Service product recommendation method and device, electronic equipment and storage medium

Abstract

The invention discloses a business product recommendation method and device, electronic equipment and a storage medium, and relates to the technical field of data processing, and the main technical scheme comprises the following steps: obtaining first business appeal information of a user; searching a first business product having correlation with the first business appeal information in a preset business knowledge base; respectively inputting the first business appeal information and the first business product into a preset question and answer model to obtain business answer information required by the user; the service answering information comprises the first service appeal information and the first service product; and calling a preset recommendation algorithm, and analyzing the service solution information to obtain a first recommended service product required by the user. Compared with the prior art, manual operation is not needed, the recommendation efficiency of the service products is improved, the appropriate first recommendation service products are recommended according to the first service appeal information input by the user, and the recommendation satisfaction degree of the service products is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to a method and device for recommending a business product, an electronic device, and a storage medium. Background Art

[0002] With the rapid development of 5G and the personal mobile business, home business, government business and emerging business (Customer, Home, Business, New, CHBN) markets, the types of CHBN business products are increasing. In order to enable users to use business products that meet their own needs, business products need to be recommended.

[0003] In the related technologies of business product recommendation, business products are usually recommended manually, or business products are displayed and recommended in a self-service terminal system, and users search for business products that meet their own needs.

[0004] Manual recommendation of business products will be affected by the subjective factors of the recommender. The recommender often recommends business products that he is familiar with to users, which may not necessarily meet the real needs of users, resulting in low satisfaction with business product recommendations. The self-service terminal system requires users to find business products that meet their own needs. Users need to spend a lot of time to understand business products, resulting in low efficiency in business product recommendations. Summary of the invention

[0005] The present disclosure provides a method and device for recommending a business product, an electronic device and a storage medium, the main purpose of which is to solve the problem of low satisfaction with the recommendation of business products and low efficiency of the recommendation of business products.

[0006] According to a first aspect of the present disclosure, a method for recommending a business product is provided, comprising:

[0007] Obtain the user's first business demand information;

[0008] Searching a preset business knowledge base for a first business product that is relevant to the first business demand information;

[0009] The first business demand information and the first business product are respectively input into a preset question-answering model to obtain business answer information required by the user; the business answer information includes the first business demand information and the first business product;

[0010] A preset recommendation algorithm is called to analyze the business answer information to obtain a first recommended business product required by the user.

[0011] Optionally, searching a preset business knowledge base for a first business product that is relevant to the first business demand information includes:

[0012] Performing text screening on the first business demand information to obtain at least one business demand phrase;

[0013] Performing vector conversion on the at least one business demand phrase to obtain a phrase vector corresponding to the at least one business demand phrase;

[0014] Calling a preset cosine similarity algorithm, respectively calculating the similarity angle between each business product vector in the preset business knowledge base and the phrase vector, to obtain the similarity angle between each business product vector and the phrase vector; each business product vector has a corresponding first business product;

[0015] The phrase vectors corresponding to the similarity angles that are smaller than a preset first threshold value among the similarity angles are converted into corresponding first business products.

[0016] Optionally, calling a preset recommendation algorithm to analyze the business answer information to obtain a first recommended business product required by the user includes:

[0017] Matching the business answer information with a preset phrase library to obtain at least two demand phrases; the demand phrases are phrases that exist in both the business answer information and the preset phrase library;

[0018] Searching for a second service product that is relevant to the at least two demand phrases in a preset service product library, and obtaining the user's historical service products;

[0019] Calculating a first recommendation score for the first service product, calculating a second recommendation score for the second service product, and calculating a third recommendation score for the historical service product;

[0020] The business product corresponding to the highest recommendation score among the first recommendation score, the second recommendation score and the third recommendation score is determined as the first recommended business product.

[0021] Optionally, the calculating of the first recommendation score of the first business product, the calculating of the second recommendation score of the second business product, and the calculating of the third recommendation score of the historical business product include:

[0022] respectively constructing a first judgment matrix between the first service product, the second service product, the historical service product and each demand phrase;

[0023] Performing weight calculation on the first judgment matrix to obtain a first weight between the first service product and each demand phrase, a second weight between the second service product and each demand phrase, and a third weight between the historical service product and each demand phrase;

[0024] Calling a preset weight algorithm to respectively calculate fourth weights of the at least two demand phrases;

[0025] The fourth weight is calculated with the first weight to obtain the first recommendation score, the fourth weight is calculated with the second weight to obtain the second recommendation score, and the fourth weight is calculated with the third weight to obtain the third recommendation score.

[0026] Optionally, the calling of a preset weight algorithm to respectively calculate the fourth weights of the at least two demand phrases includes:

[0027] Obtaining the priority order of at least two demand phrases;

[0028] According to a pre-established mapping relationship between the priority order and the importance score, searching for the importance score of the priority order;

[0029] Based on the importance scores, constructing a second judgment matrix of the at least two demand phrases;

[0030] The second judgment matrix is ​​weighted to obtain fourth weights of the at least two demand phrases.

[0031] Optionally, after calling a preset weight algorithm to respectively calculate the fourth weights of the at least two demand phrases, the method further includes:

[0032] Calling a preset consistency algorithm to calculate the fourth weight to obtain a consistency ratio of the fourth weight;

[0033] If the consistency ratio is greater than the preset second threshold, the fourth weight is recalculated until the consistency ratio is less than the preset second threshold.

[0034] Optionally, the recalculating the fourth weight until the consistency ratio is less than the preset second threshold comprises:

[0035] According to the trend of making the consistency ratio smaller than the second threshold, the mapping relationship is adjusted, and based on the adjusted mapping relationship, the fourth weight is recalculated until the consistency ratio is smaller than the preset second threshold.

[0036] Optionally, after determining the service product corresponding to the highest recommendation score among the first recommendation score, the second recommendation score and the third recommendation score as the first recommended service product, the method further includes:

[0037] Detecting whether the first recommended service product is selected;

[0038] If the first recommended business product is not selected, the second business demand information of the user is continuously obtained, and the second recommended business product of the user is searched until the second recommended business product is selected; the second business demand information is business demand information generated based on the first business demand information and the business answer information.

[0039] According to a second aspect of the present disclosure, a device for recommending a business product is provided, comprising:

[0040] An acquisition unit, used to acquire the first business demand information of the user;

[0041] A searching unit, configured to search a preset business knowledge base for a first business product that is relevant to the first business demand information;

[0042] An input unit, configured to input the first business demand information and the first business product into a preset question-answering model respectively, to obtain the business answer information required by the user; the business answer information includes the first business demand information and the first business product;

[0043] The analyzing unit is used to call a preset recommendation algorithm to analyze the business answer information to obtain a first recommended business product required by the user.

[0044] Optionally, the searching unit includes:

[0045] A screening module, configured to perform text screening on the first business demand information to obtain at least one business demand phrase;

[0046] A conversion module, configured to perform vector conversion on the at least one business demand phrase to obtain a phrase vector corresponding to the at least one business demand phrase;

[0047] A calculation module, used to call a preset cosine similarity algorithm, respectively calculate the similarity angle between each business product vector in the preset business knowledge base and the phrase vector, and obtain the similarity angle between each business product vector and the phrase vector; each business product vector has a corresponding first business product;

[0048] The conversion module is further configured to convert the phrase vector corresponding to the similarity angle less than a preset first threshold value among the similarity angles into a corresponding first business product.

[0049] Optionally, the analysis unit includes:

[0050] A matching module, used to match the business answer information with a preset phrase library to obtain at least two demand phrases; the demand phrases are phrases that exist in both the business answer information and the preset phrase library;

[0051] A search module, used to search a preset service product library for a second service product that is relevant to the at least two demand phrases, and to obtain the user's historical service products;

[0052] a calculation module, configured to calculate a first recommendation score for the first service product, calculate a second recommendation score for the second service product, and calculate a third recommendation score for the historical service product;

[0053] A determination module is used to determine the business product corresponding to the highest recommendation score among the first recommendation score, the second recommendation score and the third recommendation score as the first recommended business product.

[0054] Optionally, the calculation module is further used to:

[0055] respectively constructing a first judgment matrix between the first service product, the second service product, the historical service product and each demand phrase;

[0056] Performing weight calculation on the first judgment matrix to obtain a first weight between the first service product and each demand phrase, a second weight between the second service product and each demand phrase, and a third weight between the historical service product and each demand phrase;

[0057] Calling a preset weight algorithm to respectively calculate fourth weights of the at least two demand phrases;

[0058] The fourth weight is calculated with the first weight to obtain the first recommendation score, the fourth weight is calculated with the second weight to obtain the second recommendation score, and the fourth weight is calculated with the third weight to obtain the third recommendation score.

[0059] Optionally, the calculation module is further used to:

[0060] Obtaining the priority order of at least two demand phrases;

[0061] According to a pre-established mapping relationship between the priority order and the importance score, searching for the importance score of the priority order;

[0062] Based on the importance scores, constructing a second judgment matrix of the at least two demand phrases;

[0063] The second judgment matrix is ​​weighted to obtain fourth weights of the at least two demand phrases.

[0064] Optionally, the device further comprises:

[0065] a calculation unit, configured to, after invoking a preset weight algorithm to respectively calculate the fourth weights of the at least two demand phrases, invoke a preset consistency algorithm to calculate the fourth weights to obtain a consistency ratio of the fourth weights;

[0066] The calculation unit is further configured to, when the consistency ratio is greater than a preset second threshold, recalculate the fourth weight until the consistency ratio is less than the preset second threshold.

[0067] Optionally, the computing unit includes:

[0068] A calculation module is used to adjust the mapping relationship according to the trend of making the consistency ratio less than the second threshold, and recalculate the fourth weight based on the adjusted mapping relationship until the consistency ratio is less than the preset second threshold.

[0069] Optionally, the device further comprises:

[0070] a detection unit, configured to detect whether the first recommended service product is selected after determining the service product corresponding to the highest recommendation score among the first recommendation score, the second recommendation score and the third recommendation score as the first recommended service product;

[0071] A search unit is used to continuously obtain the user's second business demand information and search for the user's second recommended business product when the first recommended business product is not selected, until the second recommended business product is selected; the second business demand information is business demand information generated based on the first business demand information and the business answer information.

[0072] According to a third aspect of the present disclosure, there is provided an electronic device, including:

[0073] at least one processor; and

[0074] a memory communicatively connected to the at least one processor; wherein,

[0075] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect.

[0076] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the method described in the first aspect.

[0077] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the method as described in the first aspect above.

[0078] The business product recommendation method and device, electronic device and storage medium provided by the present disclosure obtain the user's first business demand information; search for the first business product that is relevant to the first business demand information in the preset business knowledge base; input the first business demand information and the first business product into the preset question-and-answer model respectively to obtain the business answer information required by the user; the business answer information includes the first business demand information and the first business product; call the preset recommendation algorithm to analyze the business answer information to obtain the first recommended business product required by the user. Compared with the related art, the embodiment of the present disclosure has the user input the first business demand information, use the preset question-and-answer model to answer the first business demand information, obtain the business answer information, use the preset recommendation algorithm to analyze the business answer information, and obtain the first recommended business product required by the user, without manual operation, thereby improving the efficiency of business product recommendation, and recommending a suitable first recommended business product based on the first business demand information input by the user, thereby improving the satisfaction of business product recommendation.

[0079] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0081] Figure 1 A flowchart of a method for recommending a business product provided by an embodiment of the present disclosure;

[0082] Figure 2 A flowchart of a method for searching a first service product provided by an embodiment of the present disclosure;

[0083] Figure 3 A hierarchical structure diagram of a preset recommendation algorithm provided in an embodiment of the present disclosure;

[0084] Figure 4 A schematic diagram of a method for dividing importance scores provided in an embodiment of the present disclosure;

[0085] Figure 5 A display diagram of a judgment matrix provided in an embodiment of the present disclosure;

[0086] Figure 6 A calculation process diagram of a sum-product method provided in an embodiment of the present disclosure;

[0087] Figure 7 A hierarchical sorting diagram of a hierarchical analysis algorithm provided by an embodiment of the present disclosure;

[0088] Figure 8 A schematic diagram of a consistency check result provided by an embodiment of the present disclosure;

[0089] Fig. 9 A recommended cyclic process for a business product provided by an embodiment of the present disclosure;

[0090] Fig.10 A schematic diagram of a recommendation process of a first recommended service product provided by an embodiment of the present disclosure;

[0091] Fig.11 A schematic diagram of the structure of a device for recommending business products provided by an embodiment of the present disclosure;

[0092] Fig.12 A schematic diagram of the structure of another device for recommending business products provided in an embodiment of the present disclosure;

[0093] Fig.13 A schematic block diagram of an exemplary electronic device provided for an embodiment of the present disclosure. DETAILED DESCRIPTION

[0094] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0095] The following describes a method and apparatus for recommending a service product, an electronic device, and a storage medium according to an embodiment of the present disclosure with reference to the accompanying drawings.

[0096] Figure 1 A flowchart of a method for recommending a business product provided by an embodiment of the present disclosure.

[0097] like Figure 1 As shown, the method is applied to a recommendation system for business products, and the method includes the following steps:

[0098] Step 101: Obtain the user's first business demand information.

[0099] The first business demand information refers to the business demand or question raised by the user and hoped to be answered. For example, the first business demand information may be that the user wants to apply for a package of about 200 yuan, requiring the traffic to reach 500G and the voice call duration to be more than 100 minutes.

[0100] The method for obtaining the user's first business demand information may be that the user inputs the first business demand information into the recommendation system of the business product through the front-end interactive interface of the recommendation system of the business product, so as to obtain the first business demand information. The input method of the first business demand information may be text input or voice input. If the input method of the first business demand information is voice input, the first business demand information in voice form is converted into the first business demand information in text form. However, it should be clear that this statement is not intended to limit the method of obtaining the user's first business demand information to only inputting the first business demand information into the recommendation system of the business product through the front-end interactive interface of the recommendation system of the business product, and the user's first business demand information may also be obtained through other methods.

[0101] By obtaining the user's first business demand information, we can understand the user's specific needs for business products. The business product recommendation system can better personalize the recommendation of related business products or services, thereby improving the satisfaction with the business product recommendations.

[0102] Step 102: searching a preset business knowledge base for a first business product that is relevant to the first business demand information.

[0103] The preset business knowledge base is a database that stores various business products and their attributes, features and related information. The preset business knowledge base is used to find business products related to the user's business demand information. The data in the preset business knowledge base is stored in the form of vectors. The preset business knowledge base is used to store the business knowledge. The first business product is a product that meets the user's communication and information service needs. The first business product includes but is not limited to communication service packages, electronic devices, virtual storage space, and virtual accounts.

[0104] By screening out the phrases in the first business demand information that are relevant to the business product, and converting the screened phrases into vectors, using the cosine similarity angle algorithm, the similarity angle between the phrase vector and the vector in the preset business knowledge base is calculated to obtain the similarity angle between the vector in the preset business knowledge base and the phrase vector, matching the vectors in the preset business knowledge base corresponding to the similarity angle less than the preset first threshold, and converting the matched vectors into text information, which is the first business product, wherein the preset first threshold can be any value, and the embodiment of the present disclosure does not limit the value of the preset first threshold. However, it should be clear that this statement is not intended to limit the method of searching for the first business product that is relevant to the first business demand information in the preset business knowledge base to only the above method, but can also be other methods.

[0105] Finding the first business product that is relevant to the first business demand information can ensure that the recommended business products are highly consistent with the user's needs, reduce misleading recommendations, and improve the accuracy and effectiveness of business product recommendations.

[0106] Step 103 , input the first business demand information and the first business product into a preset question-answer model respectively to obtain business answer information required by the user; the business answer information includes the first business demand information and the first business product.

[0107] The preset question-and-answer model is designed through the LangChian (LC) development framework and the Large Language Model (LLM) to obtain a model for understanding and answering user questions. The input of the preset question-and-answer model is divided into two parts, one part is the first business demand information input by the user, and the other part is the first business product that is related to the first business demand information. The output of the preset question-and-answer model is the business answer information required by the user obtained by the preset question-and-answer model processing the first business demand information and the first business product, wherein the business answer information refers to the information required by the user to be answered generated by the preset question-and-answer model based on the user's first business demand information and the first business product. The business answer information includes but is not limited to a detailed explanation of the first business demand, the characteristics of the first business product, the cost of the first business product, and the processing procedures of the first business product.

[0108] The preset question-answering model can improve the efficiency of processing input information and generating answers, thereby improving the efficiency of services. Compared with manual answers, the preset question-answering model can respond to user needs faster, reduce waiting time, and improve the overall service response speed.

[0109] Step 104: calling a preset recommendation algorithm to analyze the business answer information to obtain a first recommended business product required by the user.

[0110] The preset recommendation algorithm is to perform a hierarchical analysis on the business products in the preset business knowledge base, the user's historical business products and the products in the preset business product library based on the business answer information, and select the first recommended product that best meets the user's needs for recommendation.

[0111] Accurate business product recommendations can increase users' willingness to buy, thereby improving the conversion rate and sales performance of business products.

[0112] The business product recommendation method provided by the present disclosure obtains the user's first business demand information; searches for a first business product that is relevant to the first business demand information in a preset business knowledge base; inputs the first business demand information and the first business product into a preset question-and-answer model respectively to obtain the business answer information required by the user; the business answer information includes the first business demand information and the first business product; calls a preset recommendation algorithm to analyze the business answer information to obtain the first recommended business product required by the user. Compared with the related art, the embodiment of the present disclosure receives the user input of the first business demand information, uses a preset question-and-answer model to answer the first business demand information to obtain business answer information, uses a preset recommendation algorithm to analyze the business answer information to obtain the first recommended business product required by the user, and does not require manual operation, thereby improving the efficiency of business product recommendation, and recommends a suitable first recommended business product based on the first business demand information input by the user, thereby improving the satisfaction of business product recommendation.

[0113] As a refinement of step 102, when searching the preset business knowledge base for the first business product that is relevant to the first business demand information, the following methods may be used but are not limited to: Figure 2 As shown, Figure 2 A flowchart of a method for searching a first service product provided by an embodiment of the present disclosure includes:

[0114] Step 201: perform text screening on the first business demand information to obtain at least one business demand phrase.

[0115] Business demand phrases refer to keywords or phrases extracted from the user's first business demand information, which represent the user's specific needs or goals. Business demand phrases are used to express the problem the user wants to solve or the service the user expects to obtain. For example, the first business demand information is "recommend a data package of about 100 yuan", where 100 yuan, data, and package are business demand phrases.

[0116] By using term frequency-inverse document frequency (TF-IDF) to extract all phrases in the first business demand information, matching all phrases extracted from the first business demand information with phrases in a preset business phrase library, and filtering out phrases in the first business demand information that are consistent with phrases in the preset business phrase library, at least one business demand phrase is obtained. However, it should be clear that this statement is not intended to limit text filtering to TF-IDF, and text filtering can also be achieved through other methods.

[0117] To facilitate understanding, an example is provided. Suppose the first business demand information contains 5 phrases, namely a, b, c, d, and e, and the preset business phrase library contains 6 phrases, namely a, c, e, f, g, and h. Match a, b, c, d, and e with a, c, e, f, g, and h to obtain a, c, and e. Then a, c, and e are business phrases.

[0118] By focusing on key business demand phrases, it is possible to more accurately match products with the business knowledge base, avoiding interference of meaningless phrases in the first business demand information on business product recommendations, thereby improving the accuracy of business product recommendations.

[0119] Step 202: Perform vector conversion on the at least one business demand phrase to obtain a phrase vector corresponding to the at least one business demand phrase.

[0120] At least one business demand phrase is converted into a corresponding phrase vector by using TF-IDF. However, it should be clear that this statement is not intended to limit vector conversion to TF-IDF, and vector conversion can also be achieved through other methods.

[0121] Vector representation is usually more compact and efficient than text data, which can improve the efficiency of calculating similarity angles.

[0122] Step 203, calling a preset cosine similarity algorithm, respectively calculating the similarity angle between each business product vector in the preset business knowledge base and the phrase vector, to obtain the similarity angle between each business product vector and the phrase vector; each business product vector has a corresponding first business product.

[0123] The preset cosine similarity algorithm is an algorithm for measuring the similarity angle between two vectors. The preset cosine similarity algorithm determines the similarity between two vectors by calculating the similarity angle between them. The smaller the angle, the higher the similarity.

[0124] The phrase vectors corresponding to at least one business demand phrase are combined into a combination vector, and the similarity angle between each business product vector in the preset business knowledge base and the combination vector is calculated to obtain the similarity angle between each business product vector and the combination vector, that is, the similarity angle between each business product vector and the phrase vector.

[0125] The similarity angle can be used to accurately match users' business demands, which helps improve user experience, enables them to find the required business products more quickly, and increases satisfaction.

[0126] Step 204: convert the phrase vectors corresponding to the similarity angles smaller than a preset first threshold value among the similarity angles into corresponding first business products.

[0127] By using TF-IDF, the phrase vector corresponding to the similarity angle less than the preset first threshold value in the similarity angle is converted into the corresponding first business product, wherein the preset first threshold value can be any value, which is not limited in the embodiment of the present disclosure. However, it should be clear that this statement is not intended to limit the phrase vector conversion to only through TF-IDF, and phrase vector conversion can also be achieved through other methods.

[0128] To facilitate understanding, an example is provided. It is assumed that the first business demand phrase contains three business phrases, namely a, b, and c, where the phrase vector corresponding to a is A, the phrase vector corresponding to b is B, and the phrase vector corresponding to c is C. A, B, and C are combined into D. There are two business product vectors in the preset business knowledge base, namely E and F. The similarity angles of E and F with D are calculated respectively, and the similarity angle between E and D is 20, and the similarity angle between F and D is 10. The preset first threshold is 15, then F is converted into the corresponding first business product.

[0129] By setting a preset first threshold, products that are highly relevant to business demands can be screened out to ensure that the recommended business products are more in line with user needs.

[0130] As a refinement of step 104, when executing the call to the preset recommendation algorithm, analyzing the business answer information, and obtaining the first recommended business product required by the user, it can be implemented in but not limited to the following manner: matching the business answer information with a preset phrase library to obtain at least two demand phrases; the demand phrases are phrases that exist in both the business answer information and the preset phrase library; searching the preset business product library for a second business product that is correlated with the at least two demand phrases, and obtaining the user's historical business products; calculating a first recommendation score for the first business product, calculating a second recommendation score for the second business product, and calculating a third recommendation score for the historical business product; determining the business product corresponding to the highest recommendation score among the first recommendation score, the second recommendation score, and the third recommendation score as the first recommended business product.

[0131] Exemplarily, the recommendation score refers to a quantitative indicator used to evaluate and compare the matching degree between different business products and at least two demand phrases. The main difference between the preset business product library and the preset business knowledge base is that they store different contents and purposes. The preset business product library is used to store and recommend business products, while the preset business knowledge base is used to provide business-related knowledge and information. The recommendation score can be calculated by the hierarchical analysis algorithm, but it should be clear that this statement is not intended to limit the recommendation score to only be achieved by the hierarchical analysis algorithm, and it can also be achieved by other methods.

[0132] Through the hierarchical analysis algorithm, the analysis of the preset recommendation algorithm is decomposed into the target layer (optimal product), the criterion layer (factors affecting decision-making, such as price, traffic, voice, etc.) and the solution layer (referring to the solution, such as order center products, knowledge base products, product center products, etc.). In order to facilitate a better understanding of the hierarchical structure of the preset recommendation algorithm, such as Figure 3 As shown, Figure 3 A hierarchical structure diagram of a preset recommendation algorithm provided in an embodiment of the present disclosure, wherein: Solution layer: product alternatives for solving problems, including the user's historical business products, a first business product in a preset business knowledge base that is correlated with the first business demand information, and a second business product in a preset business product library that is correlated with at least two demand phrases in the business answer information. Criteria layer: obtaining at least two demand phrases in the business answer information, and conditions and solutions adopted to screen out user needs, such as price, traffic, voice, value-added services, functions, etc. Purpose layer: the best product that is screened out to best meet user needs.

[0133] The factors of the criterion layer (such as price, traffic, voice, value-added services, functions, etc.) are compared with each other. In order to determine the weight of each factor in the criterion layer for the optimal product, the importance score of the judgment matrix (1-9 importance score method) is introduced. In order to better understand the importance score, Figure 4 As shown, Figure 4 A schematic diagram of a method for dividing importance scores provided by an embodiment of the present disclosure. When the importance score is 1, the two factors are equally important. When the importance score is 3, the former is slightly more important than the latter. When the importance score is 5, the former is more important than the latter. When the importance score is 7, the former is extremely more important than the latter. When the importance score is 9, the former is strongly more important than the latter. When the importance scores are 2, 4, 6, and 8, they are the intermediate values ​​of the above adjacent judgments.

[0134] The criterion layer creates an n*n judgment matrix for judgment, which is constructed according to the key information points of user needs. For example, the user's demand is "I want to apply for a package of about 200 yuan, requiring a traffic of 500G and a voice call duration of more than 100 minutes". The key information is a 200 yuan package, 500G traffic, and 100 minutes of call duration. Priority is sorted according to the importance of the user's demand points. For example, the price is obviously important among all conditions and is set to 5 points. Other conditions are sorted according to the order proposed by the user, with traffic set to 3 points and voice set to 1 point. The angle line is the judgment of each indicator. In order to facilitate a better understanding of the judgment matrix, such as Figure 5 As shown, Figure 5 This is a display diagram of a judgment matrix provided by an embodiment of the present disclosure. For "price" and "price", the importance is 1, because the indicator itself is 1:1. For "traffic" and "price", "price" is slightly more important than "traffic", so it can be marked as 3. For "price" and "voice", "price" is obviously more important than "voice", so it can be marked as 5, and so on, until a complete judgment matrix is ​​constructed.

[0135] After constructing the judgment matrix of the criterion layer, its weight (weight vector) is calculated by the sum-product method, and the weight value of each factor (price, traffic, voice) indicator of the criterion layer is calculated. In order to better understand the sum-product method, Figure 6 As shown, Figure 6This is a calculation process diagram of a sum-product method provided by an embodiment of the present disclosure. The specific calculation can be performed based on the judgment matrix, and the calculation ensures that it can meet the eigenvalue and eigenvector conditions of the judgment matrix. Here, the maximum eigenvalue of the judgment matrix is ​​λmax, and the normalized eigenvector corresponding to λmax is W, which is the component of W. The judgment matrix is ​​used to calculate the weight (weight coefficient) of each factor on the target layer (optimal product). The calculation steps of the weight vector (W) and the maximum feature (λmax).

[0136] For ease of understanding, an example is provided. Assuming that the first recommendation score of the first business product is 0.2, the second recommendation score of the second business product is 0.3, and the third recommendation score of the third business product is 0.6, the third business product is the first recommended business product.

[0137] By calculating the recommendation score, we ensure that the final recommended business products are the most in line with user needs and historical behaviors, thereby improving conversion rate and user satisfaction.

[0138] As a refinement of the above embodiment, when executing the calculation of the first recommendation score of the first business product, the calculation of the second recommendation score of the second business product, and the calculation of the third recommendation score of the historical business product, it can be implemented in the following manner but not limited to: respectively constructing a first judgment matrix between the first business product, the second business product, and the historical business product and each demand phrase; performing weight calculation on the first judgment matrix to obtain the first weight between the first business product and each demand phrase, the second weight between the second business product and each demand phrase, and the third weight between the historical business product and each demand phrase; calling a preset weight algorithm to calculate the fourth weight of the at least two demand phrases respectively; calculating the fourth weight with the first weight to obtain the first recommendation score, calculating the fourth weight with the second weight to obtain the second recommendation score, and calculating the fourth weight with the third weight to obtain the third recommendation score. By calculating the first recommendation score, the second recommendation score, and the third recommendation score, a score can be provided for each business product to indicate its potential as a recommended product. The recommendation score takes into account the degree of matching between the business product and the user's needs and the user's historical preferences.

[0139] For example, according to each factor of the solution layer (historical products, knowledge base products, product center products), by comparing the factors in pairs, the weight values ​​(weight vectors) are calculated and consistency checks are performed. Unlike the hierarchical single sorting, the hierarchical total sorting requires multiple criterion layer factors (price, traffic, voice) to construct multiple judgment matrices for judgment, and the calculation is performed based on the criterion layer weight scores (price, traffic, voice) that have been calculated. For example, the solution layer factors are (historical products, knowledge base products, product center products), and three judgment matrix tables of solution layer factors are constructed, namely, the scores of A1 (historical products, knowledge base products, product center) for price, A2 (historical products, knowledge base products, product center) for traffic, and A3 (historical products, knowledge base products, product center) for voice. The weight calculation and consistency comparison are performed in the same way as the single-level sorting, and the scores of historical products, knowledge base products, and product center in price, traffic, and voice are obtained. Then for historical products, the total score is calculated as follows: the score of historical products in price * the weight of price + the score of historical products in traffic * the weight of traffic + the score of historical products in voice * the weight of voice = the total score of historical products. In order to better understand the hierarchical sorting of the hierarchical analysis algorithm, as follows Figure 7 As shown, Figure 7 The hierarchical ranking diagram of a hierarchical analysis algorithm provided by the embodiment of the present disclosure is performed from the highest level to the lowest level. Finally, the one with the highest score through comparison is the optimal service product of the target layer.

[0140] As a refinement of the above embodiment, when executing the call to the preset weight algorithm and calculating the fourth weights of the at least two demand phrases respectively, it can be implemented in the following manner but not limited to: obtaining the priority order of the at least two demand phrases; finding the importance score of the priority order according to the pre-established mapping relationship between the priority order and the importance score; constructing a second judgment matrix of the at least two demand phrases based on the importance score; performing weight calculation on the second judgment matrix to obtain the fourth weights of the at least two demand phrases. Among them, the construction method of the judgment matrix has been introduced in the above embodiment, and the embodiment of the present disclosure will not be elaborated. The fourth weight more accurately captures the user's demand preferences and gives appropriate weights to different demands in the recommendation process. This helps to improve the accuracy of recommendations and user satisfaction.

[0141] In actual applications, after calling the preset weight algorithm and calculating the fourth weights of the at least two demand phrases respectively, the calculation result of the fourth weight may be inaccurate. In order to make the fourth weight more accurate, it is necessary to check the consistency of the fourth weight, which can be implemented in the following way but not limited to: calling the preset consistency algorithm, calculating the fourth weight, and obtaining the consistency ratio of the fourth weight; if the consistency ratio is greater than the preset second threshold, recalculating the fourth weight until the consistency ratio is less than the preset second threshold. The preset second threshold can be any value, which is not limited in the embodiments of the present disclosure.

[0142] The principle of consistency check can be realized by formula (1):

[0143] The necessary and sufficient condition is

[0144] Among them, a ij Score the importance of the i-th row and j-th column, where n is the total number of rows or columns and k is a constant.

[0145] The calculation formula of the preset consistency algorithm can be implemented by formula (2):

[0146]

[0147] Among them, CR is the consistency ratio, CI is the consistency index, and RI is the average random consistency index.

[0148] The calculation formula of the consistency index can be realized by (3):

[0149]

[0150] Among them, CI is the consistency index, λ max is the maximum eigenvalue of the judgment matrix, and n is the total number of rows or columns of the judgment matrix.

[0151] In order to better understand the consistency test, Figure 8 As shown, Figure 8 A schematic diagram of the result of a consistency check provided in an embodiment of the present disclosure, assuming that the preset second threshold is 0.1, the consistency check is analyzed using the CR value, and a CR value less than 0.1 indicates that the consistency check has passed.

[0152] By continuously adjusting the fourth weight until the consistency requirement is met, recommendation errors caused by improper allocation of the fourth weight are reduced.

[0153] As a refinement of the above embodiment, when recalculating the fourth weight until the consistency ratio is less than the preset second threshold, it can be implemented in the following manner but not limited to: adjusting the mapping relationship according to the trend of making the consistency ratio less than the second threshold, and recalculating the fourth weight based on the adjusted mapping relationship until the consistency ratio is less than the preset second threshold. By ensuring the consistency and rationality of weight allocation, the recommendation system can provide recommendations that better meet user needs, thereby improving user satisfaction and trust.

[0154] In actual applications, after the business product corresponding to the highest recommendation score among the first recommendation score, the second recommendation score and the third recommendation score is determined as the first recommended business product, the user may be dissatisfied with the first recommended business product and need to recommend other business products to the user, which can be achieved by, but not limited to, the following methods: detecting whether the first recommended business product is selected; if the first recommended business product is not selected, continuously obtaining the second business demand information of the user, and searching for the second recommended business product of the user until the second recommended business product is selected; the second business demand information is business demand information generated based on the first business demand information and the business answer information. Continuously paying attention to user needs and providing corresponding recommendations can effectively reduce the risk of user loss due to failure to find suitable products and improve user retention rate.

[0155] In order to better understand the entire process of business product recommendation, Fig. 9 As shown, Fig. 9 A cyclic process for recommending a business product provided in an embodiment of the present disclosure continuously obtains the user's demand points during multiple rounds of user interaction, and continuously optimizes the algorithm analysis to obtain the user's potential business handling needs.

[0156] In one possible implementation of the embodiment of the present disclosure, in order to facilitate a better understanding of the recommendation process of the first recommended service product, as follows: Fig.10 As shown, Fig.10A schematic diagram of a recommendation process for a first recommended business product provided in an embodiment of the present disclosure, obtaining a user's first business demand information, performing text screening on the first business demand information to obtain at least one business demand phrase, searching a preset business knowledge base for a first business product that is relevant to the first business demand information, matching business answer information with a preset phrase library to obtain at least two demand phrases, searching a preset business product library for a second business product that is relevant to the at least two demand phrases, and obtaining the user's historical business products, calculating a first recommendation score for the first business product, calculating a second recommendation score for the second business product, and calculating a third recommendation score for the historical business products, and determining the business product corresponding to the highest recommendation score among the first recommendation score, the second recommendation score, and the third recommendation score as the first recommended business product.

[0157] In summary, the embodiments of the present disclosure can achieve the following effects:

[0158] In the disclosed embodiment, a user inputs first business demand information, uses a preset question-and-answer model to answer the first business demand information, obtains business answer information, and uses a preset recommendation algorithm to analyze the business answer information to obtain a first recommended business product required by the user. No manual operation is required, thereby improving the efficiency of business product recommendations, and recommends a suitable first recommended business product based on the first business demand information input by the user, thereby improving satisfaction with business product recommendations.

[0159] Corresponding to the above-mentioned service product recommendation method, the present invention also provides a service product recommendation device. Since the device embodiment of the present invention corresponds to the above-mentioned method embodiment, details not disclosed in the device embodiment can be referred to the above-mentioned method embodiment, and will not be repeated in the present invention.

[0160] Fig.11 Schematic diagram of a business product recommendation device provided by an embodiment of the present disclosure, wherein the device is applied to a business product recommendation system, such as Fig.11 As shown, including:

[0161] An acquisition unit 31 is used to acquire first business demand information of a user;

[0162] A searching unit 32, configured to search a preset business knowledge base for a first business product that is relevant to the first business demand information;

[0163] An input unit 33 is used to input the first business demand information and the first business product into a preset question-answering model respectively to obtain the business answer information required by the user; the business answer information includes the first business demand information and the first business product;

[0164] The analyzing unit 34 is used to call a preset recommendation algorithm to analyze the business answer information to obtain the first recommended business product required by the user.

[0165] The business product recommendation device provided by the present disclosure obtains the first business demand information of the user; searches for the first business product that is relevant to the first business demand information in the preset business knowledge base; inputs the first business demand information and the first business product into the preset question-and-answer model respectively to obtain the business answer information required by the user; the business answer information includes the first business demand information and the first business product; calls the preset recommendation algorithm to analyze the business answer information to obtain the first recommended business product required by the user. Compared with the related art, the embodiment of the present disclosure has the user input the first business demand information, uses the preset question-and-answer model to answer the first business demand information, obtains the business answer information, uses the preset recommendation algorithm to analyze the business answer information, and obtains the first recommended business product required by the user, without manual operation, thereby improving the recommendation efficiency of business products, and recommends a suitable first recommended business product based on the first business demand information input by the user, thereby improving the satisfaction of the business product recommendation.

[0166] Furthermore, in a possible implementation of the embodiment of the present disclosure, as Fig.12 As shown, the search unit 32 includes:

[0167] A screening module 321 is used to perform text screening on the first business demand information to obtain at least one business demand phrase;

[0168] A conversion module 322, configured to perform vector conversion on the at least one business demand phrase to obtain a phrase vector corresponding to the at least one business demand phrase;

[0169] The calculation module 323 is used to call a preset cosine similarity algorithm to calculate the similarity angle between each business product vector in the preset business knowledge base and the phrase vector, and obtain the similarity angle between each business product vector and the phrase vector; each business product vector has a corresponding first business product;

[0170] The conversion module 322 is further configured to convert the phrase vectors corresponding to the similarity angles less than a preset first threshold value among the similarity angles into corresponding first business products.

[0171] Furthermore, in a possible implementation of the embodiment of the present disclosure, as Fig.12 As shown, the analysis unit 34 includes:

[0172] A matching module 341 is used to match the business answer information with a preset phrase library to obtain at least two demand phrases; the demand phrase is a phrase that exists in both the business answer information and the preset phrase library;

[0173] A search module 342 is used to search for a second business product that is relevant to the at least two demand phrases in a preset business product library, and obtain the user's historical business products;

[0174] A calculation module 343, used to calculate a first recommendation score for the first service product, calculate a second recommendation score for the second service product, and calculate a third recommendation score for the historical service product;

[0175] The determination module 344 is configured to determine the service product corresponding to the highest recommendation score among the first recommendation score, the second recommendation score and the third recommendation score as the first recommended service product.

[0176] Furthermore, in a possible implementation of the embodiment of the present disclosure, as Fig.12 As shown, the calculation module 343 is also used for:

[0177] respectively constructing a first judgment matrix between the first service product, the second service product, the historical service product and each demand phrase;

[0178] Performing weight calculation on the first judgment matrix to obtain a first weight between the first service product and each demand phrase, a second weight between the second service product and each demand phrase, and a third weight between the historical service product and each demand phrase;

[0179] Calling a preset weight algorithm to respectively calculate fourth weights of the at least two demand phrases;

[0180] The fourth weight is calculated with the first weight to obtain the first recommendation score, the fourth weight is calculated with the second weight to obtain the second recommendation score, and the fourth weight is calculated with the third weight to obtain the third recommendation score.

[0181] Furthermore, in a possible implementation of the embodiment of the present disclosure, as Fig.12 As shown, the calculation module 343 is also used for:

[0182] Obtaining the priority order of at least two demand phrases;

[0183] According to a pre-established mapping relationship between the priority order and the importance score, searching for the importance score of the priority order;

[0184] Based on the importance scores, constructing a second judgment matrix of the at least two demand phrases;

[0185] The second judgment matrix is ​​weighted to obtain fourth weights of the at least two demand phrases.

[0186] Furthermore, in a possible implementation of the embodiment of the present disclosure, as Fig.12 As shown, the device also includes:

[0187] The calculation unit 35 is used to call a preset weight algorithm to respectively calculate the fourth weights of the at least two demand phrases, and then call a preset consistency algorithm to calculate the fourth weights to obtain a consistency ratio of the fourth weights;

[0188] The calculation unit 35 is further configured to, when the consistency ratio is greater than a preset second threshold, recalculate the fourth weight until the consistency ratio is less than the preset second threshold.

[0189] Furthermore, in a possible implementation of the embodiment of the present disclosure, as Fig.12 As shown, the calculation unit 35 includes:

[0190] The calculation module 351 is used to adjust the mapping relationship according to the trend of making the consistency ratio less than the second threshold, and recalculate the fourth weight based on the adjusted mapping relationship until the consistency ratio is less than the preset second threshold.

[0191] Furthermore, in a possible implementation of the embodiment of the present disclosure, as Fig.12 As shown, the device also includes:

[0192] A detection unit 36, configured to detect whether the first recommended service product is selected after determining the service product corresponding to the highest recommendation score among the first recommendation score, the second recommendation score and the third recommendation score as the first recommended service product;

[0193] The search unit 37 is used to continuously obtain the second business demand information of the user and search for the second recommended business product of the user when the first recommended business product is not selected, until the second recommended business product is selected; the second business demand information is business demand information generated based on the first business demand information and the business answer information.

[0194] It should be noted that the above explanation of the method embodiment is also applicable to the device of the embodiment of the present disclosure, and the principle is the same, which is no longer limited in the embodiment of the present disclosure.

[0195] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0196] Fig.13 A schematic block diagram of an example electronic device 400 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0197] like Fig.13 As shown, the device 400 includes a computing unit 401, which can perform various appropriate actions and processes according to a computer program stored in a ROM (Read-Only Memory) 402 or a computer program loaded from a storage unit 408 to a RAM (Random Access Memory) 403. In the RAM 403, various programs and data required for the operation of the device 400 can also be stored. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An I / O (Input / Output) interface 405 is also connected to the bus 404.

[0198] A number of components in the device 400 are connected to the I / O interface 405, including: an input unit 406, such as a keyboard, a mouse, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 408, such as a disk, an optical disk, etc.; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0199] The computing unit 401 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a CPU (Central Processing Unit), a GPU (Graphic Processing Units), various dedicated AI (Artificial Intelligence) computing chips, various computing units running machine learning model algorithms, a DSP (Digital Signal Processor), and any appropriate processor, controller, microcontroller, etc. The computing unit 401 performs the various methods and processes described above, such as a recommendation method for a business product. For example, in some embodiments, the recommendation method for a business product may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 408. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the computing unit 401, one or more steps of the method described above may be performed. Alternatively, in other embodiments, the computing unit 401 may be configured to execute the aforementioned service product recommendation method in any other appropriate manner (for example, by means of firmware).

[0200] Various embodiments of the systems and techniques described above herein may be implemented in digital electronic circuit systems, integrated circuit systems, FPGAs (Field Programmable Gate Arrays), ASICs (Application-Specific Integrated Circuits), ASSPs (Application Specific Standard Products), SOCs (System On Chips), CPLDs (Complex Programmable Logic Devices), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs that may be executed and / or interpreted on a programmable system including at least one programmable processor that may be a dedicated or general-purpose programmable processor that may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0201] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0202] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a RAM, a ROM, an EPROM (Electrically Programmable Read-Only-Memory) or a flash memory, an optical fiber, a CD-ROM (Compact Dis sc Read-Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0203] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (Cathode-Ray Tube) or LCD (Liquid Crystal Display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball), through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0204] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: LAN (Local Area Network), WAN (Wide Area Network), the Internet, and blockchain networks.

[0205] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services ("Virtual Private Server", or "VPS" for short). The server may also be a server of a distributed system, or a server combined with a blockchain.

[0206] It should be noted that artificial intelligence is a discipline that studies how computers can simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.), and includes both hardware-level and software-level technologies. Artificial intelligence hardware technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, and big data processing; artificial intelligence software technologies mainly include computer vision technology, speech recognition technology, natural language processing technology, as well as machine learning / deep learning, big data processing technology, knowledge graph technology, and other major directions.

[0207] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0208] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for recommending a business product, characterized in that: include: Obtain the user's first business demand information; Searching a preset business knowledge base for a first business product that is relevant to the first business demand information; Inputting the first business demand information and the first business product into a preset question-answering model respectively to obtain business answer information required by the user; The business answer information includes the first business demand information and the first business product; A preset recommendation algorithm is called to analyze the business answer information to obtain a first recommended business product required by the user.

2. The method according to claim 1, characterized in that The searching the preset business knowledge base for the first business product that is relevant to the first business demand information includes: Performing text screening on the first business demand information to obtain at least one business demand phrase; Performing vector conversion on the at least one business demand phrase to obtain a phrase vector corresponding to the at least one business demand phrase; Calling a preset cosine similarity algorithm, respectively calculating the similarity angle between each business product vector in the preset business knowledge base and the phrase vector, to obtain the similarity angle between each business product vector and the phrase vector; each business product vector has a corresponding first business product; The phrase vectors corresponding to the similarity angles that are smaller than a preset first threshold value among the similarity angles are converted into corresponding first business products.

3. The method according to claim 1, characterized in that The calling of a preset recommendation algorithm to analyze the business answer information to obtain the first recommended business product required by the user includes: Matching the business answer information with a preset phrase library to obtain at least two demand phrases; the demand phrases are phrases that exist in both the business answer information and the preset phrase library; Searching for a second service product that is relevant to the at least two demand phrases in a preset service product library, and obtaining the user's historical service products; Calculating a first recommendation score for the first service product, calculating a second recommendation score for the second service product, and calculating a third recommendation score for the historical service product; The business product corresponding to the highest recommendation score among the first recommendation score, the second recommendation score and the third recommendation score is determined as the first recommended business product.

4. The method according to claim 3, characterized in that The calculating of the first recommendation score of the first service product, the calculating of the second recommendation score of the second service product, and the calculating of the third recommendation score of the historical service product include: respectively constructing a first judgment matrix between the first service product, the second service product, the historical service product and each demand phrase; Performing weight calculation on the first judgment matrix to obtain a first weight between the first service product and each demand phrase, a second weight between the second service product and each demand phrase, and a third weight between the historical service product and each demand phrase; Calling a preset weight algorithm to respectively calculate fourth weights of the at least two demand phrases; The fourth weight is calculated with the first weight to obtain the first recommendation score, the fourth weight is calculated with the second weight to obtain the second recommendation score, and the fourth weight is calculated with the third weight to obtain the third recommendation score.

5. The method according to claim 4, characterized in that The calling of a preset weight algorithm to respectively calculate the fourth weights of the at least two demand phrases comprises: Obtaining the priority order of at least two demand phrases; According to a pre-established mapping relationship between the priority order and the importance score, searching for the importance score of the priority order; Based on the importance scores, constructing a second judgment matrix of the at least two demand phrases; The second judgment matrix is ​​weighted to obtain fourth weights of the at least two demand phrases.

6. The method according to claim 5, characterized in that After calling a preset weight algorithm to respectively calculate the fourth weights of the at least two demand phrases, the method further includes: Calling a preset consistency algorithm to calculate the fourth weight to obtain a consistency ratio of the fourth weight; If the consistency ratio is greater than the preset second threshold, the fourth weight is recalculated until the consistency ratio is less than the preset second threshold.

7. The method according to claim 6, characterized in that The recalculating the fourth weight until the consistency ratio is less than the preset second threshold comprises: According to the trend of making the consistency ratio smaller than the second threshold, the mapping relationship is adjusted, and based on the adjusted mapping relationship, the fourth weight is recalculated until the consistency ratio is smaller than the preset second threshold.

8. The method according to claim 3, characterized in that After determining the service product corresponding to the highest recommendation score among the first recommendation score, the second recommendation score and the third recommendation score as the first recommended service product, the method further includes: Detecting whether the first recommended service product is selected; If the first recommended business product is not selected, the second business demand information of the user is continuously obtained, and the second recommended business product of the user is searched until the second recommended business product is selected; the second business demand information is business demand information generated based on the first business demand information and the business answer information.

9. A device for recommending business products, characterized in that: include: An acquisition unit, used to acquire the first business demand information of the user; A searching unit, configured to search a preset business knowledge base for a first business product that is relevant to the first business demand information; An input unit, configured to input the first business demand information and the first business product into a preset question-answering model respectively, to obtain the business answer information required by the user; The business answer information includes the first business demand information and the first business product; The analyzing unit is used to call a preset recommendation algorithm to analyze the business answer information to obtain a first recommended business product required by the user.

10. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

11. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-8.

12. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 8.

Images