CN113076488A - Method and system for recommending information based on user data - Google Patents
Method and system for recommending information based on user data Download PDFInfo
- Publication number
- CN113076488A CN113076488A CN202110621987.9A CN202110621987A CN113076488A CN 113076488 A CN113076488 A CN 113076488A CN 202110621987 A CN202110621987 A CN 202110621987A CN 113076488 A CN113076488 A CN 113076488A
- Authority
- CN
- China
- Prior art keywords
- matrix
- neural network
- text
- budget
- characteristic vector
- 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
Images
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/12—Use of codes for handling textual entities
- G06F40/126—Character encoding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a method and a system for recommending information based on user data, wherein the method comprises the following steps: acquiring a text carrying user data, analyzing the text through a preset keyword, and acquiring a budget characteristic vector and a property characteristic vector; performing fusion processing on the budget characteristic vector and the character characteristic vector to obtain a fusion vector; loading the fusion vectors corresponding to a plurality of users into the feature matrix respectively; inputting the characteristic matrix into a first neural network for training to obtain a recommendation model; predicting a text to be processed through the recommendation model to obtain recommendation information; the sentences in the text associated with the keywords are determined by using the keywords, redundant sentences in the text are filtered, and the budget characteristic vector and the property characteristic vector are subjected to two-dimensional fusion and analysis, so that comprehensive weight consideration can be performed by combining the budget condition and the property characteristics of the user, and the precision and the personalization level of recommendation information are improved.
Description
Technical Field
The invention relates to the technical field of big data, in particular to a method and a system for recommending information based on user data.
Background
Text bearing user data often contains deep feature information, distinguished, differentiated and personalized user figures can be formed by summarizing and fusing the feature information, potential requirements of users can be predicted through the user figures, and information recommendation is carried out.
At present, the discreteness and the coupling of user data are inconvenient for accurately identifying and analyzing the user data, and key information cannot be accurately extracted, so that the problem of low accuracy of recommended information is caused.
Disclosure of Invention
In view of the above-mentioned shortcomings of the prior art, the present invention provides a method and system for recommending information based on user data, which is used to solve the problem of low accuracy of the recommended information in the prior art.
To achieve the above and other related objects, the present invention provides a method for recommending information based on user data, comprising:
acquiring a text carrying user data, analyzing the text through a preset keyword, and acquiring a budget characteristic vector and a property characteristic vector;
performing fusion processing on the budget characteristic vector and the character characteristic vector to obtain a fusion vector;
loading the fusion vectors corresponding to a plurality of users into the feature matrix respectively;
inputting the characteristic matrix into a first neural network for training to obtain a recommendation model;
and predicting the text to be processed through the recommendation model to obtain recommendation information.
Optionally, the step of analyzing the text by using a preset keyword includes:
the preset keywords comprise a first keyword and a second keyword, wherein the first keyword is associated with budget information of the user, and the second keyword is associated with property information of the user;
determining a first sentence in the text, which is associated with budget information, through the first keyword, and determining a second sentence in the text, which is associated with property information, through the second keyword;
and vectorizing the first statement and the second statement respectively to obtain the budget characteristic vector and the property characteristic vector.
Optionally, the vectorizing the first statement and the second statement respectively, and the step of obtaining the budget feature vector and the trait feature vector includes:
acquiring codes of words of the first sentence and the second sentence through a corpus;
respectively acquiring a first sentence matrix and a second sentence matrix through the coding of the words of the first sentence and the second sentence;
inputting the first statement matrix and the second statement matrix into a second neural network respectively;
convolving the first statement matrix and the second statement matrix respectively through convolution kernels with one or more sizes to obtain feature maps with one or more sizes;
and respectively performing pooling treatment on the feature maps with one or more sizes and splicing to obtain the budget feature vector and the property feature vector.
Optionally, the second neural network comprises: a second sub-neural network for processing the first statement matrix and a second sub-neural network for processing the second statement matrix;
the second sub-neural network and the second sub-neural network each include: an input layer, a convolutional layer, a pooling layer, and an output layer.
Optionally, the first neural network includes an input layer, a fully-connected layer, and an output layer, and the activation function of the output layer includes a sigmoid function.
Optionally, the step of inputting the feature matrix into a first neural network for training, and obtaining a recommendation model includes:
acquiring the feature matrix and a preset recommendation information template;
inputting the characteristic matrix into the first neural network to obtain a predicted value, and enabling the predicted value to correspond to the recommendation information template;
and through iterative training of the first neural network, when the corresponding accuracy or recall rate of the predicted value and the recommendation information template reaches or exceeds a set value, acquiring a recommendation model.
Optionally, the recommendation information template includes a consumption level sub-template and an intervention effect sub-template, where m kinds of the consumption level sub-templates, n kinds of the intervention effect sub-templates, and m × n kinds of the recommendation information templates, where m and n are positive integers.
A system for recommending information based on user data, comprising:
the preprocessing module is used for acquiring a text carrying user data, analyzing the text through a preset keyword to acquire a budget characteristic vector and a character characteristic vector, and performing fusion processing on the budget characteristic vector and the character characteristic vector to acquire a fusion vector;
the model module is used for loading fusion vectors corresponding to a plurality of users into a feature matrix respectively, inputting the feature matrix into a first neural network for training, and acquiring a recommendation model;
the processing module is used for predicting the text to be processed through the recommendation model and acquiring recommendation information;
the preprocessing module, the model module and the processing module are in signal connection.
An electronic device, comprising:
one or more processors; and
one or more machine-readable media having instructions stored thereon that, when executed by the one or more processors, cause the electronic device to perform any of the methods.
A machine-readable medium having stored thereon instructions, which when executed by one or more processors, cause an apparatus to perform any of the methods described.
As described above, the method and system for recommending information based on user data according to the present invention have the following advantages:
the sentences in the text associated with the keywords are determined by using the keywords, redundant sentences in the text are filtered, the data processing amount of the sentences in the text is reduced, the processing efficiency is improved, the interference of the sentences in the text with low correlation degree on the recommendation information is avoided, two-dimensional fusion and analysis are performed on the budget characteristic vector and the property characteristic vector, comprehensive weight consideration can be performed by combining the budget condition and the property characteristic of the user, and the precision and the individuation level of the recommendation information are improved.
Drawings
Fig. 1 is a schematic diagram illustrating a method for recommending information based on user data according to an embodiment of the present invention.
Fig. 2 is a flow chart illustrating a fusion process according to an embodiment of the present invention.
Fig. 3 is a schematic structural diagram of a second sub-neural network according to an embodiment of the present invention.
Fig. 4 is a schematic structural diagram of a first neural network according to an embodiment of the present invention.
Fig. 5 is a schematic diagram illustrating a network structure of a second sub-neural network according to an embodiment of the present invention.
Fig. 6 is a schematic diagram of a system for recommending information based on user data according to an embodiment of the present invention.
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention.
It should be noted that the drawings provided in the present embodiment are only for illustrating the basic idea of the present invention, and the components related to the present invention are only shown in the drawings rather than drawn according to the number, shape and size of the components in actual implementation, and the form, quantity and proportion of each component in actual implementation may be changed freely, and the layout of the components may be more complicated. The structures, proportions, sizes, and other dimensions shown in the drawings and described in the specification are for understanding and reading the present disclosure, and are not intended to limit the scope of the present disclosure, which is defined in the claims, and are not essential to the art, and any structural modifications, changes in proportions, or adjustments in size, which do not affect the efficacy and attainment of the same are intended to fall within the scope of the present disclosure. In addition, the terms "upper", "lower", "left", "right", "middle" and "one" used in the present specification are for clarity of description, and are not intended to limit the scope of the present invention, and the relative relationship between the terms and the terms is not to be construed as a scope of the present invention.
The information recommendation technology is found to be capable of reducing the opportunity cost of obtaining useful information by a user, but accurate information recommendation is often established on the basis of more accurate user analysis and user feature identification, and at present, deeper information cannot be obtained by processing and analyzing user data, so that the accuracy of recommended information is poor.
Referring to fig. 1, the present invention provides a method for recommending information based on user data, including:
s1: the method comprises the steps of obtaining a text carrying user data, analyzing the text through preset keywords, obtaining budget characteristic vectors and property characteristic vectors, obtaining the text through an interactive question-answer process of a user and a system, filling the text when the user registers the system, filling a text of related information in advance in a process of applying for user data recommendation information, enabling the text to correspond to each user, enabling each text to carry the characteristic information belonging to the user, enabling semantics conveyed by the characteristic information to comprise information related to user budget and information related to user property characteristics, enabling the budget characteristics to comprise but not limited to information such as income, academic calendar, social insurance, medical insurance and insurance, and enabling the property characteristics to comprise but not limited to: physical health condition, physical examination data, mental state and other properties;
s2: the budget characteristic vector and the character characteristic vector are subjected to fusion processing to obtain a fusion vector, and through two-dimensional fusion and analysis of the budget characteristic vector and the character characteristic vector, comprehensive weight consideration can be carried out by combining the budget condition and the character characteristics of the user, so that the precision and the personalization level of recommendation information are improved;
s3: loading the fusion vectors corresponding to the users into the feature matrix respectively, and performing matrixing on the fusion vectors of the users to achieve the purpose of obtaining a basic data set;
s4: inputting the characteristic matrix into a first neural network for training to obtain a recommendation model, processing the characteristic matrix through the first neural network, and obtaining a preferred training model as the recommendation model in iterative training;
s5: and predicting the text to be processed through the recommendation model to obtain recommendation information. In the embodiment, the sentences in the text associated with the keywords are determined by using the keywords, and redundant sentences in the text are filtered, so that the data processing amount of the sentences in the text needing to be processed is reduced, the processing efficiency can be improved, the interference of the sentences in the text with low correlation degree on the noise information generated by the recommendation information is avoided, the budget characteristic vector and the property characteristic vector are subjected to two-dimensional fusion and analysis, comprehensive weight consideration can be performed by combining the budget condition and the property characteristics of the user, and the precision and the personalized level of the recommendation information are improved.
In some implementations, the budget feature vector 21 and the trait feature vector 22 are fused, for example, the budget feature vector 21 and the trait feature vector 22 are fused and dimension-increased to obtain a fusion vector, a feature matrix 3 is obtained through the fusion vector, the feature matrix 3 is input into the first neural network 10 for training, a predicted value of the first neural network 10 is obtained by setting a learning rate and a number of iterations, an accuracy and a recall rate of the predicted value are observed, and when the accuracy or the recall rate reaches a set value, a recommendation model is obtained, please refer to fig. 2. The first neural network includes an input layer 110, a fully-connected layer 120, and an output layer 130, and the activation function of the output layer 130 includes a sigmoid function, and through multiple iterative training, the weight of each neural unit in the fully-connected layer 120 can be adjusted, and a training parameter that can approximate to a preferred training model is obtained, please refer to fig. 4.
In order to improve the processing efficiency of a text, a set of keywords is set, a sentence in the text associated with budget or a property is determined by the set of keywords, and the sentence with higher association degree with budget information or property information is processed, so that the processing quantity of redundant sentences is reduced, and the processing efficiency is improved, specifically, the step of analyzing the text by the preset keywords comprises the following steps of:
the preset keywords comprise a first keyword and a second keyword, wherein the first keyword is associated with budget information of the user, and the second keyword is associated with property information of the user;
determining a first sentence in the text, which is associated with budget information, through the first keyword, and determining a second sentence in the text, which is associated with property information, through the second keyword;
and vectorizing the first statement and the second statement respectively to obtain the budget characteristic vector and the property characteristic vector.
Further, the step of vectorizing the first statement and the second statement respectively to obtain the budget feature vector and the trait feature vector includes:
obtaining the codes of the words of the first sentence and the second sentence through a corpus, which exemplarily illustrates that the corpus can rely on the existing corpus, obtain the codes of the words in the first sentence and the second sentence through the corpus, and encode the sentences and the words, for example, convert each word into a 5-bit code, and for example, convert the first sentence and the second sentence into a matrix of w × 5, where w is a positive integer, i.e., a first sentence matrix and a second sentence matrix;
inputting the first sentence matrix and the second sentence matrix into a second neural network, for example, the second neural network comprises: referring to fig. 3, a network structure of the second sub-neural network includes an input layer 210, a convolutional layer 220, a pooling layer 230, and an output layer 240, and the output layer 240 may also include a fully-connected neural network structure, the network structure of the second sub-neural network is the same as that of the second sub-neural network, but an input quantity of the second sub-neural network is the first statement matrix, and an input quantity of the second sub-neural network is the second statement matrix;
the first statement matrix and the second statement matrix are convolved by convolution kernels of one or more sizes respectively to obtain feature maps of one or more sizes, please refer to fig. 5, in the second sub-neural network, the first statement matrix 200 is input, the first statement matrix 200 is a 5 × 7 matrix, the first statement matrix 200 is windowed in the input layer 210 respectively to obtain two 5 × 4 segment statement matrices, two 5 × 3 segment statement matrices and two 5 × 2 segment statement matrices 201, in the convolutional layer 220, the convolution cores with sizes of 2, 3 and 4 are used to convolve the segment statement matrices respectively to obtain two 1 × 4, two 1 × 5 and two 1 × 6 feature maps 202, in the pooling layer 230, the 1 × 4, 1 × 5 and 1 × 6 feature maps 202 are pooled and spliced respectively, and the pooling process can be performed in a maximum pooling manner, obtaining 1 × 2 eigenvectors, splicing 3 groups of 1 × 2 eigenvectors to obtain the budget eigenvector 203, then processing by the output layer 240 to obtain the processed budget eigenvector 204, and similarly, processing the second sentence matrix by the second sub-neural network according to the working mode of the second sub-neural network, fusing the processed budget eigenvector and the shape eigenvector to obtain a fused vector, arranging the fused vectors corresponding to a plurality of different texts in the form of rows or columns to obtain an eigenvector matrix, wherein the eigenvector matrix comprises texts corresponding to a plurality of users, for example, obtaining a t-dimensional eigenvector matrix by the fused vectors corresponding to the texts of t users, inputting the eigenvector matrix into the first neural network for training, obtaining a predicted value after processing by the first neural network, the predicted value is a t-dimensional vector, the t-dimensional vector can correspond to a preset recommendation information template, then the first neural network is trained in an iterative mode, and when the corresponding accuracy or recall ratio of the predicted value and the recommendation information template reaches or exceeds a set value, a recommendation model is obtained.
In order to improve the number and the types of the recommendation information templates, the recommendation information templates comprise consumption grade sub-templates and intervention effect sub-templates, wherein the types of the consumption grade sub-templates are m, the types of the intervention effect sub-templates are n, the types of the recommendation information templates are m × n, m and n are positive integers, and predicted values can be set to be t × 2 vectors which respectively correspond to the consumption grade sub-templates and the intervention effect sub-templates.
In summary, the method for recommending information based on user data provided by this embodiment can also be applied to medical question answering and insurance recommendation processes, and can provide accurate recommendation information for users by obtaining their own states, user budget capacity and character characteristics.
Referring to fig. 6, the present invention further provides a system for recommending information based on user data, including:
the preprocessing module is used for acquiring a text carrying user data, analyzing the text through a preset keyword to acquire a budget characteristic vector and a character characteristic vector, and performing fusion processing on the budget characteristic vector and the character characteristic vector to acquire a fusion vector;
the model module is used for loading fusion vectors corresponding to a plurality of users into a feature matrix respectively, inputting the feature matrix into a first neural network for training, and acquiring a recommendation model;
the processing module is used for predicting the text to be processed through the recommendation model and acquiring recommendation information;
the preprocessing module, the model module and the processing module are in signal connection. The sentences in the texts associated with the keywords are determined by using the keywords, redundant sentences in the texts are filtered, the data processing amount of the sentences in the texts needing to be processed is reduced, the processing efficiency can be improved, the interference of the sentences in the texts with low correlation degree on noise information generated by the recommendation information is avoided, two-dimensional fusion and analysis are carried out on the budget characteristic vectors and the property characteristic vectors, comprehensive weight consideration can be carried out by combining the budget condition and the property characteristic of the user, and the precision and the individuation level of the recommendation information are improved.
Optionally, the step of analyzing the text by using a preset keyword includes:
the preset keywords comprise a first keyword and a second keyword, wherein the first keyword is associated with budget information of the user, and the second keyword is associated with property information of the user;
determining a first sentence in the text, which is associated with budget information, through the first keyword, and determining a second sentence in the text, which is associated with property information, through the second keyword;
and vectorizing the first statement and the second statement respectively to obtain the budget characteristic vector and the property characteristic vector.
Optionally, the vectorizing the first statement and the second statement respectively, and the step of obtaining the budget feature vector and the trait feature vector includes:
acquiring codes of words of the first sentence and the second sentence through a corpus;
respectively acquiring a first sentence matrix and a second sentence matrix through the coding of the words of the first sentence and the second sentence;
inputting the first statement matrix and the second statement matrix into a second neural network respectively;
convolving the first statement matrix and the second statement matrix respectively through convolution kernels with one or more sizes to obtain feature maps with one or more sizes;
and respectively performing pooling treatment on the feature maps with one or more sizes and splicing to obtain the budget feature vector and the property feature vector.
Optionally, the second neural network comprises: a second sub-neural network for processing the first statement matrix and a second sub-neural network for processing the second statement matrix;
the second sub-neural network and the second sub-neural network each include: an input layer, a convolutional layer, a pooling layer, and an output layer.
Optionally, the first neural network includes an input layer, a fully-connected layer, and an output layer, and the activation function of the output layer includes a sigmoid function.
Optionally, the step of inputting the feature matrix into a first neural network for training, and obtaining a recommendation model includes:
acquiring the feature matrix and a preset recommendation information template;
inputting the characteristic matrix into the first neural network to obtain a predicted value, and enabling the predicted value to correspond to the recommendation information template;
and through iterative training of the first neural network, when the corresponding accuracy or recall rate of the predicted value and the recommendation information template reaches or exceeds a set value, acquiring a recommendation model.
Optionally, the recommendation information template includes a consumption level sub-template and an intervention effect sub-template, where m kinds of the consumption level sub-templates, n kinds of the intervention effect sub-templates, and m × n kinds of the recommendation information templates, where m and n are positive integers.
An embodiment of the present invention provides an electronic device, including: one or more processors; and one or more machine readable media having instructions stored thereon that, when executed by the one or more processors, cause the electronic device to perform one or more of the methods. The invention is operational with numerous general purpose or special purpose computing system environments or configurations. For example: personal computers, server computers, hand-held or portable devices, tablet-type devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like.
Embodiments of the invention also provide one or more machine-readable media having instructions stored thereon, which when executed by one or more processors, cause an apparatus to perform one or more of the methods described herein. The invention may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.
Claims (10)
1. A method for recommending information based on user data, comprising:
acquiring a text carrying user data, analyzing the text through a preset keyword, and acquiring a budget characteristic vector and a property characteristic vector;
performing fusion processing on the budget characteristic vector and the character characteristic vector to obtain a fusion vector;
loading the fusion vectors corresponding to a plurality of users into the feature matrix respectively;
inputting the characteristic matrix into a first neural network for training to obtain a recommendation model;
and predicting the text to be processed through the recommendation model to obtain recommendation information.
2. The method of claim 1, wherein the step of analyzing the text by a preset keyword comprises:
the preset keywords comprise a first keyword and a second keyword, wherein the first keyword is associated with budget information of the user, and the second keyword is associated with property information of the user;
determining a first sentence in the text, which is associated with budget information, through the first keyword, and determining a second sentence in the text, which is associated with property information, through the second keyword;
and vectorizing the first statement and the second statement respectively to obtain the budget characteristic vector and the property characteristic vector.
3. The method of claim 2, wherein the first sentence and the second sentence are vectorized, and the step of obtaining the budget feature vector and the trait feature vector comprises:
acquiring codes of words of the first sentence and the second sentence through a corpus;
respectively acquiring a first sentence matrix and a second sentence matrix through the coding of the words of the first sentence and the second sentence;
inputting the first statement matrix and the second statement matrix into a second neural network respectively;
convolving the first statement matrix and the second statement matrix respectively through convolution kernels with one or more sizes to obtain feature maps with one or more sizes;
and respectively performing pooling treatment on the feature maps with one or more sizes and splicing to obtain the budget feature vector and the property feature vector.
4. The method of claim 3, wherein the second neural network comprises: a second sub-neural network for processing the first statement matrix and a second sub-neural network for processing the second statement matrix;
the second sub-neural network and the second sub-neural network each include: an input layer, a convolutional layer, a pooling layer, and an output layer.
5. The method of claim 1, wherein the first neural network comprises an input layer, a fully-connected layer, and an output layer, and wherein the activation function of the output layer comprises a sigmoid function.
6. The method of claim 1 or 5, wherein the feature matrix is input into a first neural network for training, and the step of obtaining the recommendation model comprises:
acquiring the feature matrix and a preset recommendation information template;
inputting the characteristic matrix into the first neural network to obtain a predicted value, and enabling the predicted value to correspond to the recommendation information template;
and through iterative training of the first neural network, when the corresponding accuracy or recall rate of the predicted value and the recommendation information template reaches or exceeds a set value, acquiring a recommendation model.
7. The method for recommending information based on user data of claim 2, wherein the recommendation information template comprises a consumption level sub-template and an intervention effect sub-template, wherein the consumption level sub-template has m kinds, the intervention effect sub-template has n kinds, and the recommendation information template has m × n kinds, wherein m and n are positive integers.
8. A system for recommending information based on user data, comprising:
the preprocessing module is used for acquiring a text carrying user data, analyzing the text through a preset keyword to acquire a budget characteristic vector and a character characteristic vector, and performing fusion processing on the budget characteristic vector and the character characteristic vector to acquire a fusion vector;
the model module is used for loading fusion vectors corresponding to a plurality of users into a feature matrix respectively, inputting the feature matrix into a first neural network for training, and acquiring a recommendation model;
the processing module is used for predicting the text to be processed through the recommendation model and acquiring recommendation information;
the preprocessing module, the model module and the processing module are in signal connection.
9. An electronic device, comprising:
one or more processors; and
one or more machine-readable media having instructions stored thereon that, when executed by the one or more processors, cause the electronic device to perform the method recited in any of claims 1-7.
10. A machine-readable medium having stored thereon instructions, which when executed by one or more processors, cause an apparatus to perform the method of any of claims 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110621987.9A CN113076488A (en) | 2021-06-04 | 2021-06-04 | Method and system for recommending information based on user data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110621987.9A CN113076488A (en) | 2021-06-04 | 2021-06-04 | Method and system for recommending information based on user data |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113076488A true CN113076488A (en) | 2021-07-06 |
Family
ID=76617010
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110621987.9A Pending CN113076488A (en) | 2021-06-04 | 2021-06-04 | Method and system for recommending information based on user data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113076488A (en) |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107423442A (en) * | 2017-08-07 | 2017-12-01 | 火烈鸟网络(广州)股份有限公司 | Method and system, storage medium and computer equipment are recommended in application based on user's portrait behavioural analysis |
CN109255721A (en) * | 2018-10-30 | 2019-01-22 | 平安医疗健康管理股份有限公司 | Insurance recommended method, equipment, server and readable medium based on Cost Forecast |
-
2021
- 2021-06-04 CN CN202110621987.9A patent/CN113076488A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107423442A (en) * | 2017-08-07 | 2017-12-01 | 火烈鸟网络(广州)股份有限公司 | Method and system, storage medium and computer equipment are recommended in application based on user's portrait behavioural analysis |
CN109255721A (en) * | 2018-10-30 | 2019-01-22 | 平安医疗健康管理股份有限公司 | Insurance recommended method, equipment, server and readable medium based on Cost Forecast |
Non-Patent Citations (3)
Title |
---|
LEI ZHENG等: ""Joint Deep Modeling of Users and Items Using Reviews for Recommendation"", 《HTTPS://ARXIV.ORG/ABS/1701.04783》 * |
周传华等: ""融合评分矩阵和评论文本的深度神经网络推荐模型"", 《计算机应用研究》 * |
对外经济贸易大学信息学院: "《现代服务业人才培养探索与实践(2016)》", 31 August 2016, 对外经济贸易大学出版社 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110347835B (en) | Text clustering method, electronic device and storage medium | |
CN106547885A (en) | A kind of Text Classification System and method | |
CN113722438B (en) | Sentence vector generation method and device based on sentence vector model and computer equipment | |
CN110750965A (en) | English text sequence labeling method and system and computer equipment | |
CN111767375A (en) | Semantic recall method and device, computer equipment and storage medium | |
WO2017136674A1 (en) | Generating feature embeddings from a co-occurrence matrix | |
CN114358203A (en) | Training method and device for image description sentence generation module and electronic equipment | |
CN111523960A (en) | Product pushing method and device based on sparse matrix, computer equipment and medium | |
CN111507100A (en) | Convolution self-encoder and word embedding vector compression method based on same | |
EP4402612A1 (en) | Convolution attention network for multi-label clinical document classification | |
CN109359198A (en) | A kind of file classification method and device | |
CN116469110A (en) | Image classification method, device, electronic equipment and computer readable storage medium | |
CN113723077B (en) | Sentence vector generation method and device based on bidirectional characterization model and computer equipment | |
CN112307738B (en) | Method and device for processing text | |
CN113626576A (en) | Method and device for extracting relational characteristics in remote supervision, terminal and storage medium | |
CN111597336B (en) | Training text processing method and device, electronic equipment and readable storage medium | |
CN113076488A (en) | Method and system for recommending information based on user data | |
CN113076405A (en) | User portrait based solution matching method and system | |
CN113254657B (en) | User data classification method and system | |
CN114116971A (en) | Model training method and device for generating similar texts and computer equipment | |
CN114692715A (en) | Sample labeling method and device | |
CN111611379A (en) | Text information classification method, device, equipment and readable storage medium | |
CN113139382A (en) | Named entity identification method and device | |
CN116992874B (en) | Text quotation auditing and tracing method, system, device and storage medium | |
CN113283235B (en) | User label prediction method and system |
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 |