Automatic matching method for power customer engineering

Abstract

The invention relates to the technical field of power customer service and automatic matching, in particular to an automatic matching method for power customer engineering. Comprising the following steps: screening out candidate service units meeting qualification conditions by using a multi-attribute decision analysis method; based on a content-based recommendation algorithm, extracting features of customer engineering, and recommending electric power engineering service units matched with the features; according to a recommendation algorithm based on matrix decomposition, power customer engineering and power engineering service units are expressed as rows and columns of a two-dimensional matrix respectively, and service units with potential matching are recommended by calculating the incidence relation between matrixes; the recommendation results of the two algorithms are fused, other factors are considered, sorting is carried out, and the final recommendation result is comprehensively generated. According to the design, automatic matching between the customer and the electric power engineering service unit can be realized, the matching success rate is improved, the transaction cost is reduced, and the operation efficiency of the electric power market is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power customer service and automation matching, in particular to a power customer engineering automation matching method. BACKGROUND

[0002] With the online handling of business expansion, the original business process of business expansion is optimized, and the link of offline docking with customers is optimized. In the process of online business expansion, the following difficulties exist in user power service:

[0003] ① Users do not have information about business expansion engineering design / construction resources, and it is difficult to find design / construction service units with high quality, good service and low price;

[0004] ② The power supply bureau has the intention to provide better service to users, but the cost and specific construction scheme concerned by users are difficult for the power supply enterprise to accurately answer;

[0005] ③ There are some compliance and incorruptibility risks in business expansion engineering.

[0006] At present, there is a vacancy in the domestic market for solving the information asymmetry between the demand side and the market side of power engineering. In view of this, we propose a power customer engineering automation matching method. SUMMARY

[0007] The purpose of the present application is to provide a power customer engineering automation matching method to solve the problems raised in the background art.

[0008] To achieve the above technical problems, one of the purposes of the present application is to provide a power customer engineering automation matching method, comprising the following steps:

[0009] S1, using a multi-attribute decision analysis method, according to the key information of voltage level and installation capacity of power customer engineering, screening out candidate service units meeting the qualification conditions from power engineering service units;

[0010] S2, based on the content-based recommendation algorithm, by analyzing the historical data of similar power customers selecting power engineering service units, extracting the characteristics of voltage level, capacity, engineering area and other characteristics of customer engineering, and recommending power engineering service units matched with these characteristics;

[0011] S3, based on the matrix decomposition recommendation algorithm, the power customer engineering and the power engineering service unit are respectively represented as the row and column of the two-dimensional matrix, and the service units with potential matching are recommended by calculating the correlation between the matrices;

[0012] S4. Fusing the recommendation results of the content-based recommendation algorithm and the matrix factorization-based recommendation algorithm, and considering the comprehensive score of the power engineering service unit, the recent service times, the preferential activities and other data, performing sorting to generate a final recommendation result.

[0013] As a further improvement of the technical solution, in step S1, the multi-attribute decision analysis method is based on information entropy theory and determines the weight of each attribute according to the data variation degree of each attribute, and specifically includes:

[0014] S1.1. Determine attributes and weights: first, determine all attributes or indicators involved in the matching problem, and then assign a weight to each attribute according to its importance in the matching process;

[0015] S1.2. Data standardization: standardize the data through linear transformation, vector normalization, statistical processing of raw data, etc.

[0016] S1.3. Calculate weighted values: for each power engineering service unit and customer engineering, multiply the value of each attribute by the weight of the attribute to obtain the weighted value of the attribute.

[0017] S1.4. Sum to obtain the matching degree: add the weighted values of all attributes of the power engineering service unit and the customer engineering to obtain their matching degree, which is used to intuitively reflect the matching degree of the power engineering service unit and the customer engineering.

[0018] As a further improvement of the technical solution, in step S1, the multi-attribute decision analysis method further includes comprehensive consideration of factors such as project cases, capabilities, and credibility of the power engineering service unit to further screen the candidate service units.

[0019] As a further improvement of the technical solution, in step S2, the specific recommendation method process includes:

[0020] S2.1. Collect engineering feature data: collect specific information about the customer engineering, including voltage level, capacity, engineering area, and other characteristics related to engineering requirements.

[0021] S2.2. Analyze historical data: analyze historical data of similar users selecting power engineering service units to find the correlation pattern between power engineering service units and engineering features, and provide data support for subsequent recommendation algorithms.

[0022] S2.3. Extract key features: extract the most relevant key features for recommending power engineering service units from the collected engineering feature data and historical data.

[0023] S2.4, constructing a content-based recommendation algorithm: based on the extracted key features, a content-based recommendation algorithm is constructed; the algorithm uses the matching degree between engineering features and power engineering service unit features to generate a recommendation result;

[0024] S2.5, generating a recommendation result: input the features of the customer engineering into the content-based recommendation algorithm, and the algorithm calculates the matching degree of each power engineering service unit and the engineering according to the preset rules and algorithm logic; then, the power engineering service units are sorted according to the matching degree, and the top several with the highest matching degree are selected as the recommendation result.

[0025] As a further improvement of the technical solution, in step S2, the content-based recommendation algorithm further includes semantic analysis of the customer engineering description using natural language processing technology to more accurately extract features.

[0026] As a further improvement of the technical solution, the specific process of step S3 includes:

[0027] S3.1, constructing an engineering-power engineering service unit matrix: collecting historical data of customer engineering and related data of power engineering service units, taking customer engineering and power engineering service units as rows and columns of the matrix, and constructing an initial engineering-power engineering service unit matrix;

[0028] S3.2, matrix decomposition: selecting matrix decomposition technology as the decomposition method to decompose the matrix;

[0029] S3.3, calculating the correlation: extracting key feature vectors from the engineering feature matrix and the power engineering service unit feature matrix obtained by decomposition;

[0030] S3.4, calculating the correlation: using the extracted feature vectors, calculating the correlation between each customer engineering and each power engineering service unit;

[0031] S3.5, recommending power engineering service units: according to the calculated correlation, sorting the power engineering service units, and according to the sorting result, selecting the top several power engineering service units with the highest correlation as the recommendation result.

[0032] As a further improvement of the technical solution, in step S3.2, the matrix decomposition recommendation algorithm adopts non-negative matrix decomposition technology to discover the potential correlation between customer engineering and service units.

[0033] As a further improvement of the technical solution, the calculation process of the non-negative matrix decomposition technology includes:

[0034] Suppose there are n customers in the customer set, i.e. the customer matrix is U={u1,u2,u3,...,u n};Electric power engineering service unit set has m projects, namely electric power engineering service unit matrix V={v1, v2, v3,..., vm} m},The customer to the electric power engineering service unit matrix is R;

[0035] The matrix decomposition target is to decompose the matrix R into user factor matrix and electric power engineering service unit factor matrix multiplication, that is, R=UEV=UV, where R is n×m, E is a standard matrix, E is k×k, U is n×k, and V is k×m;

[0036] The objective function of matrix decomposition is as follows:

[0037]

[0038] Wherein, L represents the objective function with minimum mean square error, wherein i∈[1, n], j∈[1, m], T represents the transpose of the matrix, R ij represents the R matrix containing i rows and j columns of data, U i represents the U matrix containing i rows and k columns of data, V j represents the V matrix containing k rows and j columns of data.

[0039] As a further improvement of the technical solution, in the step S4, the aforementioned recommended results are fused to comprehensively generate the final recommended result, specifically including:

[0040] S4.1, result fusion strategy: selecting a suitable fusion strategy to assign different weights to the recommended results of the two recommendation algorithms;

[0041] Wherein, the fusion method of the fusion strategy includes weighted average, voting or other fusion strategies to comprehensively consider the output results of different recommendation algorithms;

[0042] S4.2, comprehensively generating the final recommended result.

[0043] As a further improvement of the technical solution, in the step S4.2, the final recommended result is comprehensively generated, specifically including:

[0044] S4.2.1, result merging: merging the recommended results to form a unified recommended list;

[0045] S4.2.2, sorting adjustment: according to the selected fusion strategy, the recommended results after merging are sorted and adjusted; using the weighted average strategy, the recommended results can be sorted by weight;

[0046] S4.2.3, recommended result screening: according to the actual demand, the final recommended electric power engineering service unit is screened from the recommended results after sorting adjustment; the top few can be selected as the final recommended result.

[0047] The second object of the present application is to provide an electric power customer engineering automatic matching system for implementing the electric power customer engineering automatic matching method described above, comprising a multi-attribute decision analysis module, a content-based recommendation algorithm module, a matrix decomposition-based recommendation algorithm module, and a result fusion and sorting module.

[0048] As a further improvement of the present technical solution, the system further comprises a user interface module for receiving input information of electric power customer engineering, displaying the recommended results, and receiving feedback of the recommended results from the user.

[0049] As a further improvement of the present technical solution, the system further comprises a database module for storing electric power customer engineering information, electric power engineering service unit information, and historical selection data, etc.

[0050] The third object of the present application is to provide an electric power customer engineering automatic matching device, comprising a processor, a memory, and a computer program stored in the memory and running on the processor, wherein the processor is configured to execute the computer program to implement the steps of the electric power customer engineering automatic matching method described above.

[0051] The fourth object of the present application is to provide a computer readable storage medium, which stores a computer program, wherein the computer program is configured to be executed by a processor to implement the steps of the electric power customer engineering automatic matching method described above.

[0052] Compared with the prior art, the present application has the following advantages:

[0053] 1. In the electric power customer engineering automatic matching method, the automatic matching between the customer and the electric power engineering service unit can be realized, the success rate of matching can be improved, the transaction cost can be reduced, and the operation efficiency of the electric power market can be improved.

[0054] 2. In the electric power customer engineering automatic matching method, the advantages of the two algorithms (content-based recommendation algorithm and matrix decomposition-based recommendation algorithm) are fully utilized, the historical selection preferences of the user are considered, and the potential correlation is mined, so that the comprehensiveness and accuracy of the recommendation are improved. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 The figure is an exemplary overall architecture diagram in the present application.

[0056] Figure 2 The figure is an exemplary process diagram of the matching between the customer engineering information and the electric power service engineering unit in the present application.

[0057] Figure 3This is an exemplary matrix decomposition principle diagram in this invention;

[0058] Figure 4 This is a structural diagram of an exemplary electronic computer platform device in this invention. Detailed Implementation

[0059] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0060] Example 1

[0061] like Figures 1-3 As shown in the figure, this embodiment provides an automated matching method for power customer engineering, including the following steps:

[0062] S1. Using a multi-attribute decision analysis method, based on key information such as the voltage level and installed capacity of the power customer project, candidate service providers that meet the qualification requirements (possess the corresponding qualifications and capabilities) are selected from power engineering service providers. This helps to narrow down the selection range and improve the accuracy of recommendations in subsequent steps.

[0063] In this step, based on the project type, voltage level, and the qualification level and business scope of the power engineering service provider, it is necessary to determine the weights of each attribute in the power engineering service provider's qualifications and the customer's project requirements. The multi-attribute decision analysis method, based on information entropy theory, determines the weight of each attribute according to its degree of data variation. The greater the degree of data variation, the greater the influence of that attribute on the decision, and therefore its weight should also be greater. Specifically, this includes:

[0064] S1.1 Determine Attributes and Weights: First, identify all attributes or indicators involved in the matching problem. These attributes may include the qualification level and professional field of the power engineering service unit, as well as the scale of the construction project (line, distribution project), technical requirements, and construction period requirements. Then, assign a weight to each attribute based on its importance in the matching process. The determination of weights is usually based on experience, expert opinions, or analysis of historical data.

[0065] S1.2 Data Standardization: Since different attributes may have different units of measurement or dimensions, directly performing weighted summation may lead to inaccurate results. Therefore, it is necessary to standardize the data to ensure that they have the same dimensions or range. Data standardization is achieved through methods such as linear transformation, vector normalization, and statistical processing of the original data. The specific calculation expression is as follows:

[0066] Linear transformation: Benefit-oriented attribute Cost-type attributes

[0067] Vector normalization:

[0068]

[0069] Statistical processing of raw data:

[0070]

[0071] In the formula, z ij Represents the standardized index value, y ij Representing data, 'm' represents a single indicator, and 'm' represents the number of indicators. This represents the mean of each individual indicator.

[0072] S1.3 Calculate the weighted value: For each power engineering service unit and customer project, multiply its value on each attribute by the weight of that attribute to obtain the weighted value of that attribute; this is actually an adjustment of the importance of each attribute, so that more important attributes have a greater weight in the final matching degree calculation.

[0073] S1.4 Summation to obtain the matching degree: The weighted values ​​of the power engineering service unit and the customer's project on all attributes are added together to obtain their matching degree. This matching degree is a value that comprehensively considers multiple attributes and can intuitively reflect the degree of matching between the power engineering service unit and the customer's project.

[0074] Furthermore, the multi-attribute decision analysis method also includes comprehensively considering factors such as the project cases, capabilities, and reputation of power engineering service providers in order to further screen candidate service providers.

[0075] S2. Content-based recommendation algorithm analyzes historical data on the selection of power engineering service providers by similar power customers, extracts features such as voltage level, capacity, and location of customer projects, and recommends power engineering service providers that match these features. This algorithm leverages the value of historical data to discover user preferences and patterns in service provider selection, thereby improving the accuracy of recommendations.

[0076] The specific recommended method and process in this step includes:

[0077] S2.1 Collect engineering characteristic data: Collect specific information about the customer's project, including voltage level, capacity, project location and other characteristics related to project requirements;

[0078] S2.2, Analyze historical data: Analyze historical data of similar user selection of power engineering service units, including searching for selection records of power engineering service units in similar past engineering cases, and performance evaluation (such as construction quality, progress control, cost control, etc.) of these power engineering service units; the purpose of analyzing historical data is to find the correlation pattern between power engineering service units and engineering characteristics, and to provide data support for subsequent recommendation algorithm;

[0079] S2.3, Extract key features: Extract the most relevant key features of the recommended power engineering service units from the collected engineering characteristic data and historical data; these features may include the qualification level of the power engineering service unit, the professional field, the past project experience (especially the experience of similar projects as the current project), the regional distribution, etc.;

[0080] S2.4, Build content-based recommendation algorithm: Based on the extracted key features, build a content-based recommendation algorithm; this algorithm will use the matching degree between engineering characteristics and power engineering service unit characteristics to generate recommendation results; the algorithm can include feature weight allocation, similarity calculation and other steps to ensure that the recommended power engineering service units are most matched with the needs of the current project;

[0081] The most common similarity calculation formula is cosine similarity:

[0082]

[0083] In the formula, cos(F u ,F i ) represents the similarity between data F u and F i ;

[0084] S2.5, Generate recommendation results: input the characteristics of the customer's engineering into the content-based recommendation algorithm, and the algorithm will calculate the matching degree of each power engineering service unit with the engineering according to the preset rules and algorithm logic; then sort the power engineering service units according to the matching degree, and select the top several as the recommendation results.

[0085] Further, the content-based recommendation algorithm also includes using natural language processing technology to perform semantic analysis on the customer engineering description to more accurately extract features.

[0086] S3, Matrix decomposition-based recommendation algorithm, representing power customer engineering and power engineering service units as rows and columns of two-dimensional matrix respectively, and recommending service units with potential matching by calculating the correlation between matrices; this algorithm can capture the potential correlation between customer engineering and service units, and can effectively recommend even without direct historical data.

[0087] The specific process of this step includes:

[0088] S3.1, Constructing the engineering-power engineering service unit matrix:

[0089] Data preparation: Collect historical data of customer projects and relevant data of power engineering service units. Customer project data includes voltage level, capacity, and regional characteristics; power engineering service unit data includes qualification level, professional field, and past project experience;

[0090] Matrix construction: Construct an initial engineering-power engineering service unit matrix with customer projects and power engineering service units as rows and columns. Each element in the matrix represents a preference score or correlation degree of customer projects to power engineering service units. The initial value can be set to 0 or a random decimal;

[0091] S3.2, Matrix decomposition: Select matrix decomposition technique as decomposition method for matrix decomposition;

[0092] In this step, the matrix decomposition recommendation algorithm uses non-negative matrix factorization technique to discover the potential association between customer projects and service units;

[0093] Specifically, the calculation process of non-negative matrix factorization technique includes:

[0094] Assume that there are n customers in the customer set, i.e. the customer matrix is U={u1, u2, u3,..., u n}; There are m projects in the power engineering service unit set, i.e. the power engineering service unit matrix is V={v1, v2, v3,..., v m}, then the customer to power engineering service unit matrix is R;

[0095] The mathematical theoretical basis of matrix decomposition algorithm is the row and column transformation of matrix; the row transformation of matrix A is equivalent to the left multiplication of a matrix by A, and the column transformation of matrix A is equivalent to the right multiplication of a matrix by A, so matrix A can be represented as A=REQ=PQ (E is a standard matrix);

[0096] The goal of matrix decomposition is to decompose the matrix R into the product of user factor matrix and power engineering service unit factor matrix, i.e. R=UEV=UV, where R is n×m, E is a standard matrix, E is k×k, U is n×k, and V is k×m;

[0097] The objective function of matrix decomposition is as follows:

[0098]

[0099] Where L represents the objective function with the minimum mean square error as the objective function, i∈[1,n], j∈[1,m], T represents the transpose of matrix, Rij R represents a matrix containing i rows and j columns of data, U i U represents a matrix containing i rows and k columns of data, V j V represents a matrix containing k rows and j columns of data.

[0100] Where, assuming that the customer set has 6 customers, i.e. the customer matrix is U = {u1, u2, u3, u4, u5, u6}; the power engineering service unit set has 7 projects, i.e. the power engineering service unit matrix is V = {v1, v2, v3, v4, v5, v6, v7}, then the customer-to-power engineering service unit matrix is R;

[0101] R = UEV = UV, where R is 6x7, U is 6xk, and V is kx7, and the intuitive representation is as shown in Figure 3 .

[0102] S3.3, Calculate the correlation: extract the key feature vectors from the engineering feature matrix and the power engineering service unit feature matrix obtained by decomposition; these feature vectors can reflect the representation of the engineering and the power engineering service unit in the latent feature space;

[0103] S3.4, Calculate the correlation degree: use the extracted feature vectors to calculate the correlation degree between each customer engineering and each power engineering service unit; the calculation of the correlation degree can be realized by calculating the dot product, cosine similarity or other similarity measurement methods between the feature vectors;

[0104] S3.5, Recommend the power engineering service unit:

[0105] Correlation degree ranking: according to the calculated correlation degree, rank the power engineering service units, and the power engineering service unit with higher correlation degree ranks higher;

[0106] Recommendation result generation: according to the ranking result, select the power engineering service units with the highest correlation degree as the recommendation result; the number of recommendation results can be set according to actual needs, such as the top 5 or top 10.

[0107] S4, Fuse the recommendation results of the content-based recommendation algorithm and the matrix decomposition-based recommendation algorithm, and consider the comprehensive score of the power engineering service unit, the recent service times, the preferential activities and other data, and rank them to generate the final recommendation result;

[0108] This step fuses the aforementioned recommendation results to generate the final recommendation result, specifically including:

[0109] S4.1, Result fusion strategy: select a suitable fusion strategy to assign different weights to the recommendation results of the two recommendation algorithms;

[0110] The fusion method of the fusion strategy includes weighted average, voting or other fusion strategies to comprehensively consider the output results of different recommendation algorithms.

[0111] S4.2, the final recommendation result is generated comprehensively, specifically including:

[0112] S4.2.1, result merging: merging the recommendation results to form a unified recommendation list;

[0113] S4.2.2, sorting adjustment: according to the selected fusion strategy, the sorted recommendation results are adjusted; using the weighted average strategy, the recommendation results can be weighted and sorted according to the weight;

[0114] S4.2.3, recommendation result screening: according to the actual demand, the final recommended power engineering service unit is screened from the sorted recommendation results; several top-ranked companies can be selected as the final recommendation result.

[0115] In this step, the fusion method includes assigning weights to the recommendation results of different algorithms, linearly combining them, and sorting the combined results to generate a final recommendation list. By making full use of the advantages of the two algorithms, both the historical selection preferences of users and potential association relationships are considered, thereby improving the comprehensiveness and accuracy of the recommendation.

[0116] The embodiment also provides an electric power customer engineering automation matching system, which is used to realize the electric power customer engineering automation matching method described above, and includes a multi-attribute decision analysis module, a content-based recommendation algorithm module, a matrix decomposition-based recommendation algorithm module, and a result fusion and sorting module.

[0117] As a further improvement of the technical solution, the system further includes a user interface module for receiving input information of the electric power customer engineering, displaying the recommendation results, and receiving feedback of the user on the recommendation results.

[0118] As a further improvement of the technical solution, the system further includes a database module for storing electric power customer engineering information, electric power engineering service unit information, and historical selection data.

[0119] As Figure 4 shown, the embodiment also provides an electric power customer engineering automation matching device, which includes a processor, a memory, and a computer program stored in the memory and running on the processor.

[0120] The processor includes one or more processing cores, and the processor is connected to the memory through a bus. The memory is used to store program instructions, and the processor executes the program instructions in the memory to realize the steps of the electric power customer engineering automation matching method described above.

[0121] Optionally, the memory can be implemented by any type of volatile or nonvolatile storage devices or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable programmable read only memory (EPROM), programmable read only memory (PROM), read only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0122] In addition, the present application also provides a computer readable storage medium, the computer readable storage medium stores a computer program, the computer program is executed by a processor to realize the steps of the power customer engineering automation matching method.

[0123] Optionally, the present application also provides a computer program product containing instructions, which, when running on a computer, causes the computer to execute the steps of the power customer engineering automation matching method.

[0124] Those skilled in the art can understand that the process of realizing all or part of the steps of the above-mentioned embodiments can be completed by hardware, or by program to instruct relevant hardware, the program can be stored in a computer readable storage medium, the storage medium mentioned above can be a read only memory, a magnetic disk or an optical disk, etc.

[0125] The basic principles, main features and advantages of the present application are shown and described above. Those skilled in the art should understand that the present application is not limited by the above-mentioned embodiments, the above-mentioned embodiments and the description in the specification are only preferred examples of the present application, and are not intended to limit the present application, various changes and improvements of the present application can be made without departing from the spirit and scope of the present application, and these changes and improvements all fall within the scope of the present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. An automated matching method for power customer engineering, characterized in that, Includes the following steps: S1. Using a multi-attribute decision analysis method, based on key information such as the voltage level and installed capacity of the power customer project, select candidate service providers that meet the qualification requirements from among the power engineering service providers. S2. Content-based recommendation algorithm: By analyzing historical data of similar power customers choosing power engineering service providers, extracting the characteristics of the voltage level, capacity, and location of the customer's project, and recommending power engineering service providers that match these characteristics. S3. A recommendation algorithm based on matrix factorization represents power customer projects and power project service units as rows and columns of a two-dimensional matrix, respectively, and recommends service units with potential matching by calculating the correlation between matrices. S4. The recommendation results of the content-based recommendation algorithm and the matrix factorization-based recommendation algorithm are merged, and the comprehensive score of the power engineering service unit, the number of recent services, and the data of promotional activities are taken into account and sorted to generate the final recommendation result.

2. The automated matching method for power customer engineering according to claim 1, characterized in that: In step S1, the multi-attribute decision analysis method is based on information entropy theory, and determines the weight of each attribute according to the degree of data variation. Specifically, it includes: S1.1 Determine attributes and weights: First, identify all attributes or indicators involved in the matching problem, and then assign a weight to each attribute based on its importance in the matching process; S1.2 Data Standardization: Data is standardized through linear transformation, vector normalization, and statistical processing of the original data. S1.3 Calculate the weighted value: For each power engineering service unit and customer project, multiply its value on each attribute by the weight of that attribute to obtain the weighted value of that attribute; S1.4 Summation to obtain matching degree: Summing the weighted values ​​of the power engineering service unit and the customer's project on all attributes to obtain their matching degree, which is used to intuitively reflect the degree of matching between the power engineering service unit and the customer's project.

3. The automated matching method for power customer engineering according to claim 2, characterized in that: In step S1, the multi-attribute decision analysis method also includes comprehensively considering factors such as the project cases, capabilities, and reputation of power engineering service providers to further screen candidate service providers.

4. The automated matching method for power customer engineering according to claim 1, characterized in that: In step S2, the specific recommended method flow includes: S2.1 Collect engineering characteristic data: Collect specific information about the customer's engineering projects, including voltage level, capacity, engineering location, and other characteristics related to engineering requirements; S2.2 Analyze historical data: Analyze historical data of similar users' selection of power engineering service providers to discover the correlation patterns between power engineering service providers and engineering characteristics, and provide data support for subsequent recommendation algorithms; S2.3 Extracting key features: Extracting the key features most relevant to recommending power engineering service units from the collected engineering feature data and historical data; S2.4 Constructing a content-based recommendation algorithm: Based on the extracted key features, a content-based recommendation algorithm is constructed; this algorithm uses the matching degree between engineering features and power engineering service unit features to generate recommendation results; S2.5 Generating Recommendation Results: Input the features of the customer's project into the content-based recommendation algorithm. The algorithm calculates the matching degree between each power engineering service unit and the project according to the preset rules and algorithm logic. Then, the power engineering service units are sorted according to the matching degree, and the top few matching degrees are selected as the recommendation results.

5. The automated matching method for power customer engineering according to claim 4, characterized in that: In step S2, the content-based recommendation algorithm further includes using natural language processing techniques to perform semantic analysis on the customer's engineering description in order to extract features more accurately.

6. The automated matching method for power customer engineering according to claim 1, characterized in that: The specific process of step S3 includes: S3.1 Constructing an Engineering-Power Engineering Service Provider Matrix: Collect historical data of customer projects and relevant data of power engineering service providers, and construct an initial engineering-power engineering service provider matrix by using customer projects and power engineering service providers as rows and columns of the matrix; S3.2 Matrix Decomposition: Select matrix decomposition techniques as the decomposition method to decompose the matrix; S3.3 Calculate the correlation: Extract key feature vectors from the engineering feature matrix and the power engineering service unit feature matrix obtained from the decomposition; S3.4 Calculate the correlation degree: Using the extracted feature vectors, calculate the correlation degree between each customer project and each power engineering service unit; S3.5 Recommended Power Engineering Service Providers: Based on the calculated correlation, power engineering service providers are ranked, and the top few with the highest correlation are selected as recommendations.

7. The automated matching method for power customer engineering according to claim 6, characterized in that: In step S3.2, the matrix factorization recommendation algorithm uses nonnegative matrix factorization technology to discover potential relationships between customer engineering and service units.

8. The automated matching method for power customer engineering according to claim 7, characterized in that: The calculation process of the nonnegative matrix factorization technique includes: Suppose the customer set has n customers, i.e., the customer matrix is ​​U = {u1, u2, u3, ..., u...} n The set of power engineering service providers contains m projects, i.e., the power engineering service provider matrix is ​​V = {v1, v2, v3, ..., v...}. m If}, then the customer-to-power engineering service unit matrix is ​​R; The goal of matrix factorization is to decompose matrix R into the form of the product of user factor matrix and power engineering service unit factor matrix, i.e., R = UEV = UV, where R is n×m, E is a standard matrix, E is k×k, U is n×k, and V is k×m. The objective function for matrix decomposition is as follows: Where L represents the objective function with the goal of minimizing the mean square error, i∈[1,n], j∈[1,m], T represents the transpose of the matrix, and R ij Let R be a matrix containing i rows and j columns of data. i V represents a matrix U containing i rows and k columns of data. j Let V represent a matrix containing k rows and j columns of data.

9. The automated matching method for power customer engineering according to claim 1, characterized in that: In step S4, the aforementioned recommendation results are combined to generate a final recommendation result, specifically including: S4.1 Result Fusion Strategy: Select an appropriate fusion strategy to assign different weights to the recommendation results of the two recommendation algorithms; Among them, the fusion methods of the fusion strategy include weighted average, voting or other fusion strategies, in order to comprehensively consider the output results of different recommendation algorithms; S4.2 Generate the final recommendation results.

10. The automated matching method for power customer engineering according to claim 9, characterized in that: In step S4.2, the comprehensive generation of the final recommendation result specifically includes: S4.2.1 Result Merging: Merge the recommendation results to form a unified recommendation list; S4.2.2, Ranking Adjustment: Based on the selected fusion strategy, the ranking of the merged recommendation results is adjusted; S4.2.3 Filtering of Recommendation Results: Based on actual needs, select the final recommended power engineering service providers from the adjusted recommendation results.

Images