CN112465603A - Recommendation system and method for non-high frequency consumption industry - Google Patents
Recommendation system and method for non-high frequency consumption industry Download PDFInfo
- Publication number
- CN112465603A CN112465603A CN202011457658.7A CN202011457658A CN112465603A CN 112465603 A CN112465603 A CN 112465603A CN 202011457658 A CN202011457658 A CN 202011457658A CN 112465603 A CN112465603 A CN 112465603A
- Authority
- CN
- China
- Prior art keywords
- product
- information
- preference
- user
- recommendation
- 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 25
- 230000006399 behavior Effects 0.000 claims abstract description 103
- 238000012216 screening Methods 0.000 claims description 25
- 238000012937 correction Methods 0.000 claims description 19
- 238000007405 data analysis Methods 0.000 abstract description 2
- 238000012545 processing Methods 0.000 abstract description 2
- 238000004364 calculation method Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 1
- 238000004091 panning 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
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Development Economics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of data analysis and processing, in particular to a recommendation system and a method for non-high frequency consumption industry, wherein the system comprises a server, the server is used for judging according to operation behaviors and preset ordering behaviors, when the ordering behaviors exist in the operation behaviors, service follow-up information is obtained, preference relation weight is corrected according to the service follow-up information and product characteristics, otherwise, to-be-matched characteristics of a current user are obtained, similar user groups are screened according to the to-be-matched characteristics and preset user information, and preference relation weight corresponding to the current user is corrected according to the preference relation weight in the similar user groups; the server is further used for calling the product information of the product to be sold according to the product characteristics and generating a product recommendation list according to the preference relationship weight and the product information when the purchase appeal is obtained. By adopting the scheme, the technical problem that recommendation cannot be accurately realized due to the fact that data samples collected by non-high-frequency consumption industries are low in the prior art can be solved.
Description
Technical Field
The invention relates to the technical field of data analysis and processing, in particular to a recommendation system and method for non-high-frequency consumption industry.
Background
In order to facilitate the success rate of transactions, product recommendation is often performed for the purchase appeal of users under modern society, because the product recommendation is based on attribute analysis of the users, the accuracy of the attribute analysis depends on the number of data samples, and because the high-frequency consumption industry has more corresponding user-generated operation behaviors and more data samples of user behaviors are collected, the recommendation system and the recommendation method are mostly applied to network platforms of the high-frequency consumption industry, that is, the network platforms of the users with high-frequency consumption behaviors, such as panning, shaking and the like.
For non-high frequency consumption industries, such as used cars, houses and the like, impulse consumption is rarely generated for users due to the properties and prices of products, and the consumption frequency is extremely low, which means that the user information collected by the non-high frequency consumption industries is very little. The existing recommendation system and recommendation method cannot meet the requirement of accurately recommending products under the condition of less user information, so that a recommendation system and recommendation method which are applied to the non-high-frequency consumption industry and can realize accurate recommendation of products are urgently needed.
Disclosure of Invention
One of the objectives of the present invention is to provide a recommendation system for non-high frequency consumption industry, so as to solve the technical problems in the prior art that the user information collected in the non-high frequency consumption industry is less and accurate recommendation cannot be realized.
The invention provides a basic scheme I:
the recommendation system for the non-high-frequency consumption industry comprises a server, a database and a recommendation server, wherein the server is used for acquiring product information and generating product characteristics according to the product information; the server is also used for acquiring the operation information and calculating the preference relation weight according to the operation information and the product characteristics;
the operation information comprises operation behaviors, the server is also used for judging according to the operation behaviors and preset ordering behaviors, when the ordering behaviors exist in the operation behaviors, service follow-up information is obtained, the preference relation weight is corrected according to the service follow-up information and product characteristics, when the ordering behaviors do not exist in the operation behaviors, the to-be-matched characteristics of the current user are obtained, similar user groups are screened according to the to-be-matched characteristics and the preset user information, and the preference relation weight corresponding to the current user is corrected according to the preference relation weight in the similar user groups;
the server is further used for calling the product information of the product to be sold according to the product characteristics and generating a product recommendation list according to the preference relationship weight and the product information when the purchase appeal is obtained.
The beneficial effects of the first basic scheme are as follows:
according to the scheme, the preference relation weights of the user on different characteristics of the product are calculated according to the operation behavior of the user, the user is not required to provide clear information data, the data volume is reduced, and the information overload problem is relieved.
The behavior of the user is judged, the user with the ordering behavior represents that the user purchases the corresponding product once, and in the process, professional service personnel contact the user and can continuously generate service follow-up information. The service follow-up information relative operation behavior can reflect the preference of the user more accurately, so the preference relation weight is corrected based on the service follow-up information. The users without ordering behavior represent that the users have not bought corresponding products, the collected operation behaviors of the users are only data for performing operation on line, the user preference is fuzzy due to too few data samples, other users with higher similarity to the current user are calculated to form a similar user group, the potential preference of the current user is expanded based on the similar user group, and negative effects on recommendation results due to the fact that the user preference is fuzzy due to too few data samples are relieved.
And when the user has a purchase demand, generating a product recommendation list according to the preference relation weight and the product information to recommend the product. According to the scheme, the user information is collected in an implicit mode, the preference relation weight is corrected based on the service influence of professional service personnel and the similarity possibility between users, the problem of fuzzy preference of the users due to insufficient data samples is effectively solved, the price sensitivity is low in the non-high-frequency consumption industry, the knowledge of the users related to the purchased products is not professional enough, and personalized accurate recommendation is performed in non-impulsive consumption behaviors.
Further, the server comprises a feature generation module, wherein the feature generation module is used for extracting key features from the product information of each product, calculating the occurrence rate of the key features according to the key features, and screening out the product features according to the occurrence rate and the key features.
Has the advantages that: the product information of the product comprises unnecessary interference information and non-appealing hard selection criteria, so that key features are extracted based on the product information of the current product, the interference information is reduced, and the subsequent generation of preference relationship weight and the data volume of a product recommendation list are reduced. And automatically screening according to the occurrence rate of the key features, and removing non-appealing hard selection standards so as to obtain the product attributes of the universal relationship of the user. Certainly, the system management personnel can set the product characteristics by depending on years of industry experience, and compared with manual setting, the screening of the product characteristics is more objective, and the subjective influence caused by the manual setting is reduced.
Further, the service follow-up information comprises follow-up behaviors, corresponding follow-up products and object information, the server comprises a preference correction module, and the preference correction module is used for correcting preference relation weights according to the follow-up behaviors, the follow-up products, the object information and product characteristics.
Has the advantages that: compared with a user who does not have the ordering behavior, the user who has the ordering behavior can make the preference of the user more clear according to the product purchased by the user and the behavior appearing in the purchased product due to the ordering behavior, and the accuracy of the preference relationship weight is improved by correcting the preference relationship weight through the follow-up behavior, the follow-up product and the object information, so that more accurate personalized recommendation is provided for the purchase appeal appearing again by the user.
Further, the server comprises a similar screening module and a preference correction module, wherein the similar screening module is used for matching according to the characteristics to be matched and the user characteristics in the user information to generate matching degree, and screening the users according to the matching degree and a preset matching degree threshold value to generate a similar user group; the preference correction module is used for calling the preference relation weight corresponding to the user in the similar user group and correcting the preference relation weight of the current user according to the called preference relation weight.
Has the advantages that: there are no users who are ordering, there are fewer data samples, making the user preferences ambiguous. Because the requirements of the users have similarity, the users are subjected to similarity matching based on the known user characteristics, and the users with close similarity are calculated to form a similar user group. Based on the similarity of the requirements among similar user groups, the potential preferences among the similar users are expanded, so that the preference relation weight of the current user is corrected, the fuzzy user preference caused by fewer data samples is relieved, and the accurate personalized recommendation is more effectively carried out on the user.
The server further comprises a recommendation matching module and a recommendation duplication eliminating module, wherein the recommendation matching module is used for calculating the preference degree of each product to be sold according to the preference relation weight and the product information and generating a product sorting table of the products to be sold according to the preference degree; and the recommendation duplication eliminating module is used for calling the historical recommendation information, eliminating the corresponding products in the product sorting table according to the historical recommendation information and generating the product recommendation table.
Has the advantages that: the product sorting table is subjected to duplicate removal through the historical recommendation information, products which are already recommended are removed, and the products which are already recommended are checked or known by a user, so that duplicate removal is performed on the recommended products, and the experience of the user is improved.
The invention also aims to provide a recommendation method for the non-high-frequency consumption industry.
The invention provides a second basic scheme:
the recommendation method for the non-high frequency consumption industry uses the recommendation system for the non-high frequency consumption industry, and comprises the following steps:
s1: acquiring product information and operation information, generating product characteristics according to the product information, and calculating preference relation weight according to the operation information and the product characteristics;
s2: the operation information comprises operation behaviors, judgment is carried out according to the operation behaviors and preset ordering behaviors, when the ordering behaviors exist in the operation behaviors, service follow-up information is obtained, preference relation weights are corrected according to the service follow-up information and product characteristics, when the ordering behaviors do not exist in the operation behaviors, characteristics to be matched of a current user are obtained, similar user groups are screened according to the characteristics to be matched and the preset user information, and the preference relation weights corresponding to the current user are corrected according to the preference relation weights in the similar user groups;
s3: and when the purchase appeal is acquired, calling the product information of the product to be sold according to the product characteristics, and generating a product recommendation list according to the preference relation weight and the product information.
The second basic scheme has the beneficial effects that: according to the scheme, the user information is collected in an implicit mode, the preference relation weight is corrected based on the service influence of professional service personnel and the similarity possibility between users, the problem of fuzzy preference of the users due to insufficient data samples is effectively solved, the price sensitivity is low in the non-high-frequency consumption industry, the knowledge of the users related to the purchased products is not professional enough, and personalized accurate recommendation is performed in non-impulsive consumption behaviors.
Further, the step of generating the product characteristics according to the product information in S1 includes the steps of:
extracting key features from the product information of each product, calculating the occurrence rate of the key features according to the key features, and screening out the product features according to the occurrence rate and the key features.
Has the advantages that: and extracting key features based on the product information of the current product, reducing interference information, and reducing the weight of a subsequently generated preference relationship and the data volume of a product recommendation list. The method comprises the steps of automatically screening according to the occurrence rate of key features, and removing non-appealing hard selection standards, so that product attributes of the general relationship of users are obtained.
Further, the service follow-up information includes follow-up behavior and corresponding follow-up product, and object information, and in the step of S2, the step of correcting the preference relationship weight according to the service follow-up information and the product characteristics includes the steps of:
and correcting the preference relation weight according to the follow-up behavior, the follow-up product, the object information and the product characteristics.
Has the advantages that: the preference relation weight is corrected through the follow-up behavior, the follow-up product and the object information, so that the accuracy of the preference relation weight is improved, and more accurate personalized recommendation is provided for the reoccurring purchase appeal of the user.
Further, in the step S2, screening a similar user group according to the features to be matched and the preset user information, and modifying the preference relationship weight corresponding to the current user according to the preference relationship weight in the similar user group includes the following steps:
matching according to the features to be matched and the user features in the user information to generate a matching degree, and screening users to generate a similar user group according to the matching degree and a preset matching degree threshold; and calling the preference relation weight corresponding to the user in the similar user group, and correcting the preference relation weight of the current user according to the called preference relation weight.
Has the advantages that: based on the similarity of the requirements among similar user groups, the potential preferences among the similar users are expanded, so that the preference relation weight of the current user is corrected, the fuzzy user preference caused by fewer data samples is relieved, and the accurate personalized recommendation is more effectively carried out on the user.
Further, in the step S3, the step of generating the product recommendation list according to the preference relationship weight and the product information includes the following steps:
calculating the preference degree of each product to be sold according to the preference relation weight and the product information, and generating a product sorting table of the products to be sold according to the preference degree; calling historical recommendation information, and removing corresponding products in the product sorting table according to the historical recommendation information to generate a product recommendation table.
Has the advantages that: and the duplicate removal is carried out on the recommended products, so that the experience of the user is improved.
Drawings
FIG. 1 is a logic block diagram of an embodiment of a recommendation system for non-high frequency consumer industries of the present invention.
Detailed Description
The following is further detailed by way of specific embodiments:
examples
The recommendation system for the non-high-frequency consumption industry, as shown in fig. 1, comprises a server, a user terminal and a management terminal, wherein the server comprises a database, a data calling module, a feature generation module, a preference calculation module, a behavior judgment module, a preference correction module, a similarity screening module and a recommendation matching module.
The user terminal is used by a user, for example, a mobile phone, and the user terminal is loaded with corresponding service software, in this embodiment, the second-hand car industry is taken as the non-high-frequency consumption industry for example, and then the corresponding user terminal is installed with corresponding second-hand car selective purchasing service software. The user terminal is used for acquiring user information generated by user registration, uploading the user information, wherein the user information comprises a user account, gender, age, occupation and region, the user terminal is also used for acquiring operation information generated by the user for browsing vehicles, collecting vehicles, removing vehicles, calculating budget, calculating loan and the like on service software, uploading the operation information and the user account, and the database is used for storing the user information and storing the operation information according to the user account. The operation information includes an operation behavior and an operation object, for example, a favorite vehicle, the operation behavior is favorite, and the operation object is a favorite vehicle model.
The management terminal is used for managers in non-high-frequency consumption industries, such as computers, and is used for obtaining product information of products to be sold and uploading the product information to the server, and the database is also used for storing the product information. In this embodiment, the product information is description information of the vehicle to be sold.
The server is used for acquiring product information and generating product characteristics according to the product information; the server is also used for obtaining the operation information and calculating the preference relation weight according to the operation information and the product characteristics. Specifically, the data calling module is used for calling product information and operation information from the database, sending the product information to the feature generation module, and sending the operation information to the preference calculation module and the behavior judgment module. The feature generation module is used for extracting key features from product information of each product, calculating the occurrence rate of the key features according to the key features, and screening out product features according to the occurrence rate and the key features. For example, the key features include country, brand, vehicle series, vehicle type, seat number, gear box mode, production mode, vehicle age, mileage, price, vehicle color, etc., and the product features are features that are all present in a plurality of product information, that is, features that are necessary for characterizing the vehicle, such as product features including vehicle price, seat number, gear box mode, vehicle type, and vehicle color. The preference calculation module is used for calculating preference relation weight according to the operation behaviors, the operation objects and the product characteristics, and the database is also used for storing the preference relation weight according to the user account. When the preference relation weight is calculated, setting an initial weight, adjusting the preference relation weight corresponding to the product characteristic according to the operation behavior and the product characteristic corresponding to the operation object, for example, when a certain black vehicle is collected, the user has high interest in the vehicle, so that the product characteristic of the vehicle is obtained, and when the vehicle color is black, the weight that the vehicle color is black in the preference relation weight of the user is increased.
The server is further configured to perform a judgment according to the operation behavior and a preset ordering behavior, specifically, the behavior judgment module is preset with an ordering behavior, the behavior judgment module is configured to perform matching according to the operation behavior and the ordering behavior, when a matching item exists, that is, when an ordering behavior exists in the operation behavior, a service modification signal is generated, and when a matching item does not exist, that is, when an ordering behavior does not exist in the operation behavior, a similar modification signal is generated.
The management terminal is used for obtaining service follow-up information and uploading the service follow-up information, and the service follow-up information is information collected by the car purchasing service personnel in a process of assisting a user in purchasing a car. The service follow-up information comprises follow-up behaviors, corresponding follow-up products and object information, for example, when a user adds a certain vehicle into the initial selection during purchasing, the follow-up behaviors are the initial selection, the follow-up products are the vehicles added into the initial selection, and the object information is information collected by a vehicle purchasing service person during the purchase process, such as vehicle price. The database is also used for storing service follow-up information according to the user account.
The server is further used for acquiring service follow-up information when an order placing behavior exists in the operation behavior, correcting the preference relation weight according to the service follow-up information and the product characteristics, acquiring the to-be-matched characteristics of the current user when the order placing behavior does not exist in the operation behavior, screening the similar user group according to the to-be-matched characteristics and preset user information, and correcting the preference relation weight corresponding to the current user according to the preference relation weight in the similar user group. Specifically, the data calling module is further configured to call the service follow-up information to send to the preference correction module when the behavior determination module generates the service correction signal, and call the user information to send to the similarity screening module when the behavior determination module generates the similar correction signal. And the preference correction module is used for correcting the preference relation weight according to the follow-up behavior, the follow-up product, the object information and the product characteristics when receiving the service follow-up information. The similarity screening module is used for taking the gender, the age, the occupation and the area in the user information corresponding to the current user as the characteristics to be matched, taking the gender, the age, the occupation and the area in the user information of the other users as the user characteristics, respectively matching according to the characteristics to be matched and the user characteristics, generating the matching degree of the current user and the other users, and screening the users according to the matching degree and a preset matching degree threshold value to generate a similarity user group. For example, in this embodiment, the threshold of the matching degree is 80%, and when the matching degree between the remaining users and the current user is greater than 80%, the user is taken as a similar user of the current user, and all similar users are screened to generate a similar user group. The data calling module is also used for calling the preference relation weight corresponding to the user account in the similar user group according to the similar user group and sending the preference relation weight to the preference correction module. The preference correction module is also used for receiving the preference relation weight corresponding to the user account in the similar user group and correcting the preference relation weight of the current user according to the called preference relation weight.
The user terminal is further used for obtaining the purchase appeal, and uploading the purchase appeal and the user account. The server is further used for calling the product information of the product to be sold according to the product characteristics and generating a product recommendation list according to the preference relationship weight and the product information when the purchase appeal is obtained. Specifically, the data calling module is further configured to receive the purchase appeal and the user account, call preference relationship weight and product information according to the user account, and send the preference relationship weight and the product information to the recommendation matching module, and the recommendation matching module is configured to calculate preference degrees of products to be sold according to the preference relationship weight and the product information, perform descending order according to the preference degrees to generate a product recommendation table of the products to be sold, and feed back the product recommendation table. In other embodiments, the database is further configured to store historical recommendation information, the historical recommendation information is used to record a vehicle recommended to a user once, the server further includes a recommendation deduplication module, the recommendation matching module generates a product recommendation table as a product sorting table, the recommendation deduplication module is used to call the historical recommendation information, corresponding products in the product sorting table are removed according to the historical recommendation information, the product recommendation table is generated, and the product recommendation table is fed back.
The recommendation method for the non-high frequency consumption industry comprises the following steps:
s1: and acquiring product information and operation information, generating product characteristics according to the product information, and calculating preference relation weight according to the operation information and the product characteristics.
S2: the operation information comprises operation behaviors, judgment is carried out according to the operation behaviors and preset ordering behaviors, when the ordering behaviors exist in the operation behaviors, service follow-up information is obtained, preference relation weights are corrected according to the service follow-up information and product characteristics, when the ordering behaviors do not exist in the operation behaviors, characteristics to be matched of a current user are obtained, similar user groups are screened according to the characteristics to be matched and the preset user information, and the preference relation weights corresponding to the current user are corrected according to the preference relation weights in the similar user groups.
S3: and when the purchase appeal is acquired, calling the product information of the product to be sold according to the product characteristics, and generating a product recommendation list according to the preference relation weight and the product information.
The user registers and conducts operations of browsing vehicles, collecting vehicles, removing collections, adding primary selections, removing primary selections, purchasing budgets, loan details, appeal details and the like through the user terminal. And acquiring user information generated by user registration, and uploading the user information, wherein the user information comprises a user account, gender, age, occupation and region. The method comprises the steps of obtaining operation information generated by operations of browsing vehicles, collecting vehicles, removing vehicles, calculating budget, calculating loan and the like performed by a user on service software, and uploading the operation information and a user account. And storing the user information and storing the operation information according to the user account. The operation information includes an operation behavior and an operation object, for example, a favorite vehicle, the operation behavior is favorite, and the operation object is a favorite vehicle model.
And a manager in the non-high-frequency consumption industry uploads and manages the product information through the management terminal. And acquiring product information of the product to be sold, uploading the product information, and storing the product information. In this embodiment, the product information is description information of the vehicle to be sold. The vehicle purchasing service personnel who offer help buy uploads the service follow-up information through the management terminal, and the service follow-up information is the information collected by the vehicle purchasing service personnel in the process of assisting the user in purchasing the vehicle. And acquiring service follow-up information, uploading the service follow-up information, and storing the service follow-up information according to the user account. The service follow-up information comprises follow-up behaviors, corresponding follow-up products and object information, for example, when a user adds a certain vehicle into the initial selection during purchasing, the follow-up behaviors are the initial selection, the follow-up products are the vehicles added into the initial selection, and the object information is information collected by a vehicle purchasing service person during the purchase process, such as vehicle price.
Step S1, specifically including the steps of:
calling product information and operation information from a database, extracting key features from the product information of each product, calculating the occurrence rate of the key features according to the key features, and screening out the product features according to the occurrence rate and the key features. For example, the key features include country, brand, vehicle series, vehicle type, seat number, gear box mode, production mode, vehicle age, mileage, price, vehicle color, etc., and the product features are features that are all present in a plurality of product information, that is, features that are necessary for characterizing the vehicle, such as product features including vehicle price, seat number, gear box mode, vehicle type, and vehicle color.
And calculating preference relationship weight according to the operation behavior, the operation object and the product characteristics, and storing the preference relationship weight according to the user account. When the preference relation weight is calculated, setting an initial weight, adjusting the preference relation weight corresponding to the product characteristic according to the operation behavior and the product characteristic corresponding to the operation object, for example, when a certain black vehicle is collected, the user has high interest in the vehicle, so that the product characteristic of the vehicle is obtained, and when the vehicle color is black, the weight that the vehicle color is black in the preference relation weight of the user is increased.
Step S2, specifically including the steps of:
presetting ordering behavior, matching according to the operation behavior and the ordering behavior, generating a service correction signal when a matching item exists, namely ordering behavior exists in the operation behavior, and generating a similar correction signal when no matching item exists, namely ordering behavior does not exist in the operation behavior.
And when the service correction signal is generated, calling the service follow-up information, and correcting the preference relation weight according to the follow-up behavior, the follow-up product, the object information and the product characteristics.
And when the similar correction signal is generated, calling the user information, taking the gender, age, occupation and area in the user information corresponding to the current user as the characteristics to be matched, and taking the gender, age, occupation and area in the user information of other users as the user characteristics. And respectively matching according to the features to be matched and the user features to generate the matching degree of the current user and the rest users, and screening the users according to the matching degree and a preset matching degree threshold value to generate a similar user group. For example, in this embodiment, the threshold of the matching degree is 80%, and when the matching degree between the remaining users and the current user is greater than 80%, the user is taken as a similar user of the current user, and all similar users are screened to generate a similar user group. And calling the preference relation weight corresponding to the user account in the similar user group according to the similar user group, and correcting the preference relation weight of the current user according to the called preference relation weight.
Step S3, specifically including the steps of:
and acquiring a purchase demand, and uploading the purchase demand and the user account. Calling preference relationship weight and product information according to the user account, calculating preference degrees of products to be sold according to the preference relationship weight and the product information, performing descending order according to the preference degrees to generate a product recommendation table of the products to be sold, and feeding back the product recommendation table.
In other embodiments, historical recommendation information is stored, and the historical recommendation information is used for recording vehicles recommended to the user once, and the step S3 further includes generating a product recommendation table of products to be sold as a product sorting table according to descending sorting of the preference degrees. Calling historical recommendation information, removing corresponding products in the product sorting table according to the historical recommendation information, generating a product recommendation table, and feeding back the product recommendation table.
The foregoing is merely an example of the present invention, and common general knowledge in the field of known specific structures and characteristics is not described herein in any greater extent than that known in the art at the filing date or prior to the priority date of the application, so that those skilled in the art can now appreciate that all of the above-described techniques in this field and have the ability to apply routine experimentation before this date can be combined with one or more of the present teachings to complete and implement the present invention, and that certain typical known structures or known methods do not pose any impediments to the implementation of the present invention by those skilled in the art. It should be noted that, for those skilled in the art, without departing from the structure of the present invention, several changes and modifications can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicability of the patent. The scope of the claims of the present application shall be determined by the contents of the claims, and the description of the embodiments and the like in the specification shall be used to explain the contents of the claims.
Claims (10)
1. The recommendation system for the non-high-frequency consumption industry comprises a server, a database and a recommendation server, wherein the server is used for acquiring product information and generating product characteristics according to the product information; the server is also used for acquiring the operation information and calculating the preference relation weight according to the operation information and the product characteristics; the method is characterized in that:
the operation information comprises operation behaviors, the server is also used for judging according to the operation behaviors and preset ordering behaviors, when the ordering behaviors exist in the operation behaviors, service follow-up information is obtained, the preference relation weight is corrected according to the service follow-up information and product characteristics, when the ordering behaviors do not exist in the operation behaviors, the to-be-matched characteristics of the current user are obtained, similar user groups are screened according to the to-be-matched characteristics and the preset user information, and the preference relation weight corresponding to the current user is corrected according to the preference relation weight in the similar user groups;
the server is further used for calling the product information of the product to be sold according to the product characteristics and generating a product recommendation list according to the preference relationship weight and the product information when the purchase appeal is obtained.
2. The recommendation system for a non-high frequency consumer industry according to claim 1, wherein: the server comprises a feature generation module, wherein the feature generation module is used for extracting key features from the product information of each product, calculating the occurrence rate of the key features according to the key features and screening out the product features according to the occurrence rate and the key features.
3. The recommendation system for a non-high frequency consumer industry according to claim 1, wherein: the server comprises a preference correction module, and the preference correction module is used for correcting preference relation weight according to the follow-up behavior, the follow-up product, the object information and product characteristics.
4. The recommendation system for a non-high frequency consumer industry according to claim 1, wherein: the server comprises a similarity screening module and a preference correction module, wherein the similarity screening module is used for matching according to the characteristics to be matched and the user characteristics in the user information to generate a matching degree, and screening the user according to the matching degree and a preset matching degree threshold value to generate a similar user group; the preference correction module is used for calling the preference relation weight corresponding to the user in the similar user group and correcting the preference relation weight of the current user according to the called preference relation weight.
5. The recommendation system for a non-high frequency consumer industry according to claim 1, wherein: the server comprises a recommendation matching module and a recommendation duplication eliminating module, wherein the recommendation matching module is used for calculating the preference degree of each product to be sold according to the preference relation weight and the product information and generating a product sorting table of the product to be sold according to the preference degree; and the recommendation duplication eliminating module is used for calling the historical recommendation information, eliminating the corresponding products in the product sorting table according to the historical recommendation information and generating the product recommendation table.
6. A recommendation method for non-high frequency consumer industries, characterized by: use of a recommendation system for non-high frequency consumer industries according to any of claims 1-5, comprising the steps of:
s1: acquiring product information and operation information, generating product characteristics according to the product information, and calculating preference relation weight according to the operation information and the product characteristics;
s2: the operation information comprises operation behaviors, judgment is carried out according to the operation behaviors and preset ordering behaviors, when the ordering behaviors exist in the operation behaviors, service follow-up information is obtained, preference relation weights are corrected according to the service follow-up information and product characteristics, when the ordering behaviors do not exist in the operation behaviors, characteristics to be matched of a current user are obtained, similar user groups are screened according to the characteristics to be matched and the preset user information, and the preference relation weights corresponding to the current user are corrected according to the preference relation weights in the similar user groups;
s3: and when the purchase appeal is acquired, calling the product information of the product to be sold according to the product characteristics, and generating a product recommendation list according to the preference relation weight and the product information.
7. A recommendation method for non-high frequency consumer industries according to claim 6 wherein: in the step S1, generating the product feature from the product information includes the steps of:
extracting key features from the product information of each product, calculating the occurrence rate of the key features according to the key features, and screening out the product features according to the occurrence rate and the key features.
8. A recommendation method for non-high frequency consumer industries according to claim 6 wherein: the service follow-up information comprises follow-up behaviors and corresponding follow-up products and object information, and the step of S2 comprises the following steps of correcting preference relation weights according to the service follow-up information and the product characteristics:
and correcting the preference relation weight according to the follow-up behavior, the follow-up product, the object information and the product characteristics.
9. A recommendation method for non-high frequency consumer industries according to claim 6 wherein: in the step S2, screening a similar user group according to the features to be matched and the preset user information, and modifying the preference relationship weight corresponding to the current user according to the preference relationship weight in the similar user group includes the following steps:
matching according to the features to be matched and the user features in the user information to generate a matching degree, and screening users to generate a similar user group according to the matching degree and a preset matching degree threshold; and calling the preference relation weight corresponding to the user in the similar user group, and correcting the preference relation weight of the current user according to the called preference relation weight.
10. A recommendation method for non-high frequency consumer industries according to claim 6 wherein: in the step of S3, generating the product recommendation list according to the preference relationship weight and the product information includes the following steps:
calculating the preference degree of each product to be sold according to the preference relation weight and the product information, and generating a product sorting table of the products to be sold according to the preference degree; calling historical recommendation information, and removing corresponding products in the product sorting table according to the historical recommendation information to generate a product recommendation table.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011457658.7A CN112465603A (en) | 2020-12-10 | 2020-12-10 | Recommendation system and method for non-high frequency consumption industry |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011457658.7A CN112465603A (en) | 2020-12-10 | 2020-12-10 | Recommendation system and method for non-high frequency consumption industry |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112465603A true CN112465603A (en) | 2021-03-09 |
Family
ID=74802838
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011457658.7A Pending CN112465603A (en) | 2020-12-10 | 2020-12-10 | Recommendation system and method for non-high frequency consumption industry |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112465603A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102208086A (en) * | 2010-03-31 | 2011-10-05 | 北京邮电大学 | Field-oriented personalized intelligent recommendation system and implementation method |
| CN103914783A (en) * | 2014-04-13 | 2014-07-09 | 北京工业大学 | E-commerce website recommending method based on similarity of users |
| CN108320213A (en) * | 2018-01-31 | 2018-07-24 | 深圳春沐源控股有限公司 | Electric business Method of Commodity Recommendation and electric business Platform Server |
| CN110473040A (en) * | 2018-05-10 | 2019-11-19 | 北京三快在线科技有限公司 | A kind of Products Show method and device, electronic equipment |
| CN110648188A (en) * | 2018-06-27 | 2020-01-03 | 江苏特易信息科技有限公司 | Cross-border trade BPO service recommendation system based on user preference |
| CN111882399A (en) * | 2020-07-31 | 2020-11-03 | 平安国际融资租赁有限公司 | Service information recommendation method, device, computer system and readable storage medium |
-
2020
- 2020-12-10 CN CN202011457658.7A patent/CN112465603A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102208086A (en) * | 2010-03-31 | 2011-10-05 | 北京邮电大学 | Field-oriented personalized intelligent recommendation system and implementation method |
| CN103914783A (en) * | 2014-04-13 | 2014-07-09 | 北京工业大学 | E-commerce website recommending method based on similarity of users |
| CN108320213A (en) * | 2018-01-31 | 2018-07-24 | 深圳春沐源控股有限公司 | Electric business Method of Commodity Recommendation and electric business Platform Server |
| CN110473040A (en) * | 2018-05-10 | 2019-11-19 | 北京三快在线科技有限公司 | A kind of Products Show method and device, electronic equipment |
| CN110648188A (en) * | 2018-06-27 | 2020-01-03 | 江苏特易信息科技有限公司 | Cross-border trade BPO service recommendation system based on user preference |
| CN111882399A (en) * | 2020-07-31 | 2020-11-03 | 平安国际融资租赁有限公司 | Service information recommendation method, device, computer system and readable storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 鲁城华;高升;郭倩倩;: "基于大数据技术的云制造服务个性化智能推荐", 信息系统工程, no. 08, 20 August 2020 (2020-08-20), pages 33 - 37 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11250453B2 (en) | System and method for sales generation in conjunction with a vehicle data system | |
| US10115074B1 (en) | Predictive conversion systems and methods | |
| US10546337B2 (en) | Price scoring for vehicles using pricing model adjusted for geographic region | |
| CN106779809B (en) | Price information optimization combination method and system for big data platform | |
| US20220318858A1 (en) | Systems and methods for transformation of raw data to actionable data | |
| US10504159B2 (en) | Wholesale/trade-in pricing system, method and computer program product therefor | |
| CN108710634B (en) | Protocol file pushing method and terminal equipment | |
| US11830017B2 (en) | System and method for providing a price for a vehicle | |
| CN116452261B (en) | Advertisement delivery data processing method based on cross-border E-commerce service platform | |
| US11966933B2 (en) | System and method for correlating and enhancing data obtained from distributed sources in a network of distributed computer systems | |
| CN117151830A (en) | Commodity recommendation method and system based on big data | |
| Pratama et al. | Product recommendation in offline retail industry by using collaborative filtering | |
| CN115511582B (en) | Commodity recommendation system and method based on artificial intelligence | |
| CN112465603A (en) | Recommendation system and method for non-high frequency consumption industry | |
| CN116503092A (en) | User reservation intention recognition method and device, electronic equipment and storage medium | |
| CN110910163A (en) | Automobile distribution customer demand analysis method and system based on customer relationship management system | |
| CN113744030A (en) | Recommendation method, device, server and medium based on AI user portrait | |
| CN113486086A (en) | Data mining method and system based on feature engineering | |
| CN112819570B (en) | Intelligent commodity collocation recommendation method based on machine learning | |
| CN115375397A (en) | Business opportunity recommendation method based on combined model, electronic equipment and storage medium | |
| CN113886450A (en) | User matching method, related device, equipment and storage medium | |
| CN115526690A (en) | Commodity ordering system and method based on artificial intelligence | |
| CN112116178A (en) | Ordering method and device for offline stores | |
| CN117829965A (en) | Product recommendation system and method for cross-border e-commerce platform | |
| CN117788104A (en) | Big data-based computer information processing system and method thereof |
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 |