CN110704726B - Data pushing method based on neural network and related equipment thereof - Google Patents
Data pushing method based on neural network and related equipment thereof Download PDFInfo
- Publication number
- CN110704726B CN110704726B CN201910763035.3A CN201910763035A CN110704726B CN 110704726 B CN110704726 B CN 110704726B CN 201910763035 A CN201910763035 A CN 201910763035A CN 110704726 B CN110704726 B CN 110704726B
- Authority
- CN
- China
- Prior art keywords
- user
- information
- preset
- basic
- feature
- 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.)
- Active
Links
Classifications
-
- 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
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Feedback Control In General (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the technical field of artificial intelligence, and provides a data pushing method based on a neural network and related equipment thereof, wherein the data pushing method based on the neural network comprises the following steps: when the fact that the user behavior of the operation user meets the preset condition is detected, acquiring keywords; performing priority level matching on the keywords to obtain corresponding priority levels; matching the keywords carrying the priority level with the description information in the user collection library, and selecting the user collection with the same matching as the type of the data to be identified; importing the data type to be identified into a pre-trained vendor recommendation model for identification to obtain a recommended vendor carrying vendor information; and sending the information of the recommended vehicle manufacturer to an operation user, and sending the basic information of the user to the recommended vehicle manufacturer. According to the technical scheme, data pushing is automatically carried out on the vehicle manufacturer and the operation user according to the keywords, manual intervention is avoided, and therefore data pushing efficiency and accuracy are improved.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a data pushing method based on a neural network and related equipment thereof.
Background
The data pushing method between the user and the vehicle merchant on the market at present is single, is mainly random pushing, has weak pertinence, and cannot be combined with the self demand of the user to push, so that the pushing accuracy is low, the situation of inaccurate pushing exists, the vehicle merchant information meeting the self demand cannot be accurately obtained by the user when the user needs to purchase, the vehicle merchant cannot accurately obtain the user meeting the demand, and the vehicle merchant cannot accurately push proper vehicle information for the user.
Disclosure of Invention
The embodiment of the invention provides a data pushing method based on a neural network and related equipment thereof, which are used for solving the problem that users and vendors cannot accurately acquire information meeting the requirements of the users and vendors in a random data pushing mode.
A data pushing method based on a neural network comprises the following steps:
when the fact that the user behavior of the operation user meets the preset condition is detected in the car selling webpage, keywords in the car selling webpage are obtained, wherein the operation user comprises user basic information;
performing priority level matching on the keywords to obtain priority levels corresponding to the keywords;
matching the keywords carrying the priority level with the description information in the user collection library, and selecting a user collection corresponding to the description information with the same matching as the type of the data to be identified;
Importing the data type to be identified into a pre-trained vendor recommendation model for identification to obtain a recommended vendor carrying vendor information;
establishing an association relation between the recommended vehicle merchant and the user basic information, sending the vehicle merchant information of the recommended vehicle merchant to the operation user, and sending the user basic information to the recommended vehicle merchant.
A data pushing device based on between a user and a vehicle merchant, comprising:
the first acquisition module is used for acquiring keywords in the car-selling webpage when the fact that the user behavior of the operating user meets the preset condition is detected in the car-selling webpage, wherein the operating user comprises user basic information;
the priority level determining module is used for carrying out priority level matching on the keywords to obtain priority levels corresponding to the keywords;
the first matching module is used for matching the keywords carrying the priority level with the description information in the user collection library, and selecting a user collection corresponding to the description information with the same matching as the type of the data to be identified;
the identification module is used for importing the data type to be identified into a pre-trained vehicle manufacturer recommendation model to identify, so as to obtain a recommended vehicle manufacturer carrying vehicle manufacturer information;
And the sending module is used for establishing an association relation between the recommended vehicle manufacturer and the user basic information, sending the vehicle manufacturer information of the recommended vehicle manufacturer to the operating user, and sending the user basic information to the recommended vehicle manufacturer.
A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the neural network based data pushing method described above when the computer program is executed.
A computer readable storage medium storing a computer program which, when executed by a processor, implements the steps of the neural network-based data push method described above.
According to the data pushing method based on the neural network and the related equipment, the obtained keywords are subjected to priority level matching to obtain the priority levels corresponding to the keywords, the keywords carrying the priority levels are matched with the description information in the user collection library, the data types to be identified are obtained, the data types to be identified are imported into a pre-trained vendor recommendation model, the recommended vendors carrying the vendor information are output, the association relation between the recommended vendors and the user basic information is established, the vendor information of the recommended vendors is sent to the operation user, and the user basic information is sent to the recommended vendors. Therefore, data pushing is automatically carried out on the vehicle operators and the operation users according to the keywords, basic information of the users can be quickly and accurately sent to the proper recommended vehicle operators, and information of the vehicle operators is sent to the operation users, so that manual intervention is avoided, the efficiency and the accuracy of data pushing can be effectively improved, and the searching efficiency and the accuracy of the operation users and the pushing accuracy of the vehicle operators are further improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments of the present invention will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a data pushing method based on a neural network according to an embodiment of the present invention;
fig. 2 is a flowchart of step S1 in a data pushing method based on a neural network according to an embodiment of the present invention;
fig. 3 is a flowchart of step S12 in a data pushing method based on a neural network according to an embodiment of the present invention;
fig. 4 is a flowchart of combining multidimensional features and composite features as basic features in a data pushing method based on a neural network according to an embodiment of the present invention;
fig. 5 is a flowchart of step S71 in a data pushing method based on a neural network according to an embodiment of the present invention;
fig. 6 is a flowchart of step S2 in a data pushing method based on a neural network according to an embodiment of the present invention;
fig. 7 is a schematic diagram of a data pushing device provided by an embodiment of the present invention, where the data pushing device is based on a user and a vendor;
Fig. 8 is a block diagram of the basic mechanism of a computer device according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The data processing method provided by the application is applied to the server, and the server can be realized by an independent server or a server cluster formed by a plurality of servers. In an embodiment, as shown in fig. 1, a data pushing method based on a neural network is provided, including the following steps:
s1: when the fact that the user behavior of the operation user meets the preset condition is detected in the car-selling webpage, keywords in the car-selling webpage are obtained, wherein the operation user comprises user basic information.
In an embodiment of the present invention, the user behavior may include, but is not limited to, a user browsing a web page on a computer device, which may be a smart phone, a notebook, a palm computer, or an advertiser; the preset condition may be that the duration of time the user browses the car-selling web pages reaches a set threshold value or the number of links of the browsed cars reaches a set threshold value, etc. The user basic information includes user personal information and user location information.
Specifically, when the fact that the duration of the user behavior of the operation user for browsing the vehicle-selling webpage reaches a set threshold or the number of vehicle-selling links reaches the set threshold is detected, keywords in the vehicle-selling webpage are obtained from a preset database. The preset database is a database specially used for storing keywords in a car selling webpage browsed by an operation user.
It should be noted that, before selecting the cardiology vehicle product, the operation user without explicit intention may compare some types of vehicle-selling web page contents in the vehicle-selling web page for a long time, or browse various vehicle-selling links in a short time to select the cardiology vehicle product, etc., so that such user behavior may be used as the operation user with intention to purchase vehicles, and by identifying the user behavior, some unrelated users may be effectively screened out, thereby improving the mining efficiency of the operation user.
S2: and carrying out priority level matching on the keywords to obtain the priority levels corresponding to the keywords.
Specifically, the frequency of occurrence corresponding to each keyword is obtained, and the priority corresponding to each keyword is obtained by matching the frequency of occurrence corresponding to each keyword with the frequency interval corresponding to the priority.
S3: and matching the keywords carrying the priority level with the description information in the user collection library, and selecting a user collection corresponding to the description information with the same matching as the type of the data to be identified.
In the embodiment of the invention, the keywords carrying the priority level refer to keywords containing the priority level, for example, class A brands. Different user sets and description information corresponding to the user sets are stored in the user set library in advance, wherein the user sets comprise at least 2 keywords with priority levels preset by a user. For example, there are user sets of "class a brands, class B price levels, class C X cities", and their corresponding descriptive information is "class a brands, class B price levels, class C X cities".
Specifically, matching the keywords carrying the priority with the description information in the user collection library, and if the keywords carrying the priority are matched with the description information in the user collection library, taking the user collection corresponding to the description information as the data type to be identified.
For example, keywords carrying priority levels are respectively: grade a brands, grade B price levels; the user set inventory stores the description information corresponding to the user set Q1 as the 'A-level brand and the B-level brand', the description information corresponding to the user set Q2 is the 'A-level brand and the B-level brand', keywords carrying priority levels are matched with the description information, and the keywords carrying the priority levels are the same as the description information corresponding to the user set Q1, so that the user set Q1 is used as the type of data to be identified.
S4: and importing the data type to be identified into a pre-trained vendor recommendation model for identification to obtain a recommended vendor carrying vendor information.
In the embodiment of the invention, the vehicle manufacturer recommendation model refers to a convolutional neural network model for matching a recommended vehicle manufacturer. The pre-trained driver recommendation model can rapidly and accurately output the recommended driver carrying the driver information corresponding to the data type to be identified according to the input data type to be identified. The information of the vehicle manufacturer comprises recommended vehicle types, network selling links of the vehicle manufacturer, vehicle manufacturer addresses, vehicle manufacturer contact ways and the like.
Specifically, the data type to be identified is directly imported into a pre-trained vehicle manufacturer recommendation model, and the vehicle manufacturer recommendation model can rapidly and accurately judge the recommended vehicle manufacturer corresponding to the data type to be identified and carrying vehicle manufacturer information according to the input data type to be identified and output the recommended vehicle manufacturer.
Further, when the recommendation models of the vehicle operators identify the recommendation vehicle operators with a plurality of different vehicle operator addresses, further screening is carried out according to the user position information in the user basic information, and if the description information corresponding to the user set serving as the data type to be identified does not contain the position information, the identified recommendation vehicle operators are randomly selected and output; and if the description information corresponding to the user set serving as the data type to be identified contains the position information, selecting a recommended driver of the driver address closest to the user position information for output.
It should be noted that, in the training process of the vehicle manufacturer recommendation model, the convolutional neural network model is initialized to obtain an initial model, then sample data prepared in advance for training is imported into the initial model to calculate forward output, then a prediction error between the forward output and a preset target value is calculated, and finally initial parameters of each network layer in the initial model are adjusted by using an error back propagation algorithm according to the prediction error, so as to obtain the vehicle manufacturer recommendation model.
S5: and establishing an association relation between the recommended vehicle manufacturer and the user basic information, sending the vehicle manufacturer information of the recommended vehicle manufacturer to an operation user, and sending the user basic information to the recommended vehicle manufacturer.
In the embodiment of the invention, the service end can execute information pushing according to the association relation by establishing the association relation between the recommended vehicle manufacturer and the user basic information, and the accuracy and the efficiency of the information pushing are improved. The information of the recommended vehicle manufacturer is sent to the operation user, so that the operation user can be helped to know the information of the vehicle manufacturer suitable for the operation user in time, and the success rate of purchasing vehicles of the operation user is improved; the basic information of the user is sent to the recommended vehicle manufacturer, so that the recommended vehicle manufacturer can be helped to make a vehicle selling service guide for an operating user in advance, and the selling success rate of the recommended vehicle manufacturer is improved.
Specifically, according to the recommended vendor obtained in step S4, an association relationship is established between the recommended vendor and the user basic information, the vendor information of the recommended vendor is sent to the operating user according to a preset pushing mode, and the user basic information is sent to the recommended vendor according to the preset pushing mode. The preset pushing mode can be specifically in the form of a short message, and can also be set according to the actual requirement of a user.
In this embodiment, the priority level corresponding to the keyword is obtained by performing priority level matching on the obtained keyword, the keyword carrying the priority level is matched with the description information in the user collection library, the data type to be identified is obtained, the data type to be identified is imported into a pre-trained vendor recommendation model, a vendor carrying vendor information is output, an association relationship between the vendor and user basic information is established, vendor information of the vendor is sent to an operation user, and user basic information is sent to the vendor. Therefore, data pushing is automatically carried out on the vehicle operators and the operation users according to the keywords, basic information of the users can be quickly and accurately sent to the proper recommended vehicle operators, and information of the vehicle operators is sent to the operation users, so that manual intervention is avoided, the efficiency and the accuracy of data pushing can be effectively improved, and the searching efficiency and the accuracy of the operation users and the pushing accuracy of the vehicle operators are further improved.
In an embodiment, as shown in fig. 2, in step S1, that is, when it is detected in the car-selling web page that the user behavior of the operating user satisfies the preset condition, the step of obtaining the keyword in the car-selling web page includes the following steps:
s11: and when the user behavior of the operating user is detected to meet the preset condition to be detected, acquiring the vehicle purchase intention information of the user.
Specifically, when the user behavior of the operating user is detected to meet the preset condition to be detected, acquiring user shopping intention information of the operating user from a preset intention library, wherein the preset intention library is a database specially used for storing the user shopping intention information selected by the operating user when the operating user browses a vehicle selling webpage.
The user's intention information is the intention information of purchasing the vehicle, which is queried by the operating user on the web page of the selling vehicle, and the intention information of purchasing the vehicle may be, but is not limited to, budget of purchasing the vehicle, distance of 4s store, intention brand preference, vehicle type preference, service desire, etc.
S12: and cleaning data of the purchase intention information of the user to obtain basic characteristics.
In the embodiment of the invention, the user purchase intention information has invalid data irrelevant to subsequent calculation, so that the user purchase intention information is required to be subjected to data cleaning, the invalid data is removed, and the invalid data is prevented from being subjected to redundant calculation, thereby improving the efficiency of the subsequent calculation. Basic features typically include branding, price, offers, and the like.
Specifically, data cleaning is performed on the user purchasing intention information according to preset filtering conditions, deleting processing is performed on the user purchasing intention information meeting the preset filtering conditions, the user purchasing intention information which does not meet the preset filtering conditions is reserved, and the reserved user purchasing intention information is used as a basic feature. The preset filtering conditions may be, but not limited to, a browsing time of the operation user on the vehicle-selling web page, a browsing content of the vehicle-selling web page, vehicle-purchasing intention information of the user filled in the vehicle-selling web page, and the like, and the specific preset filtering conditions may be set according to actual conditions of specific services.
S13: and matching the basic characteristics with preset characteristics.
Specifically, the basic features are matched with preset features, wherein the preset features can be brand offers, price levels and the like.
S14: if the basic feature is the same as the preset feature, the basic feature which is the same as the preset feature is determined as the target feature.
In the embodiment of the present invention, the basic feature is matched with the preset feature according to step S13, and if the basic feature is the same as the preset feature, the basic feature is determined as the target feature.
For example, if the basic features are "brand information", "selling price range" and "comfort level", respectively, the preset features are "brand information", and the "brand information", "selling price range" and "comfort level" are respectively matched with "brand information", so as to obtain "brand information" as the target feature.
S15: and integrating all the target features to obtain keywords.
Specifically, according to the target features obtained in step S14, integrating all the target features according to a preset rule to obtain the keywords after integrating. The preset rule refers to a set rule for synthesizing the target features according to the actual demands of the user.
For example, if the target features are brand a and brand B, the keyword obtained by integrating the target features according to the preset rule is brand a of brand B.
In this embodiment, the obtained user intention information is subjected to data cleaning to obtain basic features, the basic features are matched with preset features to obtain target features, and finally the target features are integrated to obtain keywords. Therefore, the corresponding keywords can be accurately extracted according to the intention information of the user to purchase the vehicle, the accuracy of the keywords to be used subsequently is guaranteed, and the accuracy of the follow-up data pushing is further guaranteed.
In one embodiment, as shown in fig. 3, in step S12, data cleaning is performed on the user intention information of purchasing the vehicle, and the obtaining basic features includes the following steps:
s121: and matching the user intention information with preset filtering conditions.
Specifically, the purchase intention information is matched with preset filtering conditions.
S122: if the user intention information is the same as the preset filtering condition, deleting the user intention information.
In the embodiment of the present invention, matching is performed according to the purchase intention information and the preset filtering condition in step S121, if the matching result is that the purchase intention information of the user is identical to the preset filtering condition, the purchase intention information of the user meets the preset filtering condition, that is, the qualification of filtering is achieved, so that the purchase intention information of the user identical to the preset filtering condition is deleted.
S123: if the user intention information is different from the preset filtering condition, determining the user intention information as a basic characteristic.
Specifically, the matching is performed according to the purchase intention information and the preset filtering condition in step S121, if the matching result is that the purchase intention information of the user is different from the preset filtering condition, the user does not meet the preset filtering condition, that is, the qualification of filtering is not achieved, so that the purchase intention information of the user which is different from the preset filtering condition is reserved, and the reserved purchase intention information of the user is determined as the basic feature.
In this embodiment, the user purchase intention information is filtered by matching the user purchase intention information with a preset filtering condition, and the retained user purchase intention information is used as a basic feature. Therefore, effective screening of the user shopping intention information is achieved, the extracted basic features are ensured to be accurate, and the accuracy of follow-up data pushing and the accuracy of operation user searching are further ensured.
In this embodiment, after step S13, the data pushing method based on the neural network further includes the following steps:
s6: if the basic feature is different from the preset feature, replacing the basic feature with the preset feature, and taking the replaced basic feature as the target feature.
In the embodiment of the invention, the basic feature is matched with the preset feature according to the step S13, if the basic feature is different from the preset feature as the result of the matching, the basic feature is imported into a preset replacement library for replacement processing, the basic feature after the replacement processing is obtained, and the basic feature is used as the target feature.
When the basic features are detected by the preset substitution library, judging whether basic features different from the preset features exist in all the basic features, and if so, deleting the basic features; meanwhile, judging whether preset features different from the basic features exist in all the preset features, and if the preset features different from the basic features exist, generating the basic features identical to the preset features; the basic characteristics are subjected to replacement processing through a preset replacement library, and then the basic characteristics identical to the preset characteristics are obtained.
The preset replacement library is a database specially used for carrying out replacement processing on basic features.
For example, if the basic features are respectively: A. b, C, D, E, the preset characteristics are: C. d, E, F, G, importing the basic features into a preset replacement library, judging whether basic features different from the preset features exist in all the basic features when the basic features are detected by the preset replacement library, and deleting the basic features A and B if the basic features A and B are different from the preset features; meanwhile, judging whether preset features different from the basic features exist in all the preset features or not to obtain preset features F and preset features G which are different from the basic features, and generating basic features identical to the preset features F and the preset features G to obtain the basic features F and the basic features G; finally, the basic characteristics are imported into a preset replacement library for replacement processing, and the obtained basic characteristics are respectively as follows: C. d, E, F, G.
In this embodiment, when the basic feature is different from the preset feature, the basic feature is replaced with the preset feature, and the replaced basic feature is used as the target feature. Therefore, the further processing of the basic features is realized under the condition that the basic features are different from the preset features, and the accuracy of obtaining the target features is ensured.
In an embodiment, the basic features include a user feature and a non-user feature, as shown in fig. 4, after step S12 and before step S13, the data pushing method based on the neural network further includes the following steps:
s71: and carrying out dimension lifting processing on the user characteristics to obtain multi-dimensional characteristics.
In the embodiment of the invention, the user characteristic refers to characteristic information related to the user in the vehicle-selling webpage, such as vehicle purchase intention information of the user, browsing content of the user in the vehicle-selling webpage and the like. Because the user features belong to low-dimensional features, the pre-trained vehicle merchant recommendation model is not suitable for recognizing the low-dimensional features, if the low-dimensional features are directly used for being imported into the pre-trained vehicle merchant recommendation model for recognition, the recognition effect of the pre-trained vehicle merchant recommendation model can be influenced, and the recognition accuracy is affected, so that the user features need to be processed into the multi-dimensional features in order to improve the accuracy.
Specifically, user characteristics are obtained from a user characteristic library, and the user characteristics are imported into a preset dimension-increasing port for dimension-increasing processing, so that the multidimensional characteristics after dimension-increasing processing are obtained. The preset dimension increasing port is a processing port specially used for processing the user characteristics into the multidimensional characteristics. A user profile library refers to a database dedicated to storing user profiles.
S72: and carrying out data combination on the non-user features according to the preset related features to obtain the composite features.
In the embodiment of the invention, the non-user features refer to feature information which is irrelevant to users in the vehicle-selling web pages, such as preferential tendency, service grading, service period matching, price grade grading and the like.
Specifically, non-user features are obtained from a non-user feature library, all the non-user features are matched with preset related features, the non-user features matched with the preset related features are selected to be imported into a preset combination port for data combination, and the composite features after data combination are obtained.
A non-user feature library refers to a database dedicated to storing non-user features.
The preset relevant features refer to relevant features for confirming user service requirement setting, and the preset non-user feature library comprises non-user features identical to the preset relevant features.
The preset combination port refers to an execution port specially used for data combination of the non-user features.
S73: the multidimensional feature and the composite feature are combined as basic features.
Specifically, the multidimensional feature in step S71 and the composite feature in step S72 are combined to obtain a combined base feature. For example, the multi-dimensional feature J and the composite feature K are combined, and the resulting base feature comprises J and K.
In this embodiment, the user features are processed into multi-dimensional features, the non-user features are processed into composite features, and the multi-dimensional features and the composite features are combined as basic features. Therefore, the preprocessing of the user features and the non-user features is realized, the preprocessed features are combined as basic features, the data types of the subsequent basic features can be ensured to be suitable for the vehicle manufacturer recommendation model, and the accuracy of recognition of the vehicle manufacturer recommendation model is ensured.
In one embodiment, as shown in fig. 5, in step S71, the step of performing the dimension-increasing process on the user feature to obtain the multi-dimensional feature includes the following steps:
s711: and carrying out nonlinear transformation on the user characteristics by using a preset kernel function to obtain nonlinear characteristics.
In the embodiment of the invention, the nonlinear characteristics after nonlinear transformation processing are obtained by directly utilizing the preset kernel function to perform nonlinear transformation on the user characteristics.
Wherein the preset kernel function refers to a function dedicated to performing a nonlinear transformation on the user characteristics.
S712: and performing coding operation on the nonlinear characteristics to obtain multidimensional characteristics.
Specifically, the encoding operation is performed on the nonlinear feature obtained in step S711 according to the preset encoding mode, a dummy variable of the nonlinear feature is obtained by performing the encoding operation, and the dummy variable is set to a corresponding multidimensional vector, that is, the set multidimensional vector is a multidimensional feature. The preset encoding mode may specifically be one-hot encoding (one-hot), or may be set according to the actual requirement of the user.
Preferably, the preset coding mode is adopted in the proposal to be one-hot coding.
Note that, one-hot coding is also called one-hot coding and one-bit efficient coding. The method is to encode N states using N-bit state registers, each with its own register bit, and at any time only one of the bits is valid. For example, assuming we have four samples, three for each sample, and assuming a user feature has 4 state values, we use 4 state bits to represent the feature, one-hot encoding ensures that only 1 bit is in state 1 for a single feature in each sample, and the others are all 0. Dummy variable coding refers to any removal of one state bit, for example, when 4 state bits are known to a certain user feature, for example, the values of the identities of the user positions are respectively housewives, white collars, workers, farmers and individual users, and when the first four state bits [0, 0] are used, the user positions can be expressed as individual users. Simply because for one sample of the study he is neither a housewife, white collar, worker, farmer, then it may default to an individual household.
In this embodiment, the user features are subjected to nonlinear transformation to obtain nonlinear features, and then the nonlinear features are subjected to encoding operation to obtain multidimensional features. Therefore, the user characteristics are processed into the multidimensional characteristics, the multidimensional characteristics are ensured to be suitable for the application of the follow-up vehicle manufacturer recommendation model, and the accuracy of recognition of the vehicle manufacturer recommendation model is further improved.
In an embodiment, as shown in fig. 6, in step S2, performing priority matching on the keywords to obtain the priority corresponding to the keywords includes the following steps:
s21: and obtaining the occurrence frequency of the keywords from the historical data table.
In the embodiment of the invention, according to the keywords obtained in the step S1, each keyword is matched with the description information in the historical data table, and if the matching is successful, the frequency corresponding to the description information which is successfully matched is obtained from the historical data table as the occurrence frequency of the keywords. The historical data table is a database which is specially used for storing description information and the occurrence frequency of the description information corresponding to the description information, and the description information which corresponds to the keywords exists in the historical data table.
S22: and matching the occurrence frequency with a frequency interval in a preset frequency library, and taking the priority corresponding to the frequency interval as the priority corresponding to the keyword if the occurrence frequency is in the frequency interval.
Specifically, according to the occurrence frequency of the keyword obtained in step S21, the occurrence frequency of the keyword is matched with a frequency interval in a preset frequency library, if the occurrence frequency of the keyword exists in the frequency interval, the occurrence frequency of the keyword is matched with the frequency interval, and the priority level corresponding to the frequency interval is used as the priority level corresponding to the keyword. The preset frequency library is a database specially used for storing frequency intervals and priority levels corresponding to the frequency intervals.
For example, if the frequency intervals in the preset frequency library are respectively [1, 30), [30, 100), and [100, + ], the corresponding priority levels are respectively one, two, three, etc.; the occurrence frequency of the keyword is 25, the occurrence frequency 25 is matched with each frequency interval, and the occurrence frequency 25 is obtained in the frequency interval [1, 30 ], so that the priority class corresponding to the frequency interval [1, 30) is used as the priority class corresponding to the keyword, namely the priority class corresponding to the keyword is equal.
In this embodiment, the priority level corresponding to the keyword is obtained by matching the frequency of occurrence of the keyword with the frequency interval. Therefore, the priority level corresponding to the keyword can be obtained rapidly and accurately, and the accuracy of matching the keyword carrying the priority level with the description information is ensured.
It should be understood that the sequence number of each step in the foregoing embodiment does not mean that the execution sequence of each process should be determined by the function and the internal logic, and should not limit the implementation process of the embodiment of the present invention.
In an embodiment, an apparatus is provided, where the data pushing apparatus based on the user and the vendor corresponds to the data pushing method based on the neural network in the foregoing embodiment one by one. As shown in fig. 7, the data pushing device based on the data between the user and the vehicle manufacturer includes a first acquisition module 71, a priority level determination module 72, a first matching module 73, an identification module 74 and a transmission module 75. The functional modules are described in detail as follows:
A first obtaining module 71, configured to obtain keywords in a vehicle-selling webpage when it is detected in the vehicle-selling webpage that a user behavior of an operating user meets a preset condition, where the operating user includes user basic information;
the priority level determining module 72 is configured to perform priority level matching on the keywords to obtain priority levels corresponding to the keywords;
a first matching module 73, configured to match the keyword carrying the priority level with the description information in the user set library, and select a user set corresponding to the description information with the same matching level as the type of the data to be identified;
the identification module 74 is configured to import the data type to be identified into a pre-trained vendor recommendation model for identification, so as to obtain a recommended vendor carrying vendor information;
and the sending module 75 is used for establishing an association relationship between the recommended vehicle manufacturer and the user basic information, sending the vehicle manufacturer information of the recommended vehicle manufacturer to the operating user, and sending the user basic information to the recommended vehicle manufacturer.
Further, the first acquisition module 71 includes:
the second acquisition sub-module is used for acquiring the user shopping intention information when detecting that the user behavior of the operating user meets the preset condition to be detected;
The data cleaning sub-module is used for cleaning data of the intention information of the purchasing of the user to obtain basic characteristics;
the second matching sub-module is used for matching the basic characteristics with preset characteristics;
the target feature determination submodule is used for determining the basic feature which is the same as the preset feature as the target feature if the basic feature is the same as the preset feature;
and the integration processing sub-module is used for integrating all the target features to obtain keywords.
Further, the data cleaning submodule includes:
the third matching unit is used for matching the user shopping intention information with preset filtering conditions;
the third matching same unit is used for deleting the user purchase intention information if the user purchase intention information is the same as the preset filtering condition;
and the third matching different units are used for determining the purchase intention information of the user as basic characteristics if the purchase intention information of the user is different from the preset filtering conditions.
Further, the data pushing device based on the user and the vehicle manufacturer further comprises:
and the second matching different module is used for replacing the basic feature with the preset feature if the basic feature is different from the preset feature, and taking the replaced basic feature as the target feature.
Further, the data pushing device based on the user and the vehicle manufacturer further comprises:
the dimension lifting module is used for carrying out dimension lifting processing on the user characteristics to obtain multidimensional characteristics;
the data combination module is used for carrying out data combination on the non-user characteristics according to the preset related characteristics to obtain composite characteristics;
and the combination module is used for combining the multidimensional feature and the composite feature as basic features.
Further, the dimension increasing module includes:
the nonlinear transformation submodule is used for carrying out nonlinear transformation on the user characteristics by utilizing a preset kernel function to obtain nonlinear characteristics;
and the coding sub-module is used for executing coding operation on the nonlinear characteristics to obtain multidimensional characteristics.
Further, the priority level determination module 72 includes:
the third acquisition sub-module is used for acquiring the occurrence frequency of the keywords from the historical data table;
and the fourth matching sub-module is used for matching the occurrence frequency with the frequency interval in the preset frequency library, and if the occurrence frequency is in the frequency interval, the priority corresponding to the frequency interval is used as the priority corresponding to the keyword.
Some embodiments of the present application disclose a computer device. Referring specifically to FIG. 8, a basic block diagram of a computer device 90 in one embodiment of the present application is shown.
As illustrated in fig. 8, the computer device 90 includes a memory 91, a processor 92, and a network interface 93 communicatively coupled to each other via a system bus. It should be noted that only computer device 90 having components 91-93 is shown in FIG. 8, but it should be understood that not all of the illustrated components need be implemented and that more or fewer components may alternatively be implemented. It will be appreciated by those skilled in the art that the computer device herein is a device capable of automatically performing numerical calculations and/or information processing in accordance with predetermined or stored instructions, the hardware of which includes, but is not limited to, microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASICs), programmable gate arrays (fields-Programmable Gate Array, FPGAs), digital processors (Digital Signal Processor, DSPs), embedded devices, etc.
The computer equipment can be a desktop computer, a notebook computer, a palm computer, a cloud server and other computing equipment. The computer equipment can perform man-machine interaction with a user through a keyboard, a mouse, a remote controller, a touch pad or voice control equipment and the like.
The memory 91 includes at least one type of readable storage medium including flash memory, hard disk, multimedia card, card memory (e.g., SD or DX memory, etc.), random Access Memory (RAM), static Random Access Memory (SRAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), programmable Read Only Memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the storage 91 may be an internal storage unit of the computer device 90, such as a hard disk or a memory of the computer device 90. In other embodiments, the memory 91 may also be an external storage device of the computer device 90, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card) or the like, which are provided on the computer device 90. Of course, the memory 91 may also include both an internal memory unit and an external memory device of the computer device 90. In this embodiment, the memory 91 is generally used to store an operating system and various application software installed on the computer device 90, such as program codes of the data pushing method based on a neural network. Further, the memory 91 may be used to temporarily store various types of data that have been output or are to be output.
The processor 92 may be a central processing unit (Central Processing Unit, CPU), controller, microcontroller, microprocessor, or other data processing chip in some embodiments. The processor 92 is generally used to control the overall operation of the computer device 90. In this embodiment, the processor 92 is configured to execute a program code stored in the memory 91 or process data, for example, a program code of the data pushing method based on a neural network.
The network interface 93 may include a wireless network interface or a wired network interface, the network interface 93 typically being used to establish communication connections between the computer device 90 and other electronic devices.
The present application also provides another embodiment, namely, a computer readable storage medium, where a recommended vendor information input program is stored, where the recommended vendor information input program is executable by at least one processor, so that the at least one processor performs the steps of any of the data pushing methods based on the neural network.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk), including several instructions for causing a computer device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method described in the embodiments of the present application.
Finally, it should be noted that the above-described embodiments are merely some, but not all, embodiments of the present application, and that the preferred embodiments of the present application are shown in the drawings and do not limit the scope of the patent. This application may be embodied in many different forms, but rather, embodiments are provided in order to provide a more thorough understanding of the present disclosure. Although the present application has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that modifications may be made to the embodiments described in the foregoing, or equivalents may be substituted for elements thereof. All equivalent structures made by the specification and the drawings of the application are directly or indirectly applied to other related technical fields, and are also within the protection scope of the application.
Claims (7)
1. The data pushing method based on the neural network is characterized by comprising the following steps of:
when the fact that the user behavior of the operation user meets the preset condition is detected in the car selling webpage, keywords in the car selling webpage are obtained, wherein the operation user comprises user basic information;
Performing priority level matching on the keywords to obtain priority levels corresponding to the keywords;
matching the keywords carrying the priority level with the description information in the user collection library, and selecting a user collection corresponding to the description information with the same matching as the type of the data to be identified;
importing the data type to be identified into a pre-trained vendor recommendation model for identification to obtain a recommended vendor carrying vendor information;
establishing an association relationship between the recommended vehicle merchant and the user basic information, sending the vehicle merchant information of the recommended vehicle merchant to the operation user, and sending the user basic information to the recommended vehicle merchant;
when the user behavior of the operating user is detected to meet the preset condition in the car-selling webpage, the step of acquiring the keywords in the car-selling webpage comprises the following steps:
when the user behavior of the operating user is detected to meet the preset condition to be detected, acquiring the vehicle purchase intention information of the user;
data cleaning is carried out on the user shopping intention information to obtain basic characteristics;
matching the basic features with preset features;
if the basic feature is the same as the preset feature, determining the basic feature which is the same as the preset feature as a target feature;
Integrating all the target features to obtain the keywords;
the step of cleaning the data of the user purchasing intention information to obtain basic characteristics comprises the following steps:
matching the user shopping intention information with preset filtering conditions;
if the user purchase intention information is the same as the preset filtering condition, deleting the user purchase intention information;
if the user intention information is different from the preset filtering condition, determining the user intention information as a basic characteristic;
the step of performing priority level matching on the keywords to obtain the priority levels corresponding to the keywords comprises the following steps:
acquiring the occurrence frequency of the keywords from a historical data table;
and matching the occurrence frequency with a frequency interval in a preset frequency library, and taking a priority corresponding to the frequency interval as a priority corresponding to the keyword if the occurrence frequency is in the frequency interval.
2. The neural network-based data pushing method of claim 1, wherein after the matching of the basic feature with the preset feature, the neural network-based data pushing method further comprises:
And if the basic feature is different from the preset feature, replacing the basic feature with the preset feature, and taking the replaced basic feature as a target feature.
3. The neural network-based data pushing method as claimed in claim 1, wherein the basic features include user features and non-user features, the data cleaning is performed on the user intention information to obtain the basic features, and before the matching is performed between the basic features and the preset features, the neural network-based data pushing method further comprises:
performing dimension lifting processing on the user characteristics to obtain multidimensional characteristics;
carrying out data combination on the non-user features according to preset related features to obtain composite features;
the multi-dimensional feature and the composite feature are combined as the base feature.
4. The data pushing method based on a neural network as claimed in claim 3, wherein the step of performing the up-scaling process on the user feature to obtain the multi-dimensional feature includes:
carrying out nonlinear transformation on the user characteristics by using a preset kernel function to obtain nonlinear characteristics;
and executing coding operation on the nonlinear characteristic to obtain the multidimensional characteristic.
5. A neural network based data pushing device, wherein the neural network based data pushing device is configured to implement the neural network based data pushing method according to any one of claims 1 to 4, the neural network based data pushing device comprising:
the first acquisition module is used for acquiring keywords in the car-selling webpage when the fact that the user behavior of the operating user meets the preset condition is detected in the car-selling webpage, wherein the operating user comprises user basic information;
the priority level determining module is used for carrying out priority level matching on the keywords to obtain priority levels corresponding to the keywords;
the first matching module is used for matching the keywords carrying the priority level with the description information in the user collection library, and selecting a user collection corresponding to the description information with the same matching as the type of the data to be identified;
the identification module is used for importing the data type to be identified into a pre-trained vehicle manufacturer recommendation model to identify, so as to obtain a recommended vehicle manufacturer carrying vehicle manufacturer information;
and the sending module is used for establishing an association relation between the recommended vehicle manufacturer and the user basic information, sending the vehicle manufacturer information of the recommended vehicle manufacturer to the operating user, and sending the user basic information to the recommended vehicle manufacturer.
6. Computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the neural network based data push method according to any of claims 1 to 4 when the computer program is executed.
7. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the steps of the neural network based data push method of any of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910763035.3A CN110704726B (en) | 2019-08-19 | 2019-08-19 | Data pushing method based on neural network and related equipment thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910763035.3A CN110704726B (en) | 2019-08-19 | 2019-08-19 | Data pushing method based on neural network and related equipment thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110704726A CN110704726A (en) | 2020-01-17 |
| CN110704726B true CN110704726B (en) | 2023-07-28 |
Family
ID=69193221
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910763035.3A Active CN110704726B (en) | 2019-08-19 | 2019-08-19 | Data pushing method based on neural network and related equipment thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110704726B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104933172A (en) * | 2015-06-30 | 2015-09-23 | 百度在线网络技术（北京）有限公司 | Information pushing method and device based on user searching behavior |
| US9531827B1 (en) * | 2011-06-14 | 2016-12-27 | Urban Airship, Inc. | Push notification delivery system with feedback analysis |
| CN107657480A (en) * | 2017-09-08 | 2018-02-02 | 广州悦鸿方生物科技有限公司 | Based on the system for selling the progress information gathering of equipment user's information and/or push |
| CN108710634A (en) * | 2018-04-08 | 2018-10-26 | 平安科技（深圳）有限公司 | A kind of method for pushing and terminal device of document of agreement |
-
2019
- 2019-08-19 CN CN201910763035.3A patent/CN110704726B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9531827B1 (en) * | 2011-06-14 | 2016-12-27 | Urban Airship, Inc. | Push notification delivery system with feedback analysis |
| CN104933172A (en) * | 2015-06-30 | 2015-09-23 | 百度在线网络技术（北京）有限公司 | Information pushing method and device based on user searching behavior |
| CN107657480A (en) * | 2017-09-08 | 2018-02-02 | 广州悦鸿方生物科技有限公司 | Based on the system for selling the progress information gathering of equipment user's information and/or push |
| CN108710634A (en) * | 2018-04-08 | 2018-10-26 | 平安科技（深圳）有限公司 | A kind of method for pushing and terminal device of document of agreement |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110704726A (en) | 2020-01-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110825957B (en) | Deep learning-based information recommendation method, device, equipment and storage medium | |
| WO2019169756A1 (en) | Product recommendation method and apparatus, and storage medium | |
| CN111798273A (en) | Training method of purchase probability prediction model of product and purchase probability prediction method | |
| WO2019061994A1 (en) | Electronic device, insurance product recommendation method and system, and computer readable storage medium | |
| CN106919625B (en) | Internet user attribute identification method and device | |
| CN111459993B (en) | Configuration updating method, device, equipment and storage medium based on behavior analysis | |
| CN108492138B (en) | Product purchase prediction method, server and storage medium | |
| CN113220657B (en) | Data processing method and device and computer equipment | |
| CN110147502B (en) | Product recommendation method, device, medium and server based on big data analysis | |
| CN116402539B (en) | Method, system, equipment and storage medium for guest development based on big data | |
| CN111680165A (en) | Information matching method and device, readable storage medium and electronic equipment | |
| JP2011227721A (en) | Interest extraction device, interest extraction method, and interest extraction program | |
| CN111882399A (en) | Service information recommendation method, device, computer system and readable storage medium | |
| CN111858922A (en) | Service side information query method and device, electronic equipment and storage medium | |
| CN111680213B (en) | Information recommendation method, data processing method and device | |
| CN111311381A (en) | Commodity recommendation method and system | |
| CN108595580B (en) | News recommendation method, device, server and storage medium | |
| CN111444447A (en) | Content recommendation page display method and device | |
| CN110515929B (en) | Book display method, computing device and storage medium | |
| CN110704726B (en) | Data pushing method based on neural network and related equipment thereof | |
| CN111767457A (en) | Recommendation method and device | |
| CN116168403A (en) | Medical data classification model training method, classification method, device and related medium | |
| KR100888329B1 (en) | System and method for automatically detecting information in real-time using rule | |
| CN111049894B (en) | Cloud Internet of things data processing method and device and electronic equipment | |
| CN111401395B (en) | Data processing method, terminal equipment and storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |