CN112287222B - Content recommendation method based on heterogeneous characteristic depth residual error network - Google Patents
Content recommendation method based on heterogeneous characteristic depth residual error network Download PDFInfo
- Publication number
- CN112287222B CN112287222B CN202011180629.0A CN202011180629A CN112287222B CN 112287222 B CN112287222 B CN 112287222B CN 202011180629 A CN202011180629 A CN 202011180629A CN 112287222 B CN112287222 B CN 112287222B
- Authority
- CN
- China
- Prior art keywords
- data
- content
- depth residual
- user
- obtaining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000012545 processing Methods 0.000 claims abstract description 23
- 238000006243 chemical reaction Methods 0.000 claims abstract description 19
- 238000004422 calculation algorithm Methods 0.000 claims description 47
- 239000011159 matrix material Substances 0.000 claims description 16
- 230000009467 reduction Effects 0.000 claims description 12
- 238000001914 filtration Methods 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 abstract description 6
- 239000013598 vector Substances 0.000 description 26
- 238000010586 diagram Methods 0.000 description 9
- 238000012549 training Methods 0.000 description 9
- 230000008859 change Effects 0.000 description 6
- 238000000354 decomposition reaction Methods 0.000 description 5
- 238000000513 principal component analysis Methods 0.000 description 5
- 238000004590 computer program Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000006872 improvement Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000001965 increasing effect Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 230000002195 synergetic effect Effects 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 210000004205 output neuron Anatomy 0.000 description 1
- 238000012847 principal component analysis method Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 235000019640 taste Nutrition 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
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
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The invention discloses a content recommendation method based on a heterogeneous characteristic depth residual error network, which comprises the following steps: acquiring source data of content to be recommended, and performing data conversion processing on the source data to obtain weighted mixed embedded data; obtaining prediction scoring data according to the weighted mixed embedded data and the depth residual error network model; and obtaining a recommendation result of the content according to the prediction scoring data. According to the method, the source data of the field content are processed, the processed data are input into the residual network model to obtain the prediction grading data, then the field content is accurately recommended according to the prediction grading data, the computing method is high in efficiency and low in resource occupancy rate, and meanwhile cold start of diversified data can be avoided when new users are recommended.
Description
Technical Field
The invention relates to the technical field of communication, in particular to a content recommendation method based on a heterogeneous characteristic depth residual error network.
Background
With the advent of the big data age, people have great demands for recommending field contents such as movies, articles, medical treatment and the like in the big data age, in order to recommend the field contents well, a plurality of methods are developed in the prior art, but the uncertainty of decomposition and verification steps exists in the prior art method, the time complexity of an algorithm is increased by a decomposed low-dimensional matrix, so that the calculation consumption is high, the efficiency is low, a cold start problem can occur when a new user is encountered, and in addition, the prior art method directly carries out operation on sparse data, the resource occupancy rate is high, and the algorithm efficiency is low.
Accordingly, there is a need for improvement and development in the art.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a content recommendation method based on a heterogeneous characteristic depth residual error network aiming at the defects of the prior art, and aims to solve the problems that cold start occurs when a new user is encountered in the prior art, and the prior art method directly carries out operation on sparse data, so that the resource occupancy rate is high and the calculation method is low in efficiency.
The technical scheme adopted by the invention for solving the problems is as follows:
in a first aspect, an embodiment of the present invention provides a content recommendation method based on a heterogeneous feature depth residual network, where the method includes:
acquiring source data of content to be recommended, and performing data conversion processing on the source data to obtain weighted mixed embedded data;
obtaining prediction scoring data according to the weighted mixed embedded data and the depth residual error network model;
and obtaining a recommendation result of the content according to the prediction scoring data.
In one implementation manner, the performing data conversion processing on the source data to obtain weighted mixed embedded data includes:
acquiring cooperative relationship data of a user and a project;
extracting characteristic values of source data of the content to be recommended to obtain characteristic value data;
performing content-based embedded construction on the characteristic value data and the cooperative relationship data of the user and the item to obtain embedded data;
and generating a first triplet loss trainer, and inputting the embedded data into the first triplet loss trainer to obtain weighted mixed embedded data.
In one implementation manner, the extracting the feature value of the source data of the content to be recommended, and obtaining feature value data includes:
according to the mapping matrix, carrying out linear dimension reduction on the source data of the content to be recommended to obtain dimension reduction matrix data;
and extracting the eigenvalue of the dimension reduction matrix data to obtain eigenvalue data.
In one implementation, the obtaining the prediction scoring data according to the weighted hybrid embedded data and the depth residual network model includes:
extracting heterogeneous characteristic data of the weighted mixed embedded data according to the weighted mixed embedded data;
linearly splicing the weighted mixed embedded data and the heterogeneous characteristic data to obtain depth residual error network input data;
and inputting the depth residual error network input data into a depth residual error network model to obtain prediction scoring data.
In one implementation manner, the depth residual network model generation manner is as follows:
acquiring input sample data and output sample data;
generating a second triplet loss trainer, and generating a network model according to the second triplet loss trainer;
inputting the input sample data into a network model to obtain network model output data;
and stopping iteration when the average absolute error rate of the network model output data and the output sample data is smaller than a preset value, and obtaining a depth residual error network model.
In one implementation manner, the obtaining the recommendation result of the content according to the prediction scoring data includes:
obtaining a Baseline algorithm result according to a Baseline algorithm;
and obtaining a recommendation result of the content according to the baseline algorithm result and the prediction scoring data.
In one implementation, the obtaining the recommendation result of the content according to the baseline algorithm result and the prediction score data includes:
obtaining baseline scoring data, user deviation data and project deviation data according to a baseline algorithm result;
according to the baseline scoring data, the user deviation data and the project deviation data, the prediction scoring data is adjusted, and target prediction scoring data is obtained;
and obtaining a recommendation result of the content according to the target prediction scoring data.
In one implementation manner, the obtaining the recommendation result of the content according to the target prediction scoring data further includes:
and when a new user and a new project are introduced, taking the baseline scoring data as target prediction scoring data to obtain a recommendation result of the content.
In a second aspect, an embodiment of the present invention further provides a content recommendation device based on a heterogeneous feature depth residual network, where the device includes:
the system comprises a source data acquisition unit of content to be recommended, a data conversion unit and a data processing unit, wherein the source data acquisition unit is used for acquiring source data of the content to be recommended and performing data conversion processing on the source data to obtain weighted mixed embedded data;
the prediction scoring data acquisition unit is used for acquiring prediction scoring data according to the weighted mixed embedded data and the depth residual error network model;
and the content recommendation result acquisition unit is used for acquiring a recommendation result of the content according to the prediction scoring data.
In a third aspect, an embodiment of the present invention further provides an intelligent terminal, including a memory, and one or more programs, where the one or more programs are stored in the memory, and configured to be executed by the one or more processors, where the one or more programs include a content recommendation method based on the heterogeneous feature depth residual network according to any one of the above.
In a fourth aspect, embodiments of the present invention further provide a non-transitory computer-readable storage medium, which when executed by a processor of an electronic device, enables the electronic device to perform the content recommendation method based on the heterogeneous feature depth residual network as set forth in any one of the above.
The invention has the beneficial effects that: firstly, acquiring source data of content to be recommended in a certain field, and performing data conversion processing on the source data to obtain weighted mixed embedded data; obtaining content prediction scoring data of the field according to the weighted mixed embedded data and the depth residual error network model; finally, according to the predicted scoring data, the content recommendation result of the field can be given. Therefore, the method and the device have the advantages that the source data of the field content are processed, the processed data are input into the residual network model to obtain the prediction scoring data, the field content is accurately recommended according to the prediction scoring data, the computing method is high in efficiency and low in resource occupancy rate, and meanwhile cold start of diversified data can be avoided when new users are recommended.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the present invention, and other drawings may be obtained according to the drawings without inventive effort to those skilled in the art.
Fig. 1 is a schematic flow diagram of a content recommendation method based on a heterogeneous feature depth residual error network according to an embodiment of the present invention
FIG. 2 is a diagram of an HRN overall network architecture of an embodiment of the invention
FIG. 3 is a diagram of a heterogeneous feature extraction architecture of an embodiment of the present invention
FIG. 4 shows a graph of the variation trend of the vector experimental result RMSE according to the embodiment of the invention
FIG. 4 is a graph showing the trend of LOSS change of the vector experiment result according to the embodiment of the invention
FIG. 5 shows a graph of MAE trend of the result of vector experiments according to an embodiment of the present invention
FIG. 6 is a graph of variation of the depth residual network MAE according to an embodiment of the present invention
FIG. 7A variation graph of depth residual network RMSE according to an embodiment of the invention
FIG. 8 a variation MAP of depth residual network MAP according to an embodiment of the invention
FIG. 9 is a graph of variation of the depth residual network MRR of an embodiment of the invention
FIG. 10 is a diagram of a variation of the depth residual network NDCG according to an embodiment of the invention
Fig. 11 is a schematic block diagram of a content recommendation device based on a heterogeneous characteristic depth residual network according to an embodiment of the present invention.
Fig. 12 is a schematic block diagram of an internal structure of an intelligent terminal according to an embodiment of the present invention.
Detailed Description
The invention discloses a content recommendation method based on a heterogeneous characteristic depth residual error network, which is used for making the purposes, the technical scheme and the effects of the invention clearer and more definite, and is further described in detail below by referring to the accompanying drawings and the embodiments. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless expressly stated otherwise, as understood by those skilled in the art. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. It will be understood that when an element is referred to as being "connected" or "coupled" to another element, it can be directly connected or coupled to the other element or intervening elements may also be present. Further, "connected" or "coupled" as used herein may include wirelessly connected or wirelessly coupled. The term "and/or" as used herein includes all or any element and all combination of one or more of the associated listed items.
It will be understood by those skilled in the art that all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs unless defined otherwise. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
In the prior art, the content recommendation method has the problem of cold start when encountering new users, and the prior art method directly carries out operation on sparse data, so that the resource occupancy rate is high, and the algorithm efficiency is low.
In order to solve the problems in the prior art, the embodiment provides a content recommendation method based on a heterogeneous characteristic depth residual error network, which comprises the steps of firstly obtaining source data of content to be recommended, and after user data, project data and user project scoring data are obtained, performing data conversion processing on the source data to facilitate subsequent processing of the data so as to obtain weighted mixed embedded data; extracting features according to the weighted mixed embedded data, vectorizing the features, converting the features into an embedded format required by a recommendation algorithm, and inputting the weighted mixed embedded data and the extracted data into a depth residual error network model to obtain prediction scoring data; and finally, obtaining a recommendation result of the content according to the prediction scoring data. The core of the depth residual network in the embodiment of the invention is HRN, which is a Chinese abbreviation of the hybrid residual network. The HRN base network has a four-layer residual network structure, h=4 means that the network depth of Principal Network (base network) is 4, and is formed by stacking four layers of res net (residual network). The network receives the vector and the characteristic value as input, the vector and the characteristic value respectively enter a Dense layer (full connection layer), the Dense layer adopts tanh (hyperbolic tangent function) as an activation function, and output neurons of the two Dense layers are spliced and input into a mixed residual network.
In order to better extract the features of different levels of data, a deeper network is generally designed, and in addition, the more abstract the features extracted by the deeper network, the more the hidden features of the data can be captured. However, if simply relying on increasing the number of layers of the neural network, not only the purpose cannot be achieved, but also gradient explosions may be caused. In order to solve the problem, the system introduces ResNet (residual network), adds Dropout (discarding the neural network unit from the network according to a certain probability) layer in the network to prevent overfitting, introduces Batch Normalization layer to normalize the input data and the data of the middle layer, thus ensuring that the network adopts random gradient descent in back propagation, and thus the network is converged.
For example, the embodiments of the present invention may be applied to various recommendation fields, in addition to movies, articles and even medical services may be used, for example, by taking a recommended movie as an example, first, items such as a plurality of movie scoring data, user data for watching a movie may be acquired, then the scoring data of the plurality of movies and the user data for watching a movie may be subjected to data conversion processing, weighted mixed embedded data related to movie content may be obtained after the data conversion processing, feature values may be extracted according to the weighted mixed embedded data, for example, the scoring data of a movie may be extracted, viewing scoring data of N users may be extracted according to the weighted mixed embedded data, the data may be subjected to data conversion processing to obtain weighted mixed embedded data, feature value data of the weighted mixed embedded data may be extracted, and the feature value data and the weighted mixed embedded data may be input into a depth residual error network, so as to obtain prediction scoring data, that a model a user and B user may have many similar viewing tastes, and finally, according to the prediction scoring data, for example, the B user may not have a viewing content of B user that has a high viewing score.
Exemplary method
The embodiment provides a content recommendation method based on a heterogeneous characteristic depth residual error network, which can be applied to an IT intelligent terminal. As shown in fig. 1, the method includes:
and step S100, acquiring source data of the content to be recommended, and performing data conversion processing on the source data to obtain weighted mixed embedded data.
Specifically, source data of content to be recommended is obtained, and after user data, project data and user project scoring data are obtained, data conversion processing is performed on the source data to facilitate subsequent processing of the data, so that weighted mixed embedded data is obtained. For example, according to the scoring data of the movie, the viewing scoring data of N users is extracted, and these data are subjected to data conversion processing to obtain weighted mixed embedded data, and the obtained data are prepared for subsequent HRN processing. The HRN overall network structure is shown in fig. 2, and the algorithm flow is as follows: 1) Extracting and vectorizing data features; 2) Constructing a content-based embedding; 3) Enhancing content-based embedding by utilizing a user item synergistic relationship, and constructing weighted mixed embedding; 4) Constructing heterogeneous characteristics, splicing the heterogeneous characteristics and weighting, mixing and embedding; 5) Constructing a depth residual prediction network; 6) Training a residual network using heterogeneous features and weighted hybrid embedding; 7) And obtaining a final prediction result.
In order to obtain weighted mixed embedded data, the data conversion processing is performed on the source data, and the weighted mixed embedded data is obtained, which comprises the following steps:
s101: acquiring cooperative relationship data of a user and a project;
s102: extracting characteristic values of source data of the content to be recommended to obtain characteristic value data;
s103: performing content-based embedded construction on the characteristic value data and the cooperative relationship data of the user and the item to obtain embedded data;
s104: and generating a first triplet loss trainer, and inputting the embedded data into the first triplet loss trainer to obtain weighted mixed embedded data.
Specifically, collaborative relationship data of users and projects, such as user id indexes, project id indexes, user vectors, project vectors, user recurrence times, namely, each user scores each project, project recurrence times, namely, each project receives scores from each user, content-based user vectors, content-based project vectors, collaborative filtering-based user vectors, and collaborative filtering-based project vectors, are obtained. And extracting the characteristic value of the source data of the content to be recommended to obtain characteristic value data. Common statistical-based feature extraction means are: the data set is extracted for its maximum value, average value, standard deviation/variance, etc. And then carrying out content-based embedded construction on the characteristic value data and the cooperative relation data of the user and the project to obtain embedded data. That is, the score received from each user for each item, the content-based user vector, the content-based item vector, the collaborative-filtered user vector, the collaborative-filtered item vector and the data set are embedded based on the maximum value, the average value, the standard deviation/variance, etc., to obtain embedded data. Finally, a first triplet loss trainer is generated and the embedded data is input to the first triplet loss trainer, the triplet loss being a triplet loss function. The first Triplet Loss module is responsible for enhancing the generated content-based embedding with user-item system relationships, training into weighted hybrid embedding, and then generating a module in the presented content-based embedding vector, responsible for converting the input user and item data into a content-based embedding vector.
In order to obtain the proper characteristic value data, the extracting the characteristic value of the source data of the content to be recommended, and obtaining the characteristic value data comprises the following operations: according to the mapping matrix, carrying out linear dimension reduction on the source data of the content to be recommended to obtain dimension reduction matrix data; and extracting the eigenvalue of the dimension reduction matrix data to obtain eigenvalue data.
Specifically, according to the mapping matrix, the source data of the content to be recommended is subjected to linear dimension reduction processing, the dimension reduction matrix adopts PCA (principal component analysis), the PCA uses an orthogonal transformation matrix as the mapping matrix, the initial data is transformed into a group of representations with each dimension being linearly irrelevant, and the representations are used for extracting main characteristic components of the data, so that the aim of expressing the whole data by the data with fewer characteristics is fulfilled, and the method is a linear dimension reduction method. The depth residual error network uses a principal component analysis method PCA (Principle Component Analysis) to reduce the dimension of the characterized high-dimensional matrix embedding, maps the high-dimensional matrix embedding into a low-dimensional space, converts the high-dimensional sparse embedding into a low-dimensional dense embedding, improves the space utilization rate, and can solve the problems of high memory occupation and long calculation time.
The embodiment provides a content recommendation method based on a heterogeneous characteristic depth residual error network, which can be applied to an IT intelligent terminal. As shown in fig. 1, the method includes:
s200: and obtaining prediction scoring data according to the weighted mixed embedded data and the depth residual error network model.
Specifically, after the weighted mixed embedded data is obtained, the data is input into a depth residual error network model, the core of the depth residual error network model is HRN, and the design basis of the HRN is: 1) In order to reduce the influence of cold start on a recommendation result, the HRN designs a Triplet Loss trainer, trains and simultaneously takes account of weighted mixed embedding based on content and collaborative filtering as the input of a prediction network. 2) Second, because of the large number of users and items, the number of items with which the user interacts is only a very small fraction. In order to avoid directly performing operations on extremely sparse user-to-item evaluation matrices, HRN uses PCA to dimensionality-reduce the embedding, uses Triplet pass to enhance content-based embedding with user-item synergistic relationships, converts the original user and item data to low-dimensional, while maintaining a well-characterized weighted hybrid embedding vector. And (5) inputting the weighted mixed embedded data into a depth residual error network model to obtain prediction scoring data.
In order to obtain prediction scoring data, the step of obtaining the prediction scoring data according to the weighted mixed embedded data and the depth residual error network model comprises the following steps:
s201: extracting heterogeneous characteristic data of the weighted mixed embedded data according to the weighted mixed embedded data;
s202: linearly splicing the weighted mixed embedded data and the heterogeneous characteristic data to obtain depth residual error network input data;
s203: and inputting the depth residual error network input data into a depth residual error network model to obtain prediction scoring data.
In practice, after the weighted mixed embedded data is acquired, heterogeneous feature data of the weighted mixed embedded data is extracted, and heterogeneous feature extraction aims at mining implicit relations among different types of features. The heterogeneous information network HIN comprises a directed graph, wherein nodes are used for representing entities, and edges are used for representing relations, and the directed graph is similar to a knowledge graph. Inspired by the heterogeneous information network of the HIN, the system fuses various learned feature vectors into heterogeneous feature vectors in a linear splicing mode, and the heterogeneous feature vectors and the learned homogeneous vectors (sequences) are subjected to linear splicing to serve as input of the depth residual error network. Based on this, heterogeneous feature extraction as shown in fig. 3 is constructed herein. Then, the weighted mixed embedded data and the heterogeneous characteristic data are linearly spliced, and the formed data are used as depth residual error network input data; because the depth residual error network model is established, the prediction scoring data can be obtained by inputting the depth residual error network input data into the depth residual error network model.
In order to use the depth residual error network model, a depth residual error network model is firstly generated, and the depth residual error network model is generated by the following steps: acquiring input sample data and output sample data; generating a second triplet loss trainer, and generating a network model according to the second triplet loss trainer; inputting the input sample data into a network model to obtain network model output data; and stopping iteration when the average absolute error rate of the network model output data and the output sample data is smaller than a preset value, and obtaining a depth residual error network model.
In practice, the network model needs to be obtained by training a model according to a set of real sample data, so that input data and output data need to be acquired, and then a second Triplet Loss trainer is generated, wherein the Triplet Loss is a Triplet Loss function, the Triplet Loss is used for learning more refined emplacement of a face image, the emplacement is a mode of converting discrete variables into continuous vector representation, and similar images are similar in the emplacement space. Triplet Loss has finer granularity than a conventional trainer. The Triplet Loss is more suitable for accurate identification, is mainly applied to the fine granularity identification problems of face identification, identity identification, vehicle identification and the like, can greatly reduce the output dimension of a network, and can obtain better feature embedding. And inputting the input sample data into a network model to obtain network model output data, wherein the output data is obtained by the network model according to the input data, and in order to obtain a final depth residual error network model, the average absolute error rate of the network model output data and the output sample data is required to be evaluated, and when the average absolute error rate of the network model output data and the output sample data is smaller than a preset value, the model training is successful, iteration is stopped, and the obtained depth residual error network model can be used.
After the depth residual network training in the embodiment of the present invention, the variation trend of RMSE (root mean square error) along with the increase of epoch (first generation training) is shown in the left side of fig. 4, the variation trend of LOSS (LOSS) along with the increase of epoch (first generation training) is shown in the right side of fig. 4, and the variation trend of MAE (mean absolute error) along with the increase of epoch (first generation training) is shown in fig. 5.
As the number of iterations increases, the LOSS, RMSE and MAE of the network decrease gradually, and when epoch reaches 30, the LOSS, RMSE and MAE of the network decrease from 3.6843,0.8995,0.7122 to 3.6234,0.8465,0.6694, respectively. It can be seen that over 30 iterations, the LOSS, RMSE and MAE of the depth residual network all drop to a relatively low level.
As can be seen from the left part of fig. 4, the right part of fig. 4 and the right part of fig. 5, the algorithm involved in comparison is roughly divided into three camps for each index. The Normal (positive rail algorithm), baseline (Baseline algorithm), clustering algorithm are classified into a camping, SVD (singular value decomposition algorithm), SVD++ (singular value decomposition improvement algorithm) are classified into a camping, a depth residual error network based on heterogeneous characteristics, hybrid_ResNet (HRN) algorithm trained by using weighted mixed embedding, and a camping formed by using hybrid_SVD (mixed singular value decomposition algorithm) and hybrid_SVD++ (mixed singular value decomposition improvement algorithm) trained by SVD and SVD++ respectively. It can be seen that as the number of user scores increases, each index of each algorithm becomes better, which also means that the more the user interacts with the item, the better the algorithm can capture the user's preference, and a better prediction can be made. By transverse comparison, the three algorithms, HRN, HSD and HSD++ trained by the method proposed by the system are found to perform significantly better than other algorithms, such as SVD and SVD++ algorithms. Second, the Baseline, coClustering, normal algorithm performs poorly. SVD++ is used as an improved version of SVD, and performs best except for the first camping set forth by the system, and various indexes are improved compared with SVD.
For the three networks proposed by the system, HRN, HSD and HSD++, the system is built based on the same weighted mixed embedding which fuses the content-based characteristics and the collaborative filtering characteristics, and the difference is that the hybrid_ResNet is compared with the content-based characteristics and the collaborative filtering characteristics, the heterogeneous characteristics are added for mining the implicit characteristics of user projects, the heterogeneous characteristics are input into a depth residual error network, compared with a common depth network, the residual error network improves the learning efficiency, and the deeper network structure is also beneficial to learning more abstract characteristics and does not influence the network performance, so that the system is also the reason for selecting the residual error network. The HSD and HSD++ are based on weighted mixed embedding, and SVD and SVD++ algorithm are used to train the network respectively. From the performance comparison result, the HRN index is better than the HRN index and the HRN index proves that the depth residual error network is utilized to mine the implicit characteristics of the user project, which is really helpful for improving the network prediction performance.
The more interactive the user has with the item, the more the recommendation algorithm can learn the user's preferences, and in the case of a ML (canonical data set) (1M) data set, the more movies the user participates in scoring, the less prediction scoring errors the HRN gives. With the increase of the user-rating-0count user rating, fig. 6 shows MAE (mean absolute error) change, fig. 7 shows RMSE (root mean square error) change, fig. 8 shows MAP (mean average precision) change, fig. 9 shows MRR (mean reciprocal rank) change, fig. 10 shows NDCG (normalized break cumulative gain) change, and with the richer rating data of the user, the algorithm is indeed more able to capture the user's behavior feature preference, making a more accurate prediction.
The embodiment provides a content recommendation method based on a heterogeneous characteristic depth residual error network, which can be applied to an IT intelligent terminal. As shown in fig. 1, the method includes:
s300: and obtaining a recommendation result of the content according to the prediction scoring data.
Specifically, the HRN takes the weighted hybrid embedding and heterogeneous feature vector as the input of the depth residual error network, and obtains prediction scoring data through training. The HRN fuses the feature vectors with different properties into heterogeneous feature vectors, and then evaluates the similarity of the new user, the new project, other users and projects and makes content recommendation according to the similarity.
In order to obtain a detailed recommendation result, the step of obtaining a recommendation result of the content according to the prediction scoring data comprises the following steps:
s301: obtaining a Baseline algorithm result according to a Baseline algorithm;
s302: and obtaining a recommendation result of the content according to the baseline algorithm result and the prediction scoring data.
In this embodiment, the Baseline algorithm is an algorithm independent of the depth residual network, and before the depth residual network is built, a Baseline algorithm is built, and the final recommended result is the combination of the result obtained by the Baseline algorithm and the result obtained by the HRN. The introduction of the baseline algorithm ensures that the network proposed by the system has a comparison baseline, and is beneficial to the improvement of the effect of the follow-up observation algorithm. And finally, correcting the predicted scoring data according to the baseline algorithm result to obtain final scoring data, and obtaining the recommended result of the field content according to the final scoring data.
In a specific implementation, the obtaining the recommendation result of the content according to the baseline algorithm result and the prediction scoring data includes the following operations: obtaining baseline scoring data, user deviation data and project deviation data according to a baseline algorithm result; according to the baseline scoring data, the user deviation data and the project deviation data, the prediction scoring data is adjusted, and target prediction scoring data is obtained; and obtaining a recommendation result of the content according to the target prediction scoring data.
In one embodiment, the basic idea due to the Baseline algorithm is: a baseline is established, introducing user bias and project bias. For example, the user bias in the movie field is the bias of the viewing habit score of each user, and the project bias in the movie field is the bias of the work itself of each movie. According to the baseline scoring data, the user deviation data and the project deviation data, the prediction scoring data is adjusted, and target prediction scoring data is obtained; the prediction score can be adjusted according to the deviation of the viewing habit score of each user and the deviation of the work of each movie to obtain target prediction score data, and then the recommendation result of the content is obtained according to the usual viewing habit of the user to determine whether to recommend movie content to the user. For example, the film watching score is carried out on the user A, the film watching baseline score is 7, the film is a good film, the project score is increased by 0.5, but the predicted score of the user A output according to the depth residual error network is generally lower than that of other people by 0.3, namely, the user A is a harsh user, so that the score of the user A on the film is obtained by subtracting 0.3, the score of the user A on the film is 7.2, and the film with the score more than 6.5 is watched according to the general watching score of the user A in practice, for example, the user A only watches the film with the score more than 6.5, and the film A can be recommended to the user A according to the result of the score of 7.2.
In order to solve the problem of cold start when a new user is introduced, the method for obtaining the recommendation result of the content according to the target prediction scoring data further comprises the following operations: and when a new user and a new project are introduced, taking the baseline scoring data as target prediction scoring data to obtain a recommendation result of the content.
Specifically, according to the baseline algorithm, baseline scoring data can be obtained, that is, a reference score is provided for a film by the masses, so that when a new user and a new project are introduced, the baseline scoring data is used as target prediction scoring data, and content recommendation is performed according to the target prediction scoring data. For example, when a new user is introduced, the user cannot be subjected to similarity assessment with other users without any information about the past of the user, and at this time, the new user is determined to recommend what content is recommended to the new user according to the target prediction scoring data by taking the baseline scoring data as the target prediction scoring data.
Exemplary apparatus
As shown in fig. 11, an embodiment of the present invention provides a content recommendation device based on a heterogeneous feature depth residual network, the device including: a source data acquisition unit 401 of the content to be recommended, a prediction score data acquisition unit 402, a content recommendation result acquisition unit 403;
a source data obtaining unit 401 for obtaining source data of the content to be recommended, and performing data conversion processing on the source data to obtain weighted mixed embedded data;
a prediction score data obtaining unit 402, configured to obtain prediction score data according to the weighted mixed embedded data and the depth residual error network model;
and a content recommendation result obtaining unit 403, configured to obtain a recommendation result of the content according to the prediction score data.
Based on the above embodiment, the present invention also provides an intelligent terminal, and a functional block diagram thereof may be shown in fig. 12. The intelligent terminal comprises a processor, a memory, a network interface, a display screen and a temperature sensor which are connected through a system bus. The processor of the intelligent terminal is used for providing computing and control capabilities. The memory of the intelligent terminal comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface of the intelligent terminal is used for communicating with an external terminal through network connection. The computer program, when executed by a processor, implements a content recommendation method based on a heterogeneous feature depth residual network. The display screen of the intelligent terminal can be a liquid crystal display screen or an electronic ink display screen, and a temperature sensor of the intelligent terminal is arranged in the intelligent terminal in advance and used for detecting the running temperature of internal equipment.
It will be appreciated by those skilled in the art that the schematic diagram in fig. 12 is merely a block diagram of a portion of the structure related to the present invention and is not limiting of the smart terminal to which the present invention is applied, and that a specific smart terminal may include more or less components than those shown in the drawings, or may combine some components, or have a different arrangement of components.
In one embodiment, a smart terminal is provided that includes a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors, the one or more programs comprising instructions for: acquiring source data of content to be recommended, and performing data conversion processing on the source data to obtain weighted mixed embedded data; obtaining prediction scoring data according to the weighted mixed embedded data and the depth residual error network model; and obtaining a recommendation result of the content according to the prediction scoring data.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), memory bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
In summary, the invention discloses a content recommendation method, an intelligent terminal and a storage medium based on a heterogeneous characteristic depth residual error network, wherein the method comprises the following steps: acquiring source data of content to be recommended, and performing data conversion processing on the source data to obtain weighted mixed embedded data; obtaining prediction scoring data according to the weighted mixed embedded data and the depth residual error network model; and obtaining a recommendation result of the content according to the prediction scoring data. According to the method, the source data of the field content are processed, the processed data are input into the residual network model to obtain the prediction grading data, then the field content is accurately recommended according to the prediction grading data, the computing method is high in efficiency and low in resource occupancy rate, and meanwhile cold start of diversified data can be avoided when new users are recommended.
It should be understood that the present invention discloses a content recommendation method based on heterogeneous characteristic depth residual network, and it should be understood that the application of the present invention is not limited to the above examples, and that modifications and changes may be made by those skilled in the art in light of the above description, and all such modifications and changes should fall within the scope of the appended claims.
Claims (8)
1. A content recommendation method based on a heterogeneous feature depth residual network, the method comprising:
acquiring source data of content to be recommended, and performing data conversion processing on the source data to obtain weighted mixed embedded data;
obtaining prediction scoring data according to the weighted mixed embedded data and the depth residual error network model;
obtaining a recommendation result of the content according to the prediction scoring data;
the step of performing data conversion processing on the source data to obtain weighted mixed embedded data includes:
acquiring collaborative relationship data of a user and an item, wherein the collaborative relationship data of the user and the item comprises user data, item data, user item scoring data, content-based user data, content-based item data, collaborative filtering-based user data and collaborative filtering-based item data;
extracting characteristic values of source data of the content to be recommended to obtain characteristic value data;
performing content-based embedded construction on the characteristic value data and the cooperative relationship data of the user and the item to obtain embedded data;
generating a first tripleless trainer, and inputting the embedded data into the first tripleless trainer to obtain weighted mixed embedded data;
the depth residual error network model generation mode is as follows:
acquiring input sample data and output sample data;
generating a second triplet loss trainer, and generating a network model according to the second triplet loss trainer;
inputting the input sample data into a network model to obtain network model output data;
and stopping iteration when the average absolute error rate of the network model output data and the output sample data is smaller than a preset value, and obtaining a depth residual error network model.
2. The content recommendation method based on the heterogeneous feature depth residual network according to claim 1, wherein extracting the feature value of the source data of the content to be recommended, obtaining feature value data comprises:
according to the mapping matrix, carrying out linear dimension reduction on the source data of the content to be recommended to obtain dimension reduction matrix data;
and extracting the eigenvalue of the dimension reduction matrix data to obtain eigenvalue data.
3. The heterogeneous feature depth residual network-based content recommendation method according to claim 2, wherein obtaining prediction scoring data according to the weighted hybrid embedded data and depth residual network model comprises:
extracting heterogeneous characteristic data of the weighted mixed embedded data according to the weighted mixed embedded data;
linearly splicing the weighted mixed embedded data and the heterogeneous characteristic data to obtain depth residual error network input data;
and inputting the depth residual error network input data into a depth residual error network model to obtain prediction scoring data.
4. The method for recommending content based on heterogeneous feature depth residual network according to claim 1, wherein obtaining the recommendation result of the content according to the prediction scoring data comprises:
obtaining a Baseline algorithm result according to a Baseline algorithm;
and obtaining a recommendation result of the content according to the baseline algorithm result and the prediction scoring data.
5. The method for recommending content based on a heterogeneous feature depth residual network according to claim 4, wherein obtaining a recommendation result of content according to the baseline algorithm result and the prediction scoring data comprises:
obtaining baseline scoring data, user deviation data and project deviation data according to a baseline algorithm result;
according to the baseline scoring data, the user deviation data and the project deviation data, the prediction scoring data is adjusted, and target prediction scoring data is obtained;
and obtaining a recommendation result of the content according to the target prediction scoring data.
6. The method for recommending content based on a heterogeneous feature depth residual network according to claim 5, wherein obtaining a recommendation result of content according to the target prediction scoring data further comprises:
and when a new user and a new project are introduced, taking the baseline scoring data as target prediction scoring data to obtain a recommendation result of the content.
7. An intelligent terminal comprising a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors, the one or more programs comprising instructions for performing the method of any of claims 1-6.
8. A non-transitory computer readable storage medium, wherein instructions in the storage medium, when executed by a processor of an electronic device, enable the electronic device to perform the method of any one of claims 1-6.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011180629.0A CN112287222B (en) | 2020-10-29 | 2020-10-29 | Content recommendation method based on heterogeneous characteristic depth residual error network |
| PCT/CN2020/136144 WO2022088417A1 (en) | 2020-10-29 | 2020-12-14 | Content recommendation method based on heterogeneous feature deep residual network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011180629.0A CN112287222B (en) | 2020-10-29 | 2020-10-29 | Content recommendation method based on heterogeneous characteristic depth residual error network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112287222A CN112287222A (en) | 2021-01-29 |
| CN112287222B true CN112287222B (en) | 2023-12-15 |
Family
ID=74373921
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011180629.0A Active CN112287222B (en) | 2020-10-29 | 2020-10-29 | Content recommendation method based on heterogeneous characteristic depth residual error network |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN112287222B (en) |
| WO (1) | WO2022088417A1 (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107832804A (en) * | 2017-10-30 | 2018-03-23 | 上海寒武纪信息科技有限公司 | A kind of information processing method and Related product |
| CN109785062A (en) * | 2019-01-10 | 2019-05-21 | 电子科技大学 | A kind of hybrid neural networks recommender system based on collaborative filtering model |
| CN111488529A (en) * | 2020-06-28 | 2020-08-04 | 腾讯科技（深圳）有限公司 | Information processing method, information processing apparatus, server, and storage medium |
| CN111506811A (en) * | 2020-03-19 | 2020-08-07 | 上海理工大学 | Click rate prediction method based on deep residual error network |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9886945B1 (en) * | 2011-07-03 | 2018-02-06 | Reality Analytics, Inc. | System and method for taxonomically distinguishing sample data captured from biota sources |
| CN109933678B (en) * | 2019-03-07 | 2021-04-06 | 合肥工业大学 | Artwork recommendation method and device, readable medium and electronic equipment |
| CN110674265B (en) * | 2019-08-06 | 2021-03-02 | 上海孚典智能科技有限公司 | Unstructured information oriented feature discrimination and information recommendation system |
| CN111651613B (en) * | 2020-07-08 | 2021-07-27 | 海南大学 | Knowledge graph embedding-based dynamic recommendation method and system |
-
2020
- 2020-10-29 CN CN202011180629.0A patent/CN112287222B/en active Active
- 2020-12-14 WO PCT/CN2020/136144 patent/WO2022088417A1/en active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107832804A (en) * | 2017-10-30 | 2018-03-23 | 上海寒武纪信息科技有限公司 | A kind of information processing method and Related product |
| CN109785062A (en) * | 2019-01-10 | 2019-05-21 | 电子科技大学 | A kind of hybrid neural networks recommender system based on collaborative filtering model |
| CN111506811A (en) * | 2020-03-19 | 2020-08-07 | 上海理工大学 | Click rate prediction method based on deep residual error network |
| CN111488529A (en) * | 2020-06-28 | 2020-08-04 | 腾讯科技（深圳）有限公司 | Information processing method, information processing apparatus, server, and storage medium |
Non-Patent Citations (3)
| Title |
|---|
| Deep Residual Learning for Image Recognition;Kaiming He等;2016 IEEE Conference on Computer Vision and Pattern Recognition(CVPR);1-12 * |
| 基于 Resnet&FM 模型的电影个性化推荐研究;简恒;中国优秀硕士学位论文全文数据库 哲学与人文科学辑;2020-01-15;F087-452页 * |
| 深度学习框架下充电桩的个性化推荐算法研究;吴敏;中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑;第C037-207页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112287222A (en) | 2021-01-29 |
| WO2022088417A1 (en) | 2022-05-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210295162A1 (en) | Neural network model training method and apparatus, computer device, and storage medium | |
| EP3937124A1 (en) | Image processing method, device and apparatus, and storage medium | |
| CN112270196A (en) | Entity relationship identification method and device and electronic equipment | |
| Chen et al. | No-reference screen content image quality assessment with unsupervised domain adaptation | |
| Zhong et al. | A Group-Based Personalized Model for Image Privacy Classification and Labeling. | |
| CN114783514A (en) | Method for predicting binding affinity of drug molecules and target protein | |
| CN112001399A (en) | Image scene classification method and device based on local feature saliency | |
| Zhu et al. | Age estimation algorithm of facial images based on multi-label sorting | |
| CN114861757A (en) | Deep clustering method for images based on dual-correlation reduction | |
| Long et al. | Trainable subspaces for low rank tensor completion: Model and analysis | |
| Tsai et al. | Formalizing generalization and robustness of neural networks to weight perturbations | |
| CN110807693A (en) | Album recommendation method, device, equipment and storage medium | |
| CN113221003B (en) | Mixed filtering recommendation method and system based on dual theory | |
| CN115131803A (en) | Document word size identification method and device, computer equipment and storage medium | |
| Jing et al. | Knowledge-enhanced attentive learning for answer selection in community question answering systems | |
| CN112287222B (en) | Content recommendation method based on heterogeneous characteristic depth residual error network | |
| CN115525782A (en) | Video abstract generation method of self-adaptive graph structure | |
| CN116108363A (en) | Incomplete multi-view multi-label classification method and system based on label guidance | |
| Shi et al. | Face-based age estimation using improved Swin Transformer with attention-based convolution | |
| CN111651626B (en) | Image classification method, device and readable storage medium | |
| CN114491296A (en) | Proposal affiliate recommendation method, system, computer device and readable storage medium | |
| CN113590815A (en) | Method and system for classifying hierarchical multi-element label texts | |
| CN111159450A (en) | Picture classification method and device, computer equipment and storage medium | |
| Wang et al. | Multi‐Task and Attention Collaborative Network for Facial Emotion Recognition | |
| Deng et al. | ContentCTR: Frame-level Live Streaming Click-Through Rate Prediction with Multimodal Transformer |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |