. 3 is a third flow chart of a method for evaluating a maintenance manufacturer according to an embodiment of the present application;

FIG. 4 is a schematic flow chart of classifying maintenance manufacturers in the maintenance manufacturer evaluation method according to the embodiment of the present application;

FIG. 5 is a schematic diagram of a maintenance manufacturer evaluation device according to an embodiment of the present application;

fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

Fig. 1 is a schematic flow chart of a maintenance manufacturer evaluation method according to an embodiment of the present application. Referring to fig. 1, the maintenance manufacturer evaluation method provided by the embodiment of the application includes:

step 100, acquiring evaluation data of a maintenance manufacturer, and constructing a feature vector set of the maintenance manufacturer based on the evaluation data; the feature vector set comprises a plurality of first feature vectors; any one of the first feature vectors is composed of one or more evaluation index values of a maintenance service provided by the maintenance manufacturer;

in this embodiment, when evaluating a maintenance manufacturer, first, evaluation data of the maintenance manufacturer is obtained, where the evaluation data is evaluation information of service items provided by the evaluation party to the maintenance manufacturer, and specifically includes a service evaluation index and an index value, where the index value is a score of the evaluation party to the service items of the maintenance manufacturer. The evaluation data can be acquired through an automatic information acquisition platform, and each evaluation party scores service items provided by maintenance manufacturers through the automatic information acquisition platform to generate corresponding evaluation data. It should be noted that, the same evaluator may evaluate a plurality of maintenance manufacturers, or may evaluate one or more service items provided by the same maintenance manufacturer, that is, the evaluator may serve as a service sharing receiver (hereinafter referred to as a "first party" or a "office party"), and may enjoy service items provided by a plurality of maintenance manufacturers or a plurality of service items provided by the same maintenance manufacturer at the same time; correspondingly, one or more service items provided by the same maintenance manufacturer can be evaluated by one or more evaluation parties, namely, the same maintenance manufacturer can provide services for a plurality of first parties and can also provide a plurality of services for the same first party; in the first party, when evaluating a certain service provided by a maintenance manufacturer, a plurality of evaluation parties may participate in the service evaluation, which is not limited herein.

Constructing a feature vector set of a maintenance manufacturer based on evaluation data of the maintenance manufacturer, wherein the feature vector set comprises a plurality of first feature vectors corresponding to various maintenance services provided by the maintenance manufacturer or corresponding to different evaluation parties; any one of the first feature vectors is constituted by one or more evaluation index values of a maintenance service provided by a maintenance manufacturer. It is known that, the service items provided by the maintenance manufacturers are different, the indexes for evaluating the maintenance manufacturers are different, and for the evaluation data of the same maintenance manufacturer, the feature vectors are constructed based on the service items provided by the maintenance manufacturers, and different service items and/or different evaluation parties of the same maintenance manufacturer correspond to different feature vectors. It can be understood that the maintenance manufacturer can provide the same service item to the plurality of first parties, so that any service item of the maintenance manufacturer can correspond to one or more first feature vectors according to different evaluation parties, and different evaluation parties respectively correspond to different feature vectors; any evaluator may also correspond to one or more first feature vectors according to different service items; any of the first feature vectors may include an evaluation index value of one or more of the evaluators for one or more of the service items provided by the maintenance vendor.

That is, the first feature vectors in the feature vector set may be constructed from different dimensions, specifically including classifying the evaluation data from dimensions of different service items provided by a maintenance manufacturer, and constructing first feature vectors, where each service item corresponds to a first feature vector, and any first feature vector may include an evaluation index value of one or more evaluators for the same service item; classifying the evaluation data from the dimension of the evaluation party, and constructing first feature vectors, wherein each evaluation party corresponds to one first feature vector, and any first feature vector can comprise evaluation index values of the same evaluation party on one or more service items. Further, the evaluation index data can be classified from the dimensions of different service items provided by a maintenance manufacturer, and then the evaluation data can be further classified from the dimensions of an evaluation party; or classifying the evaluation data from the dimensionality of the evaluation party, further classifying the evaluation data from the dimensionality of different service items provided by the maintenance manufacturer on the basis of the classification, and finally constructing a first feature vector based on the evaluation index value of the same service item provided by the same evaluation party to the maintenance manufacturer, thereby forming a feature vector set; the first characteristic vector comprises the evaluation index value of the same service item provided by the same evaluation party to the maintenance manufacturer.

Further, the service items provided by the maintenance manufacturers are different, and the evaluation indexes are also different, for example, when the maintenance manufacturers provide development services, the evaluation indexes mainly include quality data and frequency data. The quality data comprises engineering data and fault data, wherein the engineering data specifically comprises: the system availability, the number of times of version online, the success rate of first online, the time consumption of online, and whether emergency rollback or emergency version is needed after engineering failure; the fault data specifically includes: fault rate, fault response time, fault handling quality (fault recurrence); the frequency data comprises engineering online times, and the engineering online times comprise common online times and emergency online times.

When a maintenance manufacturer provides investigation and design services (responsible for investigation and pre-review and post-evaluation), the evaluation indexes mainly comprise: investigation duration, investigation and development demand deviation degree, front review working quality (including financial benefit, construction necessity, feasibility, rationality, technical scheme and production condition) and actual deviation condition, and post-evaluation (including re-evaluation of front review content, comparison deviation, and accuracy and implementation efficiency of project decision) are also included.

When a maintenance manufacturer provides business operation and maintenance services (responsible for developing a system, maintaining code level problems and maintaining all parts of the system before operation and maintenance delivery), the evaluation indexes mainly comprise: project implementation progress, project carry-over problems at each period (including fault problems caused by maintenance manufacturers), maintenance quality, version change engineering quality (number of on-line versions, success failure rate, first on-line success rate, number of emergency version processing version faults on the next day), fault processing quality (including response time of processing, processing time, fault rate and system availability).

When a maintenance manufacturer provides basic operation and maintenance services (responsible for supporting the availability requirements of a system, including popularization and implementation of an automation tool, standardized maintenance of a universal component and the like), the evaluation indexes mainly comprise: fault handling quality (including handling response time, handling time, fault rate, system availability), active-passive fault discovery rate, security hardened service quality (including hardened service response time, handling time, problem-solving replication rate).

When a maintenance manufacturer provides a foreground operation service (responsible for supporting the service requirements of a terminal user, including service functions, processes, data maintenance, account rights and the like), the evaluation indexes mainly comprise: the work order processing quality (including work order response time, processing time, work order user score, user complaint), system availability and the like.

When a maintenance manufacturer provides a database maintenance service (responsible for supporting the demands of database availability and the like), the evaluation indexes mainly include: fault handling quality (including handling response time, handling time, fault rate, system availability), active-passive fault discovery rate, security hardened service quality (including hardened service response time, handling time, problem-solving replication rate).

When a maintenance vendor provides a host OS maintenance service (an operating system responsible for maintaining a host, including handling failures, security reinforcement, etc.), the evaluation index mainly includes: fault handling quality (including handling response time, handling time, fault rate, system availability), active-passive fault discovery rate, security hardened service quality (including hardened service response time, handling time, problem-solving replication rate).

When a maintenance manufacturer provides a host hardware maintenance service (is responsible for maintaining a host hardware layer), the evaluation indexes mainly comprise: fault handling quality (including handling response time, handling time, fault rate, system availability), active-passive fault discovery rate, security hardened service quality (including hardened service response time, handling time, problem-solving replication rate).

Based on this, in step 100, the acquiring the evaluation data of the maintenance manufacturer may specifically include:

step 101, acquiring an evaluation record of each evaluation party on a maintenance manufacturer; the evaluation record comprises a manufacturer code, a manufacturer name and a plurality of evaluation index data;

step 102, identifying an abnormal evaluation record from the evaluation records; the abnormality evaluation record includes abnormality data; the abnormal data comprise index missing data and index repeated data;

and step 103, cleaning the abnormal evaluation record to obtain the evaluation data of the maintenance manufacturer.

When acquiring evaluation data, firstly acquiring an evaluation record of a maintenance manufacturer, wherein the evaluation record is original evaluation data of the maintenance manufacturer by an evaluation party, and comprises a manufacturer code, a manufacturer name and a plurality of evaluation index data; the manufacturer codes are maintenance manufacturer IDs, are unique identity marks of maintenance manufacturers, and can be used for distinguishing different maintenance manufacturers; the evaluation index data includes an index name and an index value. Since there are many evaluation indexes, the acquired evaluation records may have abnormal data, and thus, the acquired evaluation records need to be preprocessed before the feature vectors of the maintenance manufacturer are constructed.

Specifically, the obtained evaluation record is a real record of the service provider in providing the actual service, and one evaluation record of the same evaluation party comprises: a maintenance vendor ID, a maintenance vendor name, a number of evaluation index data, which may include an index to one or more service items provided by the maintenance vendor. Identifying an abnormal evaluation record from the acquired evaluation records, the abnormal evaluation record being an evaluation record containing abnormal data; the abnormal data includes index missing data and index repeated data, and may further include index data in which the index value exceeds a preset interval standard. The index missing data specifically refers to that an evaluation party does not evaluate certain indexes of one or more service items to be evaluated; the index repetition data specifically means that the same evaluation party repeatedly evaluates the same index of a certain service item, thereby generating repeated evaluation index data. And cleaning the abnormal evaluation record to obtain the evaluation data of a maintenance manufacturer.

Cleaning the abnormal evaluation records, specifically deleting the whole evaluation record for the abnormal evaluation records containing index missing data; or when the index deletion ratio reaches a set ratio threshold, for example, when the deletion rate of the evaluation index data is greater than 20%, deleting the whole corresponding evaluation record. And (3) for abnormal evaluation records containing index repeated data, performing index deletion on the repeated evaluation index data in the evaluation records, and only keeping one record of the evaluation index data.

Further, a numerical interval may be set for each item of evaluation index data in advance according to an actual requirement, after the evaluation record is obtained, each item of evaluation index data in each complete record is checked, and when the index value of the evaluation index data is out of the preset numerical interval, the item of evaluation index data is removed and marked as an abnormal data item. Since the elimination of the index may cause index deletion, after one evaluation record is inspected, the deletion rate of the evaluation index data in the finally obtained evaluation record is detected, the deletion rate satisfies the condition, and the index value of the evaluation index data also conforms to a preset value interval, for example, an evaluation record in which the index value of 80% or more of the evaluation index data is within the preset value interval is used as the evaluation data for providing the corresponding service by the maintenance manufacturer.

And constructing a data set data= { maintenance manufacturer ID, maintenance manufacturer name and a plurality of evaluation index data } based on the evaluation data obtained through preprocessing, and mapping the evaluation index data in the data set into specific numerical values to obtain corresponding first feature vectors. Wherein, the specific value of each evaluation index data map may be an index value of the evaluation index data, the index value may be a service score of the evaluation party to the maintenance manufacturer, and the first feature vector is, for example, x= { X ₁ ,x ₂ ,x ₃ ,…x _n And n is the index number of the evaluation index data. When constructing the feature vector, the missing index data may be skipped or replaced by a special value such as "-1" when mapping the index value, and it is known that the score of the evaluator to the service manufacturer is generally a value greater than or equal to 0, and the missing index data may be replaced by a value less than 0 to indicate distinction.

Step 200, performing principal component analysis on the feature vector set to obtain a key index data set of the maintenance manufacturer;

and (3) carrying out principal component analysis on the feature vector set based on the constructed feature vector set, namely carrying out principal component analysis on each first feature vector in the feature vector set, and screening based on the feature value in each first feature vector, so as to screen out main influence factors of service evaluation and obtain a key index data set of a maintenance manufacturer. Because of more evaluation index items, some unimportant index items need to be removed from each first feature vector, so that the data dimension reduction of the first feature vector is realized, and the obtained key index data set comprises the evaluation index values of one or more service items provided by a maintenance manufacturer.

Step 300, determining the manufacturer type of the maintenance manufacturer according to the key index data set;

And 400, determining service scores of the maintenance manufacturer according to the manufacturer type of the maintenance manufacturer and the key index data set.

Classifying maintenance manufacturers based on a key index data set obtained after the principal component analysis of the feature vector set, so as to determine manufacturer types of the maintenance manufacturers; the vendor type is determined by the service type of the maintenance service that the maintenance vendor can provide, for example, the maintenance type of the maintenance vendor that provides the development service is a developer, the vendor type of the maintenance vendor that provides the front-end operation service is a front-end operation vendor, and so on. Since the same maintenance vendor can provide one or more types of maintenance services, the vendor type of the same maintenance vendor is uncertain, or the vendor type of a service integrator (maintenance vendor that can provide various types of maintenance services) can be said to be diverse. The maintenance manufacturer is classified based on the index data set obtained by the principal component analysis, so that the better maintenance service provided by the maintenance manufacturer can be highlighted.

And comprehensively evaluating the maintenance manufacturer according to the manufacturer type and the key index data set, and determining service scores of the maintenance manufacturer, wherein the service scores are comprehensive evaluation results of the maintenance manufacturer. Specifically, the key index data set is used for scoring evaluation indexes of services provided by maintenance manufacturers, and different weight values are given to different service items, namely evaluation indexes corresponding to different types of services; and combining the manufacturer types, determining service items provided by the maintenance manufacturer and corresponding to the manufacturer types, adding weight values, and finally taking a weighted average value or a weighted sum of all evaluation indexes in the key index data set as an evaluation score of the maintenance manufacturer to obtain service scores of the maintenance manufacturer. Or determining an evaluation index corresponding to the manufacturer type and an index value thereof from the key index data set based on the manufacturer type; different weight values are respectively given to the evaluation indexes corresponding to the manufacturer types and other evaluation indexes in the key index data set; finally, a weighted average or a weighted sum of the evaluation indexes in the key index data set is calculated and used as a service score of the maintenance manufacturer, which is not limited herein.

In the embodiment, the key index data set is obtained by performing principal component analysis on the feature vector set formed by the evaluation index values of the maintenance manufacturer, and the manufacturer type of the maintenance manufacturer is determined according to the key index data set, so that the service type of the better maintenance service provided by the maintenance manufacturer can be determined, the subjective influence of evaluation data is reduced, and the accuracy of the evaluation data is improved; according to the key index data set and the manufacturer type, service scores of maintenance manufacturers are determined, various maintenance services provided by the maintenance manufacturers can be comprehensively evaluated, and accuracy of evaluation results is improved.

Further, the obtained original evaluation record is preprocessed, the abnormal evaluation record containing abnormal data is cleaned, the abnormal data is removed, and the usability and the effectiveness of the evaluation data are ensured.

In one embodiment, in step 100, the constructing a feature vector set of the maintenance manufacturer based on the obtained evaluation data may specifically further include:

step 104, classifying each evaluation index in the evaluation data according to the evaluation party of the evaluation data to obtain a second index data set corresponding to each evaluation party;

Step 105, classifying each evaluation index in the second index data set according to the service type of each maintenance service provided by the maintenance manufacturer, to obtain a third index data set corresponding to each service type; and constructing a first feature vector based on the third index data set;

and step 106, constructing a feature vector set of the maintenance manufacturer according to each first feature vector.

In this embodiment, when the feature vector set of the maintenance manufacturer is constructed based on the evaluation data, each evaluation index in the evaluation data is first classified according to each evaluation party of the evaluation data, and the evaluation indexes of the same evaluation party are classified into the same class, so as to obtain an index data set corresponding to each evaluation party, that is, a second index data set. Based on the index data sets corresponding to all the evaluation parties, further classifying the index data set of any evaluation party according to different service types of maintenance service provided by a maintenance manufacturer to obtain index data sets corresponding to the maintenance service of all the service types, namely a third index data set; constructing a first feature vector based on the third index data set; and generating a feature vector set according to each constructed first feature vector. The first characteristic vector comprises all evaluation index values of the same maintenance service provided by the same evaluation party for the maintenance manufacturer.

It can be understood that the evaluation data may be classified according to different service types of each maintenance service provided by the maintenance manufacturer, so as to obtain an index data set corresponding to each service type, namely, a second index data set; then, classifying each evaluation index in the index data set corresponding to any service type based on the index data set corresponding to each service type according to the difference of the evaluation parties, and obtaining an index data set corresponding to each evaluation party, namely a third index data set; and constructing first feature vectors based on the third index data set, and constructing a feature set of a maintenance manufacturer according to each first feature vector. The first characteristic vector comprises all evaluation index values of the same maintenance service provided by the same evaluation party for the maintenance manufacturer.

Further, step 200 may further include:

step 201, calculating the average value of each first feature vector in the feature vector set to obtain a second feature vector;

step 202, based on the second feature vector, performing a decentration process on any one of the first feature vectors to obtain a first index data set;

step 203, calculating a covariance matrix of any third eigenvector in the first index data set, and performing eigenvalue decomposition on the covariance matrix to obtain a fourth eigenvector of the covariance matrix; the third feature vector is any feature vector in the first index data set;

Step 204, sorting the feature values in the fourth feature vector, and selecting a preset number of feature values according to the sorting order to generate a target feature vector;

and step 205, determining a key index data set of the maintenance manufacturer according to each target feature vector.

When principal component analysis is performed on the feature vector set, the average value of each first feature vector is calculated based on each first feature vector in the feature vector set, and a second feature vector is obtained. Specifically, if the first feature vector is: x= { X ₁ ,x ₂ ,x ₃ ,…x _n N evaluation index values, based on the same evaluation index value x in each first feature vector _i Calculating the average valueObtain a second feature vector->Based on the obtained second feature vector, performing decentration treatment on any one of the first feature vectors to obtain a first index data set; wherein the decentralization process is to subtract the average value of each item corresponding to the first feature vector from each item of the second feature vector, namely +.>Obtaining a corresponding index data set according to a third feature vector M obtained after the decentration treatment of each first feature vector; calculating covariance matrix of any third eigenvector M in the first index data set, namely M after decentralization processing and transpose matrix M thereof ^T Dividing the multiplied value by the number n of index values; the covariance matrix Cov is calculated as shown in the following equation 1:

wherein n is a positive integer greater than 0, and performing eigenvalue decomposition on the calculated covariance matrix Cov to obtain the characteristics of the covariance matrixA value and a fourth feature vector; sorting all the characteristic values in the fourth characteristic vector, and selecting a preset number of characteristic values according to the sorting order to generate a target characteristic vector; and determining a key index data set of a maintenance manufacturer according to the obtained target feature vectors. The feature values in the fourth feature vector may be in a positive order or in a negative order, which is not particularly limited herein. Taking reverse order sorting as an example, sorting all feature values of the fourth feature vector according to the reverse order, and screening out the first k feature values according to the sorting order to generate a target feature vector V= { x ₁ ,x ₂ ,x ₃ ,…x _k And k is a positive integer less than n.

After obtaining the key index data set of the maintenance manufacturer, determining the manufacturer type of the maintenance manufacturer according to the key index data set, specifically, in this embodiment, a pre-trained manufacturer classification detection model is provided, where the manufacturer classification detection model includes a plurality of classifiers, and the classifiers may be weak classifiers, which are not limited in detail herein. The vendor classification detection model is pre-trained based on historical data of the key index data set and is used for classifying maintenance vendors according to evaluation indexes corresponding to different service types. Based on this, step 300 may further include:

Step 301, inputting the key index data set into a pre-trained vendor classification detection model; the vendor classification detection model comprises a plurality of classifiers;

step 302, based on the service types of each maintenance service corresponding to the key index data set, classifying and deciding the key index data set by utilizing each classifier to obtain the prediction result of each classifier on the service type of the key index data set; the prediction result of the jth classifier is the accumulated value of the prediction value of the jth classifier and the prediction value of the previous j-1 classifiers, and j is a positive integer greater than 1;

step 303, calculating residual values of the prediction result of any target classifier and the prediction result of the adjacent classifier; the target classifier is any classifier in the manufacturer classification detection model;

and step 304, determining an optimal decision result of the vendor classification detection model on the service type of the key index data set based on the gradient descent principle of the residual error value, and determining the vendor type of the maintenance vendor according to the optimal decision result.

When determining the manufacturer type of a maintenance manufacturer, inputting a key index data set into a pre-trained manufacturer classification detection model, and classifying and deciding each evaluation index in the key index data set by utilizing a plurality of classifiers in the manufacturer classification detection model in a CART (Classification and Regression Tree, classification and regression decision tree) decision tree mode and the like based on the service type of each maintenance service corresponding to the key index data set to obtain a prediction result of each classifier on the service type of the key index data set; the prediction result of the j-th classifier in the plurality of classifiers is an accumulated value of the prediction value of the j-th classifier and the prediction value of the previous j-1 classifiers, and j is a positive integer greater than 1. Calculating residual values of the prediction result of any target classifier and the prediction result of the adjacent classifier based on the prediction results of all the classifiers; the target classifier is any one of a plurality of classifiers in the vendor classification detection model. Preferably, a plurality of classifiers in the vendor classification detection model form a CART decision tree in a serial form, and the prediction result of the current j-th classifier is the sum of the accumulated value of the prediction values of the previous j-1 classifiers and the prediction value of the current classifier; based on the prediction results of all the classifiers, calculating residual values of the prediction results of all the classifiers and the prediction results of the classifiers adjacent to each other before and after each other, optimizing the residual values based on a gradient descent principle of the residual values, obtaining an optimal decision result of the service type of the maintenance service provided by a maintenance manufacturer, and determining the manufacturer type of the maintenance manufacturer based on the optimal decision result.

In the vendor classification detection model, decision results of the key index data set are determined among a plurality of classifiers through GBDT (Gradient Boosting Decision Tree, gradient lifting decision tree) and other modes. Specifically, firstly, a key index data set is sent to different classifiers to make decisions, and residual errors of classification results of the front classifier and the rear classifier are fitted; then optimizing residual errors to a final model of an acceptable predicted result by using a gradient descent principle, and taking the predicted result of the model as an optimal decision result; therefore, the classification of the maintenance manufacturer is realized according to the prominent service items in the maintenance services of various types provided by the maintenance manufacturer.

Further, the step 400 may further include classifying the maintenance manufacturer according to the prominent service item of the maintenance manufacturer, determining the manufacturer type of the maintenance manufacturer, and determining the service score of the maintenance manufacturer according to the manufacturer type of the maintenance manufacturer and the key index data set, specifically:

step 401, obtaining an optimal service data set and a weight value corresponding to the manufacturer type; the optimal service data set comprises optimal scores of service evaluation indexes corresponding to the manufacturer types;

step 402, extracting a service evaluation index value corresponding to the manufacturer type from the key index data set to obtain a target index data set of the maintenance manufacturer;

Step 403, determining the similarity between the target index data set and the best service data set, and determining an initial score of the maintenance manufacturer according to the similarity;

and step 404, determining the service score of the maintenance manufacturer according to the weight value and the initial score.

Acquiring a corresponding weight value and an optimal service data set according to the manufacturer type of a maintenance manufacturer; the weight values of different types of maintenance manufacturers can be the same or different; the optimal service data set comprises the optimal scores of all evaluation indexes corresponding to manufacturer types, and is an optimal service data set constructed based on evaluation index data of service items provided by different types of maintenance manufacturers such as development service, investigation and design service, business operation and maintenance service, basic operation and maintenance service and the like according to actual requirements; i.e. by calculating the maximum service indicator value for each service vendor, a scoring dataset is constructed. Extracting an evaluation index value corresponding to the manufacturer type from the key index data set to obtain a target index data set of a maintenance manufacturer; determining an initial score of a maintenance manufacturer according to the similarity between the target index data set and the optimal service data set; and determining the final service score of the maintenance manufacturer according to the obtained weight value and the initial score, and realizing comprehensive evaluation of the maintenance manufacturer. The similarity of the target index data set to the best service data set may be cosine similarity, and the higher the similarity of the target index data set of the maintenance manufacturer to the best service data set, the higher the initial score.

Further, after obtaining the service score of the maintenance manufacturer, verification of the validity of the service score is required, specifically, after step 400, the method may further include:

step 501, obtaining a history score of the maintenance manufacturer;

step 502, calculating Euclidean distance between the service score and the history score, and verifying the validity of the service score according to the Euclidean distance;

step 503, if the verification is passed, outputting and displaying the service score to the evaluator of the maintenance manufacturer to obtain feedback information of the evaluator.

Acquiring a historical score of the maintenance manufacturer, and calculating the Euclidean distance between the current service score and the historical score; verifying the validity of the service score according to the calculated Euclidean distance; if the verification is passed, outputting and displaying the service score to an evaluator of a maintenance manufacturer, thereby providing a reference basis for decision making of the evaluator and acquiring feedback information of the evaluator.

Further, the same maintenance vendor may correspond to one or more vendor types, i.e., in the vendor classification detection model, the optimal decision result for the key index dataset may include multiple vendor types; after the initial scores of the service items of the maintenance manufacturer are calculated, the initial scores can be arranged in a descending order, so that the scores of the service items of the maintenance manufacturer are intuitively displayed. And after calculating a final score for the prominent service item of the maintenance manufacturer according to the classification result of the maintenance manufacturer, replacing the initial score of the service item with the final score so as to highlight the prominent service item of the maintenance manufacturer, and obtaining a comprehensive evaluation result.

When the service score is validated, the obtained historical score may be the last service score of the same maintenance manufacturer, or may be multiple service scores within a period of time in the near future, which is not limited herein. And calculating Euclidean distance between the current service score and the historical score, determining the difference between the current service score and the historical score, comparing the difference with a set threshold value, and determining whether the validity verification of the current service score of the maintenance manufacturer is passed or not according to the difference between the current service score and the historical score.

In this embodiment, by constructing the best service data sets of different types of maintenance manufacturers, the similarity between the target index data set of the maintenance manufacturer and the best service data set is used as the initial score of the maintenance manufacturer; determining a corresponding weight value by combining the manufacturer type of the maintenance manufacturer, and determining a final service score of the maintenance manufacturer according to the weight value and the initial score; based on various services provided by a maintenance manufacturer, comprehensive evaluation is performed on the maintenance manufacturer, and accuracy of an evaluation result is improved. According to the similarity between the target index data set and the optimal service data set of the maintenance manufacturer, the final score of each service item of the maintenance manufacturer is obtained, and then the comprehensive evaluation result of the maintenance manufacturer is obtained, so that the operation complexity in the evaluation process can be reduced, the network load is reduced, and the evaluation efficiency of the comprehensive evaluation of the maintenance manufacturer is improved; the weight value can be used for highlighting the better service item of the maintenance manufacturer, and the evaluation result of the maintenance manufacturer can be provided for the office more intuitively.

Further, the historic evaluation result of the same maintenance manufacturer and the distance of the evaluation result are calculated through Euclidean distance, validity verification is carried out on the evaluation result, the accuracy and the validity of the output comprehensive evaluation result can be ensured, and valuable reference data are provided for the office party.

In one embodiment, referring to FIG. 2, the comprehensive evaluation of the maintenance manufacturer is performed based on a preset comprehensive evaluation model; the comprehensive evaluation model comprises a manufacturer classification detection model and a service scoring model, wherein after a basic model of the comprehensive evaluation model is pre-trained based on historical key index data sets of all maintenance manufacturers, key index data sets of one or more maintenance manufacturers are input into the pre-trained comprehensive evaluation model, and then comprehensive evaluation results of all maintenance manufacturers can be obtained. Specifically, referring to another flow chart of the maintenance manufacturer evaluation method shown in fig. 3, an evaluation process of the maintenance manufacturer evaluation method provided in the embodiment of the present application is described in detail below with reference to fig. 3.

Firstly, acquiring original evaluation records of maintenance manufacturers by each evaluation party, and cleaning the abnormal evaluation records containing abnormal data in the original evaluation records to obtain final evaluation data; and constructing a feature vector set of a maintenance manufacturer based on the evaluation data obtained through preprocessing, carrying out principal component analysis on the feature vector set, removing some unimportant evaluation index data, screening out main influence factors of evaluation, and carrying out data dimension reduction on feature vectors in the feature vector set to obtain a key index data set. Then, inputting the key index data set into a comprehensive evaluation model, classifying the maintenance manufacturer according to the key index data set by utilizing a manufacturer classification detection model in the comprehensive evaluation model, and determining the manufacturer type of the maintenance manufacturer; acquiring a weight value corresponding to the manufacturer type, and extracting a target index data set corresponding to the manufacturer type from the key index data set; and inputting the weight value and the target index data set corresponding to the manufacturer type into a service scoring model, calculating the similarity of the target index data set best service data set based on the constructed best service data set, and comprehensively evaluating the manufacturer to obtain the service score of the comprehensive evaluation result. And finally, based on the historical scores, verifying the validity of the comprehensive evaluation result, outputting a display service score to the office party when the verification passes, providing a reference basis for the office party to the relevant decisions of maintenance manufacturers, acquiring feedback information of the office party to the comprehensive evaluation result, and adjusting the comprehensive evaluation model according to the feedback information to improve the subsequent evaluation accuracy.

When a feature vector set is constructed based on evaluation data obtained through preprocessing, a first feature vector can be constructed based on evaluation index data of the same evaluation party according to different evaluation parties, and a plurality of first feature vectors are constructed according to the evaluation index data of different evaluation parties to obtain the feature vector set; the first feature vector can also be constructed based on evaluation index data of the same type of maintenance service provided by a maintenance manufacturer according to different service types of the maintenance service provided by the maintenance manufacturer; constructing a plurality of first feature vectors according to different types of maintenance services provided by a maintenance manufacturer to obtain a feature vector set; the first feature vector may be constructed based on evaluation index data of the same type of maintenance service provided by the same evaluator for the maintenance manufacturer according to different service types of the maintenance service provided by the evaluator and the maintenance manufacturer; constructing a plurality of first feature vectors according to evaluation index data of each evaluation party for each maintenance service provided by a maintenance manufacturer to obtain a feature vector set; the first feature vector constructed is not particularly limited herein.

Further, when the principal component analysis is performed on the feature vector set, data dimension reduction is performed on each first feature vector in the feature vector set, main factors affecting evaluation are screened out, and some unimportant index data in each feature vector are removed. Specifically, in the feature vector set of the same maintenance manufacturer, the feature vector set may include evaluation index data of multiple evaluators for the same type of maintenance service, and when performing principal component analysis, an average value is calculated first based on index values of the same evaluation index in each feature vector in the feature vector set, that is, the first feature vector is constructed as x= { X ₁ ,x ₂ ,x ₃ ,…x _n For example, the index values of n evaluation indexes are included, and the index value of each evaluation index is mapped to a feature value in a feature vector to obtain a corresponding feature vector. Based on the same evaluation index value X in each first feature vector X _i Calculating the average valueObtain a second feature vector->Based on the calculated second feature vector, performing decentralization treatment on X, and subtracting each item of mean value from each item of data in X, namely +.>And obtaining a corresponding first index data set according to each characteristic vector M. The covariance matrix of each eigenvector M is calculated according to the following formula:

performing eigenvalue decomposition on the covariance matrixes obtained through calculation to obtain eigenvalues and fourth eigenvectors of each covariance matrix; and (3) carrying out reverse order sequencing on the characteristic values in the fourth characteristic vector, screening out the first k characteristic values according to the sequencing order, generating a target characteristic vector V, and obtaining a key index data set PCAdata (PCA, principal Component Analysis, principal component analysis) of a maintenance manufacturer according to each screened target characteristic vector V. Wherein k is a positive integer much smaller than n; the target feature vectors V and PCAdata are shown in the following formulas:

V＝{x ₁ ,x ₂ ,x ₃ ,…x _k }

the PCAdata in the above formula is a 5*4 matrix, i.e., 5 service items, each with 4 key indicators. Inputting PCAdata into a pre-trained comprehensive evaluation model, classifying maintenance vendors by using a vendor classification detection model in the comprehensive evaluation model, and determining the vendor type of the maintenance vendors, thereby determining better service items in various maintenance services provided by the maintenance vendors. Specifically, referring to fig. 4, fig. 4 is a flow chart illustrating a classification of a maintenance manufacturer by using a manufacturer classification detection model in the present embodiment, where the manufacturer classification detection model includes a plurality of weak classifiers connected in series, each feature vector in the pcadat is input into the classifier, and based on evaluation index data of each maintenance service of the maintenance manufacturer, classification decision is made on the type of the maintenance manufacturer by using the plurality of classifiers, and the manufacturer type of the maintenance manufacturer is determined according to an optimal decision result. The constructed weak classifier is C (x), the predicted value of each classifier is f (C (x)), if the prediction is performed in the way of CART decision tree integration, the predicted result of the current jth weak classifier is the sum of the accumulated value of the predicted values of the previous j-1 weak classifiers and the predicted value of the current jth weak classifier, namely:

In the case of the formula 2 of the present invention,an accumulated value representing the predicted values of the first j-1 classifiers, C _j (x) Representing the predicted value of the current jth classifier. Calculating residual values P of the prediction results of the classifiers and the prediction results of the front and rear adjacent classifiers based on the prediction results of the classifiers, optimizing P to a preset threshold (can be the minimum value which can be reached) according to the principle of gradient descent of the residual P, and finally taking the output of an optimal classifier model as the prediction result; i.e. the vendor type to which the value of f (c (x)) of PCAdata corresponds. And (3) making decisions by utilizing different weak classifiers, fitting residual errors of classification results of the weak classifiers before and after the decision, and then gradient descent optimizing the residual errors to a final model of an acceptable prediction result, determining manufacturer types of maintenance manufacturers according to prominent service items of the maintenance manufacturers, and realizing classification of the maintenance manufacturers.

Further, an optimal service data set is constructed based on evaluation index data corresponding to each maintenance service provided by maintenance manufacturers such as development service, investigation and design service, business operation and maintenance service, basic operation and maintenance service and the like, and an optimal service data set can be respectively constructed based on the evaluation index data corresponding to each maintenance service; an optimal service data set may be jointly constructed based on evaluation index data corresponding to a plurality of maintenance services, which is not particularly limited herein. The optimal service data set can calculate the maximum service evaluation index value max (PCAdata) of each maintenance vendor by a max function _x ) Construction is performed, namely, the best service data set DataSet is as shown in the following equation 3The illustration is:

DataSet＝max{PCAdata ₁ ,PCAdata ₂ ,PCAdata ₃ ,…PCAdata _x } (3)

taking the constructed optimal service data set DataSet as a data set of a service scoring model, and extracting a target index data set PCAdata of a maintenance manufacturer i from the key index data set PCAdata _i Obtaining an optimal service data set PCAdata corresponding to the maintenance manufacturer i from the DataSet _ji The method comprises the steps of carrying out a first treatment on the surface of the PCAdata of current maintenance vendor i is calculated through service scoring model _i PCAdata corresponding to DataSet _ji Cosine similarity of (c); and taking cosine similarity as an initial score of service items of maintenance manufacturers, namely:

acquiring a weight value weight corresponding to a manufacturer type of a maintenance manufacturer:

adding a weight to the initial score of the service item corresponding to the maintenance manufacturer to obtain a final score, namely taking the sum of the initial score and the weight value as the final service score to obtain the comprehensive evaluation result of the maintenance manufacturer. Wherein the weight value is the label value corresponding to the manufacturer type and the target index data set PCAdata _x Inverse of the sum of matrix values of (2), then:

PCAdata _x PCAdata matrix column number of PCAdata matrix E (6)

E represents an identity matrix of arbitrary dimension.

Further, after the initial scores of the various maintenance services of the maintenance manufacturer are calculated, the initial scores can be arranged in a descending order, so that the scores of the various maintenance services of the maintenance manufacturer are intuitively displayed. And after calculating a final score for the prominent service item of the maintenance manufacturer according to the classification result of the maintenance manufacturer, replacing the initial score of the service item with the final score so as to highlight the prominent service item of the maintenance manufacturer, and obtaining a comprehensive evaluation result.

After the comprehensive evaluation result of the maintenance manufacturer is obtained, the effectiveness of the comprehensive evaluation result is verified, specifically, the history score of the maintenance manufacturer is obtained, the difference between the current comprehensive evaluation result and the history evaluation result is calculated, and the difference is determined by the Euclidean distance of the service score; namely:

wherein F is _score Representing the service score of the current evaluation, F ^′ _score Representing historical scores, wherein D is the Euclidean distance between the current service score and the historical scores, and if a threshold Y is set, calculating the difference H between the Euclidean distance D and the set threshold Y according to the comprehensive evaluation result; namely, based on the difference D between the current service score and the historical score, H= |D-Y| is calculated; if the gap H exceeds the preset range, the validity verification is not passed, which means that a problem may occur in a certain service item of the maintenance manufacturer. For example, if the H of the maintenance manufacturer a on the software development service item exceeds the preset range, it indicates that the performance of the software development service of the maintenance manufacturer a is not in accordance with the requirements, a plurality of maintenance manufacturers (for example, maintenance manufacturer B, C, D) of which H on the software development service item is in the preset range may be screened based on the comprehensive evaluation result of each maintenance manufacturer, and the H scores of the screened maintenance manufacturer B, maintenance manufacturer C, and maintenance manufacturer D on the software development service item are arranged in descending order, and assuming that the H score of the maintenance manufacturer B on the software development service item is the highest, the office may be recommended to use the maintenance manufacturer B instead of the maintenance manufacturer a to provide the software development service.

Outputting and displaying the comprehensive evaluation result to the office party to obtain feedback information of the office party on the comprehensive evaluation result; judging the usefulness of the comprehensive evaluation result to the office according to the opinion of the office on the comprehensive evaluation result and the feedback of the comprehensive evaluation result obtained by operation, for example, whether the service quality is improved or whether the service quality is reduced after the office replaces some maintenance manufacturers according to the comprehensive evaluation result; therefore, the evaluation process of the comprehensive evaluation model can be adjusted according to the feedback of the comprehensive evaluation result, for example, the threshold value of the optimization residual error of the manufacturer classification detection model, the calculation mode of the similarity is adjusted, or the magnitude of the maintenance service weights of different types is adjusted, so that the subsequent evaluation precision and the accuracy of the comprehensive evaluation result are improved.

In the embodiment, the evaluation index data of each maintenance service provided by a maintenance manufacturer is obtained as evaluation data, so that comprehensive evaluation of the maintenance manufacturer can be realized based on more various and comprehensive evaluation data; meanwhile, the original evaluation record is preprocessed, and the abnormal evaluation record containing abnormal data is removed, so that the accuracy of the evaluation data can be improved. And then processing the feature vector constructed based on the preprocessed evaluation data through data dimension reduction, decentralization, covariance matrix calculation, eigenvalue decomposition and the like, and carrying out principal component analysis according to the constructed feature vector to obtain a key index data set PCAdata so as to improve the evaluation precision of subsequent evaluation maintenance manufacturers and the accuracy of the evaluation result.

Further, based on a comprehensive evaluation model formed by a manufacturer classification detection model and a service scoring model, predicting the type of a maintenance manufacturer according to PCAdata in a CART decision tree integration mode; by calculating the current PCAdata _i Corresponding to PCAdata in DataSet _ji The cosine similarity of each service item of the maintenance manufacturer is used as an initial score; and adding a weight to the initial score of the corresponding service item according to the classification result of the maintenance manufacturer to obtain the final score of each service item of the maintenance manufacturer, and then obtaining the comprehensive evaluation result of the maintenance manufacturer. The method not only can reduce the operation complexity in the evaluation process, reduce the network load and improve the evaluation efficiency of comprehensively evaluating maintenance manufacturers; the weight can be used for highlighting the salient service items of the maintenance manufacturer, so that the evaluation result of the maintenance manufacturer can be provided for the office more intuitively.

Further, the current comprehensive evaluation result and the current comprehensive evaluation result of the maintenance manufacturer are calculated through the Euclidean distance; the distance of the comprehensive evaluation result is historic, the validity of the comprehensive evaluation result is verified, and the accuracy and the validity of the comprehensive evaluation result can be ensured; and the comprehensive evaluation model is adjusted according to the feedback information of the office party, so that the accuracy of the subsequent comprehensive evaluation result can be further improved.

The description of the maintenance manufacturer evaluation device provided by the embodiment of the present application is provided below, and the maintenance manufacturer evaluation device described below and the maintenance manufacturer evaluation method described above may be referred to correspondingly.

Referring to fig. 5, a maintenance manufacturer evaluation device provided in an embodiment of the present application includes:

the feature construction module 10 is configured to acquire evaluation data of a maintenance manufacturer, and construct a feature vector set of the maintenance manufacturer based on the evaluation data; the feature vector set comprises a plurality of first feature vectors; any one of the first feature vectors is composed of one or more evaluation index values of a maintenance service provided by the maintenance manufacturer;

the feature dimension reduction module 20 is configured to perform principal component analysis on the feature vector set to obtain a key index data set of the maintenance manufacturer;

a classification detection module 30, configured to determine a vendor type of the maintenance vendor according to the key index data set;

and the comprehensive evaluation module 40 is configured to determine a service score of the maintenance manufacturer according to the manufacturer type of the maintenance manufacturer and the key index data set.

In one embodiment, the feature construction module 10 is further configured to:

In one embodiment, the feature dimension reduction module 20 is further configured to:

In one embodiment, the classification detection module 30 is further configured to:

In one embodiment, the comprehensive evaluation module 40 is further configured to:

In one embodiment, the maintenance manufacturer evaluation device further includes a post-evaluation module for:

Acquiring historical scores of the maintenance manufacturer;

Fig. 6 illustrates a physical structure of an electronic device, as shown in fig. 6, where the electronic device may include: processor 610, communication interface (Communication Interface) 620, memory 630, and communication bus 640, wherein processor 610, communication interface 620, and memory 630 communicate with each other via communication bus 640. The processor 610 may call a computer program in the memory 630 to perform the steps of the maintenance vendor evaluation method, including, for example:

Further, the logic instructions in the memory 630 may be implemented in the form of software functional units and stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

In another aspect, embodiments of the present application further provide a computer program product, where the computer program product includes a computer program, where the computer program may be stored on a non-transitory computer readable storage medium, where the computer program when executed by a processor is capable of executing the steps of the maintenance manufacturer assessment method provided in the foregoing embodiments, for example, including:

In another aspect, embodiments of the present application further provide a processor-readable storage medium storing a computer program for causing a processor to execute the steps of the method provided in the above embodiments, for example, including:

The processor-readable storage medium may be any available medium or data storage device that can be accessed by a processor, including, but not limited to, magnetic storage (e.g., floppy disks, hard disks, magnetic tape, magneto-optical disks (MOs), etc.), optical storage (e.g., CD, DVD, BD, HVD, etc.), semiconductor storage (e.g., ROM, EPROM, EEPROM, nonvolatile storage (NAND FLASH), solid State Disk (SSD)), and the like.

The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.

From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present application, and are not limiting; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. The maintenance manufacturer evaluation method is characterized by comprising the following steps of:

2. The maintenance vendor evaluation method according to claim 1, wherein the step of determining a service score of the maintenance vendor from the vendor type of the maintenance vendor and the key index data set comprises:

3. The maintenance manufacturer assessment method according to claim 1, wherein the step of performing principal component analysis on the feature vector set to obtain a key index data set of the maintenance manufacturer comprises:

4. The maintenance vendor evaluation method according to claim 1, wherein the step of determining the vendor type of the maintenance vendor from the key index data set comprises:

5. The maintenance vendor evaluation method according to claim 1, wherein the step of constructing the feature vector set of the maintenance vendor based on the evaluation data comprises:

6. The maintenance vendor evaluation method according to claim 1, wherein the step of acquiring evaluation data of a maintenance vendor comprises:

7. The maintenance vendor evaluation method according to claim 1, wherein after determining the service score of the maintenance vendor according to the vendor type of the maintenance vendor and the key index data set, further comprising:

acquiring historical scores of the maintenance manufacturer;

8. A maintenance manufacturer evaluation device, comprising:

9. An electronic device comprising a processor and a memory storing a computer program, characterized in that the processor implements the steps of the maintenance vendor evaluation method of any one of claims 1 to 7 when executing the computer program.

10. A non-transitory computer readable storage medium having stored thereon a computer program, characterized in that the computer program when executed by a processor implements the steps of the maintenance vendor evaluation method of any one of claims 1 to 7.