WO2019227560A1 - 信息推荐方法、装置、系统、设备及可读存储介质 - Google Patents

信息推荐方法、装置、系统、设备及可读存储介质 Download PDF

Info

Publication number
WO2019227560A1
WO2019227560A1 PCT/CN2018/093217 CN2018093217W WO2019227560A1 WO 2019227560 A1 WO2019227560 A1 WO 2019227560A1 CN 2018093217 W CN2018093217 W CN 2018093217W WO 2019227560 A1 WO2019227560 A1 WO 2019227560A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
recommendation information
information
candidate
recommendation
Prior art date
Application number
PCT/CN2018/093217
Other languages
English (en)
French (fr)
Inventor
占吉清
刘权
陈志刚
Original Assignee
科大讯飞股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 科大讯飞股份有限公司 filed Critical 科大讯飞股份有限公司
Publication of WO2019227560A1 publication Critical patent/WO2019227560A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/953Querying, e.g. by the use of web search engines
    • G06F16/9535Search customisation based on user profiles and personalisation

Definitions

  • chat process between the two through the instant messaging application is as follows:
  • chat process between the two is as follows:
  • the method adopted is to temporarily retrieve the corresponding information according to the demand when the need arises, such as searching through a browser. Obviously, this method consumes a large amount of retrieval time for users, and cannot recommend information to other users in a timely manner.
  • this application provides an information recommendation method, device, system, device, and readable storage medium for solving the existing information recommendation methods.
  • users face the need to recommend information to other users, they need to temporarily retrieve , Leading to problems that take a long time and cannot recommend information in a timely manner.
  • An information recommendation method includes:
  • the obtaining the intention of the first user includes:
  • determining the intention of the first user based on the input data and historical portrait data of the first user includes:
  • the intent recognition model is obtained by training input data of the user and its historical portrait data as training samples, and training the user's intention as sample labels.
  • it further comprises:
  • the method before sending the candidate recommendation information set to the second user client, the method further includes:
  • the second user is determined according to the contact information in the associated social application of the first user.
  • the determining the second user according to the communication record of the first user includes:
  • a user who meets a set condition is selected as the second user, and the set condition includes:
  • the intimacy with the first user reaches any one or more of a set intimacy condition, is currently online, and is currently in a state of communication with the first user.
  • the determining the second user according to the contact information in the associated social application of the first user includes:
  • the contact of the first user's associated social application is selected as a second user who meets a set condition, where the set condition includes:
  • the intimacy with the first user reaches any one or more of a set intimacy condition, is currently online, and is currently in a state of communication with the first user.
  • the sending the candidate recommendation information set to a second user client includes:
  • it further comprises:
  • a virtual reward asset is issued to the account of the second user.
  • An information recommendation method includes:
  • target recommendation information is obtained, where the target recommendation information is information that needs to be sent to the first user client.
  • it further comprises:
  • it further comprises:
  • obtaining the target recommendation information in response to the second user's operation on the candidate recommendation information set includes:
  • the candidate recommendation information set or the candidate recommendation information selected by the second user in the candidate recommendation information set is obtained, and the second user obtains based on the candidate recommendation information set.
  • the obtained recommendation information is determined as the target recommendation information.
  • the sending the target recommendation information to the first user client includes:
  • An information recommendation device includes:
  • An intent acquisition unit configured to acquire the intent of the first user
  • a candidate set obtaining unit configured to obtain a candidate recommendation information set that matches the intention of the first user, where the candidate recommendation information set includes at least one candidate recommendation information;
  • a candidate set sending unit is configured to send the candidate recommendation information set to a second user client for the second user to determine target recommendation information that needs to be recommended to the first user based on the candidate recommendation information set.
  • An information recommendation device includes:
  • a candidate set receiving unit configured to receive a candidate recommendation information set sent by the server, the candidate recommendation information set matching the intention of the first user, and including at least one candidate recommendation information;
  • the target recommendation information determining unit is configured to obtain target recommendation information in response to a second user's operation on the candidate recommendation information set, where the target recommendation information is information that needs to be sent to the first user client.
  • it further comprises:
  • the target recommendation information sending unit is configured to send the target recommendation information to the first user client.
  • An information recommendation system includes: a first user client, a second user client, and a server, wherein the server and the second client respectively implement related steps in the foregoing information recommendation method.
  • An information recommendation device including a memory and a processor
  • the memory is used to store a program
  • the processor is configured to execute the program to implement each step of the foregoing information recommendation method.
  • a readable storage medium stores a computer program thereon.
  • the computer program is executed by a processor, each step of the foregoing information recommendation method is implemented.
  • the server obtains the intention of the first user, and obtains a set of candidate recommendation information that matches the intention of the first user.
  • the set includes at least one candidate recommendation information.
  • the candidate recommendation information set is further sent to the second user client for the second user to determine the target recommendation information to be recommended to the first user based on the candidate recommendation information set.
  • the server can obtain the intention of the first user in time, and obtain a set of candidate recommendation information matching the intention, and send it to the client of the second user, so when the second user needs to make information recommendation to the first user, it can
  • the target recommendation information is determined directly based on the set of candidate recommendation information matched with the first user sent by the server, which reduces information retrieval time, takes less time, and makes information recommendation more timely.
  • the server of the solution of this application does not directly send the obtained candidate recommendation information set that matches the intention of the first user to the first user, but sends it to the second user for the second user to recommend information to the first user.
  • This method of information recommendation by the user to the user is easier for the user to accept the recommended information and improves the acceptance of the information recommendation compared to the conventional scheme of recommending information to the user based on the user's intention.
  • FIG. 1 is a system architecture diagram for implementing information recommendation according to an embodiment of the present application
  • FIG. 2 is an optional signaling flow of the information recommendation method provided by the embodiment of this application.
  • FIG. 3 is another optional signaling flow of the information recommendation method provided by the embodiment of this application.
  • FIGS. 4-9 are diagrams illustrating examples of several application scenarios according to examples of the embodiments of the present application.
  • FIG. 10 is a schematic structural diagram of an information recommendation device disclosed in an embodiment of the present application.
  • FIG. 11 is a schematic structural diagram of another information recommendation device disclosed in an embodiment of the present application.
  • FIG. 12 is a block diagram of a hardware structure of an information recommendation device disclosed in an embodiment of the present application.
  • FIG. 1 is an optional system architecture for implementing information recommendation according to an embodiment of the present application.
  • the system architecture may include a first user client 10, a second user client 20, and a server 30.
  • the first user client 10 corresponds to the first user
  • the second user client 20 corresponds to the second user.
  • the first user and the second user are two different users.
  • the server 30 may be a service device provided on a network side.
  • the server 30 performs data interaction with the first user client 10 and the second user client 20 through a network.
  • the server 30 may be implemented by a single server, or It is a server cluster implementation composed of multiple servers.
  • the server 30 may be a server provided by a recommendation information provider. In the embodiment of the present application, the server 30 can obtain the intention of the first user, obtain a matching candidate recommendation information set based on the intention, and send the candidate recommendation information set to Second user client.
  • the first user client 10 and the second user client 20 may be terminal devices such as a television, a mobile phone, and a computer.
  • the first user client 10 and the second user client 20 can support the display of the recommended information for the user to watch and operate the recommended information. Further, through the first user client 10 and the second user client 20, the user can browse web pages, watch videos, chat, and other operations.
  • the second user client 20 can receive the candidate recommendation information set sent by the server 30 that matches the intention of the first user.
  • the second user client 20 supports the second user's operation on the candidate recommendation information set to obtain target recommendation information recommended to the first user.
  • the second user may choose to send the target recommendation information to the first user. Since the second user can determine the target recommendation information recommended to the first user based on the candidate recommendation information set, the search time consuming of the second user is reduced, and the second user can make information recommendation to the first user in a more timely manner.
  • the server 30 does not send the retrieved candidate recommendation information set directly to the first user client 10, but sends it to the second user client 20 for the second user client 30 to recommend to the first user, Compared with the traditional method of directly sending the recommendation information to the recommendation object, the present application makes it easier for the recommendation object to accept the recommended information by using other users to send the recommendation information to the recommendation object, thereby improving the acceptance of the information recommendation.
  • FIG. 2 shows an optional signaling flow of the information recommendation method provided by the embodiment of the present application.
  • the flow may include:
  • Step S10 The server 30 obtains the intention of the first user.
  • the server 30 may obtain the intention of the first user through multiple channels. For example, to determine the intention of the first user by analyzing the operation of the first user on the first user client 10, the operations may include: inputting data, browsing content, and the like. In addition, the server 30 may also analyze the intention of the first user in other ways, such as obtaining the location information and somatosensory information of the first user through a wearable device equipped by the first user, and analyzing the intention of the first user.
  • the first user For example, if it is acquired that the first user is browsing a gourmet food store, it may be determined that the first user has an intention to purchase gourmet food. For another example, when it is acquired that the first user discusses travel related content with a friend, it can be determined that the first user has an intention to travel. For another example, by analyzing the recent location trajectory of the first user and finding that the first user frequently visits a car sales store, it can be determined that the user has an intention to purchase a car, and so on.
  • the first user's intention may include various types, such as consumption intention, interest, and hobby intention.
  • consumption intention as an example, it can be further divided into: whether there is willingness to consume, and information about products to be purchased.
  • Step S11 The server 30 obtains a candidate recommendation information set that matches the intention of the first user.
  • the candidate recommendation information set includes at least one candidate recommendation information.
  • the server 30 may retrieve matching candidate recommendation information in a related database according to the intention of the first user. Taking the server 30 as the recommendation information provider as an example, the server 30 may retrieve the candidate recommendation information set that matches the intention of the first user in the information to be recommended.
  • step S12 the server 30 sends the candidate recommendation information set to the second user client 20.
  • the server 30 does not directly send the candidate recommendation information set to the first user client 10 corresponding to the first user, but instead selects the candidate recommendation information set. Sending to the second user client for the second user to determine the target recommendation information to be recommended to the first user based on the candidate recommendation information set.
  • the second user may be any one or more users different from the first user.
  • the second user may also be a user who has an association relationship with the first user.
  • the second user is a user who has a positive relationship with the first user, and the positive relationship can improve information recommendation. Relations among users of acceptance.
  • Step S13 The second user client 20 obtains target recommendation information in response to the second user's operation on the candidate recommendation information set.
  • the second user client 20 may display the candidate recommendation information set to the second user through the display component.
  • the second user may operate the candidate recommendation information set through the second user client 20 to obtain target recommendation information, where the target recommendation information is information to be recommended to the first user.
  • the second user may store the target recommendation information locally on the second user client 20 and wait for information recommendation to the first user before sending the target recommendation information to the first user client 10.
  • the second user client 20 may also send the target recommendation information to the first user client 10 directly.
  • the target recommendation information when there is a need for information recommendation by the second user, the target recommendation information can be directly determined based on the candidate recommendation information set matching the first user sent by the server 30, thereby reducing the information retrieval time. It takes less time to make information recommendations more timely.
  • the server 30 does not directly send the set of candidate recommendation information retrieved based on the intention of the first user to the first user client 10, but sends it to the second user client 20 for the second user client 30 to send to the first user recommend.
  • the present application makes it easier for the recommendation object to accept the recommended information by using other users to send the recommendation information to the recommendation object, thereby improving the acceptance of the information recommendation.
  • FIG. 3 shows another optional signaling flow of the information recommendation method provided by the embodiment of the present application.
  • the flow may include:
  • Step S20 The server 30 obtains the intention of the first user.
  • Step S21 The server 30 obtains a candidate recommendation information set that matches the intention of the first user.
  • step S22 the server 30 sends the candidate recommendation information set to the second user client 20.
  • Step S23 The second user client 20 obtains target recommendation information in response to the second user's operation on the candidate recommendation information set.
  • Step S24 The second user client 20 sends the target recommendation information to the first user client 10.
  • the second user client 20 in this embodiment further sends the target recommendation information to the first user client 10 after obtaining the target recommendation information.
  • the present application makes it easier for the recommendation object to accept the recommended information by using other users to send the recommendation information to the recommendation object, thereby improving the acceptance of the information recommendation.
  • the first user and the second user respectively chat with each other through their respective clients. Assume that the first user is nicknamed Liu xx and the second user is nicknamed Zhang xx.
  • the first user sends a message to the second user: “I want to change my phone recently. Is there any recommendation?”
  • the message is also uploaded to the server 30 as chat content.
  • the server 30 determines that the first user has an intention to purchase a mobile phone according to the chat content. Therefore, retrieve the latest mobile phone recommendation information and get the search result: the mobile phone recommendation information link. And send the link of the mobile phone recommendation information to the second user chatting with the first user.
  • the lower right corner of the display interface of the second user client 20 displays the information delivered by the server 30 through a floating window, including: mobile phone links and prompts: "There is a very good mobile phone here, please recommend it to your friend Liu xx ⁇ ".
  • the second user receives the mobile phone link, he can save the link of searching by himself and directly copy the mobile phone link and send it to the first user.
  • the specific effect is shown in Figure 5.
  • the second user when the second user receives the information recommendation request from the first user, the second user does not need to search it by himself, and can directly send the mobile phone recommendation information link issued by the server to the second user client, saving.
  • the retrieval time of the second user makes the information recommendation more timely, and because the mobile phone recommendation information is recommended by the second user to the first user, the mobile phone recommendation information is easier to be ranked by the first user than the server recommends directly to the first user. Acceptance by one user improves the acceptance of information recommendations.
  • the first user searches for Chengdu attractions information through the first user client 10.
  • the search information is uploaded to the server 30 as search content.
  • the server 30 determines that the first user has an intention to travel to Chengdu according to the search content. Therefore, the Chengdu attraction recommendation information is retrieved, and the retrieved attraction recommendation information link is issued to the second client 20.
  • the second user is playing the game through the second user client 20, and at this time receives the pop-up information issued by the server 30, as shown in the lower right corner of FIG.
  • the content of the pop-up window includes: a link to the recommendation information of the attraction and the prompt: "Don't play the game, go here with your friend Liu xx ⁇ ".
  • the second user After receiving the information issued by the server 30, the second user finds that the attraction is very good, and can copy the link of the attraction recommendation information and send it to the first user client 10. Specifically, information can be sent through an instant messaging application. The sent message is displayed in a pop-up in the lower right corner of the interface of the first user client 10.
  • the second user can recommend the attraction recommendation information link to the first user when receiving the attraction recommendation information link issued by the server.
  • the second user does not need to perform additional attraction search.
  • the attraction recommendation information is recommended by the second user to the first user.
  • the attraction recommendation information is more easily accepted by the first user, which improves the acceptance of the information recommendation.
  • the embodiment of the present application first introduces the information recommendation scheme from the perspective of the server 30.
  • the server 30 has described its intent to obtain the first user.
  • the process of obtaining the intention of the first user by the server 30 is described in detail in this embodiment.
  • the server 30 may obtain the input data of the first user at the current time or within a set time period before the current time, and further determine the intention of the first user based on the input data and historical portrait data of the first user.
  • input data such as the input chat content of the first user during a chat with other users, or the search content entered by the first user in a browser or an application store.
  • the historical portrait data of the user may include basic information of the user, interests and the like. By analyzing the obtained input data and historical portrait data of the first user, the intention of the first user can be determined.
  • the embodiments of the present application may construct an intent recognition model in advance, and the intent recognition model may be various forms of neural network models capable of classification.
  • the intent recognition model is trained based on the collected training samples and sample labels.
  • the input data of the first user and its historical portrait data can be input into the intent recognition model to obtain the intent of the first user output by the intent recognition model.
  • the number of the intention recognition models may be multiple.
  • some intent recognition models are used to determine whether the first user has a certain type of intent, such as consumption intent, use intent, and the like.
  • Other intent recognition models can be used to further determine detailed intent information under a certain type of intention when it is determined that the first user has a certain type of intention. Taking the consumption intention as an example, detailed information such as the product model, name, and size of the item that the first user wants to purchase can be determined.
  • the server 30 obtains a candidate recommendation information set that matches the intention of the first user.
  • This process can include:
  • the server 30 retrieves a candidate recommendation information set that matches the intention of the first user from the information source to be recommended based on the intention of the first user.
  • the popularity information of each piece of to-be-recommended information in the to-be-recommended information source, user attention, and other information can be considered simultaneously, that is, according to the popularity information of each piece of to-be-recommended information in the to-be-recommended information source, user attention, etc.
  • the server 30 may choose to send it to the second user client 20 in the form of links, text content, pictures, videos, or pop-up windows. As shown in Figure 4, the example is the form of pop-up window.
  • the server 30 may agree with the second user in advance, and the candidate recommendation information sets sent by the server 30 to the second user client 20 are all matched with the first user. In this way, when the second user client 20 receives the candidate recommendation information set sent by the server 30, it can directly determine that the candidate recommendation information set matches the first user, and the target recommendation information determined based on the candidate recommendation information set is also Need to be recommended to the first user.
  • the server 30 sends the candidate recommendation information set to the second user client 20, it can also send recommendation object guidance information, which includes the identification of the first user for the second user to The identification of the first user included in the recommendation object guide information determines the first user to be recommended.
  • the information sent by the server 30 to the second client 20 includes the guidance information of the recommendation object: “A mobile phone here is very good, please recommend it to your friend Liu xx ⁇ ”.
  • the recommendation object guidance information includes the identifier of the first user: Liu xx.
  • FIG. 4 only illustrates an optional composition structure of the recommended object guide information.
  • other forms of recommended object guide information can be set, as long as the second user can clearly recommend the recommended object as the first Users can.
  • a process for determining the second user may be added.
  • the server 30 may determine all other users other than the first user as the second user, or randomly select one or more from the other users who are not the first user as the second user.
  • the server 30 may determine the second user according to the communication record of the first user.
  • the server 30 may determine the second user according to the contact information in the associated social application of the first user.
  • the communication record of the first user may include a chat communication record performed by the first user through an instant messaging application. It may also include the record of communication between the first user and the customer service personnel, such as the communication record with the online customer service personnel.
  • the user who has a communication record with the first user may be determined as the second user.
  • some users may be further selected as the second user.
  • users who meet the set conditions can be selected as the second users by setting the conditions.
  • Setting conditions can include:
  • the intimacy with the first user reaches any one or more of a set intimacy condition, is currently online, and is currently in a state of communication with the first user.
  • intimacy characterizes the possibility of generating potential information recommendations among users. The greater the intimacy between two users, the more likely it is that the two users will recommend information to each other.
  • the intimacy of the first user and the second user can be set equal to the intimacy of the second user and the first user, or the intimacy of the first user and the second user can be determined by setting the intimacy determination method, and the second The intimacy of the user and the first user does not necessarily have an equal relationship.
  • the embodiment of the present application discloses a specific implementation manner for determining the intimacy of the first user and the second user.
  • this embodiment can analyze several factors that affect intimacy, and these influencing factors can include any one or more of the following:
  • the number of separate interactions between the second user and the first user, the existence status (including presence and absence) of the second user in the first user friend list, the current time difference between the second user and the first user, the second user, and the second user The number of groups coexisting with the first user, the number of interactions between the second user and the first user in the group, and the like.
  • this embodiment may determine the intimacy of the second user and the first user according to the value of any one or more influencing factors described above.
  • the above-mentioned multiple influencing factors may have the same or different influence weights for determining intimacy.
  • each influencing factor may be assigned a corresponding weight, and the influence of multiple influencing factors on intimacy is comprehensively considered in a linear weighting manner.
  • a model prediction method may also be used to determine the intimacy of the second user and the first user. That is, in this embodiment, an intimacy determination model can be trained in advance, and the values of the above-mentioned influencing factors between two training users are determined in advance during training, and the values of these influencing factors are used as training samples, while the intimacy between the labeled training users Degree values are used as sample labels to train the intimacy determination model.
  • the values of the influencing factors of the second user and the first user can be input into the intimacy determination model to obtain the intimacy values of the second user and the first user output by the model.
  • the preset intimacy conditions in this embodiment may be: selecting a user whose intimacy value exceeds the intimacy threshold as the second user, or selecting the first N users with the highest intimacy value as the second user, or , Select the top M% users in the ranking from high to low as the second user.
  • the current online status indicates that it is online in a designated application, such as being online in a designated instant messaging application.
  • the embodiment of the present application introduces a process in which the server 30 determines the second user according to the contact information in the associated social application of the first user.
  • the contacts of the first user's associated social application may be determined as the second user.
  • the number of contacts associated with the social application of the first user may be excessive, and in this embodiment, some users may be further filtered as the second user.
  • users who meet the set conditions can be selected as the second users by setting the conditions.
  • Setting conditions can include:
  • the intimacy with the first user reaches any one or more of a set intimacy condition, is currently online, and is currently in a state of communication with the first user.
  • a reward mechanism is provided for the information recommender method. That is, the server 30 may issue a virtual reward asset to the account of the second user after detecting that the second user client 20 sends the target recommendation information to the first user client 10.
  • the server 30 can monitor whether the second user client 20 has sent the target recommendation information, or can monitor whether the first user client 10 has received or opened the target recommendation information, and determine the second user client based on this Whether the terminal 20 wants the first user client 10 to send target recommendation information.
  • This embodiment can further motivate the second user to recommend information to the first user by setting a reward mechanism, thereby increasing the number of information recommendations.
  • the embodiment of the present application further introduces the information recommendation method from the perspective of the second user client 20.
  • the second user client 20 receives the candidate recommendation information set sent by the server 30, and the candidate recommendation information set matches the intention of the first user. Further, the second user client 20 obtains target recommendation information in response to the second user's operation on the candidate recommendation information set, and the target recommendation information is used as information that needs to be sent to the first user client.
  • the second user may store the target recommendation information locally on the second user client 20 and wait for information recommendation to the first user before sending the target recommendation information to the first user. Client 10.
  • the second user client 20 may also directly send the target recommendation information to the first user client 10.
  • the target recommendation information can be determined directly based on the set of candidate recommendation information matching the first user sent by the server 30, without the need for temporary information retrieval. The time is shorter, and the information recommendation is more timely.
  • the target user can be recommended by the second user to the first user instead of being recommended by the server 30 to the first user directly. This makes it easier for the first user to accept the recommended information and improves the acceptance of the information recommendation.
  • the second user client 20 may send the target recommendation information to the first user client 10 in the form of a link, text content, picture, video, or popup window.
  • the second user client 20 may also receive the recommendation target guidance information sent by the server 30, where the recommendation target guidance information includes the identifier of the first user.
  • the second user client 20 may display the recommendation object guidance information for the second user to determine the first user to be recommended according to the identifier of the first user included in the recommendation object guidance information.
  • the recommended object guidance information includes the identifier “Liu xx” of the first user.
  • This embodiment illustrates several ways of obtaining target recommendation information in response to a user operation, as follows:
  • the candidate recommendation information set or the candidate recommendation information selected by the second user in the candidate recommendation information set is selected, and the second user based on the candidate recommendation
  • the recommendation information obtained by the information collection is determined as the target recommendation information.
  • the above three ways exemplify three ways to obtain target recommendation information based on the candidate recommendation information set.
  • the second user does not need to additionally retrieve the information recommended to the first user, and the target recommendation information can be obtained by performing a simple operation based on the candidate recommendation information set issued by the server 30.
  • the first user and the second user respectively chat with each other through their respective clients. Assume that the first user is nicknamed Liu xx and the second user is nicknamed Zhang xx.
  • the first user sends a message to the second user: “I want to change my phone recently. Is there any recommendation?”
  • the message is also uploaded to the server 30 as chat content.
  • the server 30 determines that the first user has an intention to purchase a mobile phone according to the chat content. Therefore, the latest mobile phone recommendation information is retrieved, and multiple search results are obtained to form a mobile phone recommendation information link set, which contains three mobile phone recommendation information links, which are: https://shouji.com/search1, https: // shouji. com / search2, https://shouji.com/search3.
  • the server 30 further sends the mobile phone recommendation information link set to the second user client 20.
  • the lower right corner of the display interface of the second user client 20 displays information delivered by the server 30 in a floating window form, including: three mobile phone recommendation information links and prompts: "Several mobile phones are very good here, recommend them to your friends. Liu xx ⁇ ". The specific effect is shown in Figure 8.
  • the second user When the second user receives the message, they can analyze the three links one by one, and based on their own knowledge of the first user, determine that the first two links are more suitable for the first user. Therefore, the first two links can be copied and sent to the first user client 10. The specific effect is shown in Figure 9.
  • FIG. 8 to FIG. 9 corresponds to the above solution 2), that is, the user selects part of the candidate recommendation information from the candidate recommendation information set as target recommendation information and sends it to the first user client.
  • the other two schemes are not illustrated by the drawings.
  • the following describes the information recommendation device provided in the embodiment of the present application.
  • the information recommendation device described below and the information recommendation device described above can be referred to each other.
  • the information recommendation device may include:
  • An intent obtaining unit 100 configured to obtain the intent of the first user
  • the candidate set obtaining unit 110 is configured to obtain a candidate recommendation information set that matches the intention of the first user, where the candidate recommendation information set includes at least one candidate recommendation information;
  • the candidate set sending unit 120 is configured to send the candidate recommendation information set to a second user client for the second user to determine target recommendation information to be recommended to the first user based on the candidate recommendation information set.
  • the intent acquisition unit may include:
  • An input data acquisition unit configured to acquire input data of the first user at a current time or within a set time period before the current time
  • a data application unit is configured to determine the intention of the first user according to the input data and historical portrait data of the first user.
  • the data application unit may include:
  • An intent recognition model prediction unit is configured to input the input data and the historical portrait data into a preset intent recognition model to obtain the intent of the first user output by the intent recognition model; the intent recognition model is based on The input data of the training user and its historical portrait data are used as training samples, and the intention of the training user is used as training sample labels.
  • the information recommendation device of this application may further include:
  • the recommendation object guidance information sending unit is configured to send the recommendation object guidance information to the second user client, so that the second user determines the first recommended recommendation based on the identifier of the first user included in the recommendation object guidance information. A user.
  • the information recommendation device of this application may further include:
  • An communication record using unit configured to determine the second user according to the communication record of the first user
  • the contact using unit is configured to determine the second user according to the contact information in the associated social application of the first user.
  • the communication record using unit may include:
  • the second communication record uses a sub-unit, and is configured to screen, as a second user, users who meet a set condition among users who have an exchange record with the first user, and the set condition includes:
  • the intimacy with the first user reaches any one or more of a set intimacy condition, is currently online, and is currently in a state of communication with the first user.
  • the contact using unit may include:
  • a first contact using a sub-unit configured to determine a contact of an associated social application of the first user as the second user
  • the second contact using subunit is configured to filter, as the second user, a user who meets a set condition among the contacts of the first user's associated social application, where the set condition includes:
  • the intimacy with the first user reaches any one or more of a set intimacy condition, is currently online, and is currently in a state of communication with the first user.
  • the candidate set sending unit may include:
  • the first candidate set sending subunit is configured to send the candidate recommendation information set to a second user client in a form of link, text content, picture, video, or popup.
  • the information recommendation device may further include:
  • the reward unit is configured to issue a virtual reward asset to the account of the second user after detecting that the second user client sends the target recommendation information to the first user client.
  • the information recommendation device may include:
  • the candidate set receiving unit 200 is configured to receive a candidate recommendation information set sent by the server, where the candidate recommendation information set matches the intention of the first user and includes at least one candidate recommendation information;
  • the target recommendation information determining unit 210 is configured to obtain target recommendation information in response to a second user's operation on the candidate recommendation information set, where the target recommendation information is information that needs to be sent to the first user client.
  • the information recommendation device may further include:
  • the target recommendation information sending unit is configured to send the target recommendation information to the first user client.
  • the information recommendation device may further include:
  • the recommendation object guidance information receiving unit is configured to receive the recommendation object guidance information sent by the server for the second user to determine the first recommended recommendation based on the identifier of the first user included in the recommendation object guidance information. user.
  • the target recommendation information determination unit may include:
  • a first response subunit configured to respond to a second user's operation on the candidate recommendation information set, and determine the candidate recommendation information set as target recommendation information
  • a second response subunit configured to respond to a second user's operation on the candidate recommendation information set, and determine the candidate recommendation information selected by the second user in the candidate recommendation information set as the target recommendation information;
  • a third response subunit configured to respond to a second user's operation on the candidate recommendation information set, and include the candidate recommendation information set or the candidate recommendation information selected by the second user in the candidate recommendation information set, and the second user
  • the recommendation information obtained based on the candidate recommendation information set is determined as target recommendation information.
  • the target recommendation information sending unit may include:
  • the first target recommendation information sending subunit is configured to send the target recommendation information to the first user client in the form of a link, text content, picture, video, or popup.
  • the information recommendation device provided in the embodiment of the present application can be applied to an information recommendation device.
  • the information recommendation device may be the server 30 or the second user client 20.
  • FIG. 12 shows a hardware structure block diagram of the information recommendation device.
  • the hardware structure of the information recommendation device may include: at least one processor 1, at least one communication interface 2, at least one memory 3, and at least one communication bus 4.
  • the number of the processor 1, the communication interface 2, the memory 3, and the communication bus 4 is at least one, and the processor 1, the communication interface 2, and the memory 3 complete communication with each other through the communication bus 4.
  • the processor 1 may be a central processing unit CPU, or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention
  • the memory 3 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), for example, at least one disk memory;
  • the memory stores a program
  • the processor may call the program stored in the memory, and the program is configured to implement each processing procedure of the foregoing server 30 in the information recommendation scheme, or implement the foregoing second user client 20 in information recommendation Each process in the scheme.
  • An embodiment of the present application further provides a storage medium that may store a program suitable for execution by a processor, the program being configured to implement each processing flow of the foregoing server 30 in the information recommendation scheme, or implement the foregoing Each processing flow of the second user client 20 in the information recommendation scheme.
  • the embodiment of the present application also discloses an information recommendation system, and the information recommendation information includes the first user client 10, the second user client 20, and the server 30, and the specific implementation logic of the above three units can refer to the aforementioned new recommendation.
  • the relevant introduction of the method part will not be repeated here.

Abstract

本申请公开了一种信息推荐方法、装置、系统、设备及可读存储介质,其中服务端获取第一用户的意图,获取与第一用户的意图匹配的候选推荐信息集合,集合包含至少一条候选推荐信息,进一步将候选推荐信息集合发送给第二用户客户端,以供第二用户基于候选推荐信息集合确定所需向第一用户推荐的目标推荐信息。由于服务端能够及时获取第一用户的意图,并获取到与该意图匹配的候选推荐信息集合发送给第二用户客户端,这样当第二用户存在向第一用户进行信息推荐的需求时,可以直接基于服务端发送的与第一用户匹配的候选推荐信息集合,确定目标推荐信息,减少了信息检索耗时,占用时间更短,进而更加及时的进行信息推荐。

Description

信息推荐方法、装置、系统、设备及可读存储介质 技术领域
本申请要求于2018年05月29日提交中国专利局、申请号为201810531062.3、发明名称为“信息推荐方法、装置、系统、设备及可读存储介质”的国内申请的优先权,其全部内容通过引用结合在本申请中。
背景技术
随着信息技术的快速发展,越来越多的信息通过互联网传播。如广告营销、多媒体推荐等等。
对于用户而言,在信息时代都会存在主动或被动的向其他用户推荐信息的需求。如用户A与用户B是好友关系,二者通过即时通讯应用聊天过程如下:
用户A“合肥天气怎么样,我想去旅游”;
用户B:“最近一周天气不错,适合出游”;
用户A:“合肥有什么好玩的地方,给我推荐下”。
此时对于用户B而言,其存在向用户A推荐合肥旅游景点的需要。
再比如,仍以用户A和用户B为例,二者聊天过程如下:
用户A“新搬的家距离公司好远啊”;
用户B:“买个车吧,这样就方便多了”;
用户A:“一直想着买呢,不知道什么品牌的好啊”。
此时对于用户B而言,其可能主动想要向用户A推荐几款品牌的汽车。
现有技术中当用户在面临存在主动或被动的向其他用户推荐信息的需求时,采取的方式是当出现该需求时,根据需求临时去检索相应的信息,如通过浏览器搜索等。显然,这种方式会消耗用户大量的检索时间,且无法及时向其他用户进行信息推荐。
发明内容
有鉴于此,本申请提供了一种信息推荐方法、装置、系统、设备及可读存储介质,用于解决现有信息推荐方式,当用户面临向其他用户推荐信息的需求时,需要临时进行检索,导致耗时长、无法及时进行信息推荐的 问题。
为了实现上述目的,现提出的方案如下:
一种信息推荐方法,包括:
获取第一用户的意图;
获取与所述第一用户的意图匹配的候选推荐信息集合,所述候选推荐信息集合包含至少一条候选推荐信息;
将所述候选推荐信息集合发送给第二用户客户端,以供第二用户基于所述候选推荐信息集合确定所需向所述第一用户推荐的目标推荐信息。
优选地,所述获取第一用户的意图,包括:
获取第一用户当前时刻或当前时刻之前设定时间段内的输入数据;
根据所述输入数据及所述第一用户的历史画像数据,确定所述第一用户的意图。
优选地,所述根据所述输入数据及所述第一用户的历史画像数据,确定所述第一用户的意图,包括:
将所述输入数据及所述历史画像数据输入预置的意图识别模型,得到意图识别模型输出的所述第一用户的意图;
所述意图识别模型为,预先以训练用户的输入数据及其历史画像数据作为训练样本,以训练用户的意图作为样本标签训练得到。
优选地,还包括:
将推荐对象引导信息发送给第二用户客户端,以供第二用户基于所述推荐对象引导信息包含的所述第一用户的标识,确定所需推荐的第一用户。
优选地,在将所述候选推荐信息集合发送给第二用户客户端之前,该方法还包括:
根据所述第一用户的交流记录,确定所述第二用户;
或,
根据所述第一用户的关联社交应用中的联系人信息,确定所述第二用户。
优选地,所述根据所述第一用户的交流记录,确定所述第二用户,包括:
将与所述第一用户存在交流记录的用户确定为第二用户;
或,
在与所述第一用户存在交流记录的用户中,筛选满足设定条件的用户作为第二用户,所述设定条件包括:
与所述第一用户的亲密度达到设定亲密度条件、当前处于在线状态、当前处于与所述第一用户交流状态中的任意一个或多个。
优选地,所述根据所述第一用户的关联社交应用中的联系人信息,确定所述第二用户,包括:
将所述第一用户的关联社交应用的联系人均确定为所述第二用户;
或,
将所述第一用户的关联社交应用的联系人中,筛选满足设定条件的用户作为第二用户,所述设定条件包括:
与所述第一用户的亲密度达到设定亲密度条件、当前处于在线状态、当前处于与所述第一用户交流状态中的任意一个或多个。
优选地,所述将所述候选推荐信息集合发送给第二用户客户端,包括:
将所述候选推荐信息集合通过链接、文本内容、图片、视频或弹窗的形式发送给第二用户客户端。
优选地,还包括:
在检测到所述第二用户客户端向第一用户客户端发送所述目标推荐信息后,向所述第二用户的账号发放虚拟奖励资产。
一种信息推荐方法,包括:
接收服务端发送的候选推荐信息集合,所述候选推荐信息集合与第一用户的意图相匹配,且包含至少一条候选推荐信息;
响应第二用户对所述候选推荐信息集合的操作,得到目标推荐信息,所述目标推荐信息为需要向第一用户客户端发送的信息。
优选地,还包括:
向所述第一用户客户端发送所述目标推荐信息。
优选地,还包括:
接收服务端发送的推荐对象引导信息,以供所述第二用户基于所述推 荐对象引导信息包含的所述第一用户的标识,确定所需推荐的第一用户。
优选地,所述响应第二用户对所述候选推荐信息集合的操作,得到目标推荐信息,包括:
响应第二用户对所述候选推荐信息集合的操作,将所述候选推荐信息集合确定为目标推荐信息;
或,
响应第二用户对所述候选推荐信息集合的操作,将所述候选推荐信息集合中第二用户选中的候选推荐信息确定为目标推荐信息;
或,
响应第二用户对所述候选推荐信息集合的操作,将所述候选推荐信息集合或所述候选推荐信息集合中第二用户选中的候选推荐信息,以及第二用户基于所述候选推荐信息集合获取到的推荐信息,确定为目标推荐信息。
优选地,所述向所述第一用户客户端发送所述目标推荐信息,包括:
通过链接、文本内容、图片、视频或弹窗的形式,向所述第一用户客户端发送所述目标推荐信息。
一种信息推荐装置,包括:
意图获取单元,用于获取第一用户的意图;
候选集合获取单元,用于获取与所述第一用户的意图匹配的候选推荐信息集合,所述候选推荐信息集合包含至少一条候选推荐信息;
候选集合发送单元,用于将所述候选推荐信息集合发送给第二用户客户端,以供第二用户基于所述候选推荐信息集合确定所需向所述第一用户推荐的目标推荐信息。
一种信息推荐装置,包括:
候选集合接收单元,用于接收服务端发送的候选推荐信息集合,所述候选推荐信息集合与第一用户的意图相匹配,且包含至少一条候选推荐信息;
目标推荐信息确定单元,用于响应第二用户对所述候选推荐信息集合的操作,得到目标推荐信息,所述目标推荐信息为需要向第一用户客户端发送的信息。
优选地,还包括:
目标推荐信息发送单元,用于向所述第一用户客户端发送所述目标推荐信息。
一种信息推荐系统,其特征在于,包括:第一用户客户端、第二用户客户端及服务端,其中,所述服务端和所述第二客户端分别实现前述信息推荐方法中相关步骤。
一种信息推荐设备,包括存储器和处理器;
所述存储器,用于存储程序;
所述处理器,用于执行所述程序,实现前述信息推荐方法的各个步骤。
一种可读存储介质,其上存储有计算机程序,所述计算机程序被处理器执行时,实现前述信息推荐方法的各个步骤。
从上述的技术方案可以看出,本申请实施例提供的信息推荐方法,服务端获取第一用户的意图,获取与第一用户的意图匹配的候选推荐信息集合,集合包含至少一条候选推荐信息,进一步将候选推荐信息集合发送给第二用户客户端,以供第二用户基于候选推荐信息集合确定所需向第一用户推荐的目标推荐信息。由于服务端能够及时获取第一用户的意图,并获取到与该意图匹配的候选推荐信息集合发送给第二用户客户端,这样当第二用户存在向第一用户进行信息推荐的需求时,可以直接基于服务端发送的与第一用户匹配的候选推荐信息集合,确定目标推荐信息,减少了信息检索耗时,占用时间更短,进而更加及时的进行信息推荐。
并且,本申请方案服务端并非直接将获取的与第一用户意图匹配的候选推荐信息集合发送给第一用户,而是发送给第二用户,供第二用户向第一用户进行信息推荐,这种由用户向用户进行信息推荐的方式,相比于传统的机器基于用户意图向用户推荐信息的方案,更容易让用户接受推荐的信息,提高了信息推荐的接受度。
附图说明
为了更清楚地说明本申请实施例中的技术方案,下面将对实施例描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅 是本申请的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动性的前提下,还可以根据这些附图获得其他的附图。
图1为本申请实施例提供的实现信息推荐的一种系统架构图;
图2为本申请实施例提供的信息推荐方法的一种可选信令流程;
图3为本申请实施例提供的信息推荐方法的另一种可选信令流程;
图4-图9为本申请实施例示例的若干种应用场景示例图;
图10为本申请实施例公开的一种信息推荐装置结构示意图;
图11为本申请实施例公开的另一种信息推荐装置结构示意图;
图12为本申请实施例公开的一种信息推荐设备的硬件结构框图。
具体实施方式
下面将结合本申请实施例中的附图,对本申请实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本申请一部分实施例,而不是全部的实施例。基于本申请中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本申请保护的范围。
图1为本申请实施例提供的实现信息推荐的一种可选系统架构,如图1所示,该系统架构可以包括:第一用户客户端10、第二用户客户端20和服务端30。其中第一用户客户端10对应第一用户,第二用户客户端20对应第二用户。第一用户和第二用户为两个不同的用户。
其中,服务端30可以是设置于网络侧的服务设备,服务端30与第一用户客户端10和第二用户客户端20通过网络进行数据交互,服务端30可以是单台服务器实现,也可以是多台服务器组成的服务器集群实现。服务端30可以是由推荐信息供应商提供的服务器,在本申请实施例中服务端30能够获取第一用户的意图,并基于该意图获取匹配的候选推荐信息集合,将候选推荐信息集合发送给第二用户客户端。
第一用户客户端10和第二用户客户端20可以是电视、手机、电脑等终端设备。第一用户客户端10、第二用户客户端20能够支持对推荐信息的展示,以供用户观看以及对推荐信息的操作。进一步地,通过第一用户客户 端10、第二用户客户端20,用户可以浏览网页、观看视频、聊天等多种操作。
本申请所要实现的即是,第二用户客户端20能够接收服务端30发送的与第一用户的意图匹配的候选推荐信息集合。并且,第二用户客户端20支持第二用户对候选推荐信息集合的操作,得到向第一用户推荐的目标推荐信息。在此基础上,第二用户可以选择将目标推荐信息发送给第一用户。由于第二用户可以基于候选推荐信息集合来确定向第一用户推荐的目标推荐信息,减少了第二用户的搜索耗时,且使得第二用户更加及时的向第一用户进行信息推荐。
在此基础上,服务端30并非将检索的候选推荐信息集合直接发送给第一用户客户端10,而是发送给第二用户客户端20,供第二用户客户端30向第一用户推荐,相比传统机器直接向推荐对象发送推荐信息的方式,本申请借助其他用户向推荐对象发送推荐信息,更容易让推荐对象接受推荐的信息,提高了信息推荐的接受度。
基于图1所示系统架构,图2示出了本申请实施例提供的信息推荐方法的一种可选信令流程,参照图2,该流程可以包括:
步骤S10、服务端30获取第一用户的意图。
具体地,服务端30可以通过多种途径,获取第一用户的意图。如通过分析第一用户在第一用户客户端10上的操作,来确定第一用户的意图,操作可以包括:输入数据、浏览内容等。此外,服务端30还可以通过其它方式分析第一用户的意图,如通过第一用户配备的可穿戴设备,获取第一用户的位置信息、体感信息等,分析第一用户的意图。
示例如,获取到第一用户在浏览美食网店,则可以确定第一用户存在购买美食的意图。再比如,获取到第一用户与好友讨论旅游相关内容,则可以确定第一用户存在外出旅游的意图。又比如,通过分析第一用户近期位置轨迹,发现第一用户频繁逛汽车销售店,则可以确定用户存在购买汽车的意图,等等。
其中,第一用户的意图可以包括多种类型,如消费意图、兴趣、爱好意图等。以消费意图为例,其可以进一步分为:是否有消费意愿、所需购 买产品信息等。
步骤S11、服务端30获取与所述第一用户的意图匹配的候选推荐信息集合。
其中,所述候选推荐信息集合包含至少一条候选推荐信息。
服务端30在获取到第一用户的意图之后,可以根据第一用户的意图,在相关数据库检索匹配的候选推荐信息。以服务端30为推荐信息供应商提供为例,服务端30可以在待推荐信息中检索与第一用户的意图匹配的候选推荐信息集合。
步骤S12、服务端30将所述候选推荐信息集合发送给第二用户客户端20。
具体地,服务端30在获取到第一用户意图匹配的候选推荐信息集合之后,并非直接将候选推荐信息集合发送给第一用户对应的第一用户客户端10,而是选择将候选推荐信息集合发送给第二用户客户端,以供第二用户基于所述候选推荐信息集合确定所需向所述第一用户推荐的目标推荐信息。
其中,第二用户可以是不同于第一用户的任意一个或多个用户。当然,第二用户也可以是与第一用户具备关联关系的用户,当然,较优的情况下,第二用户是与第一用户具备正向关系的用户,该正向关系为能够提高信息推荐接受度的用户间关系。
步骤S13、第二用户客户端20响应第二用户对候选推荐信息集合的操作,得到目标推荐信息。
具体地,第二用户客户端20在收到候选推荐信息集合时可以通过显示组件向第二用户展示候选推荐信息集合。第二用户可以通过第二用户客户端20对候选推荐信息集合进行操作,得到目标推荐信息,该目标推荐信息为待向第一用户推荐的信息。
第二用户在得到目标推荐信息之后,可以将目标推荐信息存储在第二用户客户端20本地,等待需要向第一用户进行信息推荐时,再将目标推荐信息发送给第一用户客户端10。当然,第二用户在得到目标推荐信息时,还可以直接通过第二用户客户端20,将目标推荐信息发送给第一用户客户 端10。
本实施例提供的信息推荐方案,第二用户存在进行信息推荐的需求时,可以直接基于服务端30发送的与第一用户匹配的候选推荐信息集合,确定目标推荐信息,减少了信息检索时间,占用时间更短,进而更加及时的进行信息推荐。
进一步,服务端30并非将基于第一用户意图检索的候选推荐信息集合直接发送给第一用户客户端10,而是发送给第二用户客户端20,供第二用户客户端30向第一用户推荐。相比传统机器直接向推荐对象发送推荐信息的方式,本申请借助其他用户向推荐对象发送推荐信息,更容易让推荐对象接受推荐的信息,提高了信息推荐的接受度。
进一步参见图3,图3示出了本申请实施例提供的信息推荐方法的另一种可选信令流程,参照图3,该流程可以包括:
步骤S20、服务端30获取第一用户的意图。
步骤S21、服务端30获取与所述第一用户的意图匹配的候选推荐信息集合。
步骤S22、服务端30将所述候选推荐信息集合发送给第二用户客户端20。
步骤S23、第二用户客户端20响应第二用户对候选推荐信息集合的操作,得到目标推荐信息。
上述步骤S20-S23与前述实施例中步骤S10-S13一一对应,详细参照前述介绍,此处不再赘述。
步骤S24、第二用户客户端20将目标推荐信息发送给第一用户客户端10。
相比于上一实施例,本实施例中第二用户客户端20在得到目标推荐信息之后,进一步将目标推荐信息发送给第一用户客户端10。相比传统机器直接向推荐对象发送推荐信息的方式,本申请借助其他用户向推荐对象发送推荐信息,更容易让推荐对象接受推荐的信息,提高了信息推荐的接受度。
接下来,通过几个实际应用场景对本申请的信息推荐方法进行介绍。
参见图4和图5,其示例了本申请信息推荐方法的一种实现场景。
第一用户与第二用户分别通过各自的客户端与对方聊天。假定第一用户昵称为刘xx,第二用户昵称为张xx。
如图4所示,聊天过程中,第一用户向第二用户发送消息:“最近想换个手机,有没有推荐的?”。
该消息同时作为聊天内容上传到服务端30。服务端30根据聊天内容确定第一用户存在购买手机的意图。因此,检索最新手机推荐信息,得到检索结果:手机推荐信息链接。并将该手机推荐信息链接下发给与第一用户聊天的第二用户。
第二用户客户端20显示界面的右下角通过浮窗形式显示服务端30下发的信息,包括:手机链接及提示语:“这里有款手机很不错,快推荐给你的朋友刘xx吧~”。第二用户在收到该手机链接时,可以省去自己搜索的环节,直接复制手机链接发送给第一用户。具体效果如图5所示。
显然,通过本申请方案,第二用户在收到第一用户的信息推荐请求时,不需要自行去检索,可以直接将服务端下发的手机推荐信息链接发送给第二用户客户端,节省了第二用户的检索时间,信息推荐更加及时,且由于手机推荐信息是由第二用户推荐给第一用户的,相比于由服务端直接向第一用户推荐,该手机推荐信息更加容易被第一用户接受,提高了信息推荐的接受度。
进一步参见图6和图7,其示例了本申请信息推荐方法的另一种实现场景。
假定第一用户昵称为刘xx,第二用户昵称为张xx。
如图6所示,第一用户通过第一用户客户端10搜索成都景点信息。该搜索信息作为搜索内容上传到服务端30。
服务端30根据搜索内容确定第一用户存在去成都旅游的意图。因此,检索成都景点推荐信息,并将检索到的景点推荐信息链接下发给第二客户端20。
第二用户正通过第二用户客户端20在玩游戏,此时收到服务端30下发的弹窗信息,如图7右下角所示。弹窗内容包括:景点推荐信息链接及提示语:“别玩游戏啦,快和你的朋友刘xx去这里旅游吧~”。
第二用户收到服务端30下发的信息后,觉得这个景点很不错,可以复制该景点推荐信息链接,并发送给第一用户客户端10。具体可以是通过即时通讯应用来进行信息发送。该发送消息在第一用户客户端10界面右下角弹出显示。
显然,通过本申请方案,第二用户在收到服务端下发的景点推荐信息链接时,可以将该景点推荐信息链接推荐给第一用户。第二用户不需要额外进行景点检索。并且,景点推荐信息是由第二用户推荐给第一用户的,相比于由服务端直接向第一用户推荐,该景点推荐信息更加容易被第一用户接受,提高了信息推荐的接受度。
接下来,本申请实施例首先从服务端30的角度,对信息推荐方案做进一步介绍。
对于服务端30而言,前文已经介绍其能够获取第一用户的意图。本实施例中详细介绍服务端30获取第一用户的意图的过程。
服务端30可以获取第一用户当前时刻或当前时刻之前设定时间段内的输入数据,进一步根据输入数据及第一用户的历史画像数据,确定第一用户的意图。
其中,输入数据的类型可以有多种,如第一用户在与其他用户聊天过程的输入聊天内容,或第一用户在浏览器、应用商店内输入的搜索内容等。
用户的历史画像数据可以包括用户的基本信息、兴趣爱好等等。通过对获取的输入数据与第一用户的历史画像数据进行分析,可以确定第一用户的意图。
一种可选的方式下,本申请实施例可以预先构建意图识别模型,该意图识别模型可以是能够进行分类的各形式的神经网络模型。
预先收集训练用户的输入数据及其历史画像数据,作为训练样本,并获取训练用户在输入数据时的真实意图,作为样本标签。基于收集的训练 样本及样本标签训练意图识别模型。
在训练得到意图识别模型后,可以将获取的第一用户的输入数据及其历史画像数据输入该意图识别模型,得到意图识别模型输出的第一用户的意图。
需要说明的是,为了获取第一用户更全面的意图,该意图识别模型的个数可以是多个。多个意图识别模型中,某些意图识别模型用于确定第一用户是否有某种类型的意图,如消费意图、使用意图等。另外一些意图识别模型可以用于在确定第一用户存在某种类型的意图时,进一步确定该类型意图下的详细意图信息。以消费意图为例,可以确定第一用户所要购买物品的详细信息,如产品型号、名称、尺寸等。
以第一用户购买手机为例:
1)确定第一用户是否有购买手机的意愿。
2)在确定第一用户有购买手机的意愿时,进一步通过分析第一用户当前使用手机的机型,以及第一用户对当前机型及其它机型的评价,确定第一用户所要购买手机的机型。
进一步地,服务端30在确定了第一用户的意图之后,获取与第一用户的意图匹配的候选推荐信息集合。该过程可以包括:
服务端30基于第一用户的意图,从待推荐信息源中检索出与第一用户的意图匹配的候选推荐信息集合。在检索过程中,可以同时考虑待推荐信息源中每一条待推荐信息的热度信息、用户关注程度等信息,也即,根据待推荐信息源中每一条待推荐信息的热度信息、用户关注程度等信息,检索与第一用户意图匹配的候选推荐信息集合。
服务端30在得到候选推荐信息集合之后,可以选择通过链接、文本内容、图片、视频或弹窗等形式发送给第二用户客户端20。如图4,其示例的即是通过弹窗的形式。
再进一步的,服务端30可以和第二用户预先约定好,由服务端30向第二用户客户端20发送的候选推荐信息集合,均是与第一用户匹配的。这样,第二用户客户端20在收到服务端30发送的候选推荐信息集合时,可以直接确定该候选推荐信息集合是与第一用户匹配的,基于候选推荐信息集合所 确定的目标推荐信息也是需要向第一用户推荐的。
除此之外,服务端30在向第二用户客户端20发送候选推荐信息集合的同时,还可以发送推荐对象引导信息,该推荐对象引导信息包含第一用户的标识,以供第二用户基于推荐对象引导信息包含的第一用户的标识,确定所需推荐的第一用户。如图4示例的,服务端30向第二客户端20发送的信息中包含推荐对象引导信息:“这里有款手机很不错,快推荐给你的朋友刘xx吧~”。该推荐对象引导信息中包含了第一用户的标识:刘xx。
当然可以理解的是,图4仅仅示例了推荐对象引导信息的一种可选组成结构,除此之外还可以设置其他形式的推荐对象引导信息,只要能够向第二用户明确推荐对象为第一用户即可。
可选的,服务端30在向第二用户客户端20发送候选推荐信息集合之前,可以增加确定第二用户的过程。
一种可选的情况下,服务端30可以将非第一用户的所有其他用户均确定为第二用户,或从非第一用户的其它用户中随机选择一个或多个作为第二用户。
另一种可选的情况下,服务端30可以根据第一用户的交流记录,确定第二用户。或者,服务端30可以根据第一用户的关联社交应用中的联系人信息,确定第二用户。
首先介绍服务端30根据第一用户的交流记录,确定第二用户的过程。
第一用户的交流记录可以包括第一用户通过即时通讯应用进行的聊天交流记录。还可以包括第一用户与客服人员的交流记录,如与网店客服人员的交流记录等。
可以理解的是,与第一用户存在交流记录的用户,均与第一用户存在一定的关联关系,因此本实施例中可以将与第一用户存在交流记录的用户确定为第二用户。
进一步考虑,与第一用户存在交流记录的用户可能数量过多,本实施例可以从中进一步筛选部分用户作为第二用户。具体地可以通过设定条件,筛选满足设定条件的用户作为第二用户。设定条件可以包括:
与第一用户的亲密度达到设定亲密度条件、当前处于在线状态、当前 处于与第一用户交流状态中的任意一个或多个。
其中,亲密度表征了用户间产生潜在信息推荐的可能性。两个用户间的亲密度越大,代表两个用户互相进行信息推荐的可能性越大。当然,可以设置第一用户与第二用户的亲密度等于第二用户与第一用户的亲密度,或者,通过设置亲密度确定方式分别确定第一用户与第二用户的亲密度,以及第二用户与第一用户的亲密度,二者并不存在必然的相等关系。
本申请实施例公开了确定第一用户与第二用户亲密度的具体实现方式。
首先,本实施例可以分析影响亲密度的若干因素,这些影响因素可以包括以下任意一个或多个:
第二用户与第一用户的单独互动次数、第二用户在第一用户好友列表中的存在状态(包括存在和不存在)、第二用户与第一用户最近一次交流距离当前时间差、第二用户与第一用户共同存在的群组数目、第二用户与第一用户在群组中互动次数等。
基于此,本实施例可以根据上述任意一个或多个影响因素的值,来确定第二用户与第一用户的亲密度。其中,上述多个影响因素对确定亲密度的影响权重可以相同也可以不同,具体可以为每个影响因素分配对应的权重,通过线性加权的方式综合考虑多个影响因素对亲密度的影响。
当然,本实施例中还可以采用模型预测的方式来确定第二用户与第一用户的亲密度。也即,本实施例可以预先训练亲密度确定模型,训练时预先确定两个训练用户间的上述各影响因素的值,将这些影响因素的值作为训练样本,同时将标注的训练用户间的亲密度值作为样本标签,对亲密度确定模型进行训练。
基于训练好的亲密度确定模型,可以将第二用户与第一用户的各影响因素的值输入亲密度确定模型,得到模型输出的第二用户与第一用户的亲密度值。
本实施例中预先设定的亲密度条件可以是,选取亲密度值超过亲密度阈值的用户作为第二用户,或者是,选取亲密度值最高的前N个用户作为第二用户,再或者是,选取亲密度值由高至低排序中前M%个用户作为第 二用户。
进一步地,前述设定条件中:当前处于在线状态,表示在指定应用中处于在线状态,如在指定即时通讯应用中处于在线状态。
再进一步地,前述设定条件中:当前处于与第一用户交流状态,表示当前时刻之前设定时间段内与第一用户存在交流记录。
可以理解的是,前述设定条件还可以包括其他条件,本实施例中并未穷举。
进一步地,本申请实施例介绍服务端30根据第一用户的关联社交应用中的联系人信息,确定第二用户的过程。
可选的,本实施例中可以将第一用户的关联社交应用的联系人均确定为第二用户。
进一步考虑,第一用户的关联社交应用的联系人可能数量过多,本实施例可以从中进一步筛选部分用户作为第二用户。具体地可以通过设定条件,筛选满足设定条件的用户作为第二用户。设定条件可以包括:
与第一用户的亲密度达到设定亲密度条件、当前处于在线状态、当前处于与第一用户交流状态中的任意一个或多个。
其中,上述各设定条件已经在前文进行了介绍,详细参照前文,此处不再赘述。
进一步可选的,在本申请的另一个实施例中,针对信息推荐方方法提供了一种奖励机制。即,服务端30在检测到第二用户客户端20向第一用户客户端10发送所述目标推荐信息后,可以向第二用户的账号发放虚拟奖励资产。
具体地,服务端30可以监控第二用户客户端20是否对目标推荐信息进行了发送操作,或者可以监控第一用户客户端10是否接收到或打开了目标推荐信息,基于此确定第二用户客户端20是否想第一用户客户端10发送目标推荐信息。
本实施例通过设置奖励机制,能够进一步调动第二用户向第一用户进行信息推荐的积极性,从而提高信息推荐的数量。
接下来本申请实施例进一步从第二用户客户端20的角度对信息推荐方法进行介绍。
第二用户客户端20接收服务端30发送的候选推荐信息集合,该候选推荐信息集合与第一用户的意图相匹配。进一步,第二用户客户端20响应第二用户对候选推荐信息集合的操作,得到目标推荐信息,该目标推荐信息作为需要向第一用户客户端发送的信息。
可选的,第二用户在得到目标推荐信息之后,可以将目标推荐信息存储在第二用户客户端20本地,等待需要向第一用户进行信息推荐时,再将目标推荐信息发送给第一用户客户端10。当然,第二用户在得到目标推荐信息时,还可以直接通过第二用户客户端20,将目标推荐信息发送给第一用户客户端10。
本实施例提供的信息推荐方案,第二用户进行信息推荐的需求时,可以直接基于服务端30发送的与第一用户匹配的候选推荐信息集合,确定目标推荐信息,无需临时进行信息检索,占用时间更短,进而更加及时的进行信息推荐。
本实施例中可以由第二用户向第一用户推荐目标推荐信息,而非由服务端30直接向第一用户推荐,更容易让第一用户接受推荐的信息,提高了信息推荐的接受度。
进一步地,结合图5和图7所示,第二用户客户端20可以通过链接、文本内容、图片、视频或弹窗的形式,向第一用户客户端10发送目标推荐信息。
进一步地,第二用户客户端20在接收服务端30发送的候选推荐信息集合的同时,还可以接收服务端30发送的推荐对象引导信息,该推荐对象引导信息包含第一用户的标识。第二用户客户端20可以展示该推荐对象引导信息,以供第二用户根据推荐对象引导信息包含的第一用户的标识,确定所需推荐的第一用户。详细参照图4-图7示例的,推荐对象引导信息均包含第一用户的标识“刘xx”。
进一步地,对第二用户客户端20响应第二用户对所述候选推荐信息集合的操作,得到目标推荐信息的过程进行介绍。
本实施例示例了几种响应用户操作得到目标推荐信息的方式,分别如下:
1)、响应第二用户对所述候选推荐信息集合的操作,将所述候选推荐信息集合确定为目标推荐信息。
2)、响应第二用户对所述候选推荐信息集合的操作,将所述候选推荐信息集合中第二用户选中的候选推荐信息确定为目标推荐信息。
3)、响应第二用户对所述候选推荐信息集合的操作,将所述候选推荐信息集合或所述候选推荐信息集合中第二用户选中的候选推荐信息,以及第二用户基于所述候选推荐信息集合获取到的推荐信息,确定为目标推荐信息。
上述三种方式示例了三种基于候选推荐信息集合得到目标推荐信息的方式。该三种方式中,每一种方式,均不需要第二用户额外去检索向第一用户推荐的信息,仅仅基于服务端30下发的候选推荐信息集合进行简单操作即可得到目标推荐信息。
参见图8和图9,其示例了本申请信息推荐方法的又一种实现场景。
第一用户与第二用户分别通过各自的客户端与对方聊天。假定第一用户昵称为刘xx,第二用户昵称为张xx。
如图4所示,聊天过程中,第一用户向第二用户发送消息:“最近想换个手机,有没有推荐的?”。
该消息同时作为聊天内容上传到服务端30。服务端30根据聊天内容确定第一用户存在购买手机的意图。因此,检索最新手机推荐信息,得到多条检索结果,组成手机推荐信息链接集合,该集合包含了三条手机推荐信息链接,分别为:https://shouji.com/search1、https://shouji.com/search2、https://shouji.com/search3。服务端30进一步将手机推荐信息链接集合发送给第二用户客户端20。
第二用户客户端20显示界面的右下角通过浮窗形式显示服务端30下发的信息,包括:三条手机推荐信息链接及提示语:“这里有几款手机很不错,快推荐给你的朋友刘xx吧~”。具体效果如图8所示。
第二用户在收到该消息时,可以逐条分析三个链接,并依据自身对第 一用户的了解,从中确定前两条链接比较适合第一用户。因此可以复制前两条链接发送给第一用户客户端10。具体效果如图9所示。
图8-图9示例的情况下与上述方案2)对应,即用户从候选推荐信息集合中选中部分候选推荐信息作为目标推荐信息,发送给第一用户客户端。其余两种方案并未通过附图示例。
下面对本申请实施例提供的信息推荐装置进行描述,下文描述的信息推荐装置与上文描述的信息推荐装置可相互对应参照。
首先,结合图10,对应用于服务端30的信息推荐装置进行介绍,如图10所示,该信息推荐装置可以包括:
意图获取单元100,用于获取第一用户的意图;
候选集合获取单元110,用于获取与所述第一用户的意图匹配的候选推荐信息集合,所述候选推荐信息集合包含至少一条候选推荐信息;
候选集合发送单元120,用于将所述候选推荐信息集合发送给第二用户客户端,以供第二用户基于所述候选推荐信息集合确定所需向所述第一用户推荐的目标推荐信息。
可选的,所述意图获取单元可以包括:
输入数据获取单元,用于获取第一用户当前时刻或当前时刻之前设定时间段内的输入数据;
数据应用单元,用于根据所述输入数据及所述第一用户的历史画像数据,确定所述第一用户的意图。
可选的,所述数据应用单元可以包括:
意图识别模型预测单元,用于将所述输入数据及所述历史画像数据输入预置的意图识别模型,得到意图识别模型输出的所述第一用户的意图;所述意图识别模型为,预先以训练用户的输入数据及其历史画像数据作为训练样本,以训练用户的意图作为样本标签训练得到。
可选的,本申请的信息推荐装置还可以包括:
推荐对象引导信息发送单元,用于将推荐对象引导信息发送给第二用户客户端,以供第二用户基于所述推荐对象引导信息包含的所述第一用户 的标识,确定所需推荐的第一用户。
可选的,本申请的信息推荐装置还可以包括:
交流记录使用单元,用于根据所述第一用户的交流记录,确定所述第二用户;
或,
联系人使用单元,用于根据所述第一用户的关联社交应用中的联系人信息,确定所述第二用户。
可选的,所述交流记录使用单元可以包括:
第一交流记录使用子单元,用于将与所述第一用户存在交流记录的用户确定为第二用户;
或,
第二交流记录使用子单元,用于在与所述第一用户存在交流记录的用户中,筛选满足设定条件的用户作为第二用户,所述设定条件包括:
与所述第一用户的亲密度达到设定亲密度条件、当前处于在线状态、当前处于与所述第一用户交流状态中的任意一个或多个。
可选的,所述联系人使用单元可以包括:
第一联系人使用子单元,用于将所述第一用户的关联社交应用的联系人均确定为所述第二用户;
或,
第二联系人使用子单元,用于将所述第一用户的关联社交应用的联系人中,筛选满足设定条件的用户作为第二用户,所述设定条件包括:
与所述第一用户的亲密度达到设定亲密度条件、当前处于在线状态、当前处于与所述第一用户交流状态中的任意一个或多个。
可选的,所述候选集合发送单元可以包括:
第一候选集合发送子单元,用于将所述候选推荐信息集合通过链接、文本内容、图片、视频或弹窗的形式发送给第二用户客户端。
可选的,所述信息推荐装置还可以包括:
奖励单元,用于在检测到所述第二用户客户端向第一用户客户端发送所述目标推荐信息后,向所述第二用户的账号发放虚拟奖励资产。
进一步,结合图11,对应用于第二用户客户端20的信息推荐装置进行介绍,如图11所示,该信息推荐装置可以包括:
候选集合接收单元200,用于接收服务端发送的候选推荐信息集合,所述候选推荐信息集合与第一用户的意图相匹配,且包含至少一条候选推荐信息;
目标推荐信息确定单元210,用于响应第二用户对所述候选推荐信息集合的操作,得到目标推荐信息,所述目标推荐信息为需要向第一用户客户端发送的信息。
可选的,所述信息推荐装置还可以包括:
目标推荐信息发送单元,用于向所述第一用户客户端发送所述目标推荐信息。
可选的,所述信息推荐装置还可以包括:
推荐对象引导信息接收单元,用于接收服务端发送的推荐对象引导信息,以供所述第二用户基于所述推荐对象引导信息包含的所述第一用户的标识,确定所需推荐的第一用户。
可选的,所述目标推荐信息确定单元可以包括:
第一响应子单元,用于响应第二用户对所述候选推荐信息集合的操作,将所述候选推荐信息集合确定为目标推荐信息;
或,
第二响应子单元,用于响应第二用户对所述候选推荐信息集合的操作,将所述候选推荐信息集合中第二用户选中的候选推荐信息确定为目标推荐信息;
或,
第三响应子单元,用于响应第二用户对所述候选推荐信息集合的操作,将所述候选推荐信息集合或所述候选推荐信息集合中第二用户选中的候选推荐信息,以及第二用户基于所述候选推荐信息集合获取到的推荐信息,确定为目标推荐信息。
可选的,所述目标推荐信息发送单元可以包括:
第一目标推荐信息发送子单元,用于通过链接、文本内容、图片、视频或弹窗的形式,向所述第一用户客户端发送所述目标推荐信息。
本申请实施例提供的信息推荐装置可应用于信息推荐设备。信息推荐设备可以是服务端30或第二用户客户端20。图12示出了信息推荐设备的硬件结构框图,参照图12,信息推荐设备的硬件结构可以包括:至少一个处理器1,至少一个通信接口2,至少一个存储器3和至少一个通信总线4;
在本申请实施例中,处理器1、通信接口2、存储器3、通信总线4的数量为至少一个,且处理器1、通信接口2、存储器3通过通信总线4完成相互间的通信;
处理器1可能是一个中央处理器CPU,或者是特定集成电路ASIC(Application Specific Integrated Circuit),或者是被配置成实施本发明实施例的一个或多个集成电路等;
存储器3可能包含高速RAM存储器,也可能还包括非易失性存储器(non-volatile memory)等,例如至少一个磁盘存储器;
其中,存储器存储有程序,处理器可调用存储器存储的程序,所述程序用于:实现前述服务端30在信息推荐方案中的各个处理流程,或,实现前述第二用户客户端20在信息推荐方案中的各个处理流程。
本申请实施例还提供一种存储介质,该存储介质可存储有适于处理器执行的程序,所述程序用于:实现前述服务端30在信息推荐方案中的各个处理流程,或,实现前述第二用户客户端20在信息推荐方案中的各个处理流程。
本申请实施例还公开了一种信息推荐系统,该信息推荐信息包括第一用户客户端10、第二用户客户端20和服务端30,其中上述三个单元的具体实现逻辑可以参照前述新推荐方法部分的相关介绍,此处不再赘述。
最后,还需要说明的是,在本文中,诸如第一和第二等之类的关系术语仅仅用来将一个实体或者操作与另一个实体或操作区分开来,而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而 且,术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含,从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素,而且还包括没有明确列出的其他要素,或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的情况下,由语句“包括一个……”限定的要素,并不排除在包括所述要素的过程、方法、物品或者设备中还存在另外的相同要素。
本说明书中各个实施例采用递进的方式描述,每个实施例重点说明的都是与其他实施例的不同之处,各个实施例之间相同相似部分互相参见即可。
对所公开的实施例的上述说明,使本领域专业技术人员能够实现或使用本申请。对这些实施例的多种修改对本领域的专业技术人员来说将是显而易见的,本文中所定义的一般原理可以在不脱离本申请的精神或范围的情况下,在其它实施例中实现。因此,本申请将不会被限制于本文所示的这些实施例,而是要符合与本文所公开的原理和新颖特点相一致的最宽的范围。

Claims (20)

  1. 一种信息推荐方法,其特征在于,包括:
    获取第一用户的意图;
    获取与所述第一用户的意图匹配的候选推荐信息集合,所述候选推荐信息集合包含至少一条候选推荐信息;
    将所述候选推荐信息集合发送给第二用户客户端,以供第二用户基于所述候选推荐信息集合确定所需向所述第一用户推荐的目标推荐信息。
  2. 根据权利要求1所述的方法,其特征在于,所述获取第一用户的意图,包括:
    获取第一用户当前时刻或当前时刻之前设定时间段内的输入数据;
    根据所述输入数据及所述第一用户的历史画像数据,确定所述第一用户的意图。
  3. 根据权利要求2所述的方法,其特征在于,所述根据所述输入数据及所述第一用户的历史画像数据,确定所述第一用户的意图,包括:
    将所述输入数据及所述历史画像数据输入预置的意图识别模型,得到意图识别模型输出的所述第一用户的意图;
    所述意图识别模型为,预先以训练用户的输入数据及其历史画像数据作为训练样本,以训练用户的意图作为样本标签训练得到。
  4. 根据权利要求1所述的方法,其特征在于,还包括:
    将推荐对象引导信息发送给第二用户客户端,以供第二用户基于所述推荐对象引导信息包含的所述第一用户的标识,确定所需推荐的第一用户。
  5. 根据权利要求1所述的方法,其特征在于,在将所述候选推荐信息集合发送给第二用户客户端之前,该方法还包括:
    根据所述第一用户的交流记录,确定所述第二用户;
    或,
    根据所述第一用户的关联社交应用中的联系人信息,确定所述第二用户。
  6. 根据权利要求5所述的方法,其特征在于,所述根据所述第一用户 的交流记录,确定所述第二用户,包括:
    将与所述第一用户存在交流记录的用户确定为第二用户;
    或,
    在与所述第一用户存在交流记录的用户中,筛选满足设定条件的用户作为第二用户,所述设定条件包括:
    与所述第一用户的亲密度达到设定亲密度条件、当前处于在线状态、当前处于与所述第一用户交流状态中的任意一个或多个。
  7. 根据权利要求5所述的方法,其特征在于,所述根据所述第一用户的关联社交应用中的联系人信息,确定所述第二用户,包括:
    将所述第一用户的关联社交应用的联系人均确定为所述第二用户;
    或,
    将所述第一用户的关联社交应用的联系人中,筛选满足设定条件的用户作为第二用户,所述设定条件包括:
    与所述第一用户的亲密度达到设定亲密度条件、当前处于在线状态、当前处于与所述第一用户交流状态中的任意一个或多个。
  8. 根据权利要求1-7任一项所述的方法,其特征在于,所述将所述候选推荐信息集合发送给第二用户客户端,包括:
    将所述候选推荐信息集合通过链接、文本内容、图片、视频或弹窗的形式发送给第二用户客户端。
  9. 根据权利要求1-7任一项所述的方法,其特征在于,还包括:
    在检测到所述第二用户客户端向第一用户客户端发送所述目标推荐信息后,向所述第二用户的账号发放虚拟奖励资产。
  10. 一种信息推荐方法,其特征在于,包括:
    接收服务端发送的候选推荐信息集合,所述候选推荐信息集合与第一用户的意图相匹配,且包含至少一条候选推荐信息;
    响应第二用户对所述候选推荐信息集合的操作,得到目标推荐信息,所述目标推荐信息为需要向第一用户客户端发送的信息。
  11. 根据权利要求10所述的方法,其特征在于,还包括:
    向所述第一用户客户端发送所述目标推荐信息。
  12. 根据权利要求10或11所述的方法,其特征在于,还包括:
    接收服务端发送的推荐对象引导信息,以供所述第二用户基于所述推荐对象引导信息包含的所述第一用户的标识,确定所需推荐的第一用户。
  13. 根据权利要求10或11所述的方法,其特征在于,所述响应第二用户对所述候选推荐信息集合的操作,得到目标推荐信息,包括:
    响应第二用户对所述候选推荐信息集合的操作,将所述候选推荐信息集合确定为目标推荐信息;
    或,
    响应第二用户对所述候选推荐信息集合的操作,将所述候选推荐信息集合中第二用户选中的候选推荐信息确定为目标推荐信息;
    或,
    响应第二用户对所述候选推荐信息集合的操作,将所述候选推荐信息集合或所述候选推荐信息集合中第二用户选中的候选推荐信息,以及第二用户基于所述候选推荐信息集合获取到的推荐信息,确定为目标推荐信息。
  14. 根据权利要求11所述的方法,其特征在于,所述向所述第一用户客户端发送所述目标推荐信息,包括:
    通过链接、文本内容、图片、视频或弹窗的形式,向所述第一用户客户端发送所述目标推荐信息。
  15. 一种信息推荐装置,其特征在于,包括:
    意图获取单元,用于获取第一用户的意图;
    候选集合获取单元,用于获取与所述第一用户的意图匹配的候选推荐信息集合,所述候选推荐信息集合包含至少一条候选推荐信息;
    候选集合发送单元,用于将所述候选推荐信息集合发送给第二用户客户端,以供第二用户基于所述候选推荐信息集合确定所需向所述第一用户推荐的目标推荐信息。
  16. 一种信息推荐装置,其特征在于,包括:
    候选集合接收单元,用于接收服务端发送的候选推荐信息集合,所述候选推荐信息集合与第一用户的意图相匹配,且包含至少一条候选推荐信息;
    目标推荐信息确定单元,用于响应第二用户对所述候选推荐信息集合的操作,得到目标推荐信息,所述目标推荐信息为需要向第一用户客户端发送的信息。
  17. 根据权利要求16所述的装置,其特征在于,还包括:
    目标推荐信息发送单元,用于向所述第一用户客户端发送所述目标推荐信息。
  18. 一种信息推荐系统,其特征在于,包括:第一用户客户端、第二用户客户端及服务端,其中,所述服务端用于实现权利要求1-9任一项的信息推荐方法的各个步骤,所述第二客户端用于实现权利要求10-14任一项的信息推荐方法的各个步骤。
  19. 一种信息推荐设备,其特征在于,包括存储器和处理器;
    所述存储器,用于存储程序;
    所述处理器,用于执行所述程序,实现如权利要求1-9,或权利要求10-14任一项的信息推荐方法的各个步骤。
  20. 一种可读存储介质,其上存储有计算机程序,其特征在于,所述计算机程序被处理器执行时,实现如权利要求1-9,或权利要求10-14任一项的信息推荐方法的各个步骤。
PCT/CN2018/093217 2018-05-29 2018-06-28 信息推荐方法、装置、系统、设备及可读存储介质 WO2019227560A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810531062.3 2018-05-29
CN201810531062.3A CN108446410B (zh) 2018-05-29 2018-05-29 信息推荐方法、装置、系统、设备及可读存储介质

Publications (1)

Publication Number Publication Date
WO2019227560A1 true WO2019227560A1 (zh) 2019-12-05

Family

ID=63205086

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/093217 WO2019227560A1 (zh) 2018-05-29 2018-06-28 信息推荐方法、装置、系统、设备及可读存储介质

Country Status (2)

Country Link
CN (1) CN108446410B (zh)
WO (1) WO2019227560A1 (zh)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111026971A (zh) * 2019-12-25 2020-04-17 腾讯科技(深圳)有限公司 内容的推送方法及装置、计算机存储介质
CN111368219A (zh) * 2020-02-27 2020-07-03 广州腾讯科技有限公司 信息推荐方法、装置、计算机设备以及存储介质
CN112449002A (zh) * 2020-10-19 2021-03-05 微民保险代理有限公司 一种待推送对象的推送方法、装置、设备及存储介质
CN112465555A (zh) * 2020-12-04 2021-03-09 北京搜狗科技发展有限公司 一种广告信息推荐的方法及相关装置
CN113645474A (zh) * 2021-07-26 2021-11-12 阿里巴巴(中国)有限公司 互动信息的处理方法、显示方法及电子设备
CN113674833A (zh) * 2021-08-23 2021-11-19 成都拟合未来科技有限公司 健身视频生成方法、系统、终端及存储介质
CN113779369A (zh) * 2020-07-15 2021-12-10 北京沃东天骏信息技术有限公司 匹配方法、匹配装置、电子设备及存储介质
CN116362848A (zh) * 2023-06-03 2023-06-30 成都豪杰特科技有限公司 基于人工智能的电子商务的推荐方法、系统、设备和介质
CN112465555B (zh) * 2020-12-04 2024-05-14 北京搜狗科技发展有限公司 一种广告信息推荐的方法及相关装置

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109241425B (zh) * 2018-08-31 2022-02-18 腾讯科技(深圳)有限公司 一种资源推荐方法、装置、设备及存储介质
CN110231967A (zh) * 2019-06-06 2019-09-13 北京字节跳动网络技术有限公司 一种提及人组合显示方法、装置、终端设备及存储介质
CN110400185A (zh) * 2019-07-31 2019-11-01 中国工商银行股份有限公司 产品推荐方法及系统
CN110781354B (zh) * 2019-10-24 2022-06-10 北京齐尔布莱特科技有限公司 一种对象选择方法、装置、系统及计算设备
CN112351056B (zh) * 2019-11-15 2022-09-30 北京沃东天骏信息技术有限公司 用于分享信息的方法和装置
CN110941766B (zh) * 2019-12-10 2023-10-20 北京字节跳动网络技术有限公司 一种信息推送的方法、装置、计算机设备及存储介质
CN112132628A (zh) * 2020-09-28 2020-12-25 科大讯飞股份有限公司 一种用户意图预测方法、信息推荐方法及相关设备
CN112256719A (zh) * 2020-10-20 2021-01-22 北京字节跳动网络技术有限公司 实体查询方法、装置、可读介质与电子设备
CN112418402B (zh) * 2020-11-24 2023-08-11 百度在线网络技术(北京)有限公司 推荐对象的方法、神经网络及其训练方法、计算设备
CN113434643A (zh) * 2021-05-20 2021-09-24 华为技术有限公司 一种信息推荐方法以及相关设备
CN113393300B (zh) * 2021-06-23 2022-11-18 未鲲(上海)科技服务有限公司 基于人工智能的产品推荐方法、装置、设备及存储介质
CN113505315B (zh) * 2021-09-09 2021-12-07 环球数科集团有限公司 多用户的旅行攻略制定方法、装置和计算机设备

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104731870A (zh) * 2015-03-02 2015-06-24 百度在线网络技术(北京)有限公司 一种用于提供推荐信息的方法与设备
CN105117418A (zh) * 2015-07-30 2015-12-02 百度在线网络技术(北京)有限公司 基于搜索的服务信息管理系统及方法
CN106716418A (zh) * 2016-10-28 2017-05-24 达闼科技(北京)有限公司 软件推荐的方法、装置和终端以及服务器

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
RU2014118941A (ru) * 2011-10-13 2015-11-20 Конинклейке Филипс Н.В. Автоматическая генерация рейтинговых запросов от рекомендательной системы
US9262175B2 (en) * 2012-12-11 2016-02-16 Nuance Communications, Inc. Systems and methods for storing record of virtual agent interaction
CN104639420B (zh) * 2013-11-15 2019-06-07 腾讯科技(深圳)有限公司 即时通讯的信息处理方法和系统
CN104679536A (zh) * 2013-11-28 2015-06-03 索尼公司 应用程序的启动方法、装置以及电子设备
CN104836720B (zh) * 2014-02-12 2022-02-25 北京三星通信技术研究有限公司 交互式通信中进行信息推荐的方法及装置
CN104038909B (zh) * 2014-06-16 2017-08-29 浙江翼信科技有限公司 一种信息交互方法和设备
US10671679B2 (en) * 2014-12-30 2020-06-02 Oath Inc. Method and system for enhanced content recommendation
CN107220899B (zh) * 2016-03-21 2020-06-26 阿里巴巴集团控股有限公司 社交网络构建、信息推荐方法、装置及服务器
CN106934689A (zh) * 2017-02-25 2017-07-07 杭州领娱科技有限公司 基于即时通讯聊天室的商品推送系统及方法
CN106993048B (zh) * 2017-04-13 2018-09-14 腾讯科技(深圳)有限公司 确定推荐信息的方法及装置、信息推荐方法及装置
CN107249137A (zh) * 2017-06-28 2017-10-13 张迅 信息推送的方法、装置及系统

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104731870A (zh) * 2015-03-02 2015-06-24 百度在线网络技术(北京)有限公司 一种用于提供推荐信息的方法与设备
CN105117418A (zh) * 2015-07-30 2015-12-02 百度在线网络技术(北京)有限公司 基于搜索的服务信息管理系统及方法
CN106716418A (zh) * 2016-10-28 2017-05-24 达闼科技(北京)有限公司 软件推荐的方法、装置和终端以及服务器

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111026971A (zh) * 2019-12-25 2020-04-17 腾讯科技(深圳)有限公司 内容的推送方法及装置、计算机存储介质
CN111026971B (zh) * 2019-12-25 2023-05-02 腾讯科技(深圳)有限公司 内容的推送方法及装置、计算机存储介质
CN111368219A (zh) * 2020-02-27 2020-07-03 广州腾讯科技有限公司 信息推荐方法、装置、计算机设备以及存储介质
CN111368219B (zh) * 2020-02-27 2024-04-26 广州腾讯科技有限公司 信息推荐方法、装置、计算机设备以及存储介质
CN113779369A (zh) * 2020-07-15 2021-12-10 北京沃东天骏信息技术有限公司 匹配方法、匹配装置、电子设备及存储介质
CN112449002B (zh) * 2020-10-19 2022-08-12 微民保险代理有限公司 一种待推送对象的推送方法、装置、设备及存储介质
CN112449002A (zh) * 2020-10-19 2021-03-05 微民保险代理有限公司 一种待推送对象的推送方法、装置、设备及存储介质
CN112465555A (zh) * 2020-12-04 2021-03-09 北京搜狗科技发展有限公司 一种广告信息推荐的方法及相关装置
CN112465555B (zh) * 2020-12-04 2024-05-14 北京搜狗科技发展有限公司 一种广告信息推荐的方法及相关装置
CN113645474A (zh) * 2021-07-26 2021-11-12 阿里巴巴(中国)有限公司 互动信息的处理方法、显示方法及电子设备
CN113674833A (zh) * 2021-08-23 2021-11-19 成都拟合未来科技有限公司 健身视频生成方法、系统、终端及存储介质
CN113674833B (zh) * 2021-08-23 2024-02-06 成都拟合未来科技有限公司 健身视频生成方法、系统、终端及存储介质
CN116362848A (zh) * 2023-06-03 2023-06-30 成都豪杰特科技有限公司 基于人工智能的电子商务的推荐方法、系统、设备和介质
CN116362848B (zh) * 2023-06-03 2023-10-27 广州爱特安为科技股份有限公司 基于人工智能的电子商务的推荐方法、系统、设备和介质

Also Published As

Publication number Publication date
CN108446410A (zh) 2018-08-24
CN108446410B (zh) 2022-05-17

Similar Documents

Publication Publication Date Title
WO2019227560A1 (zh) 信息推荐方法、装置、系统、设备及可读存储介质
CN106302085B (zh) 即时通讯群组的推荐方法及系统
US8468143B1 (en) System and method for directing questions to consultants through profile matching
US10469275B1 (en) Clustering of discussion group participants
US10909601B2 (en) Providing product advice recommendation
TWI549079B (zh) 用於導引內容至一社群網路引擎之使用者的系統及方法
US20170308539A1 (en) Predictive Generation of Search Suggestions
US20150188959A1 (en) Techniques for populating a content stream on a mobile device
US20140214489A1 (en) Methods and systems for facilitating visual feedback and analysis
US9436766B1 (en) Clustering of documents for providing content
US10210429B2 (en) Image based prediction of user demographics
KR101559719B1 (ko) 효과적인 마케팅을 도출하는 자동학습 시스템 및 방법
US20180302761A1 (en) Recommendation System for Multi-party Communication Sessions
CN110059256B (zh) 用于展示信息的系统、方法及装置
US11430049B2 (en) Communication via simulated user
US20170364822A1 (en) Optimizing content distribution using a model
US20160253684A1 (en) Systems and methods of structuring reviews with auto-generated tags
US20210365995A1 (en) Advertisement and reward system based on instant messenger
WO2022247666A1 (zh) 一种内容的处理方法、装置、计算机设备和存储介质
CN110475158B (zh) 视频学习素材的提供方法、装置、电子设备及可读介质
US9569465B2 (en) Image processing
JP2013214133A (ja) 情報処理装置、情報処理方法及びプログラム
US11019379B2 (en) Stage-based content item selection and transmission
US9392041B2 (en) Delivery of two-way interactive content
JP6702625B2 (ja) 情報処理装置、情報処理方法及び情報処理プログラム

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18921259

Country of ref document: EP

Kind code of ref document: A1