CN116664239A - Product recommendation method, device, equipment and medium based on artificial intelligence - Google Patents
Product recommendation method, device, equipment and medium based on artificial intelligence Download PDFInfo
- Publication number
- CN116664239A CN116664239A CN202310656206.9A CN202310656206A CN116664239A CN 116664239 A CN116664239 A CN 116664239A CN 202310656206 A CN202310656206 A CN 202310656206A CN 116664239 A CN116664239 A CN 116664239A
- Authority
- CN
- China
- Prior art keywords
- product
- data
- target
- vector
- user
- 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 62
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 39
- 238000005516 engineering process Methods 0.000 claims abstract description 14
- 239000013598 vector Substances 0.000 claims description 181
- 238000012549 training Methods 0.000 claims description 45
- 238000007781 pre-processing Methods 0.000 claims description 25
- 238000004590 computer program Methods 0.000 claims description 17
- 238000012360 testing method Methods 0.000 claims description 17
- 238000003860 storage Methods 0.000 claims description 16
- 238000012795 verification Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 14
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 230000004927 fusion Effects 0.000 claims description 10
- 238000012216 screening Methods 0.000 claims description 9
- 238000005457 optimization Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000009191 jumping Effects 0.000 claims description 5
- 239000000047 product Substances 0.000 description 394
- 238000000605 extraction Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 9
- 230000003190 augmentative effect Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000003672 processing method Methods 0.000 description 3
- 230000002776 aggregation Effects 0.000 description 2
- 238000004220 aggregation Methods 0.000 description 2
- 239000006227 byproduct Substances 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/06—Asset management; Financial planning or analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/08—Insurance
Abstract
The application relates to the technical field of artificial intelligence and financial science and technology, and discloses a product recommendation method, device, equipment and medium based on artificial intelligence, wherein the method comprises the following steps: acquiring target product searching data, target product browsing data and target product purchasing data corresponding to a target user; extracting key features of the relation between the user and the product according to the target product searching data, the target product browsing data and the target product purchasing data to obtain target feature data; and determining a product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target characteristic data. Therefore, product recommendation based on key characteristics of the relation between the user and the product is realized, the interest points of the target user on the product are fully considered, and the accuracy of the determined product recommendation result is improved.
Description
Technical Field
The application relates to the technical fields of artificial intelligence, financial science and technology and intelligent medical treatment, in particular to a product recommendation method, device, equipment and medium based on artificial intelligence.
Background
With the development of artificial intelligence, product recommendation is performed based on the artificial intelligence, so that the product selling is facilitated, and the product recommendation method is widely applied. According to the existing method for recommending the product based on the artificial intelligence, similar users corresponding to the target users are searched based on the artificial intelligence, and the products purchased by the similar users are recommended to the target users.
Disclosure of Invention
Based on the above, it is necessary to solve the technical problem that the accuracy of the recommended product is not high due to the fact that the interest point of the target user on the product is ignored in the method for recommending the product based on the artificial intelligence in the prior art, and a product recommending method, device, equipment and medium based on the artificial intelligence are provided.
In a first aspect, there is provided an artificial intelligence based product recommendation method, the method comprising:
acquiring target product searching data, target product browsing data and target product purchasing data corresponding to a target user;
extracting key features of the relation between the user and the product according to the target product searching data, the target product browsing data and the target product purchasing data to obtain target feature data;
And determining a product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target characteristic data.
In a second aspect, there is provided an artificial intelligence based product recommendation device, the device comprising:
the data acquisition module is used for acquiring target product search data, target product browsing data and target product purchase data corresponding to a target user;
the target feature data determining module is used for extracting key features of the relation between the user and the product according to the target product searching data, the target product browsing data and the target product purchasing data to obtain target feature data;
and the product recommendation result determining module is used for determining a product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target characteristic data.
In a third aspect, a computer device is provided, 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 artificial intelligence based product recommendation method described above when the computer program is executed by the processor.
In a fourth aspect, a computer readable storage medium is provided, the computer readable storage medium storing a computer program which, when executed by a processor, implements the steps of the artificial intelligence based product recommendation method described above.
According to the artificial intelligence-based product recommendation method, target product search data, target product browsing data and target product purchasing data corresponding to a target user are firstly obtained, then key feature extraction of the relation between the user and the product is carried out according to the target product search data, the target product browsing data and the target product purchasing data, target feature data is obtained, and finally a product recommendation result corresponding to the target user is determined according to a pre-trained product preference classification prediction model and the target feature data. Therefore, product recommendation based on key characteristics of the relation between the user and the product is realized, the interest points of the target user on the product are fully considered, the accuracy of the determined product recommendation result is improved, and the recommendation purchase rate is improved.
Drawings
In order to more clearly illustrate the embodiments of the application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Wherein:
FIG. 1 is an application environment diagram of an artificial intelligence based product recommendation method in one embodiment;
FIG. 2 is a flow diagram of a method of product recommendation based on artificial intelligence in one embodiment;
FIG. 3 is a block diagram of an artificial intelligence based product recommendation device in one embodiment;
FIG. 4 is a block diagram of a computer device in one embodiment;
FIG. 5 is another block diagram of a computer device in one embodiment.
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 product recommendation method based on artificial intelligence provided by the embodiment of the invention can be applied to an application environment as shown in fig. 1, wherein a client 110 communicates with a server 120 through a network. The server 120 may receive, through the client 110, a user identifier corresponding to the target user; the server 120 obtains target product search data, target product browse data and target product purchase data corresponding to the target user according to the user identifier corresponding to the target user; extracting key features of the relation between the user and the product according to the target product searching data, the target product browsing data and the target product purchasing data to obtain target feature data; and determining a product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target characteristic data. Therefore, product recommendation based on key characteristics of the relation between the user and the product is realized, the interest points of the target user on the product are fully considered, and the accuracy of the determined product recommendation result is improved. Further, the server 120 sends page resource data to the client 110 according to the product recommendation result; the client 110 performs a product recommendation page (i.e., web page) presentation based on the page resource data.
web pages, all called web pages in english, are a computer noun, referring to a file on the world wide web that is organized in HTML format.
When the application is applied to the field of financial science and technology, the product recommendation method based on artificial intelligence can be utilized for recommending the target user aiming at various insurance schemes in insurance industry, and the recommendation of the insurance scheme is performed on the basis of the key characteristics of the relationship between the user and the product, so that the interest points of the target user on the product are fully considered, and the accuracy of the recommended insurance scheme is improved.
When the application is applied to the field of financial science and technology, the product recommendation method based on artificial intelligence can be utilized for recommending the financial product to the target user based on the key characteristics of the relationship between the user and the product aiming at various financial schemes in banking industry, thereby fully considering the interest points of the target user on the financial product and improving the accuracy of the recommended financial product.
The user identifier may be a user name, a user ID, or the like, which uniquely identifies a user.
Optionally, the client 110 obtains target product search data, target product browsing data and target product purchasing data corresponding to a target user from the server 120, then obtains the target product search data, the target product browsing data and the target product purchasing data corresponding to the target user according to a user identifier corresponding to the target user, extracts key features of a relationship between the user and the product according to the target product search data, the target product browsing data and the target product purchasing data, obtains target feature data, and determines a product recommendation result corresponding to the target user according to a pre-trained product preference classification prediction model and the target feature data. Further, the client 110 performs interface display according to the product recommendation result.
Among other things, the client 110 may be, but is not limited to, various personal computers, notebook computers, smartphones, tablet computers, and portable wearable devices. The server 120 may be implemented by a stand-alone server or a server cluster formed by a plurality of servers. The present invention will be described in detail with reference to specific examples.
Referring to fig. 2, fig. 2 is a schematic flow chart of a product recommendation method based on artificial intelligence according to an embodiment of the invention, which includes the following steps:
s1: acquiring target product searching data, target product browsing data and target product purchasing data corresponding to a target user;
the target user is an object that wants to recommend a product.
The target product search data is product search data that the target user has historically. The product search data is used for storing search time, product identification, search process data and associated data corresponding to the search application record. The product identification may be data uniquely identifying a product, such as a product name, a product ID, etc. The search process data is operation data in the search process. Searching the application record includes: clicked or not clicked product detail page.
The target product browsing data is product browsing data of a target user in history. The product browsing data are used for storing browsing time, product identification, browsing process data and associated data corresponding to browsing application records. The browsing process data is operation data in the browsing process. Browsing application records includes: purchased or not purchased.
The target product purchase data is product purchase data of the target user in history. The product purchase data is used for storing association data corresponding to the purchase time and the product identification.
For example, when the application is applied to the technical field of financial science and technology, products corresponding to target product search data, target product browsing data and target product purchase data can be insurance schemes or financial products.
For example, when the application is applied to the technical field of intelligent medical treatment, products corresponding to target product search data, target product browsing data and target product purchase data can be medicines or medical appliances.
For example, when the application is applied to an online furniture mall, products corresponding to target product search data, target product browsing data and target product purchase data can be furniture.
Specifically, the target product search data, the target product browse data and the target product purchase data corresponding to the target user input by the user (that is, any user) may be obtained, or the target product search data, the target product browse data and the target product purchase data corresponding to the target user may be obtained from a preset storage space (may be a local storage space of a server where a program file implementing the present application is located or may be a cloud server), or the target product search data, the target product browse data and the target product purchase data corresponding to the target user may be obtained from a third party application (for example, an e-commerce platform).
S2: extracting key features of the relation between the user and the product according to the target product searching data, the target product browsing data and the target product purchasing data to obtain target feature data;
the key features of the relationship between the user and the product are features that are of interest to the user. The user can click on the product detail page of the searched product, purchase the browsed product and re-purchase the purchased product.
Specifically, based on artificial intelligence, key feature extraction of the relation between the user and the product is performed according to the target product search data, the target product browsing data and the target product purchase data, and the extracted data is used as target feature data. Thus obtaining the feature data of interest of the target user to the product.
Optionally, the step of extracting key features of the relationship between the user and the product according to the target product search data, the target product browsing data and the target product purchase data to obtain target feature data includes: sequentially performing vector generation and feature extraction on the associated data recorded as the clicked product detail page by the search application in the target product search data to obtain first data; sequentially carrying out vector generation and feature extraction on the related data recorded as purchase by the browsing application in the target product browsing data to obtain second data; sequentially carrying out vector generation and feature extraction on the associated data with multiple purchases in the target product purchase data to obtain third data; and splicing the first data, the second data and the third data to obtain the target characteristic data.
S3: and determining a product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target characteristic data.
Specifically, the target feature data is input into a pre-trained product preference classification prediction model to perform product preference classification prediction, and a product recommendation result corresponding to the target user is determined according to a vector obtained through product preference classification prediction.
Optionally, each vector element in the vector obtained by product preference classification prediction corresponds to a product, so that a plurality of vector elements with top ranks of the vector elements can be screened out from the vector obtained by product preference classification prediction, and all products corresponding to the screened vector elements are used as product recommendation results corresponding to the target user.
The pre-trained product preference classification prediction model is a pre-trained multi-classifier. The specific training method of the pre-trained product preference classification prediction model can be selected from the prior art, and will not be described in detail herein.
According to the method, target product search data, target product browsing data and target product purchasing data corresponding to a target user are firstly obtained, key feature extraction of the relation between the user and the product is then carried out according to the target product search data, the target product browsing data and the target product purchasing data, target feature data are obtained, and finally a product recommendation result corresponding to the target user is determined according to a pre-trained product preference classification prediction model and the target feature data. Therefore, product recommendation based on key characteristics of the relation between the user and the product is realized, the interest points of the target user on the product are fully considered, the accuracy of the determined product recommendation result is improved, and the recommendation purchase rate is improved.
In one embodiment, the step of extracting key features of the relationship between the user and the product according to the target product search data, the target product browse data and the target product purchase data to obtain target feature data includes:
s21: denoising processing and repeating data deleting processing are respectively carried out on the target product searching data, the target product browsing data and the target product purchasing data, so as to obtain preprocessing searching data, preprocessing browsing data and preprocessing purchasing data;
specifically, a preset denoising processing method is adopted to denoise the target product search data, and the data obtained by denoising processing is subjected to repeated data deleting processing to obtain preprocessed search data; denoising the target product browsing data by adopting a preset denoising processing method, and repeating data deleting processing on the data obtained by denoising processing to obtain preprocessed browsing data; and denoising the purchase data of the target product by adopting a preset denoising processing method, and repeating data deleting processing on the data obtained by denoising processing to obtain preprocessed purchase data.
The preset denoising method may be selected from the prior art, for example, one or more of statistical model, binning, clustering, and regression, which are not limited herein.
The data de-duplication process, i.e., the process of reserving only one copy of data repeatedly occurring multiple times, is performed.
S22: establishing a feature vector of a relation between a user and a product according to the preprocessing search data, the preprocessing browse data and the preprocessing purchase data to obtain a target feature vector;
specifically, feature vectors of the relation between the user and the product are established for the preprocessing search data, the preprocessing browse data and the preprocessing purchase data respectively, and all the established feature vectors are spliced to obtain the target feature vector.
It will be appreciated that in another embodiment of the present application, step S21 is not required, and step S22 is replaced with: and establishing a feature vector of a relation between a user and a product according to the target product searching data, the target product browsing data and the target product purchasing data to obtain a target feature vector.
S23: and carrying out feature engineering on the target feature vector to extract key features and obtain the target feature data.
Specifically, feature engineering is performed on the target feature vector to extract key features, and the extracted data is used as target feature data. Therefore, the characteristic data of interest of the target user to the product is obtained based on artificial intelligence.
The feature engineering, english full name is feature engineering, is to utilize domain knowledge and existing data to create new features for machine learning algorithm; may be manual or automatic. Feature engineering includes feature selection and feature transformation to extract features of interest to a target user in a product through analysis and mining of data.
According to the embodiment, the denoising processing and the repeated data deleting processing are respectively carried out on the target product searching data, the target product browsing data and the target product purchasing data, so that the quality of data for constructing the feature vector is improved, the accuracy of target feature data is improved, and the accuracy of a determined product recommendation result is further improved; and establishing a feature vector of the relation between the user and the product according to the preprocessing search data, the preprocessing browse data and the preprocessing purchase data, and carrying out feature engineering based on the feature vector to extract key features, thereby providing a basis for recommending the product based on the key features of the relation between the user and the product.
In one embodiment, before the step of determining the product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target feature data, the method further includes:
s311: acquiring an initial model, a training sample set and a test sample set, wherein the initial model sequentially comprises: an input layer, a plurality of hidden layers and an output layer;
specifically, the initial model, the training sample set and the test sample set input by the user (any user) can be obtained, the initial model, the training sample set and the test sample set can be obtained from a preset storage space, and the initial model, the training sample set and the test sample set can be obtained from a third party application.
The initial model is a model obtained based on a BP (back propagation) neural network algorithm. Hidden layers are also known as hidden layers. Each hidden layer acquires data input by an input layer and outputs the data to an output layer. The number of neurons of the input layer is equal to the number of variables of the input model, and the number of neurons of the output layer is equal to the total number of products.
The training sample set contains a plurality of training samples. The training samples include: the method comprises the steps of obtaining a key feature data sample and a sample label, wherein the data sample comprises a key feature of a relation between a user corresponding to the user and a product, and the sample label is a true value of a vector corresponding to the data sample and aiming at the product preference.
The test sample set includes a plurality of test samples. The test sample includes: key feature data samples and sample tags.
S312: the model parameter optimization method based on the gradient descent algorithm adopts the training sample set to carry out product preference classification prediction training on the initial model to obtain a model to be verified;
specifically, the training sample set is adopted to conduct product preference classification prediction training on the initial model, a gradient descent algorithm is adopted to conduct model parameter optimization on the initial model in the training process, and when a preset training ending condition is reached, the initial model is used as a model to be verified.
And aiming at the same training sample, calculating the obtained loss value according to the predicted value of the initial model and the sample label.
The preset training ending condition is that the loss value converges to a preset first value, and/or each training sample of the training sample set is trained once on the initial model.
S313: verifying the model to be verified according to the test sample set to obtain a verification result;
specifically, inputting key characteristic data samples of each test sample in the test sample set into the model to be verified for product preference classification prediction; for the same test sample, if the predicted data and the sample label are the same, determining that the single sample result corresponding to the test sample is correct, and if the predicted data and the sample label are different, determining that the single sample result corresponding to the test sample is wrong; calculating verification accuracy according to each single sample result; if the verification accuracy is larger than the preset second numerical value, determining that the verification result is passed; if the verification accuracy is smaller than or equal to the preset second value, determining that the verification result is not passed.
S314: if the verification result is not passed, taking the model to be verified as the initial model, randomly adjusting the sequence of each training sample in the training sample set, jumping to the model parameter optimization method based on the gradient descent algorithm, and training the initial model by adopting the training sample set to carry out product preference classification prediction, so that the step of obtaining the model to be verified is continuously executed;
specifically, if the verification result is not passed, it means that the model to be verified does not meet the performance requirement, so that the model to be verified is used as the initial model, and the sequence of each training sample in the training sample set is randomly adjusted to prepare for the next training; and (3) jumping to the model parameter optimization method based on the gradient descent algorithm, training the initial model by adopting the training sample set to carry out product preference classification prediction training, and continuing the step of obtaining the model to be verified, namely jumping to the step S312, and re-executing the step S312.
S315: and if the verification result is passed, taking the model to be verified as the product preference classification prediction model.
Specifically, if the verification result is that the model to be verified meets the performance requirement, the model to be verified is used as the product preference classification prediction model.
According to the embodiment, the initial model formed by the input layer, the plurality of hidden layers and the output layer is used for training the product preference classification prediction, so that a basis is provided for predicting based on the product preference classification prediction model.
In one embodiment, the step of determining the product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target feature data includes:
s321: carrying out product preference classification prediction according to the product preference classification prediction model and the target characteristic data to obtain a classification prediction vector;
specifically, the target characteristic data is input into the product preference classification prediction model to perform product preference classification prediction, and predicted obtained data is used as classification prediction vectors.
Optionally, each vector element in the classification prediction vector corresponds to a classification tag. Each class label corresponds to a product category.
S322: obtaining product searching data to be analyzed, product browsing data to be analyzed and product purchasing data to be analyzed, which correspond to each user in a preset time period;
Specifically, the product search data to be analyzed, the product browse data to be analyzed and the product purchase data to be analyzed obtained in the step are data of all users, and the current time is taken as the end time, and the data is within the preset duration.
The product search data to be analyzed is all product search data corresponding to each user within a preset time period.
The product browsing data to be analyzed is all product browsing data corresponding to each user in a preset duration.
The product purchase data to be analyzed is all product purchase data corresponding to each user within a preset time period.
S323: inputting the product searching data to be analyzed, the product browsing data to be analyzed and the product purchasing data to be analyzed into a preset product sales trend prediction model for prediction to obtain a trend prediction vector;
specifically, the product search data to be analyzed, the product browse data to be analyzed and the product purchase data to be analyzed are input into a preset product sales trend prediction model to predict future trend, and the predicted data are used as trend prediction vectors.
Optionally, the trend prediction vector includes prediction data for a plurality of time points. The predicted data for each point in time includes a probability that each category label is favored by the user.
Optionally, the trend prediction vector includes prediction data for a plurality of time points. The predicted data for each point in time includes a probability that each product label is favored by the user.
S324: and determining the product recommendation result according to the classification prediction vector and the trend prediction vector.
Specifically, product recommendation is performed according to the classification prediction vector and the trend prediction vector, so that the product recommendation results of the preference and the popular trend of the target user are considered.
Optionally, a plurality of vector elements with vector element values ranked at the front are found out from the classified prediction vectors, and all products corresponding to the found vector elements are used as a user candidate product set; a plurality of vector elements with vector element values ranked at the front are found out from the trend prediction vector, and all products corresponding to each found vector element are used as a trend candidate product set; and performing intersection calculation on the user candidate product set and the trend candidate product set, and taking the intersection obtained by calculation as the product recommendation result. And obtaining the product recommendation result which gives consideration to the preference and the popularity trend of the target user.
According to the method, the product recommendation result is determined according to the classification prediction vector and the trend prediction vector, so that the product recommendation result considering the preference and the popularity trend of the target user is achieved, the popularity trend is considered, the influence of the popularity trend on the target user is considered in advance, and the accuracy of the determined product recommendation result is further improved.
In one embodiment, the step of determining the product recommendation based on the classification prediction vector and the trend prediction vector comprises:
s32411: carrying out weighted summation on vector element values corresponding to the same classification labels on the classification prediction vector and the trend prediction vector to obtain a fusion prediction vector;
specifically, the weighted summation of vector element values corresponding to the same classification labels is carried out on the classification prediction vector and the trend prediction vector, so that the fusion of vector elements corresponding to each classification label is realized, and each vector element value in the fusion prediction vector gives consideration to the preference and the popularity trend of the target user.
S32412: n vector elements with the maximum vector element values are screened from the fusion prediction vector to be used as hit vector element sets;
specifically, N vector elements with the largest vector element values are selected from the fusion prediction vector, each selected vector element is used as a hit vector element, and all hit vector elements determined in the step are used as hit vector element sets, wherein N is an integer greater than 0.
S32413: and recommending products according to the classification labels corresponding to the hit vector element sets, and obtaining the product recommendation results.
Specifically, obtaining a product recommendation package corresponding to each classification label corresponding to the hit vector element set from a preset product recommendation pool; and carrying out aggregation and de-duplication treatment on all the product recommendation packages corresponding to the hit vector element set to obtain the product recommendation result.
According to the method, the weighted summation of vector element values corresponding to the same classification labels is carried out on the classification prediction vectors and the trend prediction vectors, so that correction of the classification prediction vectors by adopting the trend prediction vectors is achieved, and each product in the product recommendation result gives consideration to the preference and the popularity trend of the target user through the correction.
In one embodiment, the step of determining the product recommendation result according to the classification prediction vector and the trend prediction vector further comprises:
s32421: screening M vector elements with the maximum vector element values from the classified prediction vectors to be used as a first vector element set;
specifically, M vector elements with the largest vector element values are screened from the classified prediction vectors, and all the screened vector elements are used as a first vector element set, wherein M is an integer greater than 0.
S32422: recommending products according to each classification label corresponding to the first vector element set to obtain a first product set;
specifically, obtaining a product recommendation package corresponding to each classification label corresponding to the first vector element set from a preset product recommendation pool; and carrying out aggregation and de-duplication treatment on all the product recommendation packages corresponding to the first vector element set to obtain a first product set.
S32423: acquiring data in a prediction time period from the trend prediction vector to obtain a first prediction vector;
specifically, the current time is taken as the starting time, data in the prediction duration is obtained from the trend prediction vector, and the obtained data is taken as a first prediction vector.
S32424: carrying out maximum value screening of each classification label on the first prediction vector to obtain a second prediction vector;
specifically, maximum value screening of each classification label is carried out on the first prediction vector, and all screened vector element values are used as second prediction vectors.
S32425: k vector elements with the maximum vector element values are screened out from the second prediction vector and used as a second vector element set;
specifically, K vector elements with the largest vector element values are screened out from the second prediction vector, and all the screened vector elements are used as a second vector element set, wherein K is an integer greater than 0.
S32426: taking products corresponding to the product labels corresponding to the second vector element set as a second product set;
specifically, each vector element in the trend prediction vector corresponds to one product tag, and therefore, products corresponding to the respective product tags corresponding to the second vector element set are taken as a second product set.
S32427: and performing intersection calculation on the first product set and the second product set to obtain the product recommendation result.
Specifically, intersection calculation is performed on the first product set and the second product set, and the calculated intersection is used as the product recommendation result.
According to the method, the first product set is determined according to the classification prediction vector, the second product set is determined according to the trend prediction vector, and then the intersection of the first product set and the second product set is used as the product recommendation result, so that the preference and the popularity trend of the target user are considered, the influence of the popularity trend on the target user is considered in advance due to the consideration of the popularity trend, and the accuracy of the determined product recommendation result is further improved.
In one embodiment, after the step of determining the product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target feature data, the method further includes:
s4: and carrying out product display on the product recommendation result based on the AR technology.
The AR technology, the english language of AR is generally called Augmented Reality, the chinese language is generally called augmented reality technology, is a technology for calculating the position and angle of a camera image in real time and adding a corresponding image, is a new technology for integrating real world information and virtual world information in a "seamless" manner, and aims to fit the virtual world around the real world and interact with each other on a screen.
Specifically, based on the AR technology, each product in the product recommendation result is subjected to augmented reality display, and the augmented reality display is used for providing an augmented reality shopping experience; the user can watch the display result of the product through the AR equipment, so that the shopping experience of the user is improved, and the recommended purchasing rate is improved.
According to the embodiment, based on the AR technology, the product recommendation result is displayed, the augmented reality shopping experience is provided, the shopping experience of a user is improved, and the recommended purchasing rate is improved.
Referring to FIG. 3, in one embodiment, an artificial intelligence based product recommendation apparatus is provided, the apparatus comprising:
the data acquisition module 801 is configured to acquire target product search data, target product browsing data, and target product purchase data corresponding to a target user;
the target feature data determining module 802 is configured to perform key feature extraction of a relationship between a user and a product according to the target product search data, the target product browse data, and the target product purchase data, so as to obtain target feature data;
and a product recommendation result determining module 803, configured to determine a product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target feature data.
In one embodiment, the target feature data determination module 802 includes:
the preprocessing sub-module is used for respectively carrying out denoising processing and repeated data deleting processing on the target product search data, the target product browsing data and the target product purchasing data to obtain preprocessing search data, preprocessing browsing data and preprocessing purchasing data;
the feature vector establishing sub-module is used for establishing a feature vector of a relation between a user and a product according to the preprocessing search data, the preprocessing browse data and the preprocessing purchase data to obtain a target feature vector;
And the feature extraction sub-module is used for carrying out feature engineering on the target feature vector to extract key features and obtain the target feature data.
In one embodiment, the apparatus further comprises: a model training module;
the model training module is used for acquiring an initial model, a training sample set and a test sample set, wherein the initial model sequentially comprises: an input layer, a plurality of hidden layers and an output layer; the model parameter optimization method based on the gradient descent algorithm adopts the training sample set to carry out product preference classification prediction training on the initial model to obtain a model to be verified; verifying the model to be verified according to the test sample set to obtain a verification result; if the verification result is not passed, taking the model to be verified as the initial model, randomly adjusting the sequence of each training sample in the training sample set, jumping to the model parameter optimization method based on the gradient descent algorithm, and training the initial model by adopting the training sample set to carry out product preference classification prediction, so that the step of obtaining the model to be verified is continuously executed; and if the verification result is passed, taking the model to be verified as the product preference classification prediction model.
In one embodiment, the product recommendation result determining module 803 includes:
the classification prediction sub-module is used for carrying out product preference classification prediction according to the product preference classification prediction model and the target characteristic data to obtain a classification prediction vector;
the trend prediction sub-module is used for acquiring product search data to be analyzed, product browsing data to be analyzed and product purchase data to be analyzed corresponding to each user within a preset duration, inputting the product search data to be analyzed, the product browsing data to be analyzed and the product purchase data to be analyzed into a preset product sales trend prediction model for prediction, and obtaining a trend prediction vector;
and the product recommendation result determining sub-module is used for determining the product recommendation result according to the classification prediction vector and the trend prediction vector.
In one embodiment, the product recommendation result determining submodule includes:
the fusion unit is used for carrying out weighted summation on vector element values corresponding to the same classification labels on the classification prediction vector and the trend prediction vector to obtain a fusion prediction vector;
the hit vector element set determining unit is used for screening N vector elements with the maximum vector element values from the fusion prediction vector to be used as hit vector element sets;
And the product recommending unit is used for recommending the product according to each classification label corresponding to the hit vector element set to obtain the product recommending result.
In one embodiment, the product recommendation result determination sub-module further comprises:
the first product set determining unit is used for screening M vector elements with the maximum vector element values from the classification prediction vectors, and taking the M vector elements as a first vector element set, and recommending products according to each classification label corresponding to the first vector element set to obtain a first product set;
the second product set determining unit is used for obtaining data in the prediction duration from the trend prediction vector to obtain a first prediction vector, carrying out maximum value screening on each classification label on the first prediction vector to obtain a second prediction vector, screening K vector elements with the maximum vector element values from the second prediction vector to serve as a second vector element set, and taking products corresponding to each product label corresponding to the second vector element set as a second product set;
and the intersection calculating unit is used for carrying out intersection calculation on the first product set and the second product set to obtain the product recommendation result.
In one embodiment, the apparatus further comprises:
and the display module is used for displaying the product on the basis of the AR technology.
In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 4. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes non-volatile and/or volatile storage media and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface of the computer device is for communicating with an external client via a network connection. The computer program, when executed by a processor, performs functions or steps of a server side of an artificial intelligence based product recommendation method.
In one embodiment, a computer device is provided, which may be a client, the internal structure of which may be as shown in FIG. 5. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface of the computer device is for communicating with an external server via a network connection. The computer program is executed by a processor to perform functions or steps of a client side of an artificial intelligence based product recommendation method.
In one embodiment, a computer device is provided that includes a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the following steps when executing the computer program:
acquiring target product searching data, target product browsing data and target product purchasing data corresponding to a target user;
extracting key features of the relation between the user and the product according to the target product searching data, the target product browsing data and the target product purchasing data to obtain target feature data;
and determining a product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target characteristic data.
According to the method, target product search data, target product browsing data and target product purchasing data corresponding to a target user are firstly obtained, key feature extraction of the relation between the user and the product is then carried out according to the target product search data, the target product browsing data and the target product purchasing data, target feature data are obtained, and finally a product recommendation result corresponding to the target user is determined according to a pre-trained product preference classification prediction model and the target feature data. Therefore, product recommendation based on key characteristics of the relation between the user and the product is realized, the interest points of the target user on the product are fully considered, the accuracy of the determined product recommendation result is improved, and the recommendation purchase rate is improved.
In one embodiment, a computer-readable storage medium is provided, having stored thereon a computer program which, when executed by a processor, performs the steps of:
acquiring target product searching data, target product browsing data and target product purchasing data corresponding to a target user;
extracting key features of the relation between the user and the product according to the target product searching data, the target product browsing data and the target product purchasing data to obtain target feature data;
and determining a product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target characteristic data.
According to the method, target product search data, target product browsing data and target product purchasing data corresponding to a target user are firstly obtained, key feature extraction of the relation between the user and the product is then carried out according to the target product search data, the target product browsing data and the target product purchasing data, target feature data are obtained, and finally a product recommendation result corresponding to the target user is determined according to a pre-trained product preference classification prediction model and the target feature data. Therefore, product recommendation based on key characteristics of the relation between the user and the product is realized, the interest points of the target user on the product are fully considered, the accuracy of the determined product recommendation result is improved, and the recommendation purchase rate is improved.
It should be noted that, the functions or steps implemented by the computer readable storage medium or the computer device may correspond to the relevant descriptions of the server side and the client side in the foregoing method embodiments, and are not described herein for avoiding repetition.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), memory bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules according to needs, i.e. the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-described functions.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention, and are intended to be included in the scope of the present invention.
Claims (10)
1. A method of product recommendation based on artificial intelligence, the method comprising:
acquiring target product searching data, target product browsing data and target product purchasing data corresponding to a target user;
Extracting key features of the relation between the user and the product according to the target product searching data, the target product browsing data and the target product purchasing data to obtain target feature data;
and determining a product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target characteristic data.
2. The method for recommending products based on artificial intelligence according to claim 1, wherein the step of extracting key features of a relationship between a user and a product according to the target product search data, the target product browsing data, and the target product purchase data to obtain target feature data comprises the steps of:
denoising processing and repeating data deleting processing are respectively carried out on the target product searching data, the target product browsing data and the target product purchasing data, so as to obtain preprocessing searching data, preprocessing browsing data and preprocessing purchasing data;
establishing a feature vector of a relation between a user and a product according to the preprocessing search data, the preprocessing browse data and the preprocessing purchase data to obtain a target feature vector;
And carrying out feature engineering on the target feature vector to extract key features and obtain the target feature data.
3. The artificial intelligence based product recommendation method according to claim 1, wherein before the step of determining the product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target feature data, further comprising:
acquiring an initial model, a training sample set and a test sample set, wherein the initial model sequentially comprises: an input layer, a plurality of hidden layers and an output layer;
the model parameter optimization method based on the gradient descent algorithm adopts the training sample set to carry out product preference classification prediction training on the initial model to obtain a model to be verified;
verifying the model to be verified according to the test sample set to obtain a verification result;
if the verification result is not passed, taking the model to be verified as the initial model, randomly adjusting the sequence of each training sample in the training sample set, jumping to the model parameter optimization method based on the gradient descent algorithm, and training the initial model by adopting the training sample set to carry out product preference classification prediction, so that the step of obtaining the model to be verified is continuously executed;
And if the verification result is passed, taking the model to be verified as the product preference classification prediction model.
4. The artificial intelligence based product recommendation method according to claim 1, wherein the step of determining the product recommendation result corresponding to the target user based on the pre-trained product preference classification prediction model and the target feature data comprises:
carrying out product preference classification prediction according to the product preference classification prediction model and the target characteristic data to obtain a classification prediction vector;
obtaining product searching data to be analyzed, product browsing data to be analyzed and product purchasing data to be analyzed, which correspond to each user in a preset time period;
inputting the product searching data to be analyzed, the product browsing data to be analyzed and the product purchasing data to be analyzed into a preset product sales trend prediction model for prediction to obtain a trend prediction vector;
and determining the product recommendation result according to the classification prediction vector and the trend prediction vector.
5. The artificial intelligence based product recommendation method according to claim 4, wherein the step of determining the product recommendation result based on the classification prediction vector and the trend prediction vector comprises:
Carrying out weighted summation on vector element values corresponding to the same classification labels on the classification prediction vector and the trend prediction vector to obtain a fusion prediction vector;
n vector elements with the maximum vector element values are screened from the fusion prediction vector to be used as hit vector element sets;
and recommending products according to the classification labels corresponding to the hit vector element sets, and obtaining the product recommendation results.
6. The artificial intelligence based product recommendation method according to claim 4, wherein the step of determining the product recommendation result based on the classification prediction vector and the trend prediction vector further comprises:
screening M vector elements with the maximum vector element values from the classified prediction vectors to be used as a first vector element set;
recommending products according to each classification label corresponding to the first vector element set to obtain a first product set;
acquiring data in a prediction time period from the trend prediction vector to obtain a first prediction vector;
carrying out maximum value screening of each classification label on the first prediction vector to obtain a second prediction vector;
k vector elements with the maximum vector element values are screened out from the second prediction vector and used as a second vector element set;
Taking products corresponding to the product labels corresponding to the second vector element set as a second product set;
and performing intersection calculation on the first product set and the second product set to obtain the product recommendation result.
7. The artificial intelligence based product recommendation method according to claim 1, wherein after the step of determining the product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target feature data, further comprising:
and carrying out product display on the product recommendation result based on the AR technology.
8. An artificial intelligence based product recommendation device, the device comprising:
the data acquisition module is used for acquiring target product search data, target product browsing data and target product purchase data corresponding to a target user;
the target feature data determining module is used for extracting key features of the relation between the user and the product according to the target product searching data, the target product browsing data and the target product purchasing data to obtain target feature data;
and the product recommendation result determining module is used for determining a product recommendation result corresponding to the target user according to the pre-trained product preference classification prediction model and the target characteristic data.
9. A 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 artificial intelligence based product recommendation method according to any one of claims 1 to 7 when the computer program is executed.
10. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the steps of the artificial intelligence based product recommendation method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310656206.9A CN116664239A (en) | 2023-06-02 | 2023-06-02 | Product recommendation method, device, equipment and medium based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310656206.9A CN116664239A (en) | 2023-06-02 | 2023-06-02 | Product recommendation method, device, equipment and medium based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116664239A true CN116664239A (en) | 2023-08-29 |
Family
ID=87722029
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310656206.9A Pending CN116664239A (en) | 2023-06-02 | 2023-06-02 | Product recommendation method, device, equipment and medium based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116664239A (en) |
-
2023
- 2023-06-02 CN CN202310656206.9A patent/CN116664239A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111784455B (en) | Article recommendation method and recommendation equipment | |
| US20160124996A1 (en) | Image ranking based on attribute correlation | |
| US11550856B2 (en) | Artificial intelligence for product data extraction | |
| CN110019943B (en) | Video recommendation method and device, electronic equipment and storage medium | |
| CN110008397B (en) | Recommendation model training method and device | |
| WO2018014109A1 (en) | System and method for analyzing and searching for features associated with objects | |
| CN110110233B (en) | Information processing method, device, medium and computing equipment | |
| US20210390609A1 (en) | System and method for e-commerce recommendations | |
| KR102326744B1 (en) | Control method, device and program of user participation keyword selection system | |
| US11677693B2 (en) | Systems and methods for generating dynamic conversational responses using ensemble prediction based on a plurality of machine learning models | |
| CN110633421A (en) | Feature extraction, recommendation, and prediction methods, devices, media, and apparatuses | |
| CN111680213B (en) | Information recommendation method, data processing method and device | |
| CN113420203A (en) | Object recommendation method and device, electronic equipment and storage medium | |
| CN110659701B (en) | Information processing method, information processing apparatus, electronic device, and medium | |
| CN112330442A (en) | Modeling method and device based on ultra-long behavior sequence, terminal and storage medium | |
| CN112288279A (en) | Business risk assessment method and device based on natural language processing and linear regression | |
| CN116664239A (en) | Product recommendation method, device, equipment and medium based on artificial intelligence | |
| CN115222112A (en) | Behavior prediction method, behavior prediction model generation method and electronic equipment | |
| CN117011737A (en) | Video classification method and device, electronic equipment and storage medium | |
| CN114579867A (en) | Resource recommendation method and device, electronic equipment and storage medium | |
| CN111915339A (en) | Data processing method, device and equipment | |
| CN112925972B (en) | Information pushing method, device, electronic equipment and storage medium | |
| CN112445992B (en) | Information processing method and device | |
| CN116129181A (en) | Display information pushing method and device, electronic equipment and storage medium | |
| CN116881554A (en) | Medical prescription recommendation method and device, electronic equipment and readable 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 |