CN113158057A - Buddha meridian recommendation processing device, computer equipment and storage medium - Google Patents
Buddha meridian recommendation processing device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN113158057A CN113158057A CN202110465027.8A CN202110465027A CN113158057A CN 113158057 A CN113158057 A CN 113158057A CN 202110465027 A CN202110465027 A CN 202110465027A CN 113158057 A CN113158057 A CN 113158057A
- Authority
- CN
- China
- Prior art keywords
- buddha
- user
- behavior
- longitude
- behaviors
- 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
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/048—Activation functions
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the technical field of artificial intelligence, is applied to the field of Buddhist sutra recommendation application, and discloses a Buddhist sutra recommendation processing method, a device, computer equipment and a storage medium, which are used for improving the pushing accuracy of Buddhist sutra. The method comprises the following steps: responding to the Buddha warrior recommendation request, and acquiring inherent attribute information and historical behavior information of a user, wherein the historical behavior information is obtained by the user aiming at historical operating behaviors of the Buddha warrior, the historical operating behaviors comprise forward behaviors and reverse behaviors of the user on the Buddha warrior in different time periods, and the forward behaviors and the reverse behaviors represent different behaviors of the user on the interest attitude of the Buddha warrior; constructing a user behavior sequence according to the forward behavior and the reverse behavior; determining the favorite features of the user on the Buddhist scriptures; selecting a recommended Buddha course from each candidate Buddha course according to the preference characteristics of the user on the Buddha course and the Buddha course characteristics of each candidate Buddha course; and pushing the recommended Buddha channel to the client.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, is applied to Buddhist sutra recommendation, and particularly relates to Buddhist sutra recommendation processing, a device, computer equipment and a storage medium.
Background
For the field of Buddhism, a traditional recommendation algorithm model focuses on the research of user interests, and Buddhist people may have no choice in the face of numerous Buddhists, so that the requirement for scripture recommendation is derived.
Disclosure of Invention
The embodiment of the invention provides a Buddhist scripture recommendation processing method, a Buddhist scripture recommendation processing device, computer equipment and a storage medium, and aims to solve the problem of inaccurate Buddhist scripture recommendation.
A method for processing Buddha recommendation, comprising:
receiving a Buddha experience recommendation request triggered by a user through a client;
responding to the Buddha longitude recommendation request, and acquiring inherent attribute information and historical behavior information of the user, wherein the historical behavior information is obtained by the user aiming at historical operating behaviors of the Buddha longitude, the historical operating behaviors comprise forward behaviors and reverse behaviors of the Buddha longitude, which are obtained by the user, in different time periods, and the forward behaviors and the reverse behaviors represent different behaviors of the user on the interesting attitude of the Buddha longitude;
analyzing the historical behavior information to obtain the forward behavior and the reverse behavior of the Buddha warrior at different time periods of the user, and constructing a user behavior sequence according to the forward behavior and the reverse behavior;
determining the preference characteristics of the user to Buddhist scriptures according to the inherent attribute information of the user and the user behavior sequence;
selecting a recommended Buddha longitude from each candidate Buddha longitude according to the preference characteristics of the user on the Buddha longitude and the Buddha longitude characteristics of each candidate Buddha longitude;
and pushing the recommended Buddha channel to the client.
A Buddhist experience recommendation processing apparatus comprising:
the system comprises a receiving module, a recommendation module and a recommendation module, wherein the receiving module is used for receiving a Buddha longitude recommendation request triggered by a user through a client;
the obtaining module is used for responding to the Buddha longitude recommendation request and obtaining inherent attribute information and historical behavior information of the user, wherein the historical behavior information is obtained by the user aiming at historical operating behaviors of the Buddha longitude, the historical operating behaviors comprise forward behaviors and reverse behaviors of the user on the Buddha longitude in different time periods, and the forward behaviors and the reverse behaviors represent different behaviors of the user on interest attitudes of the Buddha longitude;
the analysis module is used for analyzing the historical behavior information to obtain the forward behavior and the reverse behavior of the Buddha warrior at different time periods by the user, and constructing a user behavior sequence according to the forward behavior and the reverse behavior;
the determining module is used for determining the favorite features of the user on the Buddhist scriptures according to the inherent attribute information of the user and the user behavior sequence;
the screening module is used for selecting recommended Buddhist sutra from the candidate Buddhist sutra according to the preference characteristics of the Buddhist sutra of the user and the Buddhist sutra characteristics of the candidate Buddhist sutra;
and the pushing module is used for pushing the recommended Buddha channels to the client.
A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the above described buddha recommendation processing method when executing the computer program.
A computer-readable storage medium, which stores a computer program that, when executed by a processor, implements the steps of the above-described buddha recommendation processing method.
In the above-mentioned buddha warrior scripture recommendation processing method, device, computer equipment and storage medium, because when the interest of buddha warrior scripture is judged, the historical forward behavior of the buddha warrior scripture is referred to, and the historical backward behavior of the buddha warrior, the historical forward behavior and the historical backward behavior are combined to reflect the time sequence characteristics of the user behavior and determine the favorite characteristics of the user. In the aspect of application conversion, because of the accuracy of Buddha scriptures recommendation, the click rate of the user on the recommended Buddhist scriptures can be improved, and the user experience and the popularization effect of the Buddhist scriptures are improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments of the present invention will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without inventive labor.
FIG. 1 is a schematic diagram of an application environment of a process for recommending Buddha according to an embodiment of the present invention;
FIG. 2 is a flow chart illustrating a process of a process for recommending Buddha according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of one embodiment of step S40 in FIG. 1;
FIG. 4 is a schematic diagram of one embodiment of step S41 in FIG. 3;
FIG. 5 is a schematic diagram of a feature fusion process of a Buddha recommendation processing method according to an embodiment of the invention;
FIG. 6 is a schematic diagram of one embodiment of step S50 in FIG. 1;
FIG. 7 is a schematic diagram illustrating the training of interest models for Buddha recommendation processing in an embodiment of the present invention;
FIG. 8 is a schematic diagram of an overall architecture of a process for recommending Buddha according to an embodiment of the present invention;
FIG. 9 is a schematic diagram of a device for processing Buddha recommendation according to an embodiment of the present invention;
FIG. 10 is a schematic diagram of a computer device according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The Buddhist scripture recommendation processing method provided by the embodiment of the invention can be applied to an application environment as shown in fig. 1, wherein a client can communicate with a server through a network, and the server can receive a Buddhist scripture recommendation request triggered by a user through the client; responding to the Buddha longitude recommendation request, and acquiring inherent attribute information and historical behavior information of the user, wherein the historical behavior information is obtained by the user aiming at historical operating behaviors of the Buddha longitude, the historical operating behaviors comprise forward behaviors and reverse behaviors of the Buddha longitude, which are obtained by the user, in different time periods, and the forward behaviors and the reverse behaviors represent different behaviors of the user on the interesting attitude of the Buddha longitude; the server analyzes the historical behavior information to obtain the forward behavior and the reverse behavior of the Buddha scriptures of the user in different time periods, and a user behavior sequence is constructed according to the forward behavior and the reverse behavior; then the server determines the preference characteristics of the user to Buddhist scriptures according to the inherent attribute information of the user and the user behavior sequence; selecting a recommended Buddha longitude from each candidate Buddha longitude according to the preference characteristics of the user on the Buddha longitude and the Buddha longitude characteristics of each candidate Buddha longitude; and pushing the recommended Buddha channel to the client.
Wherein the client may be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server may be implemented by an independent server or a server cluster formed by a plurality of servers, and the scheme is not limited. The following describes the recommended treatment method of Buddha provided by the present invention in detail with reference to specific examples.
In an embodiment, as shown in fig. 2, a method for recommending Buddha is provided, which is described by taking the server in fig. 1 as an example, and includes the following steps:
s10: and receiving a Buddhist scripture recommendation request triggered by a user through a client.
The scheme is based on an application program which is used for being used as a user for Buddhist people or non-Buddhist people, and the application program is installed in a client, wherein the client can be a mobile phone, a tablet personal computer and the like. The user can trigger a Buddhist scripture recommendation request through clicking and the like through an application program on the client, and the Buddhist scripture recommendation request is used for requesting the server to provide a Buddhist scripture recommendation service. For the server, a Buddhism recommendation request triggered by the user through the client can be received.
The Buddha's meridians named in this scheme include classic or Tibetan meridians, including many different types or categories of classics, including but not limited to the six-element Buddha's book of Buddha's heaven and Roue's book of Buddha, the three-in five-band Buddha's heart-shaped Yan-Yu-De-Gong Jing, the Buddha's book of Mi-En-Ming-Gong-Jing, the Jade's book of Jade women's book of Buddha's book of Ji-horizontal Jing, the Buddha's religion's book, and the classic of the plus one's book of Araliary classic, to name a simple example. Therefore, after the user triggers the Buddhist scripture recommendation request, the corresponding Buddhist scripture needs to be accurately recommended to the user in response to the Buddhist scripture recommendation request.
S20: and responding to the Buddhist channel recommendation request, and acquiring the inherent attribute information and the historical behavior information of the user.
The historical behavior information is obtained by the user according to historical operation behaviors of the Buddha scripture, the historical operation behaviors comprise forward behaviors and reverse behaviors of the Buddha scripture in different time periods, and the forward behaviors and the reverse behaviors represent different behaviors of the user in the interest attitude of the Buddha scripture.
After receiving the Buddhist scripture recommendation request, the server responds to the Buddhist scripture recommendation request, specifically, analyzes the Buddhist scripture recommendation request, and thus obtains the user triggering the Buddhist scripture recommendation request. In an application scenario, when a user registers a user account of the application program, the inherent attribute information of the user may be uploaded to the server as registration information, so that the inherent attribute information of the user and the user identifier may be stored in the server. The user identification is identification information for uniquely identifying the registered user. When a user logs in an application program and triggers a Buddha longitude recommendation request, the Buddha longitude recommendation request can carry a user identifier, and after the server receives the Buddha longitude recommendation request, the server can determine which user triggers the Buddha longitude recommendation request according to the user identifier in the Buddha longitude recommendation request. And then, the inherent attribute information of the user can be acquired by using the user identification. Illustratively, the inherent attribute information of the user includes, but is not limited to, age, sex occupation, and the like.
In the scheme, besides the inherent attribute information of the user, the historical behavior information of the user can be obtained in response to the Buddha scripture recommendation request, and it is worth explaining that the historical behavior information in the scheme is different from the historical behavior information used by a traditional recommendation algorithm, the historical behavior information not only simply reflects the behavior that the user is interested in which Buddha scripture to obtain uninteresting behavior, but also comprises the forward behavior and the reverse behavior of the user on the Buddha scripture in different time periods, and the forward behavior and the reverse behavior represent different behaviors of the user on the interesting attitude of the Buddha scripture. That is, the historical behavior information includes timing information of forward behavior and backward behavior of the user for Buddha scriptures over different time periods. For example, the forward behavior of the pair of buddies may refer to a behavior that the buddies are clicked to be interested in, focused on or scored with a higher score and the like to represent that the buddies have an interest attitude for the user, and conversely, the reverse behavior of the pair of buddies may refer to a behavior that the buddies are clicked to be uninteresting, unfocused or scored with a lower score and the like to represent that the buddies have no interest attitude for the user.
S30: analyzing the historical behavior information to obtain the forward behavior and the reverse behavior of the Buddha warrior at different time periods, and constructing a user behavior sequence according to the forward behavior and the reverse behavior.
After the historical behavior information is obtained, the server analyzes the historical behavior information to obtain the forward behavior and the reverse behavior of the Buddha warrior at different time periods, and a user behavior sequence is constructed according to the forward behavior and the reverse behavior. The user behavior sequence here refers to an information sequence that reflects the user's time-series behavior through the user's forward behavior and reverse behavior.
As an example, a user behavior sequence is constructed according to the forward behavior and the reverse behavior, specifically, a sequence is composed of Buddha feature information of each Buddha determined according to the forward behavior or the reverse behavior of the Buddha over a certain time period. For example, if the buddies clicked by the user are 1, 2, 3, 4, and 5, and the buddies not clicked by the user are 6, 7, 8, 9, and 10, the buddies feature information of the 10 buddies and whether to click or not, whether to like, are arranged in time sequence to form an information sequence. It can be seen that the information sequence carries time sequence information, and the time sequence information can reflect the interest attitude of a user changing the interest attitude of a certain Buddha scripture along with the time, wherein the Buddha scripture feature information of the Buddha scripture can refer to the content, category and other feature information representing the Buddha scripture.
It can be understood that, in the present solution, it is considered that the preference of the user may be reflected in the operation behavior of the user, for example, when the user browses a certain Buddha, it indicates that the user prefers the Buddha with a high probability. For another example, when the user collects buddies, the description shows that the buddies are interested by the user. However, the operation behavior of the user is browsing or collecting, and does not indicate that the user likes the Buddha, for example, when the user browses the Buddha carefully, the user finds that the composition, content, and the like of the Buddha are not suitable for the user. Therefore, the user behavior sequence is related to time, and the user behavior sequence is constructed for the historical operation behavior of each Buddha scripture according to the historical operation time, so that the Buddha scripture can be accurately recommended to the user subsequently.
S40: and determining the preference characteristics of the user for Buddhist scriptures according to the inherent attribute information of the user and the user behavior sequence.
After the inherent attribute information of the user and the user behavior sequence are obtained, the preference characteristics of the user on the Buddha scripture can be determined.
As an example, as shown in fig. 3, in step S40, that is, according to the inherent attribute information of the user and the user behavior sequence, determining a favorite feature of the user for Buddhist scripture, specifically includes the following steps:
s41: and extracting the characteristics of the inherent attribute information of the user to construct a user attribute characteristic vector, and extracting the characteristics of the user behavior sequence to construct a user behavior sequence characteristic vector.
It should be noted that, as described above, the inherent attribute information of the user includes the inherent attribute information of the user, such as account registration time, gender, age, occupation, and the like, and when constructing the corresponding feature vector, the inherent attribute information of the user is first converted into a numerical feature, so as to perform feature extraction on the inherent attribute information of the user to be recommended. In specific implementation, since attribute information such as registration time, age and the like is described by using numbers, the attribute information can be directly used as numerical characteristics of the user to be recommended. Aiming at attribute information of character descriptions such as gender, places and the like, the attribute information is converted into numerical features in a coding mode, and the attribute information is conveniently converted into corresponding user attribute feature vectors.
It should be noted that, the user behavior sequence mentioned in the foregoing step is constructed according to the forward behavior and the backward behavior of the user for Buddha, and the forward behavior and the backward behavior of the user for Buddha include time sequence information, so that the user behavior sequence is subjected to feature extraction to construct a user behavior sequence feature vector to embody the time sequence information. In particular, there may be one timing tag.
The following illustrates a process of constructing a user behavior sequence according to the forward behavior and the backward behavior of the Buddha warrior. In an application scenario, for example, assume that the user browses Buddha 1 and Buddha 2 during the time period t1, then clicks on "dislike" for Buddha 2 during the time period t2, and then collects and browses Buddha 3 during the time period t 3. It can be seen from the above that the buddha meridians associated with the user are buddha meridian 1, buddha meridian 2 and buddha meridian 3, the actions associated with the user include browsing, clicking "dislike" and collecting, and then the buddha meridian 1, buddha meridian 2, buddha meridian 3, browsing, clicking "dislike" and collecting are taken as the nodes of the action time sequence. The node Buddha channel 1 corresponding to the time period t1 is connected with the node browse, the node Buddha channel 2 is connected with the node browse, and the obtained edge attributes of the two connecting edges are the time period t 1. In the period t2, the node "Buddha 2" is connected with the node "dislike", and the obtained edge attribute of one connecting edge is the period t 2. At the time period of t3, the node Buddha channel 2 is connected with the node dislike, the node Buddha channel 3 is connected with the node browse and the node favorite, and the obtained edge attributes of the two connecting edges are the time period of t 3. The timing labels where t1, t2 and t3 are timing labels are defined.
It can be seen from the above example that the time period is associated with the historical behavior of the user, and according to the connection relationship and the time information, the time period can be used as a user time sequence, so that a corresponding feature vector of a constructed user behavior sequence can be created. In a specific implementation, the time period may be selected as long as possible, and the Buddha warrior can be accurately recommended to the user in the following steps, which is not limited herein. It should be noted that the above process of constructing the user behavior sequence is only an exemplary illustration, and other user behavior sequences may be generated according to different time periods and user operation behaviors, which are not described here, and it is to be understood that the user behavior sequence in the present scheme embodies time sequence information.
As an example, as shown in fig. 4, in step S41, that is, performing feature extraction on the inherent attribute information of the user to construct a user attribute feature vector, and performing feature extraction on the user behavior sequence to construct a user behavior sequence feature vector, the method specifically includes the following steps:
s411: and inputting the user inherent attribute information into an embedding layer of a Transformer model to obtain a first word vector, and taking the first word vector as the user attribute feature vector.
S412: and inputting the user behavior sequence into an embedding layer of a Transformer network to construct a second word vector, and inputting the second word vector into the Transformer layer of the Transformer network to obtain a user behavior sequence characteristic vector.
S42: and fusing the user attribute feature vector and the user behavior sequence feature vector to obtain the favorite features of the user on the Buddha scriptures.
For steps S41-S42, it can be understood that the inherent attribute information of the user and the user behavior sequence are only one type of information, and for convenience of calculation and reduction, feature extraction is performed on the inherent attribute information of the user to construct a user attribute feature vector, and feature extraction is performed on the user behavior sequence to construct a user behavior sequence feature vector, so that the user attribute feature vector and the user behavior sequence feature vector are fused to obtain the favorite features of the user on Buddha. It should be noted that, in an application scenario, the fusion refers to feature intersection processing.
Specifically, as shown in fig. 5, fig. 5 mainly includes three parts, where the first part is an embedding layer, the embedding layer mainly maps input data into a matrix vector, the input data is divided into two parts, namely, inherent attribute information of a user and a user behavior sequence, and the embedding layer trains two word embedding matrix vectors, namely, a first word vector and a second word vector, respectively, by the inherent attribute information of the user and the user behavior sequence. The first word vector is used for mapping the input of the inherent attribute information of the user, and the second word vector is used for mapping the input of the user behavior sequence, so that the construction of the word vector is completed. A second component, a transform layer, to complete the extraction of features. Specifically, the second word vector is used for being input into the transform layer, so that the feature learning is performed on the second word vector through the transform layer, and thus the user behavior sequence feature vector is obtained. It should be noted that the transform layer mainly includes an encoder (encoder) and a decoder (decoder), and a detailed description about a specific network structure of the transform layer is omitted here. Based on the network structure characteristics of the transform layer of the transform network, the characteristics of the time sequence of the user behavior sequence can be fully learned through the transform layer of the transform network.
S50: and selecting a recommended Buddha from the candidate Buddha according to the preference characteristics of the user on the Buddha and the Buddha characteristics of the candidate Buddha.
S60: and pushing the recommended Buddha channel to the client.
After the preference characteristics of the user for the Buddha are obtained, the recommended Buddha is selected from the candidate Buddha and pushed to the user side to be recommended to the user according to the preference characteristics of the user for the Buddha and the Buddha characteristics of the candidate Buddha. The Buddha characteristics of each candidate Buddha are extracted according to attribute information of the Buddha, such as content, type and the like.
In the scheme provided by the Buddha warrior recommendation processing method, the device, the computer equipment and the storage medium, when the interest of the Buddha warrior is judged, the historical forward behavior of the Buddha warrior is referred to, the historical backward behavior of the Buddha warrior is also referred to, the historical forward behavior and the historical backward behavior are combined, the time sequence characteristics of the user behavior are embodied, and the preference characteristics of the user are determined. In the aspect of application conversion, because of the accuracy of Buddha scriptures recommendation, the click rate of the user on the recommended Buddhist scriptures can be improved, and the user experience and the popularization effect of the Buddhist scriptures are improved.
As an example, as shown in fig. 6, in S50, that is, according to the favorite features of the user for buddies and the buddies features of each candidate buddies, a recommended buddies is selected from the candidate buddies, which specifically includes the following steps:
s51: and obtaining the interest degree of the user in each candidate Buddha on the basis of the preference characteristics of the user on the Buddha and the Buddha characteristics of each candidate Buddha.
S52: and selecting a recommended Buddha from the candidate Buddha according to the interest degree of the candidate Buddha.
As an example, in step S51, that is, based on the favorite features of the Buddha scriptures of the user and the Buddha scriptures features of each candidate Buddha scriptures, obtaining the interest level of the user in each candidate Buddha scriptures, specifically: analyzing the preference characteristics of the Buddhist scriptures of the user and the Buddhist scriptures of each candidate Buddhist scriptures through the trained interest degree model to obtain the interest degree of the user in each candidate Buddhist scriptures output by the interest degree model; the trained interest degree model is obtained by training according to a training sample data set marked with interest degrees, and training samples in the training sample data set comprise historical behavior information of a sample user object on Buddhist sutra and Buddhist sutra characteristic information of the sample Buddhist sutra.
In the embodiment of the invention, the interestingness labeled in the training sample is determined according to the user behavior of the sample. For a simple example, if the user clicks on the sample Buddha, the interest degree labeled on the sample Buddha is 1, and if the user has not clicked on the sample Buddha, the interest degree labeled on the sample Buddha is 0. Therefore, the labeled interest degree, that is, the labeled information of the Buddhist sutra of the sample, can divide the training samples into positive samples or negative samples through the labeled information, and the interest degree model can be trained according to the training samples. The training sample data set comprises a plurality of training samples, and each training sample is generated based on the behavior of the sample user object on the sample Buddha longitude and comprises behavior information of the sample user object on the Buddha longitude and characteristic information of the sample Buddha longitude.
Taking an actual scene as an example, how to obtain the training samples is described. For example, when a sample user reads buddhist scriptures on buddhist scriptures applications, different behaviors may be generated for the buddhist scriptures recommended for a certain application, and the behaviors of the user on the same buddhist scriptures may be different according to the passage of time. This is taken into account when acquiring training samples, taking timing states into account. Specifically, the user may click on the Buddha or click off, like or dislike (i.e. forward behavior or reverse behavior). The user is the sample object, and the Buddha currently browsed by the user is the sample Buddha. The Buddha character information of the sample Buddha meridian refers to the attribute information of the Buddha meridian. Such as the author, category, publisher, etc. of the Buddhist scripture. Therefore, the behavior information of the sample user object on the Buddhist scriptures or the Buddhist scriptures characteristic information of the sample Buddhist scriptures can be obtained according to the user side of each user.
It should be noted that the preference characteristics of the user for Buddha scripture are extracted based on the user behavior sequence of the user and the inherent attribute information of the user, and may be used to describe feedback information of the behavior preference of the target object, which may be obtained by analyzing the historical behavior of the user for Buddha scripture. In the embodiment of the application, when the interest of the user candidate Buddha is determined based on the interest degree model, firstly, the interest degree model is used for obtaining the preference characteristics of the user for the user attribute characteristic vector and the user behavior sequence characteristic vector; and then based on the model, combining the preference characteristics of the user and the Buddha characteristics of each candidate Buddha, and obtaining the interest degree of the user in the candidate Buddha. The interest degree output by the interest degree model can be a probability value with a value range of 0-1, the larger the value corresponding to the candidate Buddha warrior, the higher the probability that the user implements a forward behavior on the Buddha warrior after recommending the Buddha warrior to the user, so that the purpose of accurate recommendation is achieved, and the user experience and the Buddha warrior product popularization are improved.
As an example, as shown in fig. 7, the trained interestingness model is trained by:
s101: and selecting a training sample from the training sample data set, wherein the training sample is marked with the interest degree of the user sample object in the sample Buddha scriptures.
S102: for each training sample in the training sample data set, inputting behavior information of a user sample object and Buddha included in the training sample into an untrained interest degree model, and obtaining the interest degree of the user sample object in the Buddha included in the untrained interest degree model output by the untrained interest degree model.
S103: and optimizing parameters in the untrained interest degree model based on an objective loss function, so that the difference between the interest degree labeled by each training sample and the interest degree obtained by the untrained interest degree model is within an allowable difference range, and obtaining the trained interest degree model, wherein the objective loss function comprises a forward behavior item and a reverse behavior item corresponding to the behavior information of the user sample object.
In the embodiment of the present invention, the target loss function may be a cross entropy loss function, or may be other types of loss functions, which is not limited specifically. The following is mainly introduced by taking a cross entropy loss function as an example, and the following calculation formula is:
wherein N is the number of the training samples, S1A forward behavior loss term, S, corresponding to the training sample in (1)2A reverse behavior loss term, S, corresponding to the training sample in (1)3P (x) is the interest degree of the interest degree model aiming at the training sample x, a1、a2、a3A weight corresponding to each loss term.
That is, training samples may be divided into three groups based on historical behavior of sample user objects, where S1The forward behavior loss item (e.g. click or browse) corresponding to the training sample in (1), S2The reverse behavior loss term (e.g. no click or click "dislike") corresponding to the training sample in (1), S3The inverse of the training sample in (1)Loss of terms to behavior (e.g., bad click); and p (x) is the interestingness output by the interestingness model for the training sample x.
It should be noted that, according to the behavior of the sample user object, a plurality of forward behavior items and backward behavior items may be used. Comprises thatAndan item.
a1、a2、a3For the weight corresponding to each loss term, the weight data of each loss term may be optimized according to experience or a model training process, which is not limited herein nor described.
In the embodiment of the invention, when the target loss function is optimized through an optimization algorithm, the interestingness model is evaluated mainly according to the interestingness output by the interestingness model, the adjustment is carried out according to the evaluation result, and the interestingness model is further optimized according to the adjusted target loss function until the interestingness model converges, so that the effect that the difference value between the interestingness labeled by each training sample and the interestingness obtained through the untrained interestingness model is within the allowable difference range is achieved.
As an example, as shown in fig. 8, specifically, the interestingness model in the embodiment of the present application mainly includes a recurrent Neural Network portion, a Transformer Network portion and a Feed-Forward Neural Network (FNN) portion, where the Transformer Network portion is mainly used for performing feature crossing on a user attribute feature vector and a user behavior sequence vector of a user to obtain a favorite feature of the user, and since a behavior sequence is added, time information is reflected, and a current interest of the user can be effectively captured. After the preference characteristics are determined based on the transform network part, behavior preference characteristics and Buddha channel characteristics of all candidate Buddhist channels are processed mainly through a feedforward neural network and a full connection layer connected behind the feedforward neural network so as to obtain interest degrees, and the Buddhist channels recommended by the Buddhist channel recommendation processing method in the embodiment of the application are more accurate.
As an example, the feedforward neural network may be connected with a sigmiod activation function, and the preference feature and the Buddha channel feature of the candidate Buddha channel are connected with the sigmiod activation function after passing through the Feedforward Neural Network (FNN), so as to generate a final binary classification result, output recommended and non-recommended results, and serve as recommended Buddha channels according to the recommended results and push the recommended Buddha channels to the client for recommendation to the user. The final model structure is shown in fig. 8.
It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.
In an embodiment, a Buddha longitude recommendation processing device is provided, and the Buddha longitude recommendation processing device corresponds to the Buddha longitude recommendation processing methods in the embodiments one to one. As shown in fig. 9, the Buddha recommendation processing device includes a receiving module 101, an obtaining module 102, a parsing module 103, a determining module 104, a screening module 105, and a pushing module 106. The functional modules are explained in detail as follows:
the system comprises a receiving module, a recommendation module and a recommendation module, wherein the receiving module is used for receiving a Buddha longitude recommendation request triggered by a user through a client;
the obtaining module is used for responding to the Buddha longitude recommendation request and obtaining inherent attribute information and historical behavior information of the user, wherein the historical behavior information is obtained by the user aiming at historical operating behaviors of the Buddha longitude, the historical operating behaviors comprise forward behaviors and reverse behaviors of the user on the Buddha longitude in different time periods, and the forward behaviors and the reverse behaviors represent different behaviors of the user on interest attitudes of the Buddha longitude;
the analysis module is used for analyzing the historical behavior information to obtain the forward behavior and the reverse behavior of the Buddha warrior at different time periods by the user, and constructing a user behavior sequence according to the forward behavior and the reverse behavior;
the determining module is used for determining the favorite features of the user on the Buddhist scriptures according to the inherent attribute information of the user and the user behavior sequence;
the screening module is used for selecting recommended Buddhist sutra from the candidate Buddhist sutra according to the preference characteristics of the Buddhist sutra of the user and the Buddhist sutra characteristics of the candidate Buddhist sutra;
and the pushing module is used for pushing the recommended Buddha channels to the client.
In an embodiment, the determining module is specifically configured to:
extracting the characteristics of the inherent attribute information to construct a user attribute characteristic vector, and extracting the characteristics of the user behavior sequence to construct a user behavior sequence characteristic vector;
and fusing the user attribute feature vector and the user behavior sequence feature vector to obtain the favorite features of the user on the Buddha scriptures.
In an embodiment, the determining module is specifically configured to: inputting the user inherent attribute information into an embedding layer of a Transformer model to obtain a first word vector, and taking the first word vector as the user attribute feature vector;
and inputting the user behavior sequence into an embedding layer of a Transformer network to construct a second word vector, and inputting the second word vector into the Transformer layer of the Transformer network to obtain a user behavior sequence characteristic vector.
In an embodiment, the screening module is specifically configured to:
obtaining the interest degree of the user in each candidate Buddha longitude based on the preference characteristic of the user for the Buddha longitude and the Buddha longitude characteristic of each candidate Buddha longitude;
and selecting a recommended Buddha from the candidate Buddha according to the interest degree of the candidate Buddha.
In an embodiment, the screening module is specifically configured to:
analyzing the preference characteristics of the Buddhist scriptures of the user and the Buddhist scriptures of each candidate Buddhist scriptures through the trained interest degree model to obtain the interest degree of the user in each candidate Buddhist scriptures output by the interest degree model; the trained interest degree model is obtained by training according to a training sample data set marked with interest degrees, and training samples in the training sample data set comprise historical behavior information of a sample user object on Buddhist sutra and Buddhist sutra characteristic information of the sample Buddhist sutra.
In one embodiment, the trained interestingness model is trained by:
selecting a training sample from the training sample data set, wherein the training sample is marked with the interest degree of a user sample object in the sample Buddha warrior;
for each training sample in the training sample data set, inputting behavior information of a user sample object and Buddha included in the training sample into an untrained interest degree model, and obtaining the interest degree of the user sample object in the Buddha included in the untrained interest degree model;
and optimizing parameters in the untrained interest degree model based on an objective loss function, so that the difference between the interest degree labeled by each training sample and the interest degree obtained by the untrained interest degree model is within an allowable difference range, and obtaining the trained interest degree model, wherein the objective loss function comprises a forward behavior item and a reverse behavior item corresponding to the behavior information of the user sample object.
In one embodiment, the target loss function may be a cross-entropy loss function as follows:
wherein N is the number of the training samples, S1A forward behavior loss term, S, corresponding to the training sample in (1)2A reverse behavior loss term, S, corresponding to the training sample in (1)3P (x) is the interest degree of the interest degree model aiming at the training sample x, a1、a2、a3A weight corresponding to each loss term.
In the Buddha sutra recommendation processing device, when the interest of the Buddha sutra is judged, besides the historical forward behavior of the user on the Buddha sutra, the historical backward behavior of the user on the Buddha sutra is also referred, the historical forward behavior and the historical backward behavior are combined, the time sequence characteristics of the user behavior are reflected, and the preference characteristics of the user are determined. In the aspect of application conversion, because of the accuracy of Buddha scriptures recommendation, the click rate of the user on the recommended Buddhist scriptures can be improved, and the user experience and the popularization effect of the Buddhist scriptures are improved.
For the specific definition of the Buddhist longitude recommendation processing device, reference may be made to the above definition of the Buddhist longitude recommendation processing method, which is not described herein again. The modules in the Buddha recommendation processing device can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 10. 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 a storage medium including a nonvolatile storage medium and a volatile storage medium, and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a Buddha recommendation processing method.
In one embodiment, a computer device is provided, comprising 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:
receiving a Buddha experience recommendation request triggered by a user through a client;
responding to the Buddha longitude recommendation request, and acquiring inherent attribute information and historical behavior information of the user, wherein the historical behavior information is obtained by the user aiming at historical operating behaviors of the Buddha longitude, the historical operating behaviors comprise forward behaviors and reverse behaviors of the Buddha longitude, which are obtained by the user, in different time periods, and the forward behaviors and the reverse behaviors represent different behaviors of the user on the interesting attitude of the Buddha longitude;
analyzing the historical behavior information to obtain the forward behavior and the reverse behavior of the Buddha warrior at different time periods of the user, and constructing a user behavior sequence according to the forward behavior and the reverse behavior;
determining the preference characteristics of the user to Buddhist scriptures according to the inherent attribute information of the user and the user behavior sequence;
selecting a recommended Buddha longitude from each candidate Buddha longitude according to the preference characteristics of the user on the Buddha longitude and the Buddha longitude characteristics of each candidate Buddha longitude;
and pushing the recommended Buddha channel to the client.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
receiving a Buddha experience recommendation request triggered by a user through a client;
responding to the Buddha longitude recommendation request, and acquiring inherent attribute information and historical behavior information of the user, wherein the historical behavior information is obtained by the user aiming at historical operating behaviors of the Buddha longitude, the historical operating behaviors comprise forward behaviors and reverse behaviors of the Buddha longitude, which are obtained by the user, in different time periods, and the forward behaviors and the reverse behaviors represent different behaviors of the user on the interesting attitude of the Buddha longitude;
analyzing the historical behavior information to obtain the forward behavior and the reverse behavior of the Buddha warrior at different time periods of the user, and constructing a user behavior sequence according to the forward behavior and the reverse behavior;
determining the preference characteristics of the user to Buddhist scriptures according to the inherent attribute information of the user and the user behavior sequence;
selecting a recommended Buddha longitude from each candidate Buddha longitude according to the preference characteristics of the user on the Buddha longitude and the Buddha longitude characteristics of each candidate Buddha longitude;
and pushing the recommended Buddha channel to the client.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile 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), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions.
The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present 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 solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present invention, and are intended to be included within the scope of the present invention.
Claims (10)
1. A Buddha recommendation processing method is characterized by comprising the following steps:
receiving a Buddha experience recommendation request triggered by a user through a client;
responding to the Buddha longitude recommendation request, and acquiring inherent attribute information and historical behavior information of the user, wherein the historical behavior information is obtained by the user aiming at historical operating behaviors of the Buddha longitude, the historical operating behaviors comprise forward behaviors and reverse behaviors of the Buddha longitude, which are obtained by the user, in different time periods, and the forward behaviors and the reverse behaviors represent different behaviors of the user on the interesting attitude of the Buddha longitude;
analyzing the historical behavior information to obtain the forward behavior and the reverse behavior of the Buddha warrior at different time periods of the user, and constructing a user behavior sequence according to the forward behavior and the reverse behavior;
determining the preference characteristics of the user to Buddhist scriptures according to the inherent attribute information of the user and the user behavior sequence;
selecting a recommended Buddha longitude from each candidate Buddha longitude according to the preference characteristics of the user on the Buddha longitude and the Buddha longitude characteristics of each candidate Buddha longitude;
and pushing the recommended Buddha channel to the client.
2. The Buddha longitude recommendation processing method according to claim 1, wherein the determining of the preference characteristics of the user for Buddha longitude according to the inherent attribute information of the user and the user behavior sequence comprises:
extracting the characteristics of the inherent attribute information to construct a user attribute characteristic vector, and extracting the characteristics of the user behavior sequence to construct a user behavior sequence characteristic vector;
and fusing the user attribute feature vector and the user behavior sequence feature vector to obtain the favorite features of the user on the Buddha scriptures.
3. The method for recommending processing of Buddha according to claim 2, wherein the extracting the characteristic of the intrinsic attribute information to construct a user attribute characteristic vector and extracting the characteristic of the user behavior sequence to construct a user behavior sequence characteristic vector comprises:
inputting the user inherent attribute information into an embedding layer of a Transformer model to obtain a first word vector, and taking the first word vector as the user attribute feature vector;
and inputting the user behavior sequence into an embedding layer of a Transformer network to construct a second word vector, and inputting the second word vector into the Transformer layer of the Transformer network to obtain the user behavior sequence characteristic vector.
4. The Buddha warrior recommendation processing method according to claim 1, wherein the selecting a recommended Buddha warrior from the candidate Buddha warriors according to the preference characteristics of the user for the Buddha warrior and the Buddha warrior characteristics of the candidate Buddha warriors comprises:
obtaining the interest degree of the user in each candidate Buddha longitude based on the preference characteristic of the user for the Buddha longitude and the Buddha longitude characteristic of each candidate Buddha longitude;
and selecting a recommended Buddha from the candidate Buddha according to the interest degree of the candidate Buddha.
5. The Buddha recommendation processing method according to claim 4, wherein the obtaining of the interest degree of the user in each candidate Buddha on the basis of the preference characteristics of the user for Buddha and the Buddha characteristics of the candidate Buddha comprises:
analyzing the preference characteristics of the Buddhist scriptures of the user and the Buddhist scriptures of each candidate Buddhist scriptures through the trained interest degree model to obtain the interest degree of the user in each candidate Buddhist scriptures output by the interest degree model;
the trained interest degree model is obtained by training according to a training sample data set marked with interest degrees, and training samples in the training sample data set comprise historical behavior information of a sample user object on Buddhist sutra and Buddhist sutra characteristic information of the sample Buddhist sutra.
6. The Buddhist recommendation process according to claim 5, wherein the trained interestingness model is trained by:
selecting a training sample from the training sample data set, wherein the training sample is marked with the interest degree of a user sample object in the sample Buddha warrior;
for each training sample in the training sample data set, inputting behavior information of a user sample object and Buddha included in the training sample into an untrained interest degree model, and obtaining the interest degree of the user sample object in the Buddha included in the untrained interest degree model;
and optimizing parameters in the untrained interest degree model based on an objective loss function, so that the difference between the interest degree labeled by each training sample and the interest degree obtained by the untrained interest degree model is within an allowable difference range, and obtaining the trained interest degree model, wherein the objective loss function comprises a forward behavior item and a reverse behavior item corresponding to the behavior information of the user sample object.
7. The Buddha recommendation processing method of claim 6, wherein the objective loss function may be a cross-entropy loss function as follows:
wherein N is the number of the training samples, S1A forward behavior loss term, S, corresponding to the training sample in (1)2A reverse behavior loss term, S, corresponding to the training sample in (1)3P (x) is the interest degree of the interest degree model aiming at the training sample x, a1、a2、a3A weight corresponding to each loss term.
8. A Buddha recommendation processing apparatus, comprising:
the system comprises a receiving module, a recommendation module and a recommendation module, wherein the receiving module is used for receiving a Buddha longitude recommendation request triggered by a user through a client;
the obtaining module is used for responding to the Buddha longitude recommendation request and obtaining inherent attribute information and historical behavior information of the user, wherein the historical behavior information is obtained by the user aiming at historical operating behaviors of the Buddha longitude, the historical operating behaviors comprise forward behaviors and reverse behaviors of the user on the Buddha longitude in different time periods, and the forward behaviors and the reverse behaviors represent different behaviors of the user on interest attitudes of the Buddha longitude;
the analysis module is used for analyzing the historical behavior information to obtain the forward behavior and the reverse behavior of the Buddha warrior at different time periods by the user, and constructing a user behavior sequence according to the forward behavior and the reverse behavior;
the determining module is used for determining the favorite features of the user on the Buddhist scriptures according to the inherent attribute information of the user and the user behavior sequence;
the screening module is used for selecting recommended Buddhist sutra from the candidate Buddhist sutra according to the preference characteristics of the Buddhist sutra of the user and the Buddhist sutra characteristics of the candidate Buddhist sutra;
and the pushing module is used for pushing the recommended Buddha channels to the client.
9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for Buddhist recommendation processing according to any one of claims 1-7 when executing the computer program.
10. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the method for Buddha recommendation processing according to any one of claims 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110465027.8A CN113158057A (en) | 2021-04-28 | 2021-04-28 | Buddha meridian recommendation processing device, computer equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110465027.8A CN113158057A (en) | 2021-04-28 | 2021-04-28 | Buddha meridian recommendation processing device, computer equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113158057A true CN113158057A (en) | 2021-07-23 |
Family
ID=76871972
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110465027.8A Pending CN113158057A (en) | 2021-04-28 | 2021-04-28 | Buddha meridian recommendation processing device, computer equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113158057A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113806568A (en) * | 2021-08-10 | 2021-12-17 | 中国人民大学 | Multimedia resource recommendation method and device, electronic equipment and storage medium |
CN115374370A (en) * | 2022-10-26 | 2022-11-22 | 小米汽车科技有限公司 | Content pushing method and device based on multiple models and electronic equipment |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110489639A (en) * | 2019-07-15 | 2019-11-22 | 北京奇艺世纪科技有限公司 | A kind of content recommendation method and device |
CN111784455A (en) * | 2020-06-30 | 2020-10-16 | 腾讯科技(深圳)有限公司 | Article recommendation method and recommendation equipment |
CN112528147A (en) * | 2020-12-10 | 2021-03-19 | 北京百度网讯科技有限公司 | Content recommendation method and apparatus, training method, computing device, and storage medium |
-
2021
- 2021-04-28 CN CN202110465027.8A patent/CN113158057A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110489639A (en) * | 2019-07-15 | 2019-11-22 | 北京奇艺世纪科技有限公司 | A kind of content recommendation method and device |
CN111784455A (en) * | 2020-06-30 | 2020-10-16 | 腾讯科技(深圳)有限公司 | Article recommendation method and recommendation equipment |
CN112528147A (en) * | 2020-12-10 | 2021-03-19 | 北京百度网讯科技有限公司 | Content recommendation method and apparatus, training method, computing device, and storage medium |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113806568A (en) * | 2021-08-10 | 2021-12-17 | 中国人民大学 | Multimedia resource recommendation method and device, electronic equipment and storage medium |
CN113806568B (en) * | 2021-08-10 | 2023-11-03 | 中国人民大学 | Multimedia resource recommendation method and device, electronic equipment and storage medium |
CN115374370A (en) * | 2022-10-26 | 2022-11-22 | 小米汽车科技有限公司 | Content pushing method and device based on multiple models and electronic equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111444428B (en) | Information recommendation method and device based on artificial intelligence, electronic equipment and storage medium | |
CN111966914B (en) | Content recommendation method and device based on artificial intelligence and computer equipment | |
CN111881343A (en) | Information pushing method and device, electronic equipment and computer readable storage medium | |
CN111079056A (en) | Method, device, computer equipment and storage medium for extracting user portrait | |
CN111506820B (en) | Recommendation model, recommendation method, recommendation device, recommendation equipment and recommendation storage medium | |
CN110717099A (en) | Method and terminal for recommending film | |
CN113596130A (en) | Artificial intelligence module training method, system and server based on interest portrait | |
CN113158057A (en) | Buddha meridian recommendation processing device, computer equipment and storage medium | |
CN111506821B (en) | Recommendation model, recommendation method, recommendation device, recommendation equipment and recommendation storage medium | |
US20210035183A1 (en) | Method and system for a recommendation engine utilizing progressive labeling and user content enrichment | |
CN112347787A (en) | Method, device and equipment for classifying aspect level emotion and readable storage medium | |
CN111444419A (en) | Resource recommendation method and device, computer equipment and storage medium | |
CN111914159A (en) | Information recommendation method and terminal | |
CN113343091A (en) | Industrial and enterprise oriented science and technology service recommendation calculation method, medium and program | |
CN115879508A (en) | Data processing method and related device | |
CN110083766B (en) | Query recommendation method and device based on meta-path guiding embedding | |
CN116541517A (en) | Text information processing method, apparatus, device, software program, and storage medium | |
CN116910201A (en) | Dialogue data generation method and related equipment thereof | |
CN115730217A (en) | Model training method, material recalling method and device | |
CN116467523A (en) | News recommendation method, device, electronic equipment and computer readable storage medium | |
CN113434633B (en) | Social topic recommendation method, device, equipment and storage medium based on head portrait | |
CN115470397B (en) | Content recommendation method, device, computer equipment and storage medium | |
CN115017413A (en) | Recommendation method and device, computing equipment and computer storage medium | |
CN114238798A (en) | Search ranking method, system, device and storage medium based on neural network | |
CN115705464A (en) | Information processing method, device and equipment |
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 |