CN112418956A - Financial product recommendation method and device - Google Patents
Financial product recommendation method and device Download PDFInfo
- Publication number
- CN112418956A CN112418956A CN202011492120.XA CN202011492120A CN112418956A CN 112418956 A CN112418956 A CN 112418956A CN 202011492120 A CN202011492120 A CN 202011492120A CN 112418956 A CN112418956 A CN 112418956A
- Authority
- CN
- China
- Prior art keywords
- user
- product
- target user
- financial
- behavior
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 126
- 230000000875 corresponding effect Effects 0.000 claims description 66
- 239000011159 matrix material Substances 0.000 claims description 32
- 238000012545 processing Methods 0.000 claims description 22
- 238000004364 calculation method Methods 0.000 claims description 18
- 238000011835 investigation Methods 0.000 claims description 10
- 238000002372 labelling Methods 0.000 claims description 7
- 238000004140 cleaning Methods 0.000 claims description 5
- 230000002596 correlated effect Effects 0.000 claims description 5
- 238000012216 screening Methods 0.000 claims description 5
- 238000002790 cross-validation Methods 0.000 claims description 4
- 230000006399 behavior Effects 0.000 description 146
- 238000012797 qualification Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000003542 behavioural effect Effects 0.000 description 2
- 238000010219 correlation analysis Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 238000013475 authorization Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Physics & Mathematics (AREA)
- Marketing (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Entrepreneurship & Innovation (AREA)
- Technology Law (AREA)
- Game Theory and Decision Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a recommendation method and a device for financial products, which are characterized in that a recommendation method most suitable for a target user is determined according to user data and behavior data of the target user to be recommended, a user-based recommendation method is adopted under the condition of less behavior data, a plurality of similar core users of the target user are determined by calculating the similarity between user figures, a financial product recommendation list is generated according to the preference degree of each similar core user to different financial products and the user similarity, and the cold start problem of less behavior data is solved; and under the condition of more behavior data, a product-based recommendation method is adopted, the similarity between the product figures is calculated, the similar financial products of the concerned financial products of the target user are determined, and then a financial product recommendation list is generated according to the behavior weight of the concerned financial products of the target user and the similarity of the financial products, so that the financial products can be quickly and accurately recommended to the user.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a recommendation method and device for financial products.
Background
At present, in the field of financing service of small and medium-sized micro enterprises, a variety of financial products are provided for different user types, the requirements of the financial products on application objects are different, the difficulty and the payment time of application are also different, so that the users cannot quickly, accurately and quickly find the financial products suitable for the users, the enterprises need to spend a large amount of time for finding the financial products suitable for the users, the user experience is influenced, and the cost of the financial institution products reaching the users is increased.
Therefore, how to rapidly and accurately recommend financial products to users of small and medium-sized micro-enterprises from a large amount of financial products becomes a technical problem to be solved urgently in the field.
Disclosure of Invention
In view of the above, the invention provides a method and a device for recommending financial products, which are used for quickly and accurately recommending financial products for small and medium-sized micro-enterprise users.
In order to achieve the above purpose, the invention provides the following specific technical scheme:
a method of recommending financial products, comprising:
acquiring a target user and user data and behavior data of the target user;
performing tagging processing on the user data and the behavior data of the target user to obtain a user tag and a behavior tag of the target user;
generating a user representation of the target user based on the user tag of the target user;
determining a recommendation method corresponding to the target user according to the user data and the behavior data of the target user, wherein the recommendation method comprises the following steps: a user-based recommendation method and a product-based recommendation method;
under the condition that the recommendation method corresponding to the target user is determined to be a user-based recommendation method, obtaining a plurality of similar core users of the target user by calculating the similarity between the user portrait of the target user and the user portrait of the core user;
generating a financial product recommendation list corresponding to the target user according to the preference degree of each similar core user to different financial products and the similarity between each similar core user and the target user;
under the condition that the recommendation method corresponding to the target user is determined to be a product-based recommendation method, calculating the behavior weight of the target user on the concerned financial product according to the behavior label of the target user and the preset weight of each type of user behavior;
calculating the similarity between the product portrait of the concerned financial product and the product portraits of other financial products to obtain a plurality of similar financial products of the concerned financial product, wherein the product portrait of each financial product is generated in advance according to a product label;
and generating a financial product recommendation list corresponding to the target user based on the behavior weight of the target user on the concerned financial product and the similarity between the concerned financial product and each similar financial product.
Optionally, the performing tagging processing on the user data and the behavior data of the target user to obtain the user tag of the target user includes:
cleaning the user data and the behavior data of the target user;
performing cross validation on registration data and credit investigation data in the user data to obtain effective user data;
screening effective behavior data related to financial products in the behavior data;
and performing labeling processing on the effective user data and the effective behavior data to generate a user label and a behavior label of the target user.
Optionally, the generating a user representation of the target user based on the user tag of the target user includes:
analyzing the correlation between the user label of the target user and the financial product to obtain a user label strongly correlated with the financial product;
and dividing the user label of the target user, which is strongly related to the financial product, into a plurality of levels according to a label structure of a preset user portrait, and generating the user portrait of the target user.
Optionally, the determining, according to the user data and the behavior data of the target user, a recommendation method corresponding to the target user includes:
judging whether the behavior data volume of the target user is larger than a preset value or not;
if so, determining that the recommendation method corresponding to the target user is a product-based recommendation method;
if not, determining that the recommendation method corresponding to the target user is a user-based recommendation method.
Optionally, the obtaining of multiple similar core users of the target user by calculating a similarity between the user portrait of the target user and the user portrait of the core user includes:
respectively calculating the similarity between the user portrait of the target user and the user portrait of each core user class obtained by pre-clustering to obtain a target core user class, wherein the target core user class is the core user class with the highest similarity between the user portrait of the target user and the user portrait of the target user;
and respectively calculating the similarity between the user image of the target user and the user image of each core user in the target core user class, and determining k core users with the highest similarity as the similar core users, wherein k is a positive integer.
Optionally, the generating a recommendation list of financial products corresponding to the target user according to the preference of each similar core user to different financial products and the similarity between each similar core user and the target user includes:
determining the user behavior type and the times of each similar core user for each type of financial products according to the behavior label of each similar core user;
calculating the behavior weight of each similar core user to each type of financial products according to the user behavior type and the times of each similar core user to each type of financial products and the preset weight of each type of user behavior, and generating a user-product behavior matrix of the similar core users, wherein rows in the behavior matrix represent the similar core users and columns represent financial products, and the value of each item in the behavior matrix represents the behavior weight of the corresponding similar core user to the corresponding financial product;
calculating a product recommendation weight of each financial product in the behavior matrix relative to the target user according to the behavior matrix and the similarity between each similar core user and the target user;
and generating a financial product recommendation list corresponding to the target user according to the product recommendation weight of each financial product in the behavior matrix relative to the target user.
Optionally, the method further includes:
acquiring product data of each financial product;
respectively generating a product label of each financial product according to the product data of each financial product;
dividing the product label of each financial product into a plurality of levels according to the label structure of the preset product portrait to generate the product portrait of each financial product;
and clustering the product labels of each financial product to obtain a plurality of financial product classes and product portrayals of the financial product classes.
Optionally, the calculating a similarity between the product portrait of the financial product of interest and the product portraits of other financial products to obtain a plurality of similar financial products of the financial product of interest includes:
respectively calculating the similarity between the product image of the concerned financial product and the product image of each financial product class to obtain a target financial product class, wherein the target financial product class is the financial product class with the highest similarity between the product image of the concerned financial product and the product image of the concerned financial product;
and respectively calculating the similarity between the product image of the concerned financial product and the product image of each financial product in the target financial product class, and determining n financial products with the highest similarity as the similar financial products, wherein n is a positive integer.
Optionally, the method further includes:
acquiring the financial products applied or paid by the target user within historical preset time;
and if the applied or deposited financial products exist in the financial product recommendation list, deleting the applied or deposited financial products from the financial product recommendation list.
An apparatus for recommending financial products, comprising:
the data acquisition unit is used for acquiring a target user and user data and behavior data of the target user;
the tagging processing unit is used for performing tagging processing on the user data and the behavior data of the target user to obtain a user tag and a behavior tag of the target user;
a user representation generating unit for generating a user representation of the target user based on a user tag of the target user;
a recommendation method determining unit, configured to determine, according to the user data and the behavior data of the target user, a recommendation method corresponding to the target user, where the recommendation method includes: a user-based recommendation method and a product-based recommendation method;
the user similarity calculation unit is used for calculating the similarity between the user portrait of the target user and the user portrait of the core user to obtain a plurality of similar core users of the target user under the condition that the recommendation method corresponding to the target user is determined to be a user-based recommendation method;
the user recommendation unit is used for generating a financial product recommendation list corresponding to the target user according to the preference degree of each similar core user to different financial products and the similarity between each similar core user and the target user;
the behavior weight calculation unit is used for calculating the behavior weight of the target user on the concerned financial product according to the behavior label of the target user and the preset weight of each type of user behavior under the condition that the recommendation method corresponding to the target user is determined to be a product-based recommendation method;
a product similarity calculation unit, configured to calculate similarities between product images of the concerned financial product and product images of other financial products, to obtain a plurality of similar financial products of the concerned financial product, where the product image of each financial product is generated in advance according to a product tag;
and the product-based recommendation unit is used for generating a financial product recommendation list corresponding to the target user based on the behavior weight of the target user on the concerned financial product and the similarity between the concerned financial product and each similar financial product.
Optionally, the labeling processing unit is specifically configured to:
cleaning the user data and the behavior data of the target user;
performing cross validation on registration data and credit investigation data in the user data to obtain effective user data;
screening effective behavior data related to financial products in the behavior data;
and performing labeling processing on the effective user data and the effective behavior data to generate a user label and a behavior label of the target user.
Optionally, the user representation generating unit is specifically configured to:
analyzing the correlation between the user label of the target user and the financial product to obtain a user label strongly correlated with the financial product;
and dividing the user label of the target user, which is strongly related to the financial product, into a plurality of levels according to a label structure of a preset user portrait, and generating the user portrait of the target user.
Optionally, the recommendation method determining unit is specifically configured to:
judging whether the behavior data volume of the target user is larger than a preset value or not;
if so, determining that the recommendation method corresponding to the target user is a product-based recommendation method;
if not, determining that the recommendation method corresponding to the target user is a user-based recommendation method.
Optionally, the user similarity calculation unit is specifically configured to:
respectively calculating the similarity between the user portrait of the target user and the user portrait of each core user class obtained by pre-clustering to obtain a target core user class, wherein the target core user class is the core user class with the highest similarity between the user portrait of the target user and the user portrait of the target user;
and respectively calculating the similarity between the user image of the target user and the user image of each core user in the target core user class, and determining k core users with the highest similarity as the similar core users, wherein k is a positive integer.
Optionally, the user-based recommendation unit is specifically configured to:
determining the user behavior type and the times of each similar core user for each type of financial products according to the behavior label of each similar core user;
calculating the behavior weight of each similar core user to each type of financial products according to the user behavior type and the times of each similar core user to each type of financial products and the preset weight of each type of user behavior, and generating a user-product behavior matrix of the similar core users, wherein rows in the behavior matrix represent the similar core users and columns represent financial products, and the value of each item in the behavior matrix represents the behavior weight of the corresponding similar core user to the corresponding financial product;
calculating a product recommendation weight of each financial product in the behavior matrix relative to the target user according to the behavior matrix and the similarity between each similar core user and the target user;
and generating a financial product recommendation list corresponding to the target user according to the product recommendation weight of each financial product in the behavior matrix relative to the target user.
Optionally, the apparatus further includes a product portrait generation unit, specifically configured to:
acquiring product data of each financial product;
respectively generating a product label of each financial product according to the product data of each financial product;
dividing the product label of each financial product into a plurality of levels according to the label structure of the preset product portrait to generate the product portrait of each financial product;
and clustering the product labels of each financial product to obtain a plurality of financial product classes and product portrayals of the financial product classes.
Optionally, the product similarity calculation unit is specifically configured to:
respectively calculating the similarity between the product image of the concerned financial product and the product image of each financial product class to obtain a target financial product class, wherein the target financial product class is the financial product class with the highest similarity between the product image of the concerned financial product and the product image of the concerned financial product;
and respectively calculating the similarity between the product image of the concerned financial product and the product image of each financial product in the target financial product class, and determining n financial products with the highest similarity as the similar financial products, wherein n is a positive integer.
Optionally, the apparatus further includes a recommendation optimization processing unit, specifically configured to:
acquiring the financial products applied or paid by the target user within historical preset time;
and if the applied or deposited financial products exist in the financial product recommendation list, deleting the applied or deposited financial products from the financial product recommendation list.
Compared with the prior art, the invention has the following beneficial effects:
the invention discloses a recommendation method of financial products, which utilizes portrait technology to combine with the characteristics of financial products to generate a product portrait capable of accurately and comprehensively reflecting the characteristics of the financial products, and combines with the user characteristics and credit data of financial product users to generate a user portrait capable of accurately and comprehensively reflecting the user characteristics. On the basis, a recommendation method which is most suitable for a target user is determined according to user data and behavior data of the target user to be recommended, a user-based recommendation method is adopted under the condition of less behavior data, a plurality of similar core users of the target user are determined by calculating the similarity between user figures, a financial product recommendation list is generated according to the preference degree of each similar core user to different financial products and the user similarity, and the cold start problem of less behavior data is solved; under the condition of more behavior data, a product-based recommendation method can be adopted or a user-based recommendation method and a product-based recommendation method can be adopted at the same time, wherein the product-based recommendation method determines similar financial products of concerned financial products of a target user by calculating the similarity between product figures, and further generates a financial product recommendation list according to the behavior weight of the concerned financial products of the target user and the similarity of the financial products, so that the financial products can be quickly and accurately recommended for the user.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
FIG. 1 is a flow chart illustrating a method for recommending financial products according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart illustrating a user-based recommendation method according to an embodiment of the present invention;
FIG. 3 is a flowchart illustrating a product-based recommendation method according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of a recommendation apparatus for financial products according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The method introduces the ideas of user portrait and product portrait on the basis of the traditional collaborative filtering algorithm, adopts the similarity between the user portrait to measure the similarity between users, adopts the similarity between the product portrait to measure the similarity between financial products, covers the user behavior deviation, effectively solves the problem of user-product score sparsity, provides a financial product recommendation method based on portrait and collaborative filtering, and realizes the purpose of quickly and accurately recommending financial products for users.
Referring to fig. 1, the method for recommending financial products disclosed in this embodiment includes the following steps:
s101: acquiring a target user and user data and behavior data of the target user;
the target user is a user to be recommended, and the user data comprises: the user basic enterprise data and credit investigation data, and the behavior data comprises: browsing, collecting, applying for and feeding back financial products by users.
S102: performing tagging processing on user data and behavior data of a target user to obtain a user tag and a behavior tag of the target user;
firstly, cleaning user data and behavior data of a target user, eliminating missing, invalid and error data, and reducing the noise value of a data source.
Then, searching corresponding enterprise credit investigation data in a credit investigation database by using the user name and the registered enterprise name, and performing cross verification with the registered data in the user data by using the user data dimension with rich credit investigation data to obtain effective user data, wherein the effective user data mainly comprises the affiliated industry, the operation state, the number of employees, the establishment date, the information of responsible persons, the qualification condition of the share right, the enterprise qualification, the patent information, the trademark information, the soft literature information, the website filing information, the administrative punishment information and the judicial public indication information. Besides the public and searchable credit investigation information, if the user carries out related authorization on the platform, the platform is allowed to collect related information when inquiring information such as electric power data, enterprise financial data, enterprise tax data, Unionpay data and the like of an enterprise.
Then, by using the behavior data obtained in advance, namely the enterprise operation record of the user on the platform, screening out effective behavior data related to financial products, and generating a user behavior database after arrangement, wherein the method specifically comprises the following steps: the method comprises the steps of browsing and recording loan financing policy of a user, collecting and recording loan financing policy, browsing and recording finance or business field news, collecting and recording finance or business field news, browsing and recording loan products, collecting and recording loan products, applying and recording loan products, successfully applying loan products and refusing loan product application.
Finally, performing labeling processing on the effective user data and the effective behavior data, and converting specific unstructured data into a label index, wherein the processing mode of each data is different according to the type of the original data and the meaning in the actual service, for example: and generating an enterprise scale label according to the amount of the registered capital and related indexes such as mobile assets, mobile liabilities, fixed assets, total assets and the like of the enterprise, wherein the label takes four items of large, medium, small and tiny values. Another example is: the method comprises the steps of utilizing information such as the operation range, the product standard name, the product classification, the product introduction, the administrative license information, the main production qualification information, the patent type, the patent abstract, the trademark information, the soft literature information and the website record information registered by an enterprise to obtain an enterprise subdivided field label through the combination of manual information extraction and automatic generation rules, and further delicately depicting the key operation field of the enterprise on the basis of the industry classification.
S103: generating a user representation of the target user based on the user tag of the target user;
firstly, the correlation between the user tag of the target user and the financial product is analyzed, and the correlation analysis can be any existing correlation analysis method, such as a statistical analysis method, and the like, so that the user tag which is strongly correlated with the financial product is obtained.
Dividing user tags of a target user, which are strongly related to financial products, into a plurality of levels according to a tag structure of a preset user portrait, and generating the user portrait of the target user, wherein the tag structure of the preset user portrait can include seven major categories, namely primary tags, including: enterprise information, business information, financial information, public announcement records, power information, behavioral information, and other information. The first-level label enterprise information also comprises second-level labels such as enterprise names, operating states, legal representatives, registered capital, addresses, enterprise scale and associated enterprises; the second-level label associated enterprises also comprise third-level labels of enterprise external investment, stock control enterprises, branch office information and the like; the third-level label stock control enterprise also comprises a stock control enterprise name, an investment proportion, an investment path and other fourth-level labels, and so on.
Preferably, the user tags include a fact tag and a modeling tag, and the fact tag is used for describing fact information, and is usually text or a field; modeling tags are information used for data analysis calculations, usually numbers or options corresponding to unique codes. For example, under the three-level labeled patent information, the patent name and the patent abstract are fact labels, the patent type and the legal state are modeling labels, and under the three-level labeled patent number, the total number of patents label is the modeling label derived in the fourth step of the first embodiment.
The user portrait result generated by the method is generated by combining user enterprise data and credit investigation data oriented to financial products, can describe user characteristics completely and comprehensively, is convenient for visual display and reading, and can effectively support data modeling and analysis by extracting the modeling label.
S104: determining a recommendation method corresponding to the target user according to the user data and the behavior data of the target user;
the recommendation method comprises the following steps: a user-based recommendation method and a product-based recommendation method.
Specifically, whether the behavior data volume of the target user is larger than a preset value or not is judged, and if yes, the recommendation method corresponding to the target user is determined to be a product-based recommendation method; if not, determining that the recommendation method corresponding to the target user is a user-based recommendation method.
Optionally, when the behavior data amount of the target user is greater than the preset value, a user-based recommendation method and a product-based recommendation method can be simultaneously adopted, and finally, a financial product recommendation list generated by the two recommendation methods is combined to recommend a financial product to the target user, and the recommendation result integrates the enterprise characteristics of the user and the behavior characteristics of the user, so that the recommendation result is more accurate and comprehensive.
S105: under the condition that the recommendation method corresponding to the target user is determined to be based on the user recommendation method, a plurality of similar core users of the target user are obtained by calculating the similarity between the user portrait of the target user and the user portrait of the core user;
s106: generating a financial product recommendation list corresponding to the target user according to the preference degree of each similar core user to different financial products and the similarity between each similar core user and the target user;
s107: under the condition that the recommendation method corresponding to the target user is determined to be a product-based recommendation method, calculating the behavior weight of the target user on the concerned financial products according to the behavior label of the target user and the preset weight of each type of user behavior;
s108: calculating the similarity between the product portrait of the concerned financial product and the product portraits of other financial products to obtain a plurality of similar financial products of the concerned financial product, wherein the product portrait of each financial product is generated in advance according to the product label;
s109: and generating a financial product recommendation list corresponding to the target user based on the behavior weight of the target user on the concerned financial products and the similarity between the concerned financial products and each similar financial product.
Further, referring to fig. 2, the recommendation method based on the user disclosed in this embodiment includes the following steps:
s201: and respectively calculating the similarity between the user portrait of the target user and the user portrait of each core user class obtained by pre-clustering to obtain a target core user class, wherein the target core user class is the core user class with the highest similarity between the user portrait of the target user and the user portrait of the target user.
It should be noted that, in order to improve data quality and avoid the influence of a cold start user on a subsequent recommendation result, a core user is a user with rich data, and in order to reduce the calculation amount, firstly, the similarity between the user portrait of a target user and the user portrait of each core user class obtained by pre-clustering is respectively calculated to obtain a target core user class, and then, the similarity between the user portrait of the target user and the user portrait of each core user in the target core user class is respectively calculated.
Wherein, the user portrait used for similarity calculation is the user portrait only containing modeling labels.
In the process of calculating the similarity, firstly, the influence of different dimension data dimensions on the calculation of the user similarity result is isolated according to a normalization formula, then the distance value between different user images is calculated according to the Euclidean distance, and the reciprocal of the distance between the user images is used as the similarity between different users. The Euclidean distance formula is as follows:
in the formula: dist (X, Y) represents the distance between the user representations of the X user and the Y user.
User portrait similarity formula:
in the formula: sim (X, Y) represents the similarity between the user images of the X user and the Y user.
S202: and respectively calculating the similarity between the user portrait of the target user and the user portrait of each core user in the target core user class, and determining k core users with the highest similarity as similar core users, wherein k is a positive integer.
S203: and determining the user behavior type and times of each similar core user for each type of financial products according to the behavior label of each similar core user.
The user behavior types include: browsing, collecting, applying for and feeding back financial products by users.
S204: and calculating the behavior weight of each similar core user to each type of financial products according to the user behavior type and times of each similar core user to each type of financial products and the preset weight of each type of user behavior, and generating a user-product behavior matrix of the similar core users.
User behavior type in the actual dataset: browsing, collecting, applying and feeding back are respectively distinguished by operation type codes 0-3. In order to reflect different interests and preferences of users for products in different behaviors, different behavior weights need to be set for each user behavior type, and different behaviors of the users for the products are weighted and accumulated to serve as interest and preference degrees of enterprises for the products. The user behavior weight formula is as follows:
user behavior weight formula:
where k represents the operation type code for different behaviors, RkWeight, X, representing corresponding user behavior(N,k)And | represents the number of times of K-type operations of the X user on the N product.
The user-product behavior matrix represents the preference degree of different similar core users for different products, the rows in the behavior matrix represent similar core users, the columns represent financial products, and the value of each item in the behavior matrix represents the behavior weight of the corresponding similar core user for the corresponding financial product.
It should be noted that, in this embodiment, different weights are set for different user behavior types according to the preference degree of the user behavior type to the product, and if the preference of browsing, collecting, applying, and feeding back to the product is sequentially enhanced, a browsing weight < a collecting weight < an applying weight < a feeding back weight is set.
S205: and calculating the product recommendation weight of each financial product in the behavior matrix relative to the target user according to the behavior matrix and the similarity between each similar core user and the target user.
In the recommended product list generation stage, a plurality of products with the maximum weight are screened out as products to be recommended according to the behavior weight of the similar core users to the financial products in the user-product behavior matrix. To calculate the interest preference of the target user for the corresponding product, the similarity between the target user and the similar core user is taken as a weight, and the weight is multiplied and accumulated with the interest preference of the similar core user for the product to obtain a final result. The product recommendation weight formula is as follows:
where sim (X, Y) represents the similarity between the X user and the target user Y. XNRepresenting the behavioral weight of the X users on the N products.
S206: and generating a financial product recommendation list corresponding to the target user according to the product recommendation weight of each financial product in the behavior matrix relative to the target user.
And generating a financial product recommendation list comprising a preset number according to the ranking of the product recommendation weights.
Further, referring to fig. 3, the product-based recommendation method disclosed in this embodiment includes the following steps:
s301: and calculating the behavior weight of the target user to the concerned financial products according to the behavior label of the target user and the preset weight of each type of user behavior.
The concerned financial product is obtained from the behavior tag of the target user, and the target user refers to the user behavior weight formula in the above embodiment for the behavior weight of the concerned financial product, which is not described herein again.
When the concerned financial products are more than one, the behavior weight of the target user to each concerned financial product is calculated respectively.
S302: and respectively calculating the similarity between the product portrait of the concerned financial product and the product portrait of each financial product class to obtain a target financial product class, wherein the target financial product class is the financial product class with the highest similarity between the product portrait of the concerned financial product and the product portrait of the concerned financial product.
It should be noted that the product portrait of each financial product is generated in advance, specifically, based on the static information of the product provided by each financial institution, the characteristics for describing the product are screened in a manner of combining manual generation and automatic generation, and mainly include the product name, the financial institution to which the product belongs, the service region, the highest interest rate, the lowest interest rate, the product limit, the highest term, the lowest term, the type of the collateral, the applicable industry, and the like. And then, combining the service data to generate corresponding labels for indexes such as application amount, payment amount, average processing speed, initial trial rejection rate, adjusted power, demand satisfaction proportion and the like of specific products. And respectively dividing the product label of each financial product into a plurality of levels according to the label structure of the preset product portrait to generate the product portrait of each financial product.
It should be further noted that the financial product representation generated in the present embodiment is generated according to the characteristics of the financial product, and in consideration of the fact that the financial product has clear application conditions in the normal situation, the present embodiment can better generate the product representation expressing the product attributes, the application conditions, and the application popularity, so that different financial products can be effectively described and distinguished, the accuracy and the partitionability of the product representation are further improved, and the accurate financial product recommendation for the user in the following process is facilitated.
The product label is distinguished through the fact label and the modeling label, visual display and reading are facilitated, and data modeling and analysis can be effectively supported through extraction of the modeling label.
In order to reduce the calculation amount, the product labels of each financial product are clustered to obtain a plurality of financial product classes and the product representation of each financial product class, and in consideration of the fact that a certain financial product may belong to more than one financial product class simultaneously under the actual condition, the embodiment supports manual adjustment of clustering results, that is, the finally obtained financial product classes are obtained by combining clustering and manual classification, so that the financial products can be simultaneously divided into more than one financial product class according to actual needs.
S303: and respectively calculating the similarity between the product portrait of the concerned financial product and the product portrait of each financial product in the target financial product class, and determining the n financial products with the highest similarity as similar financial products, wherein n is a positive integer.
Firstly, similarity among different financial products is calculated according to each modeling label dimension of the financial products, the influence of data dimension is isolated through a normalization means, and the reciprocal of Euclidean distance among the products is used as the similarity among the products.
The product similarity formula is:
in the formula: sim (N)1,N2) Indicates the similarity between the two products, dist (N)1,N2) Represents N1,N2The distance between the two products can be calculated by the formula (5).
S304: and generating a financial product recommendation list corresponding to the target user based on the behavior weight of the target user on the concerned financial products and the similarity between the concerned financial products and each similar financial product.
Calculating the product recommendation weight of the target user based on the behavior weight of the target user on the concerned financial products and the similarity between the concerned financial products and each similar financial product:
wherein, sim (N, N)j) Representing the product N to be recommended and the concerned financial product NjThe similarity between them. Y isNjN representing a target user Y for a historical preset time period, e.g. used in last three monthsjBehavioral weighting of the product.
And then, generating a financial product recommendation list corresponding to the target user based on the ranking of the product recommendation weights of the target user.
Further, the embodiment can also optimize the recommendation list to obtain the financial products applied or paid by the target user within the historical preset time; and if the financial products which are applied for or placed in the financial product recommendation list exist, deleting the financial products from the financial product recommendation list.
Preferably, the embodiment further discloses a method for recommending a prominent new product, which takes the product applied and paid by the target user within the preset historical time, such as the past month, as a filter, and removes the product applied and paid by the target user from all the products to be recommended. Then, products which are never used by the target user are screened out from all recommendation lists, the relevance between the products and the target user is calculated according to a product recommendation weight formula and is ranked, and the product with the highest relevance is added into the recommendation product list, so that the aim of finding new products for enterprises as far as possible is achieved.
Based on the recommendation method for financial products disclosed in the above embodiments, the present embodiment correspondingly discloses a recommendation device for financial products, please refer to fig. 4, and the device includes:
a data acquiring unit 100, configured to acquire a target user and user data and behavior data of the target user;
a tagging processing unit 200, configured to perform tagging processing on the user data and the behavior data of the target user to obtain a user tag and a behavior tag of the target user;
a user representation generating unit 300 for generating a user representation of the target user based on a user tag of the target user;
a recommendation method determining unit 400, configured to determine, according to the user data and the behavior data of the target user, a recommendation method corresponding to the target user, where the recommendation method includes: a user-based recommendation method and a product-based recommendation method;
a user similarity calculation unit 500, configured to, when it is determined that the recommendation method corresponding to the target user is a user-based recommendation method, obtain a plurality of similar core users of the target user by calculating a similarity between a user portrait of the target user and a user portrait of a core user;
a user recommendation unit 600, configured to generate a recommendation list of financial products corresponding to the target user according to the preference of each similar core user for different financial products and the similarity between each similar core user and the target user;
a behavior weight calculation unit 700, configured to calculate, when it is determined that the recommendation method corresponding to the target user is a product-based recommendation method, a behavior weight of the target user for a concerned financial product according to a behavior tag of the target user and a preset weight of each type of user behavior;
a product similarity calculation unit 800, configured to calculate similarities between product images of the financial product of interest and product images of other financial products, to obtain a plurality of similar financial products of the financial product of interest, where a product image of each financial product is generated in advance according to a product tag;
a product recommendation unit 900, configured to generate a financial product recommendation list corresponding to the target user based on the behavior weight of the target user on the focused financial product and the similarity between the focused financial product and each similar financial product.
Optionally, the labeling processing unit 200 is specifically configured to:
cleaning the user data and the behavior data of the target user;
performing cross validation on registration data and credit investigation data in the user data to obtain effective user data;
screening effective behavior data related to financial products in the behavior data;
and performing labeling processing on the effective user data and the effective behavior data to generate a user label and a behavior label of the target user.
Optionally, the user representation generating unit 300 is specifically configured to:
analyzing the correlation between the user label of the target user and the financial product to obtain a user label strongly correlated with the financial product;
and dividing the user label of the target user, which is strongly related to the financial product, into a plurality of levels according to a label structure of a preset user portrait, and generating the user portrait of the target user.
Optionally, the recommendation method determining unit 400 is specifically configured to:
judging whether the behavior data volume of the target user is larger than a preset value or not;
if so, determining that the recommendation method corresponding to the target user is a product-based recommendation method;
if not, determining that the recommendation method corresponding to the target user is a user-based recommendation method.
Optionally, the user similarity calculation unit 500 is specifically configured to:
respectively calculating the similarity between the user portrait of the target user and the user portrait of each core user class obtained by pre-clustering to obtain a target core user class, wherein the target core user class is the core user class with the highest similarity between the user portrait of the target user and the user portrait of the target user;
and respectively calculating the similarity between the user image of the target user and the user image of each core user in the target core user class, and determining k core users with the highest similarity as the similar core users, wherein k is a positive integer.
Optionally, the user-based recommendation unit 600 is specifically configured to:
determining the user behavior type and the times of each similar core user for each type of financial products according to the behavior label of each similar core user;
calculating the behavior weight of each similar core user to each type of financial products according to the user behavior type and the times of each similar core user to each type of financial products and the preset weight of each type of user behavior, and generating a user-product behavior matrix of the similar core users, wherein rows in the behavior matrix represent the similar core users and columns represent financial products, and the value of each item in the behavior matrix represents the behavior weight of the corresponding similar core user to the corresponding financial product;
calculating a product recommendation weight of each financial product in the behavior matrix relative to the target user according to the behavior matrix and the similarity between each similar core user and the target user;
and generating a financial product recommendation list corresponding to the target user according to the product recommendation weight of each financial product in the behavior matrix relative to the target user.
Optionally, the apparatus further includes a product portrait generation unit, specifically configured to:
acquiring product data of each financial product;
respectively generating a product label of each financial product according to the product data of each financial product;
dividing the product label of each financial product into a plurality of levels according to the label structure of the preset product portrait to generate the product portrait of each financial product;
and clustering the product labels of each financial product to obtain a plurality of financial product classes and product portrayals of the financial product classes.
Optionally, the product similarity calculation unit 800 is specifically configured to:
respectively calculating the similarity between the product image of the concerned financial product and the product image of each financial product class to obtain a target financial product class, wherein the target financial product class is the financial product class with the highest similarity between the product image of the concerned financial product and the product image of the concerned financial product;
and respectively calculating the similarity between the product image of the concerned financial product and the product image of each financial product in the target financial product class, and determining n financial products with the highest similarity as the similar financial products, wherein n is a positive integer.
Optionally, the apparatus further includes a recommendation optimization processing unit, specifically configured to:
acquiring the financial products applied or paid by the target user within historical preset time;
and if the applied or deposited financial products exist in the financial product recommendation list, deleting the applied or deposited financial products from the financial product recommendation list.
The recommendation device for financial products disclosed by the embodiment determines a recommendation method most suitable for a target user according to user data and behavior data of the target user to be recommended, adopts a user-based recommendation method under the condition of less behavior data, determines a plurality of similar core users of the target user by calculating the similarity between user figures, further generates a financial product recommendation list according to the preference degree and the user similarity of each similar core user to different financial products, and solves the cold start problem of less behavior data; and under the condition of more behavior data, a product-based recommendation method is adopted, the similarity between the product figures is calculated, the similar financial products of the concerned financial products of the target user are determined, and then a financial product recommendation list is generated according to the behavior weight of the concerned financial products of the target user and the similarity of the financial products, so that the financial products can be quickly and accurately recommended to the user.
The above embodiments can be combined arbitrarily, and the features described in the embodiments in the present specification can be replaced or combined with each other in the above description of the disclosed embodiments, so that those skilled in the art can implement or use the present application.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A method for recommending financial products, comprising:
acquiring a target user and user data and behavior data of the target user;
performing tagging processing on the user data and the behavior data of the target user to obtain a user tag and a behavior tag of the target user;
generating a user representation of the target user based on the user tag of the target user;
determining a recommendation method corresponding to the target user according to the user data and the behavior data of the target user, wherein the recommendation method comprises the following steps: a user-based recommendation method and a product-based recommendation method;
under the condition that the recommendation method corresponding to the target user is determined to be a user-based recommendation method, obtaining a plurality of similar core users of the target user by calculating the similarity between the user portrait of the target user and the user portrait of the core user;
generating a financial product recommendation list corresponding to the target user according to the preference degree of each similar core user to different financial products and the similarity between each similar core user and the target user;
under the condition that the recommendation method corresponding to the target user is determined to be a product-based recommendation method, calculating the behavior weight of the target user on the concerned financial product according to the behavior label of the target user and the preset weight of each type of user behavior;
calculating the similarity between the product portrait of the concerned financial product and the product portraits of other financial products to obtain a plurality of similar financial products of the concerned financial product, wherein the product portrait of each financial product is generated in advance according to a product label;
and generating a financial product recommendation list corresponding to the target user based on the behavior weight of the target user on the concerned financial product and the similarity between the concerned financial product and each similar financial product.
2. The method according to claim 1, wherein the tagging the user data and the behavior data of the target user to obtain the user tag of the target user comprises:
cleaning the user data and the behavior data of the target user;
performing cross validation on registration data and credit investigation data in the user data to obtain effective user data;
screening effective behavior data related to financial products in the behavior data;
and performing labeling processing on the effective user data and the effective behavior data to generate a user label and a behavior label of the target user.
3. The method of claim 1, wherein generating the user representation of the target user based on the user tag of the target user comprises:
analyzing the correlation between the user label of the target user and the financial product to obtain a user label strongly correlated with the financial product;
and dividing the user label of the target user, which is strongly related to the financial product, into a plurality of levels according to a label structure of a preset user portrait, and generating the user portrait of the target user.
4. The method according to claim 1, wherein the determining the recommendation method corresponding to the target user according to the user data and the behavior data of the target user comprises:
judging whether the behavior data volume of the target user is larger than a preset value or not;
if so, determining that the recommendation method corresponding to the target user is a product-based recommendation method;
if not, determining that the recommendation method corresponding to the target user is a user-based recommendation method.
5. The method of claim 1, wherein the obtaining a plurality of similar core users of the target user by calculating a similarity between the user representation of the target user and the user representation of a core user comprises:
respectively calculating the similarity between the user portrait of the target user and the user portrait of each core user class obtained by pre-clustering to obtain a target core user class, wherein the target core user class is the core user class with the highest similarity between the user portrait of the target user and the user portrait of the target user;
and respectively calculating the similarity between the user image of the target user and the user image of each core user in the target core user class, and determining k core users with the highest similarity as the similar core users, wherein k is a positive integer.
6. The method of claim 5, wherein the generating of the recommendation list of the financial product corresponding to the target user according to the preference of each similar core user for different financial products and the similarity between each similar core user and the target user comprises:
determining the user behavior type and the times of each similar core user for each type of financial products according to the behavior label of each similar core user;
calculating the behavior weight of each similar core user to each type of financial products according to the user behavior type and the times of each similar core user to each type of financial products and the preset weight of each type of user behavior, and generating a user-product behavior matrix of the similar core users, wherein rows in the behavior matrix represent the similar core users and columns represent financial products, and the value of each item in the behavior matrix represents the behavior weight of the corresponding similar core user to the corresponding financial product;
calculating a product recommendation weight of each financial product in the behavior matrix relative to the target user according to the behavior matrix and the similarity between each similar core user and the target user;
and generating a financial product recommendation list corresponding to the target user according to the product recommendation weight of each financial product in the behavior matrix relative to the target user.
7. The method of claim 1, further comprising:
acquiring product data of each financial product;
respectively generating a product label of each financial product according to the product data of each financial product;
dividing the product label of each financial product into a plurality of levels according to the label structure of the preset product portrait to generate the product portrait of each financial product;
and clustering the product labels of each financial product to obtain a plurality of financial product classes and product portrayals of the financial product classes.
8. The method of claim 7, wherein the calculating the similarity between the product representation of the financial product of interest and the product representations of the other financial products to obtain a plurality of similar financial products of the financial product of interest comprises:
respectively calculating the similarity between the product image of the concerned financial product and the product image of each financial product class to obtain a target financial product class, wherein the target financial product class is the financial product class with the highest similarity between the product image of the concerned financial product and the product image of the concerned financial product;
and respectively calculating the similarity between the product image of the concerned financial product and the product image of each financial product in the target financial product class, and determining n financial products with the highest similarity as the similar financial products, wherein n is a positive integer.
9. The method of claim 1, further comprising:
acquiring the financial products applied or paid by the target user within historical preset time;
and if the applied or deposited financial products exist in the financial product recommendation list, deleting the applied or deposited financial products from the financial product recommendation list.
10. An apparatus for recommending financial products, comprising:
the data acquisition unit is used for acquiring a target user and user data and behavior data of the target user;
the tagging processing unit is used for performing tagging processing on the user data and the behavior data of the target user to obtain a user tag and a behavior tag of the target user;
a user representation generating unit for generating a user representation of the target user based on a user tag of the target user;
a recommendation method determining unit, configured to determine, according to the user data and the behavior data of the target user, a recommendation method corresponding to the target user, where the recommendation method includes: a user-based recommendation method and a product-based recommendation method;
the user similarity calculation unit is used for calculating the similarity between the user portrait of the target user and the user portrait of the core user to obtain a plurality of similar core users of the target user under the condition that the recommendation method corresponding to the target user is determined to be a user-based recommendation method;
the user recommendation unit is used for generating a financial product recommendation list corresponding to the target user according to the preference degree of each similar core user to different financial products and the similarity between each similar core user and the target user;
the behavior weight calculation unit is used for calculating the behavior weight of the target user on the concerned financial product according to the behavior label of the target user and the preset weight of each type of user behavior under the condition that the recommendation method corresponding to the target user is determined to be a product-based recommendation method;
a product similarity calculation unit, configured to calculate similarities between product images of the concerned financial product and product images of other financial products, to obtain a plurality of similar financial products of the concerned financial product, where the product image of each financial product is generated in advance according to a product tag;
and the product-based recommendation unit is used for generating a financial product recommendation list corresponding to the target user based on the behavior weight of the target user on the concerned financial product and the similarity between the concerned financial product and each similar financial product.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011492120.XA CN112418956A (en) | 2020-12-16 | 2020-12-16 | Financial product recommendation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011492120.XA CN112418956A (en) | 2020-12-16 | 2020-12-16 | Financial product recommendation method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112418956A true CN112418956A (en) | 2021-02-26 |
Family
ID=74775887
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011492120.XA Pending CN112418956A (en) | 2020-12-16 | 2020-12-16 | Financial product recommendation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112418956A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113344433A (en) * | 2021-06-28 | 2021-09-03 | 平安信托有限责任公司 | Product matching method and device, electronic equipment and readable storage medium |
| CN113360773A (en) * | 2021-07-07 | 2021-09-07 | 脸萌有限公司 | Recommendation method and device, storage medium and electronic equipment |
| CN113407827A (en) * | 2021-06-11 | 2021-09-17 | 广州三七极创网络科技有限公司 | Information recommendation method, device, equipment and medium based on user value classification |
| CN113487380A (en) * | 2021-06-25 | 2021-10-08 | 天元大数据信用管理有限公司 | Financial product recommendation method, device, equipment and medium |
| CN114581192A (en) * | 2022-03-08 | 2022-06-03 | 山东大学 | Financial product recommendation method and system based on user implicit data |
| CN114662007A (en) * | 2022-05-25 | 2022-06-24 | 太平金融科技服务(上海)有限公司深圳分公司 | Data recommendation method and device, computer equipment and storage medium |
| CN117390232A (en) * | 2023-11-30 | 2024-01-12 | 金网络(北京)数字科技有限公司 | Enterprise portrait construction method, system, equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102426686A (en) * | 2011-09-29 | 2012-04-25 | 南京大学 | Internet information product recommending method based on matrix decomposition |
| CN108615177A (en) * | 2018-04-09 | 2018-10-02 | 武汉理工大学 | Electric terminal personalized recommendation method based on weighting extraction interest-degree |
| CN109345348A (en) * | 2018-09-30 | 2019-02-15 | 重庆誉存大数据科技有限公司 | The recommended method of multidimensional information portrait based on travel agency user |
| WO2019061976A1 (en) * | 2017-09-28 | 2019-04-04 | 平安科技（深圳）有限公司 | Fund product recommendation method and apparatus, terminal device, and storage medium |
| CN111026966A (en) * | 2019-12-06 | 2020-04-17 | 创新奇智(成都)科技有限公司 | Search recommendation ranking method based on user, product portrait and correlation degree of user and product portrait |
-
2020
- 2020-12-16 CN CN202011492120.XA patent/CN112418956A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102426686A (en) * | 2011-09-29 | 2012-04-25 | 南京大学 | Internet information product recommending method based on matrix decomposition |
| WO2019061976A1 (en) * | 2017-09-28 | 2019-04-04 | 平安科技（深圳）有限公司 | Fund product recommendation method and apparatus, terminal device, and storage medium |
| CN108615177A (en) * | 2018-04-09 | 2018-10-02 | 武汉理工大学 | Electric terminal personalized recommendation method based on weighting extraction interest-degree |
| CN109345348A (en) * | 2018-09-30 | 2019-02-15 | 重庆誉存大数据科技有限公司 | The recommended method of multidimensional information portrait based on travel agency user |
| CN111026966A (en) * | 2019-12-06 | 2020-04-17 | 创新奇智(成都)科技有限公司 | Search recommendation ranking method based on user, product portrait and correlation degree of user and product portrait |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113407827A (en) * | 2021-06-11 | 2021-09-17 | 广州三七极创网络科技有限公司 | Information recommendation method, device, equipment and medium based on user value classification |
| CN113487380A (en) * | 2021-06-25 | 2021-10-08 | 天元大数据信用管理有限公司 | Financial product recommendation method, device, equipment and medium |
| CN113344433A (en) * | 2021-06-28 | 2021-09-03 | 平安信托有限责任公司 | Product matching method and device, electronic equipment and readable storage medium |
| CN113360773A (en) * | 2021-07-07 | 2021-09-07 | 脸萌有限公司 | Recommendation method and device, storage medium and electronic equipment |
| CN114581192A (en) * | 2022-03-08 | 2022-06-03 | 山东大学 | Financial product recommendation method and system based on user implicit data |
| CN114581192B (en) * | 2022-03-08 | 2024-01-26 | 山东大学 | Financial product recommendation method and system based on user implicit data |
| CN114662007A (en) * | 2022-05-25 | 2022-06-24 | 太平金融科技服务(上海)有限公司深圳分公司 | Data recommendation method and device, computer equipment and storage medium |
| CN117390232A (en) * | 2023-11-30 | 2024-01-12 | 金网络(北京)数字科技有限公司 | Enterprise portrait construction method, system, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112418956A (en) | Financial product recommendation method and device | |
| Verma et al. | Big data analytics: Challenges and applications for text, audio, video, and social media data | |
| CN107851097B (en) | Data analysis system, data analysis method, data analysis program, and storage medium | |
| US20140039981A1 (en) | Systems and methods for detection of satisficing in surveys | |
| CN108763362A (en) | Method is recommended to the partial model Weighted Fusion Top-N films of selection based on random anchor point | |
| CN1670733A (en) | Rendering tables with natural language commands | |
| Sudira et al. | Instagram sentiment analysis with naive bayes and KNN: exploring customer satisfaction of digital payment services in Indonesia | |
| KR20210128614A (en) | Influencer marketing service method | |
| KR20080086430A (en) | Technical document attribute association analysis supporting apparatus | |
| CN117151870B (en) | Portrait behavior analysis method and system based on guest group | |
| KR100901782B1 (en) | Method and System for Generating Marketing Information | |
| Ram et al. | Fake reviews detection using supervised machine learning | |
| CN115965468A (en) | Transaction data-based abnormal behavior detection method, device, equipment and medium | |
| CN115719183A (en) | Power customer self-feedback service evaluation method and system based on weight dynamic grading | |
| CN114331572A (en) | Potential customer determination method and device, electronic equipment and storage medium | |
| CN114119069A (en) | Resource management product recommendation method and device, electronic equipment and storage medium | |
| Roelands et al. | Classifying businesses by economic activity using web-based text mining | |
| CN113487380A (en) | Financial product recommendation method, device, equipment and medium | |
| Hadden | A customer profiling methodology for churn prediction | |
| Nam et al. | Online Document Mining Approach to Predicting Crowdfunding Success | |
| CN117668205B (en) | Smart logistics customer service processing method, system, equipment and storage medium | |
| Ayyasamy et al. | Design and implementation of residential rental rates forecast model using data mining algorithms | |
| CN112182165B (en) | New product quality planning method based on online comments | |
| Park et al. | Investigating the Effect of a Brand Factor in Product Design Based on a Data-Driven Approach Using Online Reviews | |
| Lakshmikanth | Identification of Fake Reviews Using Supervised Machine Learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210226 |