Product recommendation method, storage medium and computer program product

Abstract

The present application discloses a product recommendation method, storage medium and computer program product, which relates to the field of data processing technology, including: obtaining user information of a target user, and constructing a user portrait of the target user based on the user information; determining a first product recommendation list from a preset financial product database based on the user portrait; determining a second product recommendation list from the financial product database based on the enterprise geographical information and historical financing intention in the user information; and determining a recommended product based on the first product recommendation list and the second product recommendation list. The present application improves the accuracy of product recommendations.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a product recommendation method, a storage medium, and a computer program product. Background Art

[0002] With the development of Internet financial services, users are faced with more and more financial product choices. Existing product recommendations mainly rely on users to manually fill out forms to obtain financing needs, and then financial advisors recommend products based on financing demand information and relevant knowledge background of banks. However, this product recommendation method cannot grasp comprehensive financing information or product information in real time, which affects the accuracy of product recommendations.

[0003] The above contents are only used to assist in understanding the technical solution of the present application and do not constitute an admission that the above contents are prior art. Summary of the invention

[0004] The main purpose of this application is to provide a product recommendation method, storage medium and computer program product, aiming to solve the technical problem that the accuracy of product recommendations is affected.

[0005] To achieve the above objectives, the present application proposes a product recommendation method, which comprises:

[0006] Acquire user information of a target user, and construct a user profile of the target user based on the user information;

[0007] Determining a first product recommendation list from a preset financial product database based on the user portrait;

[0008] Determine a second product recommendation list from the financial product database based on the enterprise region information and historical financing intention in the user information;

[0009] Determine recommended products based on the first product recommendation list and the second product recommendation list.

[0010] In one embodiment, the step of determining a second product recommendation list from the financial product database based on the enterprise region information and historical financing intention in the user information includes:

[0011] Extracting a user modeling tag value of the target user under a preset modeling tag from the enterprise region information and the historical financing intention;

[0012] Determine a product modeling tag value for each financial product in the financial product database under the preset modeling tag;

[0013] Constructing a product radiation map centered on the target user based on the user modeling tag value and the product modeling tag value, wherein the central node of the product radiation map is the target user, and the radiation nodes of the product radiation map are the financial products in the financial product database;

[0014] A second product recommendation list is determined from the financial product database based on the product radiation map.

[0015] In one embodiment, the step of extracting the user modeling tag value of the target user under the preset modeling tag from the enterprise region information and the historical financing intention includes:

[0016] Determine the user registration location, user office location and user actual factory address of the target user from the enterprise geographical information, and determine the user historical financing situation of the target user from the historical financing intention, wherein the user historical financing situation is used to indicate whether the financing of the target user is successful;

[0017] The user's registered place is used as the user modeling tag value corresponding to the registered place tag in the preset modeling tag, the user's office location is used as the user modeling tag value corresponding to the office location tag in the preset modeling tag, the user's actual factory site is used as the user modeling tag value corresponding to the factory site tag in the preset modeling tag, and the user's historical financing situation is used as the user modeling tag value corresponding to the financing tag in the preset modeling tag.

[0018] In one embodiment, the step of constructing a product radiation map centered on the target user based on the user modeling tag value and the product modeling tag value includes:

[0019] Traversing each financial product in the financial product database, and determining a product score of the financial product based on the product modeling tag value and the user modeling tag value;

[0020] A product radiation diagram is constructed with the target user as the central node, each of the financial products as the radiation nodes, and the product score corresponding to each of the financial products as the length of the node connection line.

[0021] In one embodiment, the preset modeling tag includes a plurality of modeling sub-tags, and the step of determining the product score of the financial product based on the product modeling tag value and the user modeling tag value includes:

[0022] Determine a sub-tag difference value corresponding to each of the modeling sub-tags, wherein the sub-tag difference value is a difference between a product modeling tag value and a user modeling tag value under the modeling sub-tag;

[0023] Obtaining a preset first weighting coefficient group, wherein the first weighting coefficient group includes weighting coefficients corresponding to each of the modeling sub-tags;

[0024] The sub-label differences corresponding to each other are weighted by each weighting coefficient to obtain a weighted value, and the sum of each weighted value is used as the product score.

[0025] In one embodiment, the step of determining a first product recommendation list from a preset financial product database based on the user portrait includes:

[0026] For any target product among the financial products in the preset financial product database, determine the product attribute tag value of the target product under each attribute tag, wherein each attribute tag is a tag for constructing the user portrait;

[0027] Determine the user attribute tag value of the target user under each attribute tag based on the user portrait;

[0028] Traversing each of the attribute tags, calculating the attribute similarity between the user attribute tag value and the product attribute tag value, and calculating the sum of the attribute similarities to obtain a similarity sum value;

[0029] Determine first candidate products whose similarity and value meet a first preset recommendation condition from each of the financial products, and construct a first product recommendation list based on the first candidate products.

[0030] In one embodiment, the step of determining recommended products based on the first product recommendation list and the second product recommendation list includes:

[0031] Determine the regional attribute value of each list product in the first product recommendation list and the second product recommendation list;

[0032] Obtaining a preset second weighting coefficient group, and performing weighted processing on the regional attribute values ​​of each of the listed products using the second weighting coefficient group to obtain a recommendation score for each of the listed products;

[0033] The listed products whose recommendation scores satisfy the second preset recommendation condition among the listed products are taken as recommended products.

[0034] In one embodiment, the step of constructing a user profile of the target user based on the user information includes:

[0035] Extracting the target user's regional characteristics, risk preference, financing intention, and credit status from the user information, and determining an attribute label based on the regional characteristics, risk preference, financing intention, and credit status;

[0036] Determine the label index value of each of the attribute labels, and determine the index mapping value of each of the label index values ​​according to the preset portrait rules based on the label index values;

[0037] A user portrait is constructed based on the attribute label, the label indicator value and the indicator mapping value.

[0038] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the product recommendation method described above are implemented.

[0039] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the product recommendation method described above.

[0040] In the present application, user information of a target user is obtained, and a user profile of the target user is constructed based on the user information; a first product recommendation list is determined from a preset financial product database based on the user profile; a second product recommendation list is determined from the financial product database based on the enterprise geographical information and historical financing intentions in the user information; and recommended products are determined based on the first product recommendation list and the second product recommendation list.

[0041] This application builds user portraits and comprehensively considers the user's regional information and historical financing intentions, which can more accurately understand the user's real needs, thereby providing more personalized product recommendations and significantly improving the accuracy of recommendations. In addition, this application reduces the tedious process of users manually filling out forms, improving the convenience and work efficiency of product recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, they can also obtain the drawings based on the drawings without paying any creative work.

[0044] Figure 1 A flowchart of the first embodiment of the method for recommending the product of this application;

[0045] Figure 2A flowchart of the second embodiment of the method for recommending the product of this application is provided;

[0046] Figure 3 A product radiation diagram provided for an implementation method of the product recommendation method of this application;

[0047] Figure 4 A schematic diagram of the architecture of a product recommendation system provided in accordance with an embodiment of the present application;

[0048] Figure 5 This is a schematic diagram of the module structure of the product recommendation device in the embodiment of the present application;

[0049] Figure 6 A schematic diagram of the device structure of the hardware operating environment involved in the product recommendation method in the embodiment of the present application.

[0050] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0051] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0052] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0053] The main solution of the embodiment of the present application is: obtain user information of the target user, and build a user portrait of the target user based on the user information; determine a first product recommendation list from a preset financial product database based on the user portrait; determine a second product recommendation list from the financial product database based on the enterprise geographical information and historical financing intentions in the user information; determine recommended products based on the first product recommendation list and the second product recommendation list.

[0054] In this embodiment, for ease of description, the following description is made with the product recommendation device as the execution subject.

[0055] With the development of Internet financial services, users are faced with more and more financial product choices. However, the existing recommendation system ignores the user's regional characteristics and specific financing needs, resulting in low accuracy of recommendation results and low user satisfaction. The existing technology mainly relies on users to manually fill out forms to obtain financing needs, and then the financial advisor recommends products based on the financing needs information and the relevant knowledge background of familiar banks. However, this product recommendation method cannot grasp comprehensive financing information or product information in real time, resulting in the accuracy of product recommendations being affected. In addition, this method is inefficient and cannot grasp comprehensive financing information or product information in real time. During the matchmaking process, it is difficult to achieve fast and efficient financing.

[0056] This application provides a solution that can more accurately understand the user's real needs by comprehensively considering the user's regional information and historical financing intentions, thereby providing more personalized product recommendations and significantly improving the accuracy of recommendations. In addition, this application reduces the tedious process of users manually filling out forms and improves the convenience of product recommendations.

[0057] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, a product recommendation device, etc. The following takes the product recommendation device as an example to illustrate this embodiment and the following embodiments.

[0058] Based on this, the present application embodiment provides a product recommendation method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the product recommendation method of this application.

[0059] In this embodiment, the product recommendation method includes steps S10 to S40:

[0060] Step S10, obtaining user information of a target user, and constructing a user profile of the target user based on the user information;

[0061] It should be noted that target users refer to the user groups for whom product recommendations are targeted. User portraits are virtual representations built based on user information, used to describe and summarize user characteristics, preferences, needs, and behavior patterns. User portraits usually include multiple dimensions, such as demographic characteristics, psychological characteristics, behavioral characteristics, etc.

[0062] Collect user information of target users. User information refers to various data related to users, including but not limited to basic personal information, financial product transaction records, online activity records, etc., which are used to analyze user preferences, needs and behavior patterns. Based on user information, the system constructs a user portrait of the target user through data mining, machine learning and other technical means. The user portrait is a collection of user characteristics that can comprehensively and accurately reflect the user's personality and needs.

[0063] In this embodiment, the process of constructing a user portrait is not limited here and can be set according to actual needs. For example, in a feasible implementation, the step S10: the step of constructing the user portrait of the target user based on the user information includes:

[0064] Step S101, extracting the target user's regional characteristics, risk preference, financing intention and credit status from the user information, and determining an attribute label based on the regional characteristics, risk preference, financing intention and credit status;

[0065] It should be noted that the key features related to the target users extracted from user information may include regional characteristics, risk preferences, financing intentions and credit status. Among them, regional characteristics reflect the economic, cultural and social environment of the user's region, and the regional characteristics can have an impact on the user's financial needs and preferences. Risk preferences describe the user's acceptance of financial risks. Financing intentions reflect the user's capital needs and usage plans, and credit status evaluates the user's repayment ability and credit record.

[0066] The system uses data mining and analysis technology to extract key features related to the target user from user information, and then constructs the dimensions of the user portrait based on the features, ensuring that each dimension can accurately reflect a certain aspect of the user's characteristics. By constructing user portraits, a richer information basis is provided for subsequent product recommendations, making the recommendation results more personalized and accurate.

[0067] Step S102, determining the label index value of each of the attribute labels, and determining the index mapping value of each of the label index values ​​according to a preset portrait rule based on each of the label index values;

[0068] It should be noted that attribute labels refer to different aspects or attributes used to describe and classify user characteristics, preferences, behaviors and needs. The label index value is the value of the attribute label. For each attribute label, the user's characteristics are described by a specific index value.

[0069] Determine the specific indicator value for each attribute label, and the indicator value will be used to quantify the user's performance in various dimensions. For example, regional characteristics can be measured by indicators such as regional economic development level and financial activity; risk preference can be measured by indicators such as the user's past investment behavior and risk tolerance; financing intention can be measured by indicators such as the user's capital demand scale and financing period; credit status can be measured by indicators such as the user's credit score and overdue rate. Then, the indicator value is converted into an indicator mapping value according to the preset portrait rules. It should be noted that the indicator mapping value is to convert the label indicator value into a form that is easier to analyze or compare according to the preset portrait rules. For example, for the indicator value on the credit status dimension (such as credit score), it can be converted into a credit grade (such as A, B, C grade) through the mapping rule.

[0070] Understandably, by providing financial institutions with a clearer basis for user risk assessment and product recommendation, it helps to reduce risks and increase returns.

[0071] Step S103: construct a user portrait based on the attribute label, the label index value and the index mapping value.

[0072] The user portrait is constructed by determining the attribute labels, label indicator values ​​and indicator mapping values. The user portrait is a complex data structure containing multiple dimensions and indicator values, and is an important basis for subsequent product recommendations. It should be noted that the specific process of constructing the user portrait is not limited here and can be set according to actual needs, and is not limited here.

[0073] By providing a comprehensive, accurate and easy-to-understand user portrait, it provides support for subsequent product recommendations and marketing strategy formulation, enhances financial institutions' user insight capabilities, and helps improve customer experience and satisfaction.

[0074] Step S20, determining a first product recommendation list from a preset financial product database based on the user portrait;

[0075] According to the characteristic information in the user portrait, combined with the information of each financial product in the preset financial product database, an algorithm model (such as collaborative filtering, content recommendation, etc.) is used to screen and sort the user's personal characteristics, preferences, historical behaviors and other information to recommend a series of products that may be of interest to the user to obtain a first product recommendation list. The specific process of generating the first product recommendation list is not restricted here and can be set according to actual needs.

[0076] It is understandable that by providing personalized product recommendations, the personalized needs of users are met and the accuracy and coverage of the product recommendation system are improved, because the system can recommend more products that meet their needs to users.

[0077] Step S30, determining a second product recommendation list from the financial product database based on the enterprise region information and historical financing intention in the user information;

[0078] Enterprise geographical information refers to data related to the geographical location of the enterprise, including the city, region, country, etc. where the enterprise is located. Enterprise geographical information may affect the enterprise's marketing strategy and product recommendations. Historical financing intention refers to the user or enterprise's interest and preference in financing activities in the past. It may include information such as the type, amount, and term of financing that the user or enterprise has applied for, as well as the user or enterprise's acceptance and satisfaction with different financing products. The second product recommendation list is a set of product recommendations generated based on specific factors such as enterprise geographical information and historical financing intentions. Compared with the first product recommendation list, the second product recommendation list may be more focused on meeting the user's specific needs and market environment.

[0079] Based on the user's corporate geographical information and historical financing intentions, the user's position and needs in the current financial market are determined, and then, combined with factors such as the financial market environment and product characteristics, products that are more in line with the user's regional characteristics and financing needs are recommended to obtain a second recommendation list. It can be understood that by considering regional characteristics and financing needs, the accuracy of product recommendations and user satisfaction are improved.

[0080] Step S40: determining recommended products based on the first product recommendation list and the second product recommendation list.

[0081] By comprehensively considering the information of the first product recommendation list and the second product recommendation list, a comprehensive recommended product list is finally determined through certain weight allocation and algorithm optimization. It can be understood that the final recommended product list not only takes into account the personalized needs of users, but also takes into account regional characteristics and financing needs, provides a comprehensive and comprehensive recommendation result, improves user satisfaction and trust, and enhances the robustness and stability of the recommendation system.

[0082] In this embodiment, by comprehensively considering the user's regional information and historical financing intentions, the user's real needs can be understood more accurately, thereby providing more personalized product recommendations and significantly improving the accuracy of recommendations. In addition, this embodiment reduces the tedious process of users manually filling out forms and improves the convenience of product recommendations.

[0083] In one feasible implementation, the step S20: determining a first product recommendation list from a preset financial product database based on the user portrait, comprises:

[0084] Step S201, for any target product among the financial products in the preset financial product database, determine the product attribute tag value of the target product under each attribute tag, wherein each attribute tag is a tag for constructing the user portrait;

[0085] It should be noted that attribute labels are labels used to describe user portraits. They are used to extract and quantify the key features of users. Attribute labels are the same as the labels used when constructing user portraits.

[0086] Traverse each financial product in the preset financial product database, and for any target product among each financial product, extract the product attribute label value of the target product under the attribute label, that is, determine the value of each attribute label used by the financial product when constructing the user portrait, which provides basic data for subsequent similarity calculations, allowing the system to compare the degree of match between user attributes and product attributes.

[0087] Step S202, determining the user attribute tag value of the target user under each attribute tag based on the user portrait;

[0088] Determine the value of each attribute label corresponding to the target user according to the user portrait, that is, the user attribute label value. Specifically, the label data associated with the attribute label can be extracted from the user information according to the construction rules of the user portrait; and calculate the value of each attribute label corresponding to the target user based on the label data.

[0089] Step S203, traversing each of the attribute tags, calculating the attribute similarity between the user attribute tag value and the product attribute tag value, and calculating the sum of the attribute similarities to obtain a similarity sum value;

[0090] Traverse all attribute tags, calculate the similarity between the user attribute tag value and the product attribute tag value under each attribute tag, hereinafter referred to as attribute similarity for distinction, attribute similarity is a value between 0 and 1, indicating the degree of match between the two, sum up the similarities under all attribute tags, and obtain a similarity sum value. It should be noted that the similarity can be calculated using a variety of methods, such as cosine similarity, Jaccard similarity, etc. The specific choice depends on the data type of the attribute value and tag and the business requirements, and is not limited here.

[0091] Step S204: determining first candidate products whose similarities and values ​​satisfy a first preset recommendation condition from among the financial products, and constructing a first product recommendation list based on the first candidate products.

[0092] The first preset recommendation condition is set in advance. It should be noted that the first preset recommendation condition can be set according to actual needs and is not limited here. For example, the second preset condition can be that the similarity and value are greater than the preset threshold. Products that meet the first preset recommendation condition are screened out from all first candidate products, that is, the first candidate products, and then a first product recommendation list is formed based on the first candidate products. The first candidate products in the first product recommendation list can be sorted according to similarity and value, and the products with the highest similarity and value, that is, the products that best match the user portrait, are placed in front.

[0093] It is understandable that by setting the first preset recommendation condition, products that are more compatible with the user profile are screened out, thereby improving the accuracy and personalization of product recommendations. By calculating the similarity, the degree of match between the target product and the user profile can be quantified, providing a basis for subsequent product recommendations, thereby improving the accuracy and personalization of product recommendations, because recommendations can be made based on specific tags in the user profile.

[0094] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. Figure 2The step S30: determining a second product recommendation list from the financial product database based on the enterprise region information and historical financing intention in the user information, comprises:

[0095] Step S301, extracting the user modeling tag value of the target user under the preset modeling tag from the enterprise region information and the historical financing intention;

[0096] Preset modeling tags are a series of tags that are pre-set when modeling users or products to describe user characteristics or product features. They are used to extract and quantify the key features of users or products for subsequent analysis and recommendation. Preset modeling tags can be determined based on business logic, data analysis, and expert experience, and are not limited here. User modeling tag values ​​refer to specific numerical or categorical values ​​extracted and quantified from user information based on preset modeling tags. They reflect the characteristics or preferences of users under a specific tag and are used to build user portraits and make product recommendations.

[0097] The user modeling label values ​​of the target users under the preset modeling labels are extracted from the enterprise's geographical information and historical financing intentions. It can be understood that the extracted user modeling label values ​​provide basic data for the subsequent construction of the product radiation map, and improve the accuracy and personalization of recommendations by quantifying user characteristics.

[0098] Step S302, determining a product modeling tag value under the preset modeling tag for each financial product in the financial product database;

[0099] It should be noted that the product modeling label value refers to the specific numerical value or classification value extracted and quantified from the financial product information based on the preset modeling label, which reflects the characteristics or attributes of the financial product under a specific label and is used for comparison and matching with the user modeling label value. By assigning label values ​​to financial products, their characteristics can be quantified, providing a basis for subsequent product score calculation and recommendation.

[0100] Determine the label value under the preset modeling label for each financial product in the financial product database. The preset modeling label is pre-defined based on user portraits and product characteristics and is used to describe the key characteristics of users and products. The specific process of determining the product modeling label value can be: traverse each financial product in the financial product database, and for any target product in each financial product, calculate or determine its value under each label according to its attributes (such as product type, risk level, rate of return, etc.) and the definition of the preset modeling label, that is, the product modeling label value.

[0101] Step S303: constructing a product radiation map centered on the target user based on the user modeling tag value and the product modeling tag value, wherein the central node of the product radiation map is the target user, and the radiation nodes of the product radiation map are the financial products in the financial product database;

[0102] It should be noted that the product radiation map is a visualization tool used to show the relationship between users and financial products, where the target user is the central node and the financial product is the radiation node. The connection line between the central node and the radiation node represents the degree of association between the user and the product of the radiation node. Through the product radiation map, you can intuitively see which financial products the user is closer to or matches, thereby assisting in product recommendations. Based on the extracted user modeling label values ​​and product modeling label values, a product radiation map centered on the target user is constructed. The specific construction process is not limited here and can be set according to actual needs.

[0103] Step S304: determining a second product recommendation list from the financial product database based on the product radiation map.

[0104] Based on the constructed product radiation map, a second product recommendation list is determined from the financial product database, where the list includes financial products that best match the needs of the target users.

[0105] It can be understood that by extracting user modeling label values, constructing a product radiation map and determining a second product recommendation list, personalized recommendations for financial products based on user information are achieved, the accuracy and personalization of recommendations are improved, and the understanding and analysis capabilities of business personnel are improved through visualization tools.

[0106] In a feasible implementation manner, the step S301: extracting the user modeling tag value of the target user under the preset modeling tag from the enterprise region information and the historical financing intention includes:

[0107] Step S3011, determining the user registration location, user office location and user actual factory location of the target user from the enterprise regional information, and determining the user historical financing situation of the target user from the historical financing intention, wherein the user historical financing situation is used to indicate whether the financing of the target user is successful;

[0108] Extract enterprise geographical information from user information, identify the specific address of the enterprise's registered place (i.e., registered place), the place of daily operation and management of the enterprise (i.e., office location), and the specific location of production or manufacturing facilities (i.e., actual factory site) in the user information. Identify and record the enterprise's past historical financing situation, which is used to characterize whether the target user's financing is successful. Specifically, the historical financing situation can be obtained by analyzing the enterprise's financing records, loan applications, investment agreements and other documents.

[0109] It is understandable that by extracting key information such as the company's geographical information and historical financing intentions, we can have a more comprehensive understanding of the company's operating environment and financing history, providing an important basis for subsequent product recommendations and risk assessments.

[0110] Step S3012, using the user's registered place as the user modeling tag value corresponding to the registered place tag in the preset modeling tag, using the user's office location as the user modeling tag value corresponding to the office location tag in the preset modeling tag, using the user's actual factory site as the user modeling tag value corresponding to the factory location tag in the preset modeling tag, and using the user's historical financing situation as the user modeling tag value corresponding to the financing tag in the preset modeling tag.

[0111] The extracted registered place, office address, actual factory site and historical financing situation are converted into quantifiable label values. For example, the registered place, office address and actual factory site can be converted into geographic location coordinates or region codes; the historical financing situation can be converted into quantitative indicators such as the number of successful financings, the number of failures or the financing success rate; the label values ​​are integrated together to form a complete modeling label set for subsequent recommendation and risk assessment models.

[0112] It is understandable that by converting key information into modeling label values, information can be used more efficiently to provide accurate data support for subsequent product recommendations and risk assessments. The label value conversion process helps to eliminate noise and redundant information in the original data, improves the generalization ability of the model, and enables the model to focus more on key features.

[0113] In a feasible implementation manner, the step S303: the step of constructing a product radiation map centered on the target user based on the user modeling tag value and the product modeling tag value, comprises:

[0114] Step S3031, traversing each financial product in the financial product database, and determining a product score of the financial product based on the product modeling tag value and the user modeling tag value;

[0115] The product score of the target product is calculated based on the product modeling tag value and the user modeling tag value. The product score reflects the degree of match between the target product and the target user. The specific process of determining the product score can be: for each target product, obtain its product modeling tag value under all preset modeling tags; calculate the product score of the target product according to a certain algorithm (such as weighted sum, cosine similarity, etc.), which represents the similarity or match between the target product and the target user.

[0116] It can be understood that by calculating the product score, the degree of match between the target product and the target user can be quantified, providing a basis for subsequent product recommendations.

[0117] Step S3032, constructing a product radiation diagram with the target user as the central node, each of the financial products as the radiation nodes, and the product score corresponding to each of the financial products as the length of the node connection line.

[0118] Exemplarily, the process of constructing a product radiation map may be: taking the target user as the central node of the product radiation map; for each financial product in the financial product database, calculating the modeling difference between the product modeling label value of each financial product and the user modeling label value of the target user; determining the product score of the financial product based on the calculated modeling difference, wherein the modeling difference is inversely proportional to the product score; determining the length of the node connection line between the central node (target user) and the radiation node (financial product) based on the product score, it can be understood that the product score is inversely proportional to the length of the connection line, and the shorter the node connection line, the more the financial product matches the needs of the target user; using graphics software or tools, draw a product radiation map according to the length and direction of the connection line.

[0119] In one feasible implementation, the preset modeling tag includes a plurality of modeling sub-tags, and the step S3032: the step of determining the product score of the financial product based on the product modeling tag value and the user modeling tag value includes:

[0120] Step S30321, determining the sub-tag difference corresponding to each of the modeling sub-tags, wherein the sub-tag difference is the difference between the product modeling tag value and the user modeling tag value under the modeling sub-tag;

[0121] Calculate the difference between the product modeling label value and the user modeling label value under each modeling sub-label, that is, the sub-label difference. The sub-label difference reflects the difference in the degree of matching between the product and the user on the sub-label. For each modeling sub-label, obtain the corresponding value from the product modeling label value set and the user modeling label value set, and calculate the difference between the two values, that is, the sub-label difference. In the specific calculation process, if the sub-label is numeric, the subtraction operation is performed directly; if the sub-label is categorical, it may be necessary to convert it to numeric type (such as using one-hot encoding or label encoding) before calculation.

[0122] It can be understood that by calculating the sub-tag differences, the difference in the degree of matching between the product and the user on each sub-tag can be quantified, providing a basis for subsequent weighted processing.

[0123] Step S30322, obtaining a preset first weighting coefficient group, wherein the first weighting coefficient group includes weighting coefficients corresponding to each of the modeling sub-tags;

[0124] It should be noted that the weighting coefficient of each modeling sub-label is predetermined, specifically based on experience, data analysis or expert evaluation, and is not limited here; the weighting coefficient is stored as an array, namely the first weighting coefficient group.

[0125] It is understandable that by presetting the weighting coefficient, the influence of each modeling sub-label in calculating the product score can be flexibly adjusted, thereby ensuring the accuracy and rationality of the recommendation results.

[0126] Step S30323, weighting the sub-label differences corresponding to each other by using the weighting coefficients to obtain a weighted value, and taking the sum of the weighted values ​​as the product score.

[0127] For each modeling sub-label, multiply its corresponding sub-label difference with the corresponding weighting coefficient in the first weighting coefficient group to obtain a weighted processing value; add the weighted processing values ​​of all modeling sub-labels to obtain the product score of the target product.

[0128] It is understandable that through weighted processing, the impact of each modeling sub-label on the product score can be comprehensively considered to obtain a more comprehensive and accurate product score, so that the product score can reflect the overall matching degree between the product and the user, and can provide more personalized product recommendations for target users, thereby improving the accuracy of recommendations and user satisfaction.

[0129] It is understandable that the process of determining the second product recommendation list from the financial product database based on the product radiation diagram may be: determining the second candidate product whose node connection line length satisfies the third preset recommendation condition from the financial product database, and constructing the second product recommendation list based on the second candidate product, wherein the connection line length corresponding to the second candidate product is proportional to the ranking of the second candidate product in the second product recommendation list. Exemplarily, the process of constructing the second product recommendation list may be: traversing the product scores of all target products, screening out the second candidate products that meet the third preset recommendation condition; sorting the second candidate products according to the size of the connection line length, and constructing the second product recommendation list, wherein the product with a lower connection line length is ranked higher in the list.

[0130] It should be noted that when determining the second product recommendation list, the number of products in the list can be pre-set. There may be second candidate products in the second product recommendation list that exceed the preset number. In order to screen out the second candidate products that meet the preset number, in a feasible implementation, if the number of second candidate products exceeds the preset number, the preset priorities corresponding to each modeling sub-tag are determined, and the modeling sub-tag with the highest priority in the modeling sub-tag is used as the first filtering tag; the first difference of each second candidate product is determined, wherein the first difference is the difference between the modeling tag value under the first filtering tag product and the user modeling tag value; a preset number of third candidate products are determined from each second candidate product, wherein the third candidate product is the second candidate product with the smallest first difference; and a second product recommendation list is constructed based on the third candidate product.

[0131] It should be noted that there may be second candidate products with the same connecting line length in the second product recommendation list. The second candidate products with the same connecting line length have the same ranking in the second product recommendation list. In order to sort the second candidate products with the same connecting line length, in a feasible implementation manner, the second candidate products are sorted according to the connecting line length to obtain a second recommended product list; it is detected whether there are second candidate products with the same ranking in the second recommended product list, wherein the product with a shorter connecting line length is ranked higher in the list; if there are second candidate products with the same ranking, the preset priority corresponding to each modeling sub-tag is determined, and the modeling sub-tag with the highest priority among the modeling sub-tags is used as the second screening tag; the same-order products (that is, the second candidate products with the same ranking) are determined from each second candidate product, and the second A second difference value corresponding to each same-sequence product under the filtering label, wherein the second difference is the difference between the product modeling label value and the user modeling label value of the same-sequence product; sorting each same-sequence product based on the second difference value, wherein the smaller the first difference value, the higher the ranking of the same-sequence product; updating the second product recommendation list according to the ranking of each same-sequence product, and detecting whether there is a second candidate product with the same ranking in the updated second recommended product list; if there is a second candidate product with the same ranking, updating the modeling sub-label with a lower priority than the second filtering label and adjacent to the second filtering label in each modeling sub-label to the second filtering label, and returning to the step of determining the same-sequence product from each second candidate product, and determining the second difference value corresponding to each same-sequence product under the second filtering label, until there is no second candidate product with the same ranking.

[0132] It is understandable that by constructing the second product recommendation list, personalized product recommendations can be provided to target users, improving the accuracy of recommendations and user satisfaction. At the same time, sorting according to the length of the connection line can ensure that the products that best meet the needs of users are recommended first.

[0133] For example, assuming that financial products are recommended to users of Company A, the specific process of determining the second product recommendation list may be:

[0134] Data collection: First, collect Company A’s corporate geographical information and historical financing intentions. For example, Company A is located in a first-tier city, has applied for short-term loans in the past, and has repaid them on time every time.

[0135] Get predefined preset modeling tags. Preset modeling tags can be set according to business needs and data availability, without restriction, such as credit rating, income level, financing preference, geographical risk, etc.

[0136] Extract label values: Based on the data of Company A, extract the user modeling label values ​​under these modeling labels, for example, Company A's credit rating is "excellent", income level is "high", financing preference is "short-term loan", and regional risk is "low". For each financial product in the financial product database, extract its value under the preset modeling label, for example, a short-term loan product requires a credit rating of "good", an expected rate of return of 4%, and an investment period of 3 months.

[0137] Construct a product radiation map. First, take Company A as the central node of the product radiation map; calculate the difference between the product modeling label value of each financial product and the user modeling label value of Company A. For example, for the above-mentioned short-term loan product, the credit rating difference is small (because Company A's credit rating is "excellent", which is higher than the "good" requirement of the product), but the difference in its label may be larger or smaller, depending on the specific product; based on the calculated difference, the company determines the length of the connection line between the central node (Company A) and the radiation node (financial product), where the smaller the difference, the shorter the connection line, indicating that the financial product is more compatible with Company A's needs; using graphics software or tools, draw a product radiation map with Company A as the central node based on the length and direction of the connection line.

[0138] Determine the second recommended product list. According to the third preset recommendation condition and the product radiation map, select the financial products that meet the condition as the second product recommendation list. For example, select the top 5 financial products with the shortest connection line length; sort the selected products (such as from short to long connection line length). For specific product radiation maps, please refer to Figure 3 , Figure 3 Five financial products are shown in the figure. Financial products 1, 2, 3 and Company A are all in region A, financial product 4 is in region B, financial product 5 is in region C, and financial product 6 is in region C. When determining the second recommended product list based on the product radiation map, if the third preset recommendation condition requires screening three financial products with the shortest connection lines, then the second recommended product list includes financial products 1, 3, and 5. Assuming that the third preset recommendation condition stipulates that the length of the connection line is used as the first priority and the region where the product is located is used as the second priority for product sorting, since the connection lines corresponding to 1 and 5 are shorter than 3, product 1 and Company A are in the same region, so the second product recommendation list is 1, 5, and 3.

[0139] In one feasible implementation, the step S40: the step of determining recommended products based on the first product recommendation list and the second product recommendation list includes:

[0140] Step S401, determining the regional attribute value of each product in the first product recommendation list and the second product recommendation list;

[0141] Determine the geographical attribute value of each product in the first product recommendation list and the second product recommendation list. The geographical attribute value may include the product's place of registration, place of sale, service scope, etc. The attribute reflects the degree of association between the product and the user's region. Specifically, extract geographical information from the description, metadata or related database of each product; convert the extracted geographical information into quantifiable geographical attribute values. This may involve converting text descriptions into geographic coordinates, area codes or classification labels, or assigning different weight values ​​to different regions according to preset rules. It is understandable that by determining the geographical attribute value, the degree of match between the product and the user's region can be considered, thereby improving the accuracy and relevance of the recommendation, providing basic data for subsequent weighted processing, and making the recommendation process more scientific and objective.

[0142] Step S402, obtaining a preset second weighting coefficient group, and performing weighted processing on the regional attribute values ​​of each of the listed products by using the second weighting coefficient group to obtain a recommendation score for each of the listed products;

[0143] The regional attribute value of each product is weighted using the preset second weighting coefficient group to calculate the recommendation score of each product. The second weighting coefficient group reflects the importance of the regional attribute in determining the recommendation score. It should be noted that the second weighting coefficient group for each regional attribute can be determined based on business needs, user preferences, data analysis results, etc. The second weighting coefficient group can be fixed or dynamically adjusted based on user feedback or market changes, and there is no limitation here.

[0144] Multiply the regional attribute value of each product by its corresponding second weighting coefficient group to obtain the weighted regional attribute value. Then, aggregate or weighted average all weighted regional attribute values ​​to obtain the recommendation score of the product. Perform the above weighting process on each product in the first product recommendation list and the second product recommendation list to obtain the recommendation score of the respective list products. It can be understood that through weighting process, the importance of regional attributes in the recommendation process can be more accurately reflected, the personalization and accuracy of the recommendation can be improved, and the second weighting coefficient group can be adjusted according to actual needs, thereby improving the flexibility and adaptability of the model.

[0145] Step S403: taking the listed products whose recommendation scores satisfy the second preset recommendation condition among the listed products as recommended products.

[0146] The products in the two weighted lists are sorted according to the recommendation scores, and then the products that meet the second preset recommendation condition are selected as the final recommended products. For example, the top N products with the highest recommendation scores may be selected, or products with recommendation scores exceeding a certain threshold may be selected. The second preset recommendation condition may include a recommendation score threshold, a sorting requirement, etc., which are not limited here and can be set according to business needs.

[0147] It is understandable that by setting recommendation conditions and filtering recommended products, users can be provided with more accurate recommendation results that meet their needs, and suitable products can be quickly filtered out from a large number of products, thereby improving the efficiency and accuracy of product recommendations.

[0148] For example, in order to help understand the implementation process of the product recommendation method obtained by combining this embodiment with the above-mentioned embodiment 1, this embodiment provides a comprehensive intelligent recommendation system, which can automatically analyze the user's region, historical financing intention, and recommend the most suitable financial products in combination with market trends and product characteristics. Please refer to Figure 4 , Figure 4 The product recommendation system architecture diagram is provided, which specifically includes: data collection module, user portrait construction module, recommendation algorithm module, financial database, regional characteristics analysis module, historical financing intention analysis module and recommendation result optimization module. Figure 4 As shown, the process of product recommendation based on the product recommendation system can be: the data collection module is used to collect data, the collected data is input into the user portrait construction module to construct the user portrait, and then the user portrait is input into the recommendation algorithm module, the recommendation algorithm module determines the first recommended product list from the financial product database based on the user portrait; based on the financial product database, the regional characteristics analysis module and the historical financing intention analysis module obtain the second recommended product list; the first recommended product list and the second recommended product list are input into the recommendation result optimization module to obtain the recommendation result. The following is a detailed description of each module:

[0149] 1. Data collection module: collects basic information of users, including but not limited to user geographic location, historical financing records, investment preferences, etc. During the data collection process, the geographic location, registered address, office address, branch address, etc. of corporate users registered in the system are collected, and at the same time, each financing record of the user's transaction is collected. The collected data is cleaned and user data is integrated to ensure data quality and consistency (that is, to obtain user information of target users).

[0150] 2. User portrait construction module: Based on the collected data, construct user portraits, including the user's regional characteristics, financing intentions, risk preferences, etc. (that is, construct the user portrait of the target user based on the user information). Use statistical analysis and machine learning techniques to extract features from user data and construct user portraits. The user portrait construction process can be:

[0151] 2.1. Determine the four major elements of user portrait - GRFC. This embodiment constructs the four major elements of user portrait, namely, geographical region features, risk appetite, financing intention, and credit status, referred to as GRFC. Among them, geographical characteristics refer to information such as the registered address of the financing enterprise, the office address of the enterprise, the actual factory address of the manufacturing enterprise, the address of the actual business of the enterprise, and the address of the branch (subsidiary, office) of the enterprise; risk preference refers to information data such as enterprise classification, market environment, financial status, and enterprise development strategy; financing intention refers to the amount of financing of the enterprise, the category of financing objects (referring to the five major banks, joint-stock banks, city commercial banks, village banks, private funds, financing guarantees, insurance companies and other financial institutions), the proportion of financing amount to net assets, the proportion of financing finance to monthly operating income, the industry in which financing is invested, and the financing repayment plan; credit status refers to information such as the lawsuit information, enterprise punishment information, enterprise execution information, enterprise mortgage information, various rating data of the enterprise, punishment information of the directors, supervisors and senior managers of the enterprise, execution information of the directors, supervisors and senior managers of the enterprise, punishment information of the shareholders associated with the enterprise, lawsuit information of the shareholders associated with the enterprise, and other credit data of the enterprise, the related directors, supervisors and senior managers, and the shareholders.

[0152] 2.2. Construct user portraits based on determined factors. The user portrait model of this embodiment constructs the underlying labels of user portraits by establishing a first-level dimension, a second-level indicator, and a third-level standard. Among them, the first-level dimension not only includes the GRFC element, but also includes non-important factors such as the change dimension and the public opinion dimension. The second-level indicator is to create each portrait indicator under the dimension, such as whether the registered address of the enterprise is within the jurisdiction of the financial institution, whether the historical financing record of the enterprise is proportional to the positive growth of the operating income, such as the number of lawsuits and the number of failures of the enterprise in a period of time. The third-level standard refers to the establishment of a normal distribution diagram or a scatter distribution diagram or a list through a large amount of financing data when judging whether the historical financing record of the enterprise is proportional to the positive growth indicator of the operating income, and the establishment of a segmented standard for proportional positive growth, such as 5%, 15%, 35%, etc. The user portrait of this embodiment is constructed based on the underlying label of the user portrait, and then by constructing the enterprise large diamond portrait model, the enterprise diamond protruding portrait of the enterprise is constructed through the matching difference of each underlying label. Relying on this portrait, the length of the matching difference of each label can be found, and it can be quickly recommended in the subsequent docking of financial products.

[0153] 2.3. Determine the user portrait update frequency, and then update the user portrait according to the preset update frequency. The user portrait update frequency in this embodiment can be a fixed update frequency and an irregular update frequency. The fixed update frequency is initially set to 1 week by default, but the fixed update frequency may be adjusted according to the parameters. For example, the minimum adjustment value can be set to 5 days, and the minimum increase or decrease adjustment range is 1 day. The irregular update frequency means that for a certain user portrait, if there is a change in data during a business call, the latest user portrait update is performed.

[0154] 3. Financial product database: A database containing information on various financial products is established in advance, and each product is marked with attributes such as applicable region, risk level, and expected interest rate.

[0155] 4. Recommendation algorithm module: Based on the user portrait, potentially matching financial products are screened out from the financial product database, and a recommendation algorithm is used (exemplarily, a collaborative filtering algorithm and a hybrid recommendation algorithm can be used) to calculate the matching degree between the user and the financial product according to the user portrait and the financial product database, and generate a recommendation list. In this embodiment, the optimal solution of the collaborative filtering algorithm and the hybrid recommendation algorithm is used to make recommendations between financial products and financing companies. In the financial product data, each product has its own attributes such as usage region, risk level, expected interest rate, loan amount range, loan term, application scenario, etc. The value of each attribute is entered into the system when the financial product is put into storage (that is, based on the user portrait, the first product recommendation list is determined from the preset financial product database; based on the enterprise geographical information and historical financing intentions in the user information, the second product recommendation list is determined from the financial product database).

[0156] The steps of using collaborative filtering algorithm in this embodiment are as follows: Step 1: Use the cosine similarity algorithm to calculate the similarity values ​​of the financial products and the labels corresponding to the user portraits in turn through the similarity between the financial products and the labels corresponding to the user portraits, and the range of each value is between [-1, 1]. Sum the similarity values ​​of each label. Step 2: Calculate the similarity values ​​of all products and the user portraits of the enterprise. Step 3: All similarity values ​​calculated in the second step are stored in a linear table. If there is a data change for a product later, the system's automatic calculation function is triggered to calculate and update the similarity value between the financial product and the user portrait (that is, for any target product among the financial products in the preset financial product database, determine the product attribute label value of the target product under each attribute label, wherein each attribute label is a label for constructing the user portrait; determine the user attribute label value of the target user under each attribute label based on the user portrait; traverse each attribute label, calculate the attribute similarity between the user attribute label value and the product attribute label value, and calculate the sum of each attribute similarity to obtain a similarity sum value; determine the first candidate product whose similarity sum value meets the first preset recommendation condition from each financial product, and construct a first product recommendation list based on the first candidate product). In this embodiment, a hybrid recommendation algorithm is also used, which mainly models the enterprise region and the enterprise's historical financing intention. A circular model is established with the enterprise's registered place, office location, actual office factory site, and historical financing success and failure as the center point, and each financial product is placed in the circular model in turn according to its geographical distance. Through the weight and distribution distance of each regional characteristic, as well as the authority of the enterprise's historical financing intention analysis, a weighted sum parallel hybrid recommendation is performed. By assigning weights to regions and historical financing based on the top 20 optimal solutions calculated by the collaborative filtering algorithm and the hybrid recommendation algorithm, a relevant recommendation list can be obtained. Then, according to the number of financial products recommended to the enterprise by the system configuration, the corresponding financial products are selected and recommended to the enterprise, and recommended labels are given, such as high regional similarity, historical financing success rate, etc. (that is, the user modeling label value of the target user under the preset modeling label is extracted from the enterprise's regional information and the historical financing intention; the product modeling label value of each financial product in the financial product database under the preset modeling label is determined; a product radiation map centered on the target user is constructed based on the user modeling label value and the product modeling label value, wherein the central node of the product radiation map is the target user, and the radiation nodes of the product radiation map are the financial products in the financial product database; a second product recommendation list is determined from the financial product database based on the product radiation map).

[0157] 5. Regional characteristic analysis module: Analyze the financing needs and preferences of users in different regions, and provide region-related weights for the recommendation algorithm. In this embodiment, the regional characteristics and historical financing intention analysis and the result data are mainly considered for analysis. We have established a scatter plot of the financing needs and preferences of users in all regions, and set weights for different regions according to the distribution size of the financing needs and preferences of users in each region (that is, determine the regional attribute values ​​of each list product in the first product recommendation list and the second product recommendation list; obtain the preset second weighting coefficient group, and use the second weighting coefficient group to weight the regional attribute values ​​of each list product to obtain the recommendation score of each list product; and use the list products whose recommendation scores meet the second preset recommendation conditions as recommended products).

[0158] 6. Historical financing intention analysis module: Analyze the user's historical financing intention and identify the user's demand pattern and preference changes. The historical financing intention analysis module in this embodiment mainly analyzes the amount of corporate financing, usage scenarios and other factors, and directly assigns weights to the relevant tags in the corporate portrait according to the number of matches through the matching algorithm, so that when the financial product uses the collaborative filtering algorithm, the calculated cosine similarity value is closer to 1.

[0159] 7. Recommendation result optimization module: Based on user feedback, continuously optimize recommendation results to improve recommendation accuracy and user satisfaction.

[0160] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the product recommendation method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0161] This application also provides a product recommendation device, please refer to Figure 5 , the product recommendation device comprises:

[0162] An acquisition module 10 is used to acquire user information of a target user and construct a user profile of the target user based on the user information;

[0163] A first recommendation module 20, configured to determine a first product recommendation list from a preset financial product database based on the user portrait;

[0164] A second recommendation module 30, configured to determine a second product recommendation list from the financial product database based on the enterprise region information and historical financing intention in the user information;

[0165] The determination module 40 is used to determine recommended products based on the first product recommendation list and the second product recommendation list.

[0166] Optionally, the second recommendation module 30 is used to:

[0167] Extracting a user modeling tag value of the target user under a preset modeling tag from the enterprise region information and the historical financing intention;

[0168] Based on the user modeling label value, a product radiation map centered on the target user is constructed, wherein the central node of the product radiation map is the target user, the radiation nodes of the product radiation map are the financial products in the financial product database, and the length of the connection line between the central node and the radiation node is used to represent the difference between the product modeling label value and the user modeling label value, and the product modeling label value is the value corresponding to the financial product under the preset modeling label;

[0169] A second product recommendation list is determined from the financial product database based on the product radiation map.

[0170] Optionally, the second recommendation module 30 is used to:

[0171] Determine the user registration location, user office location and user actual factory address of the target user from the enterprise geographical information, and determine the user historical financing situation of the target user from the historical financing intention, wherein the user historical financing situation is used to indicate whether the financing of the target user is successful;

[0172] The user's registered place is used as the user modeling tag value corresponding to the registered place tag in the preset modeling tag, the user's office location is used as the user modeling tag value corresponding to the office location tag in the preset modeling tag, the user's actual factory site is used as the user modeling tag value corresponding to the factory site tag in the preset modeling tag, and the user's historical financing situation is used as the user modeling tag value corresponding to the financing tag in the preset modeling tag.

[0173] Optionally, the second recommendation module 30 is used to:

[0174] Traversing each financial product in the financial product database, and determining a product score of the financial product based on the product modeling tag value and the user modeling tag value;

[0175] A product radiation diagram is constructed with the target user as the central node, each of the financial products as the radiation nodes, and the product score corresponding to each of the financial products as the length of the node connection line.

[0176] Optionally, the preset modeling tag includes a plurality of modeling sub-tags, and the second recommendation module 30 is used to:

[0177] Determine a sub-tag difference value corresponding to each of the modeling sub-tags, wherein the sub-tag difference value is a difference between a product modeling tag value and a user modeling tag value under the modeling sub-tag;

[0178] Obtaining a preset first weighting coefficient group, wherein the first weighting coefficient group includes weighting coefficients corresponding to each of the modeling sub-tags;

[0179] The sub-label differences corresponding to each other are weighted by each weighting coefficient to obtain a weighted value, and the sum of each weighted value is used as the product score.

[0180] Optionally, the first recommendation module 20 is used to:

[0181] For any target product among the financial products in the preset financial product database, determine the product attribute tag value of the target product under each attribute tag, wherein each attribute tag is a tag for constructing the user portrait;

[0182] Determine the user attribute tag value of the target user under each attribute tag based on the user portrait;

[0183] Traversing each of the attribute tags, calculating the attribute similarity between the user attribute tag value and the product attribute tag value, and calculating the sum of the attribute similarities to obtain a similarity sum value;

[0184] Determine first candidate products whose similarity and value meet a first preset recommendation condition from each of the financial products, and construct a first product recommendation list based on the first candidate products.

[0185] Optionally, the determining module 40 is used to:

[0186] Determine the regional attribute value of each list product in the first product recommendation list and the second product recommendation list;

[0187] Obtaining a preset second weighting coefficient group, and performing weighted processing on the regional attribute values ​​of each of the listed products using the second weighting coefficient group to obtain a recommendation score for each of the listed products;

[0188] The listed products whose recommendation scores satisfy the second preset recommendation condition among the listed products are taken as recommended products.

[0189] Optionally, the acquisition module 10 is used to:

[0190] Extracting the target user's regional characteristics, risk preference, financing intention, and credit status from the user information, and determining an attribute label based on the regional characteristics, risk preference, financing intention, and credit status;

[0191] Determine the label index value of each of the attribute labels, and determine the index mapping value of each of the label index values ​​according to the preset portrait rules based on the label index values;

[0192] A user portrait is constructed based on the attribute label, the label indicator value and the indicator mapping value.

[0193] The product recommendation device provided by this application adopts the product recommendation method in the above embodiment, which can solve the technical problem that the accuracy of product recommendation is affected. Compared with the prior art, the beneficial effects of the product recommendation device provided by this application are the same as the beneficial effects of the product recommendation method provided by the above embodiment, and the other technical features in the product recommendation device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0194] The present application provides a product recommendation device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the product recommendation method in the above-mentioned embodiment one.

[0195] Reference below Figure 6 , which shows a schematic diagram of the structure of a product recommendation device suitable for implementing an embodiment of the present application. The product recommendation device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 6 The product recommendation device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0196] like Figure 6As shown, the product recommendation device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the product recommendation device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the product recommendation device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a product recommendation device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or provided instead.

[0197] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0198] The product recommendation device provided by this application adopts the product recommendation method in the above embodiment, which can solve the technical problem that the accuracy of product recommendation is affected. Compared with the prior art, the beneficial effects of the product recommendation device provided by this application are the same as the beneficial effects of the product recommendation method provided by the above embodiment, and the other technical features in the product recommendation device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0199] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0200] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0201] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, wherein the computer-readable program instructions are used to execute the product recommendation method in the above-mentioned embodiment.

[0202] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0203] The computer-readable storage medium may be included in the product recommendation device; or may exist independently without being assembled into the product recommendation device.

[0204] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the product recommendation device, the product recommendation device enables the product recommendation device to: obtain user information of the target user, and build a user portrait of the target user based on the user information; determine a first product recommendation list from a preset financial product database based on the user portrait; determine a second product recommendation list from the financial product database based on the enterprise geographical information and historical financing intentions in the user information; and determine recommended products based on the first product recommendation list and the second product recommendation list.

[0205] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0206] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0207] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0208] The readable storage medium provided in this application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned product recommendation method, and can solve the technical problem that the accuracy of product recommendation is affected. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the product recommendation method provided in the above-mentioned embodiment, and will not be repeated here.

[0209] The present application also provides a computer program product, including a computer program, which implements the steps of the product recommendation method as described above when executed by a processor.

[0210] The computer program product provided by this application can solve the technical problem that the accuracy of product recommendation is affected. Compared with the prior art, the beneficial effects of the computer program product provided by this application are the same as the beneficial effects of the product recommendation method provided by the above embodiment, which will not be repeated here.

[0211] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in the relevant technical fields, are included in the patent protection scope of the present application.

Claims

1. A product recommendation method, characterized in that: The product recommendation method comprises: Acquire user information of a target user, and construct a user profile of the target user based on the user information; Determining a first product recommendation list from a preset financial product database based on the user portrait; Extracting the user modeling tag value of the target user under the preset modeling tag from the enterprise region information and historical financing intention in the user information; Determine a product modeling tag value for each financial product in the financial product database under the preset modeling tag; Constructing a product radiation map centered on the target user based on the user modeling tag value and the product modeling tag value, wherein the central node of the product radiation map is the target user, and the radiation nodes of the product radiation map are the financial products in the financial product database; Determine a second product recommendation list from the financial product database based on the product radiation map; Determining recommended products based on the first product recommendation list and the second product recommendation list; The step of constructing a product radiation map centered on the target user based on the user modeling tag value and the product modeling tag value includes: Traversing each financial product in the financial product database, and determining a product score of the financial product based on the product modeling tag value and the user modeling tag value; A product radiation diagram is constructed with the target user as the central node, each of the financial products as the radiation nodes, and the product score corresponding to each of the financial products as the length of the node connection line; The preset modeling tag includes a plurality of modeling sub-tags, and the step of determining the product score of the financial product based on the product modeling tag value and the user modeling tag value includes: Determine a sub-tag difference value corresponding to each of the modeling sub-tags, wherein the sub-tag difference value is a difference between a product modeling tag value and a user modeling tag value under the modeling sub-tag; Obtaining a preset first weighting coefficient group, wherein the first weighting coefficient group includes weighting coefficients corresponding to each of the modeling sub-tags; The sub-label differences corresponding to each other are weighted by each weighting coefficient to obtain a weighted value, and the sum of each weighted value is used as the product score.

2. The product recommendation method according to claim 1, characterized in that: The step of extracting the user modeling tag value of the target user under the preset modeling tag from the enterprise region information and the historical financing intention includes: Determine the user registration place, user office place and user actual factory address of the target user from the enterprise geographical information, and determine the user historical financing situation of the target user from the historical financing intention, wherein the user historical financing situation is used to indicate whether the financing of the target user is successful; The user's registered place is used as the user modeling tag value corresponding to the registered place tag in the preset modeling tag, the user's office location is used as the user modeling tag value corresponding to the office location tag in the preset modeling tag, the user's actual factory site is used as the user modeling tag value corresponding to the factory site tag in the preset modeling tag, and the user's historical financing situation is used as the user modeling tag value corresponding to the financing tag in the preset modeling tag.

3. The product recommendation method according to claim 1, characterized in that: The step of determining a first product recommendation list from a preset financial product database based on the user portrait includes: For any target product among the financial products in the preset financial product database, determine the product attribute tag value of the target product under each attribute tag, wherein each attribute tag is a tag for constructing the user portrait; Determine the user attribute tag value of the target user under each attribute tag based on the user portrait; Traversing each of the attribute tags, calculating the attribute similarity between the user attribute tag value and the product attribute tag value, and calculating the sum of the attribute similarities to obtain a similarity sum value; Determine first candidate products whose similarity and value meet a first preset recommendation condition from each of the financial products, and construct a first product recommendation list based on the first candidate products.

4. The product recommendation method according to claim 1, characterized in that: The step of determining the recommended products based on the first product recommendation list and the second product recommendation list includes: Determine the regional attribute value of each list product in the first product recommendation list and the second product recommendation list; Obtaining a preset second weighting coefficient group, and performing weighted processing on the regional attribute values ​​of each of the listed products using the second weighting coefficient group to obtain a recommendation score for each of the listed products; The listed products whose recommendation scores satisfy the second preset recommendation condition among the listed products are taken as recommended products.

5. The product recommendation method according to any one of claims 1 to 4, characterized in that: The step of constructing a user profile of the target user based on the user information includes: Extracting the target user's regional characteristics, risk preference, financing intention, and credit status from the user information, and determining an attribute label based on the regional characteristics, risk preference, financing intention, and credit status; Determine the label index value of each of the attribute labels, and determine the index mapping value of each of the label index values ​​according to the preset portrait rules based on the label index values; A user portrait is constructed based on the attribute label, the label indicator value and the indicator mapping value.

6. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the product recommendation method according to any one of claims 1 to 5 are implemented.

7. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the product recommendation method according to any one of claims 1 to 5 are implemented.

Images