CN112561663A - Vehicle recommendation method and device, computer equipment and storage medium - Google Patents
Vehicle recommendation method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN112561663A CN112561663A CN202011573596.6A CN202011573596A CN112561663A CN 112561663 A CN112561663 A CN 112561663A CN 202011573596 A CN202011573596 A CN 202011573596A CN 112561663 A CN112561663 A CN 112561663A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- information
- vector
- feature vector
- user
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 239000013598 vector Substances 0.000 claims abstract description 213
- 238000012512 characterization method Methods 0.000 claims abstract description 49
- 238000012545 processing Methods 0.000 claims abstract description 38
- 230000006399 behavior Effects 0.000 claims description 40
- 238000012163 sequencing technique Methods 0.000 claims description 21
- 238000003062 neural network model Methods 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 12
- 238000005295 random walk Methods 0.000 claims description 9
- 238000013528 artificial neural network Methods 0.000 abstract description 5
- 230000000694 effects Effects 0.000 abstract description 5
- 230000007547 defect Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 15
- 238000004891 communication Methods 0.000 description 9
- 230000006870 function Effects 0.000 description 4
- 230000002452 interceptive effect Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000011144 upstream manufacturing Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013523 data management Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0641—Shopping interfaces
- G06Q30/0643—Graphical representation of items or shoppers
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application relates to a vehicle recommendation method, a device, a computer device and a storage medium, wherein the vehicle recommendation method comprises the following steps: acquiring first vehicle information and user pair historical flow information; processing first vehicle information and the historical flow information through a preset vector characterization model to obtain a first feature vector of the first vehicle; and reading the stored second feature vector, sorting the first vehicle based on the first feature vector and the second feature vector, and determining a recommended vehicle according to the sorting result. According to the vehicle recommendation method, the vehicle recommendation device, the computer equipment and the storage medium, the characteristics of multiple modes of the vehicle are converted into the vector statements representing the similarity through the neural network, and the vector statements are compared with the vehicle types purchased by the user for recommendation, so that the defect that the vehicle types interacted by the user can only be recommended in the original marketing process is overcome, the overall marketing effect is improved, and the vehicle recommendation accuracy is improved.
Description
Technical Field
The present application relates to the field of deep learning, and in particular, to a vehicle recommendation method, apparatus, computer device, and storage medium.
Background
With the explosion of electronic commerce, the online consumption is higher and higher in the retail sales of commodities in the whole society. The great shopping advantages provided by online shopping for consumers are mainly reflected in breaking through space-time limitation, convenience in shopping, more commodity choices, competitive price, rich commodity information, personalization and customization. The method and the system improve the operation and advertisement putting efficiency of the e-commerce by using the behavior data of the user, and are concerned by more and more e-commerce enterprises and advertisers.
In the marketing scenario of the automobile B2B, a DMP (Data Management Platform) generates preferences of a user on a model and a brand as an image of an automobile dealer according to browsing clicks of the user on the automobile, and helps the Platform perform personalized marketing on the automobile dealer. In the marketing process, when short messages or push marketing is performed, in order to promote highly-intentioned vehicle types, the following labels are often used: recent transaction vehicle types, recent browsing vehicle types and recent click vehicle types. However, these labels are not flexible enough, and the recent purchasing view does not necessarily represent the future preference, and at the same time, these labels are too simple and rough, and the long-tail preference of the user is ignored in this process, for example, the vehicle which the user does not interact with is not without the purchasing intention, but is caused by the temporary market situation, and finally, the recommendation list of the user is pushed more and more narrowly, and the exposure of other vehicle types on the B2B platform is not enough.
At present, no effective solution is provided for the problem of low vehicle recommendation accuracy in the related art.
Disclosure of Invention
The embodiment of the application provides a vehicle recommendation method and device, computer equipment and a storage medium, and aims to at least solve the problem of low vehicle recommendation accuracy in the related art.
In a first aspect, an embodiment of the present application provides a vehicle recommendation method, including:
acquiring first vehicle information of a first vehicle to be recommended and historical flow information generated by browsing and clicking second vehicle information by a user, wherein the first vehicle information comprises vehicle type information and appearance information of the first vehicle;
processing the first vehicle information and the historical flow information through a preset vector characterization model to obtain a first feature vector of the first vehicle;
and reading a stored second feature vector, sequencing the first vehicle based on the first feature vector and the second feature vector, and determining a recommended vehicle according to the sequencing result, wherein the second feature vector is generated by processing historical transaction information of a second vehicle by the vector characterization model and is used for characterizing an intention label of a user for the second vehicle.
In some embodiments, the obtaining the first vehicle information of the first vehicle to be recommended includes:
acquiring at least one of the type, the series and the brand of the first vehicle as the type information;
acquiring a vehicle image of the first vehicle;
and inputting the vehicle image into a trained neural network model to obtain appearance information of the first vehicle, wherein the appearance information comprises an appearance vector representing the appearance similarity of the vehicle.
In some embodiments, the processing the first vehicle information and the historical flow rate information through a preset vector characterization model to obtain a first feature vector of the first vehicle includes:
generating a user click behavior sequence based on the historical traffic information;
and processing the first vehicle information and the user click behavior sequence through a preset vector characterization model to obtain a first feature vector of the first vehicle.
In some of these embodiments, the generating a sequence of user click behaviors based on the historical traffic information comprises:
constructing a directed unweighted graph based on the historical traffic information, wherein the directed unweighted graph comprises vehicle information and a vehicle click sequence;
and selecting any initial node in the directed unweighted graph, and generating a user click behavior sequence in a random walk mode.
In some embodiments, the processing the first vehicle information and the user click behavior sequence through a preset vector characterization model to obtain a first feature vector of the first vehicle includes:
taking the first vehicle information as input, taking an adjacent vehicle of the first vehicle in the user click behavior sequence as output, training the vector characterization model to obtain a hidden layer vector, and taking the hidden layer vector as a first feature vector of the first vehicle.
In some embodiments, reading the stored second feature vector further comprises:
training the vector characterization model based on the user click behavior sequence and historical transaction information of a second vehicle to obtain a hidden layer vector of the second vehicle;
and acquiring the average value of the hidden layer vector of the second vehicle based on the hidden layer vector of the second vehicle to obtain the second feature vector.
In some of these embodiments, the ranking the first vehicle based on the first feature vector and the second feature vector, the determining a recommended vehicle according to the ranking result comprising:
obtaining cosine similarity of the first feature vector and the second feature vector;
ranking the first vehicle based on the cosine similarity;
and selecting a preset number of first vehicles for recommendation from high to low based on the ranking.
In a second aspect, an embodiment of the present application provides a vehicle recommendation device, including:
the system comprises an acquisition module, a recommendation module and a recommendation module, wherein the acquisition module is used for acquiring first vehicle information of a first vehicle to be recommended and historical flow information generated by browsing and clicking second vehicle information by a user, and the first vehicle information comprises vehicle type information and appearance information of the first vehicle;
the first processing module is used for processing the first vehicle information and the historical flow information through a preset vector representation model to obtain a first feature vector of the first vehicle;
and the recommending module is used for reading a stored second feature vector, sequencing the first vehicle based on the first feature vector and the second feature vector, and determining a recommended vehicle according to the sequencing result, wherein the second feature vector is generated by processing the historical transaction information of the second vehicle by the vector characterization model and is used for characterizing the intention label of the user to the second vehicle.
In a third aspect, an embodiment of the present application provides a computer device, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor, when executing the computer program, implements the vehicle recommendation method according to the first aspect.
In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the vehicle recommendation method according to the first aspect.
Compared with the related art, the vehicle recommendation method, the device, the computer device and the storage medium provided by the embodiment of the application acquire the first vehicle information of the first vehicle to be recommended and the historical flow information generated by browsing and clicking the second vehicle information by the user, wherein the first vehicle information comprises the vehicle type information and the appearance information of the first vehicle; processing the first vehicle information and the historical flow information through a preset vector characterization model to obtain a first feature vector of the first vehicle; and reading the stored second feature vector, sequencing the first vehicle based on the first feature vector and the second feature vector, and determining a recommended vehicle according to the sequencing result, wherein the second feature vector is generated by processing the historical transaction information of the second vehicle by the vector characterization model and is used for characterizing the intention label of the user for the second vehicle, so that the defect and the deficiency of low vehicle recommendation accuracy in the related technology are overcome, and the vehicle recommendation accuracy is improved.
The details of one or more embodiments of the application are set forth in the accompanying drawings and the description below to provide a more thorough understanding of the application.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
FIG. 1 is a schematic flow diagram of a vehicle recommendation method according to an embodiment of the present application;
FIG. 2 is a block diagram of a trained neural network model of a vehicle recommendation method according to an embodiment of the present application;
FIG. 3 is a schematic diagram of a directed, unweighted graph of a vehicle recommendation method according to an embodiment of the present application;
FIG. 4 is a block diagram of a vector characterization model of a vehicle recommendation method according to an embodiment of the present application;
FIG. 5 is a schematic flow chart diagram of a vehicle recommendation method in accordance with a preferred embodiment of the present application;
FIG. 6 is a block diagram of a vehicle recommendation device according to an embodiment of the present application;
fig. 7 is a hardware configuration diagram of a computer device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments provided in the present application without any inventive step are within the scope of protection of the present application.
It is obvious that the drawings in the following description are only examples or embodiments of the present application, and that it is also possible for a person skilled in the art to apply the present application to other similar contexts on the basis of these drawings without inventive effort. Moreover, it should be appreciated that in the development of any such actual implementation, as in any engineering or design project, numerous implementation-specific decisions must be made to achieve the developers' specific goals, such as compliance with system-related and business-related constraints, which may vary from one implementation to another.
Reference in the specification to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the specification. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of ordinary skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments without conflict.
Unless defined otherwise, technical or scientific terms referred to herein shall have the ordinary meaning as understood by those of ordinary skill in the art to which this application belongs. Reference to "a," "an," "the," and similar words throughout this application are not to be construed as limiting in number, and may refer to the singular or the plural. The present application is directed to the use of the terms "including," "comprising," "having," and any variations thereof, which are intended to cover non-exclusive inclusions; for example, a process, method, system, article, or apparatus that comprises a list of steps or modules (elements) is not limited to the listed steps or elements, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus. Reference to "connected," "coupled," and the like in this application is not intended to be limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The term "plurality" as referred to herein means two or more. "and/or" describes an association relationship of associated objects, meaning that three relationships may exist, for example, "A and/or B" may mean: a exists alone, A and B exist simultaneously, and B exists alone. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship. Reference herein to the terms "first," "second," "third," and the like, are merely to distinguish similar objects and do not denote a particular ordering for the objects.
Before describing and explaining embodiments of the present application, a description will be given of the related art used in the present application as follows: and the vector characterization model is a neural network model adopting a word2vec algorithm. The neural network model, which employs the word2vec algorithm, is a group of correlation models that are used to generate word vectors. These models are shallow, two-layer neural networks that are trained to reconstruct linguistic word text. The network is represented by words and the input words in adjacent positions are guessed, and the order of the words is unimportant under the assumption of the bag-of-words model in word2 vec. After training is completed, the word2vec model can be used to map each word to a vector, which can be used to represent word-to-word relationships, and the vector is a hidden layer of the neural network.
Fig. 1 is a flowchart illustrating a vehicle recommendation method according to an embodiment of the present application. As shown in fig. 1, the process includes:
step S101, obtaining first vehicle information of a first vehicle to be recommended and historical flow information generated by browsing and clicking second vehicle information by a user, wherein the first vehicle information comprises vehicle type information and appearance information of the first vehicle.
Illustratively, the historical traffic information of the user is interactive data generated when the user browses and clicks the recommended vehicle within a preset time, for example, the type, the train system, and the brand of the vehicle clicked and viewed by the user, the order in which the user clicks and views the vehicle, and the like. It is understood that the historical traffic information of the subscriber can be obtained through a buried point at the subscriber end. In this embodiment, the first vehicle is a vehicle that a vehicle dealer needs to push for marketing, and the specific range of the first vehicle may be specifically determined according to an actual situation, which is not limited herein.
Step S102, processing the first vehicle information and the historical flow information through a preset vector representation model to obtain a first feature vector of the first vehicle.
In this embodiment, a neural network model of the word2vec algorithm may be used as a vector characterization model, the first vehicle information is used as an input, and the historical traffic information of the user is used as an output to perform joint training, so as to obtain a first feature vector of the first vehicle. In other embodiments, other neural network models may also be used to train to obtain the first feature vector, which is not specifically limited herein.
And S103, reading the stored second feature vector, sequencing the first vehicle based on the first feature vector and the second feature vector, and determining a recommended vehicle according to the sequencing result, wherein the second feature vector is generated by processing the historical transaction information of the second vehicle by using a vector characterization model and is used for characterizing the intention label of the user to the second vehicle.
Illustratively, historical transaction information of the second vehicle, namely data of a user purchasing the vehicle within a preset time, wherein the data includes vehicle type information and appearance information of the purchased vehicle, the historical transaction information is used as input, historical flow information of the user is used as output, and a neural network model which also adopts a word2vec algorithm is subjected to joint training to obtain a second feature vector of the second vehicle. It is understood that the neural network model used in this step only needs to be consistent with the neural network model used in step S102.
It can be understood that the first feature vector represents the comprehensive features of the first vehicle, the second feature vector represents the comprehensive features of the second vehicle, the first feature vector and the second feature vector are compared, so that which vehicles in the first vehicle are closer to the purchasing demand and preference of the user can be obtained, the vehicles ranked in the front are ranked after quantification, and the vehicles ranked in the front are recommended to the user.
Through the steps from S101 to S103, acquiring first vehicle information of a first vehicle to be recommended and historical flow information generated by browsing and clicking second vehicle information by a user, wherein the first vehicle information comprises vehicle type information and appearance information of the first vehicle; processing the first vehicle information and the historical flow information through a preset vector representation model to obtain a first feature vector of the first vehicle; and reading the stored second feature vector, sequencing the first vehicle based on the first feature vector and the second feature vector, and determining a recommended vehicle according to the sequencing result, wherein the second feature vector is generated by processing the historical transaction information of the second vehicle by a vector characterization model and is used for characterizing the intention label of the user for the second vehicle. The mode, through the vector representation model, turn into the vector statement of representation similarity with the characteristics of a plurality of modals of vehicle to compare with user's purchase motorcycle type, in order to recommend, and not only rely on the motorcycle type to recommend, compare more comprehensively, solved in the original marketing process can only recommend shortcoming and not enough to the motorcycle type that the user interacted, promoted the marketing and the exposure of long-tailed car commodity, promoted holistic marketing effect, improved the degree of accuracy that the vehicle recommended.
In some embodiments, obtaining the first vehicle information of the first vehicle to be recommended includes the following steps:
In the embodiment, at least one of the information of the vehicle type, the vehicle series and the brand is obtained as the vehicle type information, and under the condition that the vehicle is not sold, the vehicle series or the brand can be used for replacing the vehicle type characteristics and spliced with the picture vector, so that the cold start problem of the commodity is solved; in addition, the vehicle images are preprocessed to obtain appearance vectors representing the vehicle appearance similarity as appearance information, so that vehicle recommendation is more accurate.
In this embodiment, according to the model of the first vehicle, the car image on the platform is labeled, the banner image of the car display page is standardized by the car detection model, the foreground and the background are separated, the background image of the car is used as the car image, and the car image is input into the trained neural network model to obtain the picture vector. In this embodiment, a resnet-34 deep learning model may be used as the neural network model for training. Referring to fig. 2, fig. 2 is a structural diagram of a trained neural network model of a vehicle recommendation method according to an embodiment of the present application. Under the condition of traditional softmax loss, a center loss is added into the neural network model of the embodiment as a penalty term, so that the distance of the picture vectors of the vehicles with similar appearance characteristics is short; the vehicle picture vectors with different appearance characteristics are far away. Specifically, the loss function is as follows:
wherein L isSRepresented by the conventional softmax loss function, LCRepresenting the loss due to the distance from each sample point to the class center, cyiRepresents belonging to yiThe vector of class center is the mean value of the vector as the class center, and λ represents the weight of the influence of the centerlos loss on the whole loss function.
The resulting 512-dimensional vector serves as automobile appearance information. In other embodiments, other deep learning models can be used to process the vehicle image, and only the image vector representing the vehicle similarity can be obtained.
It can be understood that the appearance information of the first vehicle may be a picture of the vehicle, or may be a picture vector processed based on the picture of the vehicle.
In some embodiments, the step of processing the first vehicle information and the historical flow rate information through a preset vector characterization model to obtain a first feature vector of the first vehicle includes:
And 2, processing the first vehicle information and the user click behavior sequence through a preset vector representation model to obtain a first feature vector of the first vehicle.
Illustratively, based on historical traffic information of the user, that is, interactive data generated when the user browses and clicks a recommended vehicle within a preset time, for example, a vehicle type, a vehicle series, and a brand of the vehicle clicked and viewed by the user, a sequence in which the user clicks and views the vehicle, and the like, a plurality of user click behavior sequences may be obtained, that is, a sequence formed by using vehicle type information of the vehicle as a node and using a click sequence of the user as a connecting line is used to represent a preference condition of the user.
In some embodiments, generating a user click behavior sequence based on historical traffic information includes:
And 2, selecting any initial node in the directed unweighted graph, and generating a user click behavior sequence in a random walk mode.
According to the embodiment, the directed weightless graph is established, the user click behavior sequence is generated in a random walk mode, the obtained data reflected by the user click behavior sequence is more comprehensive, the data size is larger, the neural network training effect is better, and the accuracy of vehicle recommendation is higher.
Referring to fig. 3, fig. 3 is a schematic diagram of a directed graph without rights according to an embodiment of the present application. Illustratively, a directed unweighted graph is defined as G ═ V, epsilon, where V and epsilon represent sets of points and edges, respectively, and the goal of graph vectorization is to learn a graph vector characterization model Φ: v → RdAnd converting each node into a vector characterization of d dimension. The method for constructing the directed weightless graph comprises the following steps:
obtaining a user's sequence of clicks on the vehicle, wherein the sequence of clicks is truncated by day, results in a daily sequence of clicks for each user [ car1, car2, car 3.
Each vehicle is embedded in the graph as node V, and through the click sequence, click car2 after clicking car1, connect node 1 and node 2, and set the direction to node 1 through node 2.
And traversing the click sequences of all users, and constructing a directed weightless graph with the click sequence as an edge and the vehicle as a node.
Illustratively, after the construction of the directed unweighted graph is completed, any starting node is selected to randomly walk, and a plurality of user click behavior sequences are generated to serve as training sets. Specifically, the flow of random walk is as follows:
inputting parameters: graph G ═ V, epsilon, number of iterations γ, per path length t
And (3) outputting: path set Path
for i＝0in range(γ):
Randomly arranging the vertices, O
Initializing Path set Path initializes Single Path
for each vi ∈ O do:
for k＝0in range(t):
from viRandomly take one of the adjacent nodes of (1) and put it into a path, pathi)))
node＝random(neighbor(vi)
end for
Paths.append(Path)
end for
end for
Exemplarily, taking any node as a starting point and t as a path length, performing random walk on the directed unweighted graph, obtaining a user click behavior sequence after the number of the walk nodes reaches t, then reselecting any node for iteration, wherein the number of iterations is gamma, and finally obtaining gamma user click behavior sequences.
In some embodiments, the step of processing the first vehicle information and the user click behavior sequence through a preset vector characterization model to obtain a first feature vector of the first vehicle includes: taking the first vehicle information as input, taking the adjacent vehicle of the first vehicle in the user click behavior sequence as output, training the vector characterization model to obtain a hidden vector, and taking the hidden vector as the first feature vector of the first vehicle.
In the embodiment, various characteristics of the vehicle are integrated through the vector characterization model to obtain the characteristic vector of the comprehensive characterization vehicle characteristics, so that the characteristic expression is more accurate, and the accuracy of vehicle recommendation is higher.
Referring to fig. 4, fig. 4 is a structural diagram of a vector characterization model of a vehicle recommendation method according to an embodiment of the present application.
The vehicle type information of the vehicle comprises vehicle characteristics, brand characteristics and vehicle series characteristics, and the picture characteristics are appearance information.
In this embodiment, the input of the vector characterization model is first vehicle information of the first vehicle, that is, vehicle type information and appearance information, specifically, the vehicle type information may be a brand, a train or a specific vehicle type of the vehicle, and may also be input together with the appearance information; and outputting two adjacent vehicles of the first vehicle as input in the user click behavior sequence, namely an upstream vehicle and a downstream vehicle of the first vehicle in the user click behavior sequence.
The vehicle information, the appearance information and the user click behavior sequence of the first vehicle are used as a training set, and the vector characterization model is trained to obtain hidden layer vectors corresponding to each input, wherein the hidden layer vectors can be used for representing the relation between word pairs and words, namely the first feature vectors.
In other embodiments, other neural network models may be used, as long as feature vectors can be obtained based on the input feature synthesis.
In some embodiments, reading the stored second feature vector further comprises the following steps:
And 2, acquiring an average value of hidden layer vectors of the second vehicle based on the hidden layer vectors of the second vehicle to obtain a second feature vector.
In the embodiment, the vehicle in the user historical transaction information is subjected to the same vector conversion mode as the first vehicle to obtain the characteristic vector representing the comprehensive characteristics of the vehicle, and the characteristic vector is matched with the first vehicle more in the subsequent characteristic comparison, so that the recommendation accuracy of the vehicle is higher.
Exemplarily, the same vector conversion mode is adopted for a second vehicle, historical transaction information, namely vehicle type information and appearance information, of the second vehicle is obtained, then the brand, the vehicle series or the vehicle type and the appearance information are used as input, a user click behavior sequence is used as output, an vector characterization model is trained to obtain hidden layer vectors corresponding to each input second vehicle, and the hidden layer vectors of all the second vehicles are averaged to obtain second feature vectors.
It can be understood that, for each vehicle, the continuous embedded features can be extracted to characterize, for example, the model, the train, the brand and the appearance information of the vehicle are 512-dimensional characterization vectors, and after weighted splicing is performed through a neural network, hidden layer vectors of 1024 dimensions, 1536 dimensions or 2048 dimensions are obtained to comprehensively represent the features of the vehicle.
In some embodiments, ranking the first vehicle based on the first eigenvector and the second eigenvector, and determining the recommended vehicle according to the ranking result includes the following steps:
And 2, sequencing the first vehicles based on the cosine similarity.
And 3, selecting a preset number of first vehicles from high to low based on the sequence for recommendation.
According to the embodiment, the similarity between the first vehicle and the second vehicle is compared by calculating the cosine similarity between the first characteristic vector and the second characteristic vector, and the characteristic vectors are quantized and compared, so that the method is more intuitive, convenient for sequencing and recommending and higher in recommending accuracy.
Illustratively, when a manufacturer of an operation B terminal carries out marketing, a recommendation list is constructed according to historical purchases. According to the recommendation list, obtaining the hidden vector of each second vehicle, averaging to obtain a central vector, then calculating the hidden vector of each first vehicle, calculating the cosine similarity of the hidden vector and the central vector of the first vehicle, sequencing the first vehicles according to the similarity, selecting the top-N vehicles to be placed in the marketing list, and increasing the robustness and the long tail effect of the recommendation list. Specifically, how many vehicles are selected for recommendation can be set according to actual conditions.
In some embodiments, the input of the training set of the first neural network model may be one or more of a vehicle type or brand or vehicle series and appearance information, the one or more of the vehicle type or brand or vehicle series and the appearance information together being a multi-modal feature. When the vehicle type of the first vehicle is a new vehicle type, certain browsing clicks are not accumulated at the moment, the feature vector cannot be obtained based on the vehicle type features, at the moment, the hidden vector can be formed by replacing the vehicle type features with brand or vehicle series features, the hidden vector is spliced with the appearance information, the 1024-dimensional vector is obtained and used as the feature vector, and therefore the cold start problem of the commodity is well solved.
Referring to fig. 5, fig. 5 is a flowchart illustrating a feature vector obtaining method according to a preferred embodiment of the present application. As shown in fig. 5, the process includes:
in step S501, vehicle appearance information is extracted.
For example, vehicle appearance information may be extracted from the vehicle image, the appearance information characterizing vehicle appearance similarity.
Step S502, obtaining the historical flow information of the user.
Illustratively, the historical traffic information of the user is interactive data generated when the user browses and clicks the recommended vehicle within a preset time, for example, the type, the train system, and the brand of the vehicle clicked and checked by the user, the order in which the user clicks and checks the vehicle, and the like.
Step S503, the presence/absence right graph is constructed.
Illustratively, based on historical traffic information of the users, click sequences of all the users are traversed, and a directed weightless graph with edges as click sequences and nodes as vehicles is constructed.
Step S504, the random walk is carried out, and a user click behavior sequence is generated.
Illustratively, any starting node is selected in the directed unweighted graph, and a user click behavior sequence is generated in a random walk manner.
In step S505, vehicle type information and appearance information of the vehicle are input.
Exemplarily, the vehicle type information and the appearance information of the vehicle are input into the vector characterization model.
Step S506, training the vector characterization model.
Illustratively, vehicle type information and appearance information of the vehicles are used as input, and an upstream vehicle and a downstream vehicle of a first vehicle in the user click behavior sequence are used as output to train the vector characterization model.
In step S507, a feature vector is generated.
Illustratively, hidden layer vectors in the vector characterization model are obtained as feature vectors.
In the embodiment, various characteristics of the vehicle are integrated through the vector characterization model to obtain the characteristic vector of the comprehensive characterization vehicle characteristics, so that the characteristic expression is more accurate, and the accuracy of vehicle recommendation is higher.
It should be noted that the steps illustrated in the above-described flow diagrams or in the flow diagrams of the figures may be performed in a computer system, such as a set of computer-executable instructions, and that, although a logical order is illustrated in the flow diagrams, in some cases, the steps illustrated or described may be performed in an order different than here.
The embodiment further provides a vehicle recommendation device, which is used for implementing the above embodiments and preferred embodiments, and the description of the device is omitted. As used hereinafter, the terms "module," "unit," "subunit," and the like may implement a combination of software and/or hardware for a predetermined function. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.
Fig. 6 is a block diagram of a vehicle recommendation apparatus according to an embodiment of the present application, and as shown in fig. 6, the apparatus includes:
the obtaining module 10 is configured to obtain first vehicle information of a first vehicle to be recommended and historical flow information generated by browsing and clicking second vehicle information by a user, where the first vehicle information includes vehicle type information and appearance information of the first vehicle.
The first processing module 20 is coupled to the obtaining module 10, and configured to process the first vehicle information and the historical flow information through a preset vector characterization model to obtain a first feature vector of the first vehicle.
And the recommending module 30 is coupled with the first processing module 20 and is used for reading the stored second feature vector, sequencing the first vehicle based on the first feature vector and the second feature vector, and determining a recommended vehicle according to the sequencing result, wherein the second feature vector is generated by processing the historical transaction information of the second vehicle by using a vector characterization model and is used for characterizing the intention label of the user for the second vehicle.
In some embodiments, the obtaining module 10 is further configured to obtain at least one of a model, a series, and a brand of the first vehicle as the model information; acquiring a vehicle image of a first vehicle; and inputting the vehicle image into the trained neural network model to obtain appearance information of the first vehicle, wherein the appearance information comprises an appearance vector representing the appearance similarity of the vehicle. In some embodiments, the first processing module 20 is configured to generate a user click behavior sequence based on the historical traffic information; and processing the first vehicle information and the user click behavior sequence through a preset vector characterization model to obtain a first feature vector of the first vehicle.
In some embodiments, the first processing module 20 is configured to construct a directed weightless graph based on historical traffic information of a user, where the directed weightless graph includes vehicle information and a vehicle click sequence; and selecting any initial node in the directed unweighted graph, and generating a user click behavior sequence in a random walk mode.
In some embodiments, the first processing module 20 is configured to use the first vehicle information as an input, use an adjacent vehicle of the first vehicle in the user click behavior sequence as an output, train the vector characterization model to obtain a hidden layer vector, and use the hidden layer vector as the first feature vector of the first vehicle.
In some embodiments, the recommending module 30 is configured to obtain cosine similarity between the first feature vector and the second feature vector; sorting the first vehicles based on the cosine similarity; and selecting a preset number of first vehicles for recommendation from high to low based on the ranking.
In some embodiments, the vehicle recommendation device is further configured to train the vector characterization model based on the user click behavior sequence and historical transaction information of the second vehicle to obtain a hidden layer vector of the second vehicle; and obtaining the average value of the hidden layer vectors of the second vehicle based on the hidden layer vectors of the second vehicle to obtain a second feature vector.
The above modules may be functional modules or program modules, and may be implemented by software or hardware. For a module implemented by hardware, the modules may be located in the same processor; or the modules can be respectively positioned in different processors in any combination.
In addition, the vehicle recommendation method of the embodiment of the present application described in conjunction with fig. 1 may be implemented by a computer device. Fig. 7 is a schematic hardware configuration diagram of a computer device according to an embodiment of the present application.
The computer device may comprise a processor 71 and a memory 72 in which computer program instructions are stored.
Specifically, the processor 71 may include a Central Processing Unit (CPU), or A Specific Integrated Circuit (ASIC), or may be configured to implement one or more Integrated circuits of the embodiments of the present Application.
The memory 72 may be used to store or cache various data files that need to be processed and/or used for communication, as well as possible computer program instructions executed by the processor 71.
The processor 71 reads and executes the computer program instructions stored in the memory 72 to implement any one of the vehicle recommendation methods in the above embodiments.
In some of these embodiments, the computer device may also include a communication interface 73 and a bus 70. As shown in fig. 7, the processor 71, the memory 72, and the communication interface 73 are connected via the bus 70 to complete mutual communication.
The communication interface 73 is used for realizing communication among modules, devices, units and/or equipment in the embodiment of the present application. The communication interface 73 may also enable communication with other components such as: the data communication is carried out among external equipment, image/data acquisition equipment, a database, external storage, an image/data processing workstation and the like.
The bus 70 comprises hardware, software, or both that couple the components of the computer device to one another. Bus 70 includes, but is not limited to, at least one of the following: data Bus (Data Bus), Address Bus (Address Bus), Control Bus (Control Bus), Expansion Bus (Expansion Bus), and Local Bus (Local Bus). By way of example, and not limitation, Bus 70 may include an Accelerated Graphics Port (AGP) or other Graphics Bus, an Enhanced Industry Standard Architecture (EISA) Bus, a Front-Side Bus (FSB), a Hyper Transport (HT) Interconnect, an ISA (ISA) Bus, an InfiniBand (InfiniBand) Interconnect, a Low Pin Count (LPC) Bus, a memory Bus, a microchannel Architecture (MCA) Bus, a PCI (Peripheral Component Interconnect) Bus, a PCI-Express (PCI-X) Bus, a Serial Advanced Technology Attachment (SATA) Bus, a Video Electronics Bus (audio Electronics Association), abbreviated VLB) bus or other suitable bus or a combination of two or more of these. Bus 70 may include one or more buses, where appropriate. Although specific buses are described and shown in the embodiments of the application, any suitable buses or interconnects are contemplated by the application.
The computer device may execute the vehicle recommendation method in the embodiment of the present application based on the acquired computer program instruction, thereby implementing the vehicle recommendation method described in conjunction with fig. 1.
In addition, in combination with the vehicle recommendation method in the foregoing embodiments, the embodiments of the present application may be implemented by providing a computer-readable storage medium. The computer readable storage medium having stored thereon computer program instructions; the computer program instructions, when executed by a processor, implement any of the vehicle recommendation methods in the above embodiments.
According to the vehicle recommendation method, the vehicle recommendation device, the computer equipment and the storage medium, the first vehicle information of the first vehicle to be recommended and the historical flow information generated by browsing and clicking the second vehicle information by the user are obtained, wherein the first vehicle information comprises the vehicle type information and the appearance information of the first vehicle; processing the first vehicle information and the historical flow information through a preset vector representation model to obtain a first feature vector of the first vehicle; and reading the stored second feature vector, sequencing the first vehicle based on the first feature vector and the second feature vector, and determining a recommended vehicle according to the sequencing result, wherein the second feature vector is generated by processing the historical transaction information of the second vehicle by a vector characterization model and is used for characterizing the intention label of the user for the second vehicle. The mode, through the vector representation model, turn into the vector statement of representation similarity with the characteristics of a plurality of modals of vehicle to compare with user's purchase motorcycle type, in order to recommend, and not only rely on the motorcycle type to recommend, compare more comprehensively, solved in the original marketing process can only recommend shortcoming and not enough to the motorcycle type that the user interacted, promoted the marketing and the exposure of long-tailed car commodity, promoted holistic marketing effect, improved the degree of accuracy that the vehicle recommended. Meanwhile, the vehicle series or the brand can be used for replacing vehicle type characteristics and can be spliced with the picture vector, and the cold start problem of the commodity is well solved.
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A vehicle recommendation method, comprising:
acquiring first vehicle information of a first vehicle to be recommended and historical flow information generated by browsing and clicking second vehicle information by a user, wherein the first vehicle information comprises vehicle type information and appearance information of the first vehicle;
processing the first vehicle information and the historical flow information through a preset vector characterization model to obtain a first feature vector of the first vehicle;
and reading a stored second feature vector, sequencing the first vehicle based on the first feature vector and the second feature vector, and determining a recommended vehicle according to the sequencing result, wherein the second feature vector is generated by processing historical transaction information of a second vehicle by the vector characterization model and is used for characterizing an intention label of a user for the second vehicle.
2. The vehicle recommendation method of claim 1, wherein the obtaining first vehicle information of a first vehicle to be recommended comprises:
acquiring at least one of the type, the series and the brand of the first vehicle as the type information;
acquiring a vehicle image of the first vehicle;
and inputting the vehicle image into a trained neural network model to obtain appearance information of the first vehicle, wherein the appearance information comprises an appearance vector representing the appearance similarity of the vehicle.
3. The vehicle recommendation method according to claim 1, wherein the processing the first vehicle information and the historical flow information through a preset vector characterization model to obtain a first feature vector of the first vehicle comprises:
generating a user click behavior sequence based on the historical traffic information;
and processing the first vehicle information and the user click behavior sequence through a preset vector characterization model to obtain a first feature vector of the first vehicle.
4. The vehicle recommendation method of claim 3, wherein said generating a sequence of user click behaviors based on said historical traffic information comprises:
constructing a directed unweighted graph based on the historical traffic information, wherein the directed unweighted graph comprises vehicle information and a vehicle click sequence;
and selecting any initial node in the directed unweighted graph, and generating a user click behavior sequence in a random walk mode.
5. The vehicle recommendation method according to claim 3, wherein the processing the first vehicle information and the user click behavior sequence through a preset vector characterization model to obtain a first feature vector of the first vehicle comprises:
taking the first vehicle information as input, taking an adjacent vehicle of the first vehicle in the user click behavior sequence as output, training the vector characterization model to obtain a hidden layer vector, and taking the hidden layer vector as a first feature vector of the first vehicle.
6. The vehicle recommendation method of claim 3, wherein said reading the stored second feature vector further comprises, prior to:
training the vector characterization model based on the user click behavior sequence and historical transaction information of a second vehicle to obtain a hidden layer vector of the second vehicle;
and acquiring the average value of the hidden layer vector of the second vehicle based on the hidden layer vector of the second vehicle to obtain the second feature vector.
7. The vehicle recommendation method of claim 1, wherein the ranking the first vehicle based on the first eigenvector and the second eigenvector, the determining a recommended vehicle according to the ranking result comprising:
obtaining cosine similarity of the first feature vector and the second feature vector;
ranking the first vehicle based on the cosine similarity;
and selecting a preset number of first vehicles for recommendation from high to low based on the ranking.
8. A vehicle recommendation device, comprising:
the system comprises an acquisition module, a recommendation module and a recommendation module, wherein the acquisition module is used for acquiring first vehicle information of a first vehicle to be recommended and historical flow information generated by browsing and clicking second vehicle information by a user, and the first vehicle information comprises vehicle type information and appearance information of the first vehicle;
the first processing module is used for processing the first vehicle information and the historical flow information through a preset vector representation model to obtain a first feature vector of the first vehicle;
and the recommending module is used for reading a stored second feature vector, sequencing the first vehicle based on the first feature vector and the second feature vector, and determining a recommended vehicle according to the sequencing result, wherein the second feature vector is generated by processing the historical transaction information of the second vehicle by the vector characterization model and is used for characterizing the intention label of the user to the second vehicle.
9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the vehicle recommendation method as claimed in any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out a vehicle recommendation method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011573596.6A CN112561663A (en) | 2020-12-24 | 2020-12-24 | Vehicle recommendation method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011573596.6A CN112561663A (en) | 2020-12-24 | 2020-12-24 | Vehicle recommendation method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112561663A true CN112561663A (en) | 2021-03-26 |
Family
ID=75033539
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011573596.6A Pending CN112561663A (en) | 2020-12-24 | 2020-12-24 | Vehicle recommendation method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112561663A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113449002A (en) * | 2021-06-28 | 2021-09-28 | 平安银行股份有限公司 | Vehicle recommendation method and device, electronic equipment and storage medium |
| CN115062230A (en) * | 2022-08-16 | 2022-09-16 | 深圳市海清视讯科技有限公司 | Data processing method, apparatus, device, medium, and program product |
| CN117688137A (en) * | 2024-01-31 | 2024-03-12 | 成都航空职业技术学院 | Data analysis method and system applied to automobile marketing management system software |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110457589A (en) * | 2019-08-19 | 2019-11-15 | 上海新共赢信息科技有限公司 | A kind of vehicle recommended method, device, equipment and storage medium |
| CN112102031A (en) * | 2020-08-24 | 2020-12-18 | 深圳市元征科技股份有限公司 | Recommendation method, recommendation device and terminal equipment |
-
2020
- 2020-12-24 CN CN202011573596.6A patent/CN112561663A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110457589A (en) * | 2019-08-19 | 2019-11-15 | 上海新共赢信息科技有限公司 | A kind of vehicle recommended method, device, equipment and storage medium |
| CN112102031A (en) * | 2020-08-24 | 2020-12-18 | 深圳市元征科技股份有限公司 | Recommendation method, recommendation device and terminal equipment |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113449002A (en) * | 2021-06-28 | 2021-09-28 | 平安银行股份有限公司 | Vehicle recommendation method and device, electronic equipment and storage medium |
| CN115062230A (en) * | 2022-08-16 | 2022-09-16 | 深圳市海清视讯科技有限公司 | Data processing method, apparatus, device, medium, and program product |
| CN115062230B (en) * | 2022-08-16 | 2023-01-06 | 深圳市海清视讯科技有限公司 | Data processing method, apparatus, device, medium, and program product |
| CN117688137A (en) * | 2024-01-31 | 2024-03-12 | 成都航空职业技术学院 | Data analysis method and system applied to automobile marketing management system software |
| CN117688137B (en) * | 2024-01-31 | 2024-04-12 | 成都航空职业技术学院 | Data analysis method and system applied to automobile marketing management system software |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112561663A (en) | Vehicle recommendation method and device, computer equipment and storage medium | |
| US8631044B2 (en) | Machine optimization devices, methods, and systems | |
| US20180342004A1 (en) | Cumulative success-based recommendations for repeat users | |
| CN110929138B (en) | Recommendation information generation method, device, equipment and storage medium | |
| CN110415032A (en) | A kind of exposure conversion ratio predictor method, device, equipment and storage medium | |
| KR20080083638A (en) | Automatic detection of online commercial intention | |
| CN111242748B (en) | Method, apparatus, and storage medium for recommending items to a user | |
| CN110781407A (en) | User label generation method and device and computer readable storage medium | |
| CN112000819A (en) | Multimedia resource recommendation method and device, electronic equipment and storage medium | |
| CN109145200A (en) | Promote method, apparatus, equipment and the computer storage medium showed | |
| CN112149003B (en) | Commodity community recommendation method and device and computer equipment | |
| CN109961351A (en) | Information recommendation method, device, storage medium and computer equipment | |
| CN111612588A (en) | Commodity presentation method and device, computing equipment and computer-readable storage medium | |
| CN111651679A (en) | Recommendation method and device based on reinforcement learning | |
| CN111967924A (en) | Commodity recommendation method, commodity recommendation device, computer device, and medium | |
| CN115358807A (en) | Article recommendation method and device, storage medium and electronic equipment | |
| CN111767457B (en) | Recommendation method and device | |
| CN113744002A (en) | Method, device, equipment and computer readable medium for pushing information | |
| CN116955591A (en) | Recommendation language generation method, related device and medium for content recommendation | |
| CN116611891A (en) | Content information recommendation method, device, server and storage medium | |
| CN115017399B (en) | Automatic recommendation method and device for online booking vehicle types | |
| CN115238188A (en) | Object recommendation method and system and object recommendation model system | |
| CN111598599B (en) | User characterization method, device, electronic equipment and computer readable medium | |
| KR102608859B1 (en) | Method and system for operating online market management integrated platform | |
| KR20240028810A (en) | Apparatus and method for recommending suitable consumer information to intellectual property rights holders |
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 |