Information processing method and related device

Abstract

The invention provides an information processing method and a related device. The method comprises the steps of performing normalization processing on a plurality of pieces of first labeling information used for labeling emotions of a first object for push information to obtain a plurality of pieces of first information; fusing the plurality of pieces of first information to obtain second information; based on first interaction information of the first object for the push information, obtaining emotion prediction information of the first object for the push information and a first weight of the emotion prediction information; based on the first weight, fusing the second information and the emotion prediction information to obtain third information; and determining the pushing frequency of the pushed information based on the third information. According to the information pushing method and device, the accuracy of the pushing frequency of the pushed information can be improved.

Description

Technical Field

[0001] The present application relates to artificial intelligence technology, and in particular to an information processing method and related devices. Background Art

[0002] For users who have developed negative emotions (such as aversion) towards pushed information, high-frequency information push can easily cause interference to users, and the method of manually marking the degree of user aversion to push information has serious uncertainty. Summary of the invention

[0003] The embodiments of the present application provide an information processing method and related devices, which can accurately determine the push frequency of pushed information.

[0004] The technical solution of the embodiment of the present application is implemented as follows:

[0005] The present application provides an information processing method, the method comprising:

[0006] Normalizing the plurality of first annotation information to obtain a plurality of first information, wherein the first annotation information is used to annotate the emotion of the first object with respect to the pushed information;

[0007] fusing the plurality of first information to obtain second information;

[0008] Based on the first interaction information of the first object with respect to the pushed information, obtaining the emotion prediction information of the first object with respect to the pushed information and a first weight of the emotion prediction information;

[0009] Based on the first weight, fusing the second information and the emotion prediction information to obtain third information;

[0010] Based on the third information, a push frequency of the push information is determined.

[0011] The present application provides an information processing device, including:

[0012] A preprocessing module, used for normalizing the plurality of first annotation information to obtain the plurality of first information, wherein the first annotation information is used for annotating the emotion of the first object with respect to the pushed information;

[0013] An information processing module, configured to fuse the plurality of first information to obtain second information; obtain, based on the first interaction information of the first object with respect to the pushed information, emotion prediction information of the first object with respect to the pushed information and a first weight of the emotion prediction information; and fuse the second information with the emotion prediction information based on the first weight to obtain third information;

[0014] The information push module is used to determine the push frequency of the push information based on the third information.

[0015] An embodiment of the present application provides an electronic device, including:

[0016] A memory for storing computer executable instructions;

[0017] The processor is used to implement the information processing method provided in the embodiment of the present application when executing the computer executable instructions stored in the memory.

[0018] An embodiment of the present application provides a computer-readable storage medium storing a computer program or computer-executable instructions for implementing the information processing method provided in the embodiment of the present application when executed by a processor.

[0019] An embodiment of the present application provides a computer program product, including a computer program or computer executable instructions. When the computer program or computer executable instructions are executed by a processor, the information processing method provided by the embodiment of the present application is implemented.

[0020] The embodiments of the present application have the following beneficial effects:

[0021] Based on the first interactive information of the first object for the pushed information, the emotion prediction information of the first object for the pushed information and the first weight of the emotion prediction information are obtained. Based on the first weight, the second information and the emotion prediction information are integrated to obtain the third information. The second information is obtained by processing multiple first annotation information. In this way, the first annotation information of the first object for the pushed information and the emotion prediction information predicted based on the first interactive information of the first object for the pushed information are combined to avoid the uncertainty problem caused by relying solely on manual annotation, thereby improving the accuracy of the third information. Based on the third information, the push frequency of the pushed information is determined. In this way, the accuracy of the push frequency determination of the pushed information is further improved, so that the first object will not have a large emotional fluctuation after receiving the push information. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic diagram of the architecture of the information processing system provided by the embodiment of the present application;

[0023] Figure 2 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application;

[0024] Figure 3A It is a first flow chart of the information processing method provided in the embodiment of the present application;

[0025] Figure 3B is a second flow chart of the information processing method provided in an embodiment of the present application;

[0026] Figure 3C is a third flow chart of the information processing method provided in the embodiment of the present application;

[0027] Figure 3D is a fourth flow chart of the information processing method provided in an embodiment of the present application;

[0028] Figure 3E is a fifth flow chart of the information processing method provided in the embodiment of the present application;

[0029] Figure 3F is a sixth flow chart of the information processing method provided in an embodiment of the present application;

[0030] Figure 3G is a seventh flow chart of the information processing method provided in the embodiment of the present application;

[0031] Figure 3H is an eighth flow chart of the information processing method provided in the embodiment of the present application;

[0032] Fig. 3I is a ninth flow chart of the information processing method provided in an embodiment of the present application;

[0033] Figure 4 This is a flowchart of sending a first SMS based on the degree of disgust provided in an embodiment of the present application;

[0034] Figure 5 is a data recording flow chart provided in an embodiment of the present application;

[0035] Figure 6 This is a second SMS sending flow chart based on the degree of disgust provided in an embodiment of the present application.

[0036] It should be pointed out that the above-mentioned "first" and "second" are only used to distinguish different solutions, and do not represent the degree of superiority or inferiority of the solutions or the priority in the implementation process. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0038] In the following description, reference is made to “some embodiments”, which describe a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0039] In the following description, the terms "first\second\third" involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It can be understood that "first\second\third" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0040] In the embodiments of the present application, the term "module" or "unit" refers to a computer program or a part of a computer program with a predetermined function, and works together with other related parts to achieve a predetermined goal, and can be implemented in whole or in part by using software, hardware (such as processing circuits or memories) or a combination thereof. Similarly, a processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be part of an overall module or unit that includes the function of the module or unit.

[0041] The relevant data collection and processing in the embodiments of this application should be strictly in accordance with the requirements of relevant national laws and regulations when applied in examples, obtain the informed consent or separate consent of the personal information subject, and carry out subsequent data use and processing within the scope of authorization of laws and regulations and the personal information subject.

[0042] Unless otherwise defined, all technical and scientific terms used in the embodiments of the present application have the same meanings as those commonly understood by those skilled in the art. The terms used in the embodiments of the present application are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0043] Before further describing the embodiments of the present application in detail, the nouns and terms involved in the embodiments of the present application are explained. The nouns and terms involved in the embodiments of the present application are subject to the following interpretations.

[0044] 1) The first object is an object used to push information. For example, the first object is an object that has been pushed information or interacted with by a robot or a business person and is ready to be pushed information again.

[0045] 2) The annotation information is information annotated with respect to the emotion of the first object with respect to the pushed information. The manually annotated annotation information is annotated with respect to the emotion of the first object with respect to the pushed information after each interaction with the first object.

[0046] 3) Emotion prediction information is obtained by performing emotion prediction on the interaction information of the first object using the emotion prediction model. The emotion prediction model predicts the emotion prediction information of the first object by learning the emotion features in the interaction information of the first object.

[0047] 4) The interaction information of the first object is the information of the first object during the interaction (such as voice, text, image, video, etc.) or the operation information of the first object on the product (ie, application) during or after the interaction (such as information on logging in or using the product).

[0048] In the related art, it is impossible to determine the emotion of the first object towards the pushed information based on the actual communication situation of the first object. The emotion of the first object towards the pushed information is only manually labeled, which reduces the accuracy of determining the push frequency of the push information. To address the above problems, the embodiments of the present application provide an information processing method, device, electronic device, computer-readable storage medium and computer program product to improve the accuracy of determining the push frequency of push information.

[0049] The information processing method recorded in the embodiments of the present application can be applied to various fields, such as determining the recommendation frequency of recommended information pushed to users in information recommendation scenarios, determining the teaching frequency of teaching information taught to students in online classroom scenarios, and other types of fields. That is, the information processing method in the embodiments of the present application is not limited to a certain field.

[0050] The following describes an exemplary application of the electronic device provided in the embodiment of the present application. The device provided in the embodiment of the present application can be implemented as a terminal or a server. The following describes an exemplary application when the device is implemented as a server.

[0051] See also Figure 1 , Figure 1 It is a schematic diagram of the architecture of the information processing system 100 provided in an embodiment of the present application. To support an information processing application, a terminal (terminal 400 is shown as an example) is connected to a server 200 via a network 300. The network 300 may be a wide area network or a local area network, or a combination of the two.

[0052] The terminal 400 is used to send multiple first annotation information for marking the emotions of the first object with respect to the pushed information to the server 200 through the network 300, and the server 200 is used to normalize the multiple first annotation information to obtain multiple first information; fuse the multiple first information to obtain second information; based on the first interaction information of the first object with respect to the pushed information, obtain the emotion prediction information of the first object with respect to the pushed information and the first weight of the emotion prediction information; based on the first weight, fuse the second information and the emotion prediction information to obtain third information; based on the third information, determine the push frequency of the push information, and return the push information and the push frequency of the push information to the terminal 400. The terminal 400 displays the push information and the push frequency of the push information through the graphical interface 410, and pushes the push information to the first object according to the push frequency.

[0053] Next, an example of information processing performed by terminal 400 will be described.

[0054] In some embodiments, the terminal 400 can independently complete information processing tasks. For example, the terminal 400 is used to normalize multiple first annotation information used to annotate the emotions of the first object with respect to the pushed information to obtain multiple first information; fuse multiple first information to obtain second information; based on the first interaction information of the first object with respect to the pushed information, obtain the emotion prediction information of the first object with respect to the pushed information and the first weight of the emotion prediction information; based on the first weight, fuse the second information and the emotion prediction information to obtain third information; based on the third information, determine the push frequency of the push information, display the push information and the push frequency of the push information through the graphical interface 410, and push the push information to the first object according to the push frequency.

[0055] In an implementation scenario, a server or terminal can determine the recommendation frequency of recommended information pushed to users in an information recommendation scenario. Taking disgust as an example, multiple annotation information used to mark the user's disgust with the recommended information are normalized to obtain multiple first disgust levels; multiple first disgust levels are integrated to obtain second disgust levels; based on the user's interaction information with the recommended information, the user's predicted disgust level with the recommended information and a first weight of the predicted disgust level are obtained; based on the first weight, the second disgust level and the predicted disgust level are integrated to obtain a normalized disgust level; based on the third disgust level, the recommendation frequency of the recommended information is determined, and the recommended information is pushed to the user according to the recommendation frequency.

[0056] In an implementation scenario, a server or terminal can determine the teaching frequency of teaching information taught to students in an online classroom scenario. Taking disgust as an example, multiple annotation information used to mark students' disgust for teaching information are normalized to obtain multiple first disgust levels; multiple first disgust levels are integrated to obtain second disgust levels; based on students' interactive information regarding the teaching information, students' predicted disgust levels for the teaching information and a first weight for the predicted disgust levels are obtained; based on the first weight, the second disgust levels and the predicted disgust levels are integrated to obtain a third disgust level; based on the third disgust level, the teaching frequency of the teaching information is determined, and the teaching information is taught to students according to the teaching frequency, where the teaching frequency can be 5 math classes a day.

[0057] In some embodiments, server 200 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), as well as big data and artificial intelligence platforms.

[0058] The terminal 400 may be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, an intelligent voice interaction device, a smart home appliance, a vehicle terminal, an aircraft, etc., but is not limited thereto. The terminal and the server may be directly or indirectly connected via wired or wireless communication, which is not limited in the embodiments of the present application.

[0059] See also Figure 2 , Figure 2 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application, Figure 2 The electronic device 500 shown may be Figure 1 In the terminal 400 or server 200, the electronic device 500 includes: at least one processor 510, a memory 550, and at least one network interface 520. The various components in the server 200 are coupled together through a bus system 540. It can be understood that the bus system 540 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 540 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, in Figure 2 Various buses are labeled as bus system 540 .

[0060] The processor 510 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., where the general-purpose processor can be a microprocessor or any conventional processor, etc.

[0061] The user interface 530 includes one or more output devices 531 that enable presentation of media content, including one or more speakers and / or one or more visual display screens. The user interface 530 also includes one or more input devices 532, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls;

[0062] In some embodiments, when the terminal 400 independently completes the information processing task, the server 200 provided in the embodiment of the present application does not include the user interface 530 .

[0063] The memory 550 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state memory, a hard disk drive, an optical disk drive, etc. The memory 550 may optionally include one or more storage devices that are physically located away from the processor 510.

[0064] The memory 550 includes a volatile memory or a non-volatile memory, and may also include both volatile and non-volatile memories. The non-volatile memory may be a read-only memory (ROM), and the volatile memory may be a random access memory (RAM). The memory 550 described in the embodiments of the present application is intended to include any suitable type of memory.

[0065] In some embodiments, the memory 550 can store data to support various operations, examples of which include programs, modules, and data structures, or a subset or superset thereof, as exemplarily described below.

[0066] Operating system 551, including system programs for processing various basic system services and performing hardware-related tasks, such as framework layer, core library layer, driver layer, etc., for implementing various basic services and processing hardware-based tasks;

[0067] A network communication module 552, for reaching other computing devices via one or more (wired or wireless) network interfaces 520, exemplary network interfaces 520 include: Bluetooth, Wireless Compatibility Certification (WiFi), and Universal Serial Bus (USB);

[0068] a presentation module 553 for enabling presentation of information via one or more output devices 531 (e.g., display screen, speaker, etc.) associated with the user interface 530 (e.g., a user interface for operating peripherals and displaying content and information);

[0069] In some embodiments, when the information processing task is completed independently by the terminal 400, the server 200 provided in the embodiment of the present application may not include the presentation module 553.

[0070] The input processing module 554 is used to detect one or more user inputs or interactions from one of the one or more input devices 532 and translate the detected inputs or interactions; in some embodiments, when the embodiment independently completes the information processing task by the terminal 400, the server 200 provided in the embodiment of the present application may not include the presentation module 553.

[0071] In some embodiments, the device provided in the embodiments of the present application can be implemented in software. Figure 2The information processing device 555 stored in the memory 550 is shown, which can be software in the form of a program and a plug-in, etc., including the following software modules: a pre-processing module 5551, an information processing module 5552, and an information push module 5553. These modules are logical, so they can be arbitrarily combined or further split according to the functions implemented. The functions of each module will be described below.

[0072] It should be noted that in the information processing example below, those skilled in the art can apply the information processing method provided in the embodiment of the present application to determine the pushing frequency of push information based on their understanding of the following.

[0073] See also Figure 3A , Figure 3A is a first flow chart of the information processing method provided in the embodiment of the present application, which will be combined with Figure 3A The steps shown are used for explanation. The information processing method provided in the embodiments of the present application can be implemented by a server or a terminal alone, or by a server and a terminal in collaboration. The following will take the collaborative implementation by a server and a terminal as an example for explanation.

[0074] In step 101, a plurality of first annotation information is normalized to obtain a plurality of first information.

[0075] The first annotation information is used to annotate the emotion of the first object with respect to the pushed information.

[0076] Here, the first annotation information is information that annotates the emotion of the first object with respect to the pushed information after the first object completes the interaction. The embodiment of the present application does not limit the object that interacts with the first object. The object that interacts with the first object can be a robot, a business person who manually pushes information, etc. The pushed information is the first pushed content that has been historically pushed to the first object. The embodiment of the present application does not limit the first pushed content. The first pushed content can be content related to the product (application), such as the product name, product function, etc. The embodiment of the present application does not limit the emotion. The emotion can be disgust, joy, etc. The first annotation information is used to characterize the emotional degree of the first object with respect to the pushed information. The emotion of the object towards the pushed information is described as disgust, and the first annotation information may be the disgust degree of the first object towards the push information, and the disgust degree may be a real number of 0-10, wherein 0 indicates that the first object does not dislike the push information pushed to the first object, and 10 indicates that the first object particularly dislikes the push information pushed to the first object; the first object is an interacted object, and after each interaction, the first annotation information is annotated for the push information of the first object. The embodiment of the present application does not limit the interaction method, and the interaction method may be voice communication, text communication, etc.; the emotion of the first object towards the push information may be annotated through multiple electronic devices (such as robots), and one first annotation information corresponds to one first information.

[0077] Before step 101, the first annotation information can be determined in the following ways: performing emotion recognition on the interactive information of the first object, and using the result of emotion recognition as the first annotation information of the first object for the pushed information; or, performing emotion prompt word extraction on the interactive information of the first object, and determining the information corresponding to the extracted emotion prompt words as the first annotation information for the pushed information.

[0078] For example, emotion recognition is performed on the interactive information of the first object (such as "I am very happy to communicate with you"), and the result of emotion recognition (such as 0.2) is used as the first annotation information of the first object for the push information; emotion prompt words are extracted from the interactive information of the first object, and the information corresponding to the emotion prompt word "happy" is 0.2, and the information corresponding to the emotion prompt word "unhappy" is 0.8 is used as an example to illustrate, and the information corresponding to the extracted emotion prompt word ("happy") (such as 0.2) is determined as the first annotation information of the first object for the push information.

[0079] Continuing from the above embodiment, determining the information corresponding to the extracted emotion prompt word can be achieved in the following way: querying the third mapping table based on the emotion prompt word, and determining the queried candidate information corresponding to the emotion prompt word as the information corresponding to the extracted emotion prompt word, wherein the third mapping table includes the correspondence between different candidate emotion prompt words and different candidate information.

[0080] It should be noted that each candidate emotion prompt word corresponds to its own candidate information, and the information corresponding to the extracted emotion prompt word can be obtained by weighted summing the information corresponding to each emotion prompt word. The third mapping table is a data structure for storing the correspondence between candidate emotion prompt words and candidate information. The embodiment of the present application does not limit the third mapping table. The data structure of the first mapping table is similar to that of the first mapping table, and will not be repeated here.

[0081] For example, determine candidate information A of candidate emotion prompt word A, determine candidate information B of candidate emotion prompt word B, and determine the queried candidate information (such as candidate information A) corresponding to the candidate emotion prompt word (such as candidate emotion prompt word A) as the first annotation information of the first object for the push information.

[0082] Continuing with the above example, determine the information of emotion prompt word A (such as 0.2), determine the information of emotion prompt word B (such as 0.4), perform weighted summation on the information corresponding to multiple emotion prompt words (such as emotion prompt word A and emotion prompt word B), and obtain the information of the emotion prompt word (such as 0.3).

[0083] In some embodiments, step 101 can be implemented in the following manner: determine the maximum annotation information and the minimum annotation information among multiple first annotation information; perform the following processing for each first annotation information to determine the first difference between the first annotation information and the minimum annotation information; determine the second difference between the maximum annotation information and the minimum annotation information, and determine the ratio of the first difference to the second difference as the normalized annotation information.

[0084] Continuing from the above embodiment, determining the maximum annotation information and the minimum annotation information among multiple first annotation information can be achieved by: sorting multiple first annotation information to obtain a sorting result; and determining some of the first annotation information in the sorting result as the maximum annotation information and the minimum annotation information.

[0085] For example, multiple first annotation information (such as [0.2, 0.5, 0.8, 0.3]) are sorted in ascending order to obtain an ascending sorting result (such as [0.2, 0.3, 0.5, 0.8]), the first first annotation information (such as 0.2) in the ascending sorting result is determined as the minimum annotation information, and the last first annotation information (such as 0.8) in the ascending sorting result is determined as the maximum annotation information, and multiple first annotation information (such as [0.2, 0.5, 0.8, 0.3]) are sorted in descending order to obtain a descending sorting result (such as [0.8, 0.5, 0.3, 0.2]), the first first annotation information (such as 0.8) in the descending sorting result is determined as the maximum annotation information, and the last first annotation information (such as 0.2) in the descending sorting result is determined as the minimum annotation information.

[0086] Continuing with the above example, determine the maximum annotation information (such as 0.8) and the minimum annotation information (such as 0.2) among multiple first annotation information (such as [0.2, 0.5, 0.8, 0.3]); perform the following processing for each first annotation information: determine the first difference (such as 0.3) between the first annotation information (such as 0.5) and the minimum annotation information; determine the second difference (such as 0.6) between the maximum annotation information and the minimum annotation information, and determine the ratio of the first difference to the second difference as the normalized annotation information (such as 0.5).

[0087] In some embodiments, see Figure 3B , Figure 3B is a second flow chart of the information processing method provided in the embodiment of the present application, in step Figure 3A Before showing 101, execute Figure 3B Steps 201 to 205 are described in detail below.

[0088] In step 201, a plurality of second annotation information is normalized to obtain a plurality of fourth information.

[0089] Here, the second annotation information is information that annotates the second object's emotion toward the pushed information after the second object completes the interaction. In an embodiment of the present application, the emotion of the second object toward the pushed information can be annotated by an electronic device (such as a robot).

[0090] It should be noted that step 201 is similar to step 101 and will not be described in detail here.

[0091] In step 202, multiple fourth information are fused to obtain fifth information.

[0092] It should be noted that step 202 is similar to the following step 102, see the description of step 102 for details.

[0093] In step 203, first object information of the first object and second object information of the second object are determined.

[0094] The second object is an object including the fifth information.

[0095] Here, the second object is an object other than the first object to which the message has been pushed, the first object information is the object information of the target object, and the second object information is the object information of other objects.

[0096] In some embodiments, the object information may be determined in the following manner: determining an object identifier of the object, and determining the object information corresponding to the object based on the object identifier.

[0097] It should be noted that the object identifier of an object is used to identify the object. The object includes a first object and a second object. Each object has a unique object identifier. The object information of the object may include the object name, object identifier, contact information, etc. of the target object, where the same object name may correspond to different object identifiers.

[0098] For example, determine the object identification (such as "2020**3021") of an object (such as object A), and based on the object identification, determine the object information corresponding to the object (such as the object identification is "2020**3021", the object name is "Object A", and the contact information is "**").

[0099] In step 204, the similarity between the first object information and the second object information is determined.

[0100] In some embodiments, step 204 can be implemented by determining the edit distance (e.g., 1) between the first object information (e.g., "Unknown") and the second object information (e.g., "Anonymous"), determining the maximum length (e.g., 3) between the length of the first object information (e.g., 2) and the length of the second object information (e.g., 3), determining the ratio of the edit distance to the maximum length (e.g., 3), and determining the difference between the preset similarity (e.g., 1) and the ratio (e.g., 0.33) as the similarity between the first object information and the second object information (e.g., 0.6).

[0101] Among them, the edit distance is used to measure how many single-character edits (insertion, deletion or replacement) are required between the first object information and the second object information to convert the first object information into the second object information; the first object information and the second object information are encoded to obtain a first object information encoding vector and a second object information encoding vector, respectively; based on the encoded first object information encoding vector and the encoded second object information encoding vector, the similarity (such as 0.4) between the first object information and the second object information is determined. The embodiment of the present application does not limit the similarity. The similarity can be the cosine similarity, Euclidean distance, etc. between the encoded first object information encoding vector and the encoded second object information encoding vector.

[0102] In step 205, when the similarity is less than the first similarity threshold, a step of normalizing the plurality of first annotation information to obtain the plurality of first information is performed.

[0103] In some embodiments, when the similarity is greater than or equal to a first similarity threshold, the fifth information of the second object is determined as the second information of the first object for fusion with the emotion prediction information.

[0104] It should be noted that the second similarity threshold may be the same as or different from the first similarity threshold.

[0105] Through the embodiments of the present application, the second information of the first object is determined based on the fifth information obtained by processing the second annotation information of the second object, which increases the correlation between objects and improves the accuracy of determining the second information of the first object.

[0106] In step 102, multiple first information are fused to obtain second information.

[0107] In some embodiments, step 102 may be implemented in the following manner: performing weighted summation on multiple pieces of first information to obtain the second information of the first object.

[0108] For example, the first information A is 0.4, and the first information B is 0.8. When the first information A and the first information B have the same weight, that is, the weights of the first information A and the first information B are both 0.5, the second information of the first object is 0.6. When the first information A and the first information B have different weights, where the weight of the first information A is 0.6 and the weight of the first information B is 0.4, the second information of the first object is 0.56.

[0109] Through the embodiments of the present application, data processing is performed on multiple first information to avoid determining the second information through a single first information, thereby increasing the credibility of the second information and improving the accuracy of the second information.

[0110] In step 103, based on the first interaction information of the first object with respect to the pushed information, the emotion prediction information of the first object with respect to the pushed information and the first weight of the emotion prediction information are obtained.

[0111] In some embodiments, see Figure 3C , Figure 3C is a third flow chart of the information processing method provided in the embodiment of the present application. Figure 3A Before step 103 shown, execute Figure 3C Steps 301 to 302 are implemented as described below.

[0112] In step 301, the similarity between the first interaction information and the second interaction information of the third object with respect to the push information is determined, and the third object is an object after the emotion of the push information is predicted.

[0113] The third object is an object other than the first object with which the interaction has been performed.

[0114] It should be noted that step 301 is similar to step 201 and will not be described in detail here.

[0115] In step 302, when the similarity is less than the second similarity threshold, it is determined to execute the step of obtaining the emotion prediction information of the first object for the pushed information and the first weight of the emotion prediction information based on the first interaction information of the first object for the pushed information.

[0116] For example, when the similarity (eg, 0.3) is less than the second similarity threshold (eg, 0.4), step 103 is performed.

[0117] Through the embodiment of the present application, when the similarity between the interaction information of the first object and the interaction information of the second object is less than the second similarity threshold, step 103 is performed, thereby avoiding determining the third information only by the first object and improving the robustness of the third information.

[0118] In some embodiments, see Figure 3D , Figure 3D is a third flow chart of the information processing method provided in the embodiment of the present application, for Figure 3A Step 103 shown can be performed by Figure 3D The step 1031 is implemented as described in detail below.

[0119] In step 1031, emotion prediction processing is performed on the first interaction information of the first object with respect to the pushed information to obtain emotion prediction information of the first object with respect to the pushed information and a first weight of the emotion prediction information.

[0120] Here, the embodiment of the present application does not impose any restriction on the first interaction information of the first object for the push information. The first interaction information of the first object for the push information may be the information of the first object during the interaction (such as voice, text, image, video, etc.), or it may be the operation information of the first object for the product (i.e., application) during or after the interaction, etc., wherein the operation information for the product (i.e., application) may be login information or usage information when logging in or using the product, and the first weight is used to measure the importance of the emotion prediction information.

[0121] In some embodiments, see Figure 3E , Figure 3E is a fifth flow chart of the information processing method provided in the embodiment of the present application, for Figure 3D Step 1031 shown may be performed by Figure 3E Step 10311 is implemented as described in detail below.

[0122] In step 10311, a sentiment prediction model is determined, and sentiment prediction is performed on the first interaction information of the first object for the pushed information through the sentiment prediction model to obtain sentiment prediction information of the first object for the pushed information and a first weight corresponding to the sentiment prediction information.

[0123] Here, the embodiments of the present application do not limit the method for determining the emotion prediction model. For example, the emotion prediction model can be obtained locally, or the emotion prediction model can be called from the cloud, etc.

[0124] In some embodiments, the above-mentioned "performing emotion prediction on the first interaction information of the first object for the pushed information through the emotion prediction model, and obtaining emotion prediction information of the first object for the pushed information and the first weight corresponding to the emotion prediction information" can be achieved in the following way: encoding the first interaction information of the first object for the pushed information through the emotion prediction model to obtain the encoded features; decoding the encoded features through the emotion prediction model to obtain the emotion prediction information of the first object for the pushed information and the first weight corresponding to the emotion prediction information.

[0125] It should be noted that the embodiment of the present application does not limit the model structure of the emotion prediction model. The emotion prediction model is taken as an example to include an encoder, a first decoder and a second decoder. The encoder of the emotion prediction model encodes the first interactive information of the first object for the pushed information to obtain the encoded features, the first decoder of the emotion prediction model decodes the encoded features to obtain the emotion prediction information of the first object for the pushed information, and the second decoder of the emotion prediction model decodes the encoded features to obtain the first weight corresponding to the emotion prediction information, wherein the encoder includes multiple encoding layers, the input of the first encoding layer is the first interactive information of the first object for the pushed information, and the input of each encoding layer except the first layer is the output of the previous encoding layer. The first coding layer is used to extract coding features of the first interactive information of the first object for the pushed information, the extracted coding features are normalized to obtain the output features corresponding to the first coding layer, and the output features of the first coding layer are extracted through the second coding layer. Multiple iterations are performed in sequence to determine the output features of the last coding layer as coding features; the first decoder and the second decoder both include multiple layers of decoding layers, the input of the first decoding layer is the coding feature, the input of each decoding layer except the first layer is the output of the previous decoding layer, and the output obtained by the last decoding layer is mapped to a real number greater than or equal to 0 and less than or equal to 1, so as to obtain the emotion prediction information of the first object for the pushed information and the first weight corresponding to the emotion prediction information, respectively.

[0126] For example, the first interaction information (such as "I like this product very much") of the first object in response to the push information (such as "Do you like this product") is encoded through the sentiment prediction model to obtain the encoded features (such as [0.2, 0.3, 0.1, 0.6, 0.1, 0.4, 0.3, 0.7, 0.1, 0.6, 0.1, 0.4]); the encoded features are decoded through the sentiment prediction model to obtain the sentiment prediction information of the first object in response to the push information (such as 0.4) and the first weight corresponding to the sentiment prediction information (such as 0.4).

[0127] Through the embodiments of the present application, a sentiment prediction model is used to perform sentiment prediction on the first interactive information of the first object with respect to the pushed information, and the sentiment prediction information and the first weight of the sentiment prediction information are determined at the same time. By estimating the importance of the sentiment prediction information, the accuracy of determining the third information is improved.

[0128] In step 104, based on the first weight, the second information and the emotion prediction information are fused to obtain the third information.

[0129] In some embodiments, see Figure 3F , Figure 3F is a sixth flow chart of the information processing method provided in the embodiment of the present application, for Figure 3A Step 104 shown may be performed by Figure 3F Steps 1041 to 1042 are implemented as described in detail below.

[0130] In step 1041, based on the first weight, a second weight of the second information is determined.

[0131] In some embodiments, step 1041 may be implemented in the following manner: determining the difference between the preset weight and the first weight as the second weight of the second information.

[0132] For example, the difference between the preset weight (eg, 1) and the first weight (eg, 0.8) is determined as the second weight (eg, 0.2) of the second information.

[0133] In step 1042, based on the first weight and the second weight, the second information and the emotion prediction information are weighted and summed to obtain the third information.

[0134] In some embodiments, step 1042 can be implemented in the following manner: determine the product of the second information and the second weight as the first emotional information; determine the product of the emotional prediction information and the first weight as the second emotional information, add the first emotional information and the second emotional information to obtain the third information.

[0135] For example, the product of the second information (such as 0.4) and the second weight (such as 0.2) is determined as the first emotional information (such as 0.08), the product of the emotional prediction information (such as 0.8) and the first weight (such as 0.8) is determined as the second emotional information (such as 0.64), and the first emotional information and the second emotional information are added to obtain the third information (such as 0.72).

[0136] In step 105, based on the third information, the push frequency of the push information is determined.

[0137] In some embodiments, see Figure 3G , Figure 3G is a seventh flow chart of the information processing method provided in the embodiment of the present application, Figure 3A Step 105 shown can be performed by Figure 3G Steps 1051 to 1052 are implemented as described in detail below.

[0138] In step 1051 , a normalized interval matching the third information is determined.

[0139] In some embodiments, step 1051 can be implemented in the following manner: determine multiple candidate normalization intervals, wherein a candidate normalization interval has a lower bound and an upper bound, and the lower bound is smaller than the upper bound; when the third information is greater than the lower bound of any candidate normalization interval and the third information is less than or equal to the upper bound of any candidate normalization interval, determine any candidate normalization interval as the normalization interval to which the third information belongs.

[0140] Here, the candidate normalized interval is a continuous or discrete numerical range. The embodiment of the present application does not limit the candidate normalized interval. The candidate normalized interval can be a continuous interval, a discrete interval, etc., where, taking the candidate normalized interval as an example, according to the type of endpoints, the candidate normalized interval can be a closed interval, an open interval, or a half-open and half-closed interval; taking the candidate normalized interval as an example, the candidate normalized interval is a discrete interval, and the candidate normalized interval is a subset of the integer set. The lower bound of the candidate normalized interval is the minimum value of the numerical value contained in the candidate normalized interval, and the upper bound of the candidate normalized interval is the maximum value of the numerical value contained in the candidate normalized interval.

[0141] For example, candidate normalization interval A has a lower bound (such as 0.3) and an upper bound (such as 0.4), and candidate normalization interval B has a lower bound (such as 0.4) and an upper bound (such as 0.5), where the upper bound is greater than or equal to the lower bound. When the third information (such as 0.35) is greater than or equal to the lower bound of candidate normalization interval A and less than the upper bound of candidate normalization interval A, any candidate normalization interval (such as candidate normalization interval A) is determined as the normalization interval to which the third information belongs.

[0142] In step 1052, based on the normalized interval, the second mapping table is queried, and the candidate push frequency corresponding to the normalized interval is determined as the push frequency of the push information.

[0143] The second mapping table is used to represent the corresponding relationship between the normalized interval and the candidate push frequency.

[0144] Here, the embodiment of the present application does not impose any restrictions on the first object. The object information of the first object includes the object name, object identifier, contact information, etc. of the first object. In addition, the embodiment of the present application does not impose any restrictions on the correspondence between different normalized intervals and different candidate push frequencies. The correspondence can be that multiple normalized intervals correspond to one candidate push frequency, or that each normalized interval corresponds to its own candidate push frequency.

[0145] It should be noted that the second mapping table is a data structure used to store the correspondence between the normalized interval and the candidate push frequency. The embodiment of the present application does not limit the second mapping table. The data structure of the second mapping table is similar to the data structure of the first mapping table, which will not be repeated here. The push information is information used to push to the first object. The embodiment of the present application does not limit the push information. The push information can be the first push content historically pushed to the first object, or the second push content used to push to the first object, so that the first object can perform relevant operations on the product (ie, application) based on the recommended information. The embodiment of the present application does not limit the push information. The push information can be product information, holiday visit information, etc. The embodiment of the present application does not limit the second push content. The second push content can be content related to the product (application), such as the product name, product function, etc. The second push content can be consistent with the first push content, or it can be different from the first push content.

[0146] It should be noted that the push method includes a push method for pushing information according to the push frequency. Pushing information to the first object according to the push frequency is one of the push methods. The embodiment of the present application does not limit the push method. The push method includes a push method for pushing information according to the push frequency, a push method for pushing information according to a preset push time (for example, 12 noon every day), etc. The medium for pushing information can be electronic media such as text messages and telephone calls.

[0147] For example, the lower bound of normalized interval A is 0, and the upper bound is 0.4. The push frequency of normalized interval A is based on 5 information pushes to the first object every day. The lower bound of normalized interval B is 0.8, and the upper bound is 1. The push frequency of normalized interval B is no information push. The push frequency of pushed information (such as no information push) is determined based on the normalized interval (normalized interval B) matched by the third information (such as 0.9).

[0148] In some embodiments, determining the push frequency of push information based on the normalized interval can also be achieved in the following way: determining the push method of push information based on the normalized interval, and when the push method is to push information according to the push frequency, executing step 1052.

[0149] Continuing from the above embodiment, based on the normalized interval, determining the push method of the push information can be achieved in the following manner: querying the fourth mapping table based on the normalized interval, and determining the candidate push method corresponding to the normalized interval as the push method, wherein the fourth mapping table includes the correspondence between the normalized interval and different candidate push methods.

[0150] It should be noted that the fourth mapping table is a data structure used to store the correspondence between different normalized intervals and different candidate push methods. The embodiment of the present application does not limit the correspondence between different normalized intervals and different candidate push methods. The correspondence can be that multiple normalized intervals correspond to one candidate push method, or each normalized interval corresponds to its own candidate push method. The embodiment of the present application does not limit the fourth mapping table. The data structure of the fourth mapping table is similar to that of the first mapping table, and will not be repeated here.

[0151] For example, candidate push mode A for normalized interval A is determined, candidate push mode B for normalized interval B is determined, and the candidate push mode (such as push mode A) corresponding to the normalized interval (such as normalized interval A) is determined as the push mode.

[0152] In the following embodiment, based on the normalized interval, determining the push mode of the push information can be implemented in the following manner: determining a label of the normalized interval; and determining the push mode of the push information based on the label.

[0153] For example, the lower bound of normalized interval A is 0, the upper bound is 0.4, the label of normalized interval A is low-offensive type, the lower bound of normalized interval B is 0.8, the upper bound is 1, the label of normalized interval B is high-offensive type, the push method of low-offensive type is to push information according to the push frequency, and the push method of high-offensive type is to push information according to the random push method. The label of the normalized interval (normalized interval B) is determined (such as the high-offensive type); based on the label, the push method of the pushed information is determined (such as pushing information according to the random push method).

[0154] In some embodiments, the push information is the second push content for being pushed to the first object. Determining the second push content for being pushed to the first object can be achieved in the following manner: querying the third mapping table based on the normalized interval, and determining the candidate push content corresponding to the normalized interval as the third push content, wherein the third mapping table includes the correspondence between different normalized intervals and different candidate push content; integrating the third push content and the object information of the first object to obtain the second push content for being pushed to the first object.

[0155] It should be noted that the third mapping table is a data structure used to store the correspondence between the normalized interval and the candidate push content. The embodiment of the present application does not limit the third mapping table. The data structure of the third mapping table is similar to the data structure of the first mapping table, and will not be repeated here. The embodiment of the present application does not limit the method of integrating the third push content and the object information of the first object to obtain the second push content for pushing to the first object. The method of integrating the third push content and the object information of the first object can be splicing, replacement, insertion, etc.

[0156] For example, candidate push content A of normalized interval A is determined, candidate push content B of normalized interval B is determined, and the queried candidate push content (such as candidate push content A) corresponding to the target normalized interval (such as normalized interval A) is determined as the third push content.

[0157] Continuing with the above example, the object information of the first object is determined (such as the object name "Unnamed"), the third push content is determined (such as "Product A is pushed to you below, and the specific usage of the pushed product is explained in detail"). The third push content and the object information of the first object are used to obtain the second push content for pushing to the first object (such as "Unnamed: Product A is pushed to you below, and the specific usage of the pushed product is explained in detail").

[0158] In some embodiments, see Figure 3H , Figure 3H is an eighth flow chart of the information processing method provided in the embodiment of the present application, Figure 3A After step 105 shown, execute Figure 3H Step 401 is described in detail below.

[0159] In step 401, push information is pushed to a first object based on a push frequency.

[0160] In some embodiments, see Fig. 3I , Fig. 3I is a ninth flow chart of the information processing method provided in the embodiment of the present application, Figure 3H Step 401 shown can be performed by Figure 3HSteps 4011 to 4013 are implemented as described in detail below.

[0161] In step 4011, based on the push frequency, the total number of times the push information is pushed to the first object is determined.

[0162] Here, push frequency refers to the frequency or number of times push information is pushed within a certain period of time (such as the number of pushes per hour, day or week); the total number of pushes refers to the total number of times push information can be pushed; push information is information used to be pushed to the first object so that the first object can perform relevant operations on the product (i.e., application) based on the recommended information.

[0163] In some embodiments, based on the push frequency, determining the total number of push times of push information to the first object can be achieved in the following way: determining the product of at least one of the push frequencies and a preset time as the total number of push times; or, determining the number of push times corresponding to any push frequency, and determining one of at least one of the push times as the total number of push times, wherein the push frequency is the ratio of the number of push times to the time.

[0164] For example, determine the push frequency (such as 5 times every 3 days, 4 times every 2 days), and take the product of the minimum push frequency (such as 4 times every 2 days, that is, 2 times every day) and the preset time (such as 1 day) as the total number of pushes (such as 2 times); or, determine the number of pushes corresponding to any push frequency (such as the number of pushes for the push frequency of "5 times every 3 days" is 5 times, and the number of pushes for the push frequency of "4 times every 2 days" is 4 times), and determine the minimum number of pushes (such as 4 times) as the total number of pushes.

[0165] In step 4012, based on the total number of pushes and the number of times the push information has been pushed to the first object, the number of times that the push information can be pushed to the first object is determined.

[0166] Here, the number of pushed times is the cumulative number of all push information sent before the query time point, wherein the query time point is the time point of the operation of determining the push frequency of the push information based on the third information; the number of pushable times is the number of push information that can be pushed to the first object, that is, the difference between the total number of pushes and the number of pushed times;

[0167] In some embodiments, the validity period of the push frequency is determined; the receiving time point of the information push request is determined, and if the receiving time point is not within the validity period, the number of times the push information has been pushed to the first object is updated, and the validity period of the push frequency is updated; in response to the information push request, based on the total number of push times of the push information to the first object and the updated number of push times of the push information to the first object, the number of times that the push information can be pushed to the first object is determined.

[0168] In the example, the validity period of the push frequency is from March 6 to March 8, 2023, and the time point of receiving the information push request is March 9, 2023. Since the receiving time point is not within the validity period, the number of pushes of the push information to the first object (such as 3) is updated to obtain the updated number of pushes of the target interactive object (such as 0), and the validity period of the push frequency is updated (such as March 9 to March 11, 2023). In response to the information push request, based on the total number of pushes of the push information to the first object (such as 5) and the updated number of pushes of the push information to the first object (such as 0), the number of pushes that can be pushed to the first object is determined (such as 5).

[0169] In some embodiments, in response to a push frequency update request, the push frequency is updated to obtain an updated push frequency.

[0170] In step 4013, the push information is pushed to the first object according to the number of possible push times.

[0171] In some embodiments, step 4013 can be implemented in the following manner: determine a normalized interval that matches the third information; query the first mapping table based on the normalized interval, and determine the candidate push channel corresponding to the normalized interval as the push channel, wherein the first mapping table is used to represent the correspondence between the normalized interval and the candidate push channel; push the push information to the first object according to the number of pushable times and the push channel.

[0172] It should be noted that the above step of “determining a normalized interval matching the third information” is similar to step 1051 and will not be repeated here.

[0173] Continuing from the above embodiment, based on the normalized interval, determining the push method of the push information can be achieved in the following way: querying the first mapping table based on the normalized interval, and determining the candidate push channel corresponding to the normalized interval as the push channel, wherein the first mapping table includes the correspondence between the normalized interval and different candidate push channels.

[0174] Here, the embodiment of the present application does not limit the correspondence between different normalized intervals and different candidate push channels. The correspondence can be that multiple normalized intervals correspond to one candidate push channel, or each normalized interval corresponds to its own candidate push channel.

[0175] It should be noted that the push channel is the medium for pushing push information. The embodiment of the present application does not limit the medium for pushing push information. The medium can be a public account, text message, etc. The first mapping table is a data structure used to store the correspondence between normalized intervals and candidate push channels. The embodiment of the present application does not limit the first mapping table. The first mapping table can be a data structure implemented as a hash table, a data structure implemented as an array, etc., which is used to quickly perform search, insertion and deletion operations on the correspondence between normalized intervals and candidate push channels, wherein the hash table uses a hash function to calculate the index value, and maps the input normalized interval to the position in the hash table to access the candidate push method.

[0176] For example, candidate push channel A of normalized interval A is determined, candidate push channel B of normalized interval B is determined, and the candidate push channel (such as push mode A) corresponding to the normalized interval (such as normalized interval A) is determined as the push channel.

[0177] Continuing with the above example, when any information in the normalized interval (such as 0.34) is less than the normalized threshold (such as 0.4), information can be pushed through the SMS push channel to increase the probability that the first object receives the push information. When any information in the normalized interval (such as 0.84) is greater than or equal to the normalized threshold, it can be pushed through the public account push channel to increase the probability that the first object will perform operations on the product (i.e., application) after receiving the push information.

[0178] Through the embodiments of the present application, the number of times that can be sent is determined based on the push frequency, that is, the number of times the push information is pushed to the first object for interaction is limited. It can also be ensured that after waiting for a period of time, the information can continue to be pushed to the first object, thereby avoiding the push information from failing to reach the first object. At the same time, the push method of pushing the push information to the first object is regulated based on the third information, thereby avoiding a single push method from reducing the user experience and improving the flexibility of information push. In addition, the labeling information of the first object and the emotion prediction information obtained based on the prediction of the interaction information of the first object for the push information are integrated, thereby avoiding the uncertainty caused by relying solely on manual labeling, improving the accuracy of determining the third information, and thus improving the accuracy of the push frequency.

[0179] The following is an explanation of an exemplary application of the information processing method provided in an embodiment of the present application in a practical application scenario.

[0180] In the related art, a fixed number of times a user can send messages is limited. Once the number of times is exhausted, the latest message cannot reach the user. In addition, it is impossible to flexibly and automatically adjust and switch the frequency of SMS sending according to the actual communication situation between the agent and the user.

[0181] In order to solve the above problems, an embodiment of the present application proposes a method for sending text messages based on disgust level, which can calculate the comprehensive disgust level (i.e., the third normalized information) according to the marked disgust level (i.e., the first marked information), the score output by the model (i.e., the emotion prediction information) and the weight (i.e., the first weight), thereby improving the accuracy of obtaining the target disgust level (i.e., the third normalized information), so as to facilitate the adjustment of the sending frequency (i.e., the push frequency) of text messages (i.e., push information) based on the target disgust level and reduce the interference to users.

[0182] Take SMS sending based on disgust as an example, see Figure 4 , Figure 4 This is a first SMS sending flow chart based on the degree of disgust provided in an embodiment of the present application. The process of SMS sending based on the degree of disgust provided in an embodiment of the present application is explained below.

[0183] In response to the SMS sending request, step 1 is executed. In step 1, the initial annotated disgust of the communication records between the agent and the user in the past M days (such as 10 days) is obtained.

[0184] Here, the initial annotated disgust recorded is saved after each communication, see Figure 5 , Figure 5 is a data recording flow chart provided in an embodiment of the present application, such as Figure 5 As shown, the process of recording data provided in the embodiment of the present application is explained below.

[0185] In response to the agent communication ending, step 501 is executed, in which the call content is saved.

[0186] Here, the content includes voice, text or images during the communication between the agent and the user.

[0187] In step 502, the initial annotation objection of the user submitted by the agent is determined.

[0188] Here, the agent submits the user's initial marked aversion (such as a number between 1-10, the larger the number, the greater the user's resistance), wherein the agent may be a virtual agent replaced by a robot.

[0189] In step 503, the antipathy model is called to obtain the antipathy model score (ie, the sentiment prediction information) and the antipathy model weight (ie, the first weight of the sentiment prediction information).

[0190] Here, the antipathy model is called to process the call content (ie, the interaction information of the first object) to obtain the antipathy model score and the antipathy model weight.

[0191] In step 504, the data is saved.

[0192] Here, the saved data includes the user's initial labeled disgust, disgust model score, and disgust model weight.

[0193] Continue to see Figure 4 In step 2, the overall annotation disgust (ie, the second information) is calculated.

[0194] Here, the initial annotation aversion of the communication records between agents and users in the past 10 days (those over 10 days have no reference value) is comprehensively calculated to obtain the overall annotation aversion.

[0195] First, the normalized value of the annotation disgust degree (ie, the first information) is calculated.

[0196] Here, the calculation is performed in sequence according to the disgust level submitted by the agent in each communication. The normalized value of the marked disgust level = (current disgust level - minimum disgust level) / (maximum disgust level - minimum disgust level), where the current disgust level is the marked disgust level submitted by the agent in each communication. The range of the marked disgust level is less than or equal to 10 and greater than or equal to 1. The minimum disgust level is a fixed value of 1, and the maximum disgust level is a fixed value of 10. Taking the disgust level submitted in this communication as an example, the normalized value of the marked disgust level 1 = (current disgust level 1-1) / (10-1); the normalized value of the marked disgust level 2 = (current disgust level 2-1) / (10-1); ...; the normalized value of the marked disgust level n = (current disgust level n-1) / (10-1).

[0197] Then, the overall annotation disgust is calculated based on the normalized value of the annotation disgust.

[0198] Here, the overall annotation offensiveness is calculated based on the normalized value of the annotation offensiveness calculated each time, where the overall annotation offensiveness = (N1*annotation offensiveness normalized value 1+N2*annotation offensiveness normalized value 2+N3*annotation offensiveness normalized value 3+...+Nn*annotation offensiveness normalized value n) / (N1+N2+N3+...+Nn), where N1, N2,...Nn are the numbers of each type of annotation offensiveness.

[0199] Continue to see Figure 4 In step 3, a weighted calculation is performed based on the overall annoyance of the annotation, the annoyance model score output by the annoyance model, and the weight to obtain the comprehensive annoyance (ie, the third information).

[0200] Here, a weighted calculation is performed based on the antipathy model score (such as any real number greater than or equal to 0 and less than or equal to 1) output by the antipathy model, the antipathy weight (such as any real number greater than or equal to 0 and less than or equal to 1), and the calculated overall antipathy of the annotation to obtain a comprehensive antipathy, where comprehensive antipathy = overall antipathy of the annotation * overall antipathy weight of the annotation (i.e., the second weight) + antipathy model weight * antipathy model score, where overall antipathy of the annotation = 1 - antipathy model weight.

[0201] Continue to see Figure 4 ,In step 4, the corresponding scenario label is obtained according to the comprehensive disgust.

[0202] Here, the comprehensive disgust degree is judged by interval, the comprehensive disgust degree interval (ie, normalized interval) is obtained, and the corresponding scenario label is obtained as shown in Table 1.

[0203] Table 1 Results of comprehensive antipathy interval division

[0204] Overall disgust normalization value interval Scenario-based tags [0-0.4) Low aversion [0.4-0.6) Medium Dislike [0.6-1] High Dislike

[0205] Continue to see Figure 4 ,In step 5, the corresponding SMS frequency limit is obtained according to the scenario tag.

[0206] Here, each scenario-based tag corresponds to a SMS frequency (i.e., push frequency) limit configuration. The SMS frequency is the number of SMS messages sent within a certain period of time. Then, the corresponding SMS frequency limit configuration (i.e., push frequency) is obtained according to the scenario-based tag, and the remaining number of SMS messages that can be sent (i.e., the number of push messages) for the day is calculated based on the configuration.

[0207] Continue to see Figure 4 , in step 6, calculate whether it exceeds the limit.

[0208] Here, the number of times that can be sent is used to calculate whether the limit is exceeded. If there are times, it means that the sending is within the limit, and step 7 is executed. Otherwise, the limit is set and the SMS cannot be sent.

[0209] Continue to see Figure 4 , in step 7, send the text message.

[0210] Here, the text message content is sent to the user.

[0211] In summary, according to the actual communication situation between the agent and the user, the corresponding SMS limit frequency and SMS content can be flexibly adjusted and switched to ensure that the message triggered to the user is what the user wants. It can be flexibly configured on a daily basis to more effectively trigger the information to the user and reduce the risk of user complaints. In addition, after the number of SMS messages exceeds the limit and the configured number of days, some SMS messages can be released gradually to prevent the information from failing to reach the user. At the same time, the calculation method of the normalized value of the disgust degree and the weighted algorithm is adopted. The disgust degree marked by the communication between the agent and the user in the past 10 days is normalized and calculated, and then the disgust degree score and weight output by the disgust degree model are combined for weighted calculation to obtain the final disgust degree, which is more valuable and accurate.

[0212] See also Figure 6 , Figure 6 This is the second flow chart of SMS sending provided by the embodiment of the present application, such as Figure 6As shown, the process of sending SMS based on disgust level provided in the embodiment of the present application is explained below.

[0213] In the server, in response to the SMS sending request, step 601 is executed. In step 601, the initial annotated disgust of the communication records between the agent and the user in the past M days (such as 10 days) is obtained.

[0214] Here, before obtaining the initial annotated antipathy of the communication records between the agent and the user in the past 10 days, various pre-configurations are performed. The pre-configuration includes the comprehensive antipathy interval division results shown in Table 1, and the scenario-based label strategy SMS frequency configuration shown in Table 2.

[0215] Table 2 Scenario-based tag strategy SMS frequency configuration

[0216] Scenario-based tags Limit frequency Low aversion 2 times a day, 3 times in 5 days Medium Dislike Once a day, twice in 5 days High Dislike 0 times per day

[0217] After completing various pre-configurations, the initial annotated dislike degree and magnitude distribution corresponding to each communication of user A in the communication record are obtained as shown in Table 3.

[0218] Table 3 Initial annotation disgust and magnitude distribution

[0219]

[0220] In step 602, a normalized value of the annotation objectionability (ie, first normalized information) is calculated.

[0221] Here, the normalized value of the annotation disgust degree = (current disgust degree (i.e., annotation information) - minimum disgust degree (i.e., minimum annotation information)) / (maximum disgust degree (i.e., maximum annotation information) - minimum disgust degree), and the normalized value of the annotation disgust degree for each communication is shown in Table 4.

[0222] Table 4 Annotation of the normalized values ​​of disgust

[0223]

[0224] In step 603, the overall annotation offensiveness is calculated based on the calculated annotation offensiveness normalization value.

[0225] Here, the overall annoyance of the annotation = (N1*annotated annoyance normalized value 1+N2*annotated annoyance normalized value 2+N3*annotated annoyance normalized value 3+...+Nn*annotated annoyance normalized value n) / (N1+N2+N3+...+Nn), and the overall annoyance of the annotation is shown in Table 5.

[0226] Table 5: Overall disgust level

[0227]

[0228] In step 604, a weighted calculation is performed based on the overall annoyance of the annotation, the annoyance model score output by the annoyance model, and the weight to obtain a comprehensive annoyance.

[0229] Here, the antipathy model score and antipathy model weight output by calling the antipathy model during the most recent outbound call are obtained as shown in Table 6.

[0230] Table 6 Antipathy model scores and antipathy model weights

[0231]

[0232] Since comprehensive disgust = marked overall disgust * (1-antipathy model weight) + disgust model weight * disgust model score, the comprehensive disgust is shown in Table 7.

[0233] Table 7 Comprehensive disgust

[0234]

[0235] In step 605, the corresponding scenario label is obtained according to the comprehensive disgust degree.

[0236] Using Table 1 as a reference indicator, according to the comprehensive disgust of user A (such as 0.37), Table 1 is searched to obtain the scenario label of user A as shown in Table 8.

[0237] Table 8 Scenario tags of user A

[0238] user Comprehensive disgust Scenario-based tags User A 0.37 Low aversion

[0239] In step 606, the corresponding SMS frequency limit configuration is obtained according to the scenario tag.

[0240] Here, the SMS limit frequency configuration is queried based on the scenario tag as shown in Table 9.

[0241] Table 9 SMS frequency limit configuration

[0242] user Scenario-based tags SMS frequency limit User A Low aversion 2 times a day, 3 times in 5 days

[0243] In step 607, the SMS frequency restriction configuration list is cycled.

[0244] Here, the SMS frequency may be updated according to a specific application scenario, so as to calculate the number of times the SMS is sent based on the new SMS frequency limit.

[0245] In step 608, according to the configured number of days, the short message records sent to the customer within the number of days are queried.

[0246] Here, the number of days and times configured for each SMS is obtained, and the number of SMS sending records within the configured number of days is queried based on the user identity code.

[0247] In step 609, the difference between the configuration times and the number of SMS sending records is recorded.

[0248] Here, the configured number of times is the number of times each SMS is sent.

[0249] After the loop is completed, step 610 is executed. In step 610, the values ​​recorded each time are compared and the minimum value is taken as the remaining number of times.

[0250] Here, the values ​​recorded each time are compared, and the smallest value is taken as the remaining number of SMS messages sent by the user on that day, as shown in Table 10.

[0251] Table 10 The remaining number of SMS messages sent on the day

[0252]

[0253] In step 611, it is determined whether the remaining number is greater than 0.

[0254] Here, when the remaining number is less than or equal to 0, the loop ends, and when the remaining number is greater than 0, step 612 is executed.

[0255] In step 612, a short message sending template (ie, push information) is obtained.

[0256] Here, the SMS template code for determining the user is shown in Table 11.

[0257] Table 11 SMS template coding

[0258] user Is it possible to send SMS messages? SMS template encoding User A yes dfg_dx_notice

[0259] Table 12 shows the text messages that can be sent by user A based on the text message template code.

[0260] Table 12 can send text messages

[0261]

[0262] In step 613, the user's mobile phone number information (ie, the object information of the first object) is queried.

[0263] Here, the user's mobile phone number information can be determined by querying the user's unique identity code.

[0264] In step 614, the SMS information is supplemented according to the SMS template.

[0265] Here, the user's name (i.e., the object information of the first object), such as "Mr. Zhang", is determined based on the user's mobile phone number information, and the SMS information is supplemented according to the SMS template (such as "Dear user Mr. Zhang, you are specially invited to participate in our company's activities"), indicating that user A can send SMS.

[0266] In step 615, the SMS platform is called to send a SMS.

[0267] Here, the SMS platform is called to execute steps 6151 and 6152.

[0268] In step 6151, the SMS message to be sent (ie, push message) is received.

[0269] Here, the SMS message to be sent can be: Dear user Mr. Zhang, hello, you are specially invited to participate in our company's activities.

[0270] In step 6152, a text message is sent to reach the user.

[0271] Here, the text message is sent to the user according to the user's phone information.

[0272] In summary, it is possible to perform real-time calculations based on the recent communication between the agent and the user, flexibly adjust and switch the corresponding SMS limit frequency, improve the value rate of SMS, limit some SMS sent to users with high annoyance, and reduce costs. In addition, after normalizing the annoyance marked by the agent's communication with the user in the past 10 days, the annoyance score and weight output by the annoyance model are combined for weighted calculation to obtain the final overall annoyance, which improves the accuracy of the annoyance. At the same time, the SMS frequency control configuration can be arbitrarily configured to achieve multi-dimensional management, effectively control the frequency of SMS sending, effectively solve the behavior that interferes with users, and reduce the risk of user complaints. It supports the configuration of allowing Y times of sending within X days. For users who exceed the SMS frequency limit, SMS can continue to be sent after the configured number of days, avoiding the failure of information to reach users in real time.

[0273] The following is a description of an exemplary structure of the information processing device 555 provided in the embodiment of the present application implemented as a software module. In some embodiments, Figure 2 As shown, the software modules stored in the information processing device 555 of the memory 550 may include:

[0274] The preprocessing module 5551 is used to normalize the multiple first annotation information to obtain multiple first information, where the first annotation information is used to annotate the emotion of the first object with respect to the pushed information.

[0275] The information processing module 5552 is used to fuse multiple first information to obtain second information; based on the first interaction information of the first object for the pushed information, obtain the emotion prediction information of the first object for the pushed information and the first weight of the emotion prediction information; based on the first weight, fuse the second information and the emotion prediction information to obtain third information.

[0276] The information push module 5553 is used to determine the push frequency of the pushed information based on the third information.

[0277] In some embodiments, the information push module 5553 is also used to determine the total number of push times for push information to the first object based on the push frequency; determine the number of times that the push information can be pushed to the first object based on the total number of push times and the number of times the push information has been pushed to the first object; and push the push information to the first object according to the number of times that can be pushed.

[0278] In some embodiments, the information push module 5553 is also used to determine a normalized interval that matches the third information; query the first mapping table based on the normalized interval, and determine the candidate push channel corresponding to the normalized interval as the push channel, wherein the first mapping table is used to represent the correspondence between the normalized interval and the candidate push channel; push the push information to the first object according to the number of pushable times and the push channel.

[0279] In some embodiments, the information push module 5553 is also used to determine a normalized interval that matches the third information; based on the normalized interval, the second mapping table is queried, and the candidate push frequency corresponding to the normalized interval is determined as the push frequency of the push information, wherein the second mapping table is used to represent the correspondence between the normalized interval and the candidate push frequency.

[0280] In some embodiments, the preprocessing module 5551 is also used to normalize multiple second annotation information to obtain multiple fourth information; fuse multiple fourth information to obtain fifth information; determine first object information of the first object and second object information of the second object, wherein the second object is an object including the fifth information; determine the similarity between the first object information and the second object information; and when the similarity is less than the first similarity threshold, perform the step of normalizing the multiple first annotation information to obtain multiple first information.

[0281] In some embodiments, the information processing module 5552 is further used to determine the similarity between the first interaction information and the second interaction information of the third object for the pushed information, where the third object is the object after the emotion of the pushed information is predicted; when the similarity is less than the second similarity threshold, the step of obtaining the emotion prediction information of the first object for the pushed information and the first weight of the emotion prediction information based on the first interaction information of the first object for the pushed information is performed. .

[0282] In some embodiments, the information processing module 5552 is also used to determine the second weight of the second information based on the first weight; based on the first weight and the second weight, perform weighted sum processing on the second information and the emotion prediction information to obtain the third information of the first object.

[0283] An embodiment of the present application provides a computer program product, which includes computer executable instructions. The computer executable instructions are stored in a computer-readable storage medium. A processor of an electronic device reads the computer executable instructions from the computer-readable storage medium, and the processor executes the computer executable instructions, so that the electronic device executes the above-mentioned information processing method of the embodiment of the present application.

[0284] The present application embodiment provides a computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions or computer programs are stored. When the computer-executable instructions or computer programs are executed by a processor, the processor will be caused to execute the information processing method provided by the present application embodiment, for example, FIG. 3A to FIG. 3I The information processing method shown.

[0285] In some embodiments, the computer-readable storage medium may be a memory such as RAM, ROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or may be various devices including one or any combination of the above memories.

[0286] In some embodiments, computer executable instructions may be in the form of a program, software, software module, script or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment.

[0287] As an example, computer-executable instructions may, but need not, correspond to a file in a file system, may be stored as part of a file that stores other programs or data, such as in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files storing one or more modules, subroutines, or code portions).

[0288] As an example, computer executable instructions may be deployed to be executed on one electronic device, or on multiple electronic devices located at one site, or on multiple electronic devices distributed at multiple sites and interconnected by a communication network.

[0289] In summary, based on the first interactive information of the first object for the push information, the emotion prediction information of the first object for the push information and the first weight of the emotion prediction information are obtained, and based on the first weight, the second normalized information and the emotion prediction information are fused to obtain the third normalized information, and the second normalized information is obtained by processing multiple first labeled information, thus, the first labeled information of the first object for the push information and the emotion prediction information predicted based on the first interactive information of the first object for the push information are combined, and the uncertainty problem caused by relying solely on manual labeling is avoided, thereby improving the accuracy of the third normalized information. Based on the third normalized information, the push frequency of the push information is determined, thus further improving the accuracy of the push frequency determination of the push information, so that the first object will not have a large emotional fluctuation after receiving the push information.

[0290] The above is only an embodiment of the present application and is not intended to limit the protection scope of the present application. Any modifications, equivalent substitutions and improvements made within the spirit and scope of the present application are included in the protection scope of the present application.

Claims

1. An information processing method, characterized in that: The method comprises: Normalizing the plurality of first annotation information to obtain a plurality of first information, wherein the first annotation information is used to annotate the emotion of the first object with respect to the pushed information; fusing the plurality of first information to obtain second information; Based on the first interaction information of the first object with respect to the pushed information, obtaining the emotion prediction information of the first object with respect to the pushed information and a first weight of the emotion prediction information; Based on the first weight, fusing the second information and the emotion prediction information to obtain third information; Based on the third information, a push frequency of the push information is determined.

2. The method according to claim 1, characterized in that The method further comprises: Based on the push frequency, determine a total number of times the push information is pushed to the first object; Determine a possible number of pushes for pushing the push information to the first object based on the total number of pushes and the number of times the push information has been pushed to the first object; The push information is pushed to the first object according to the possible push times.

3. The method according to claim 2, characterized in that The pushing the push information to the first object according to the number of possible push times includes: Determining a normalized interval matching the third information; querying a first mapping table based on the normalized interval, and determining the queried candidate push channel corresponding to the normalized interval as the push channel, wherein the first mapping table is used to represent the correspondence between the normalized interval and the candidate push channel; The push information is pushed to the first object according to the possible push times and the push channel.

4. The method according to claim 1, characterized in that: The determining, based on the third information, a push frequency of the push information includes: Determining a normalized interval matching the third information; Based on the normalized interval, a second mapping table is queried, and the candidate push frequency corresponding to the normalized interval is determined as the push frequency of the push information, wherein the second mapping table is used to represent the correspondence between the normalized interval and the candidate push frequency.

5. The method according to any one of claims 1 to 4, characterized in that: Before normalizing the plurality of first annotation information to obtain the plurality of first information, the method further includes: Normalizing the plurality of second annotation information to obtain a plurality of fourth information; fusing a plurality of the fourth information to obtain fifth information; Determine first object information of the first object and second object information of the second object, wherein the second object is an object including the fifth information; determining similarity between the first object information and the second object information; When the similarity is less than the first similarity threshold, a step of normalizing the plurality of first annotation information to obtain the plurality of first information is performed.

6. The method according to any one of claims 1 to 4, characterized in that: Before obtaining the emotion prediction information of the first object with respect to the pushed information and the first weight of the emotion prediction information based on the first interaction information of the first object with respect to the pushed information, the method further includes: Determine similarity between the first interaction information and second interaction information of a third object with respect to the push information, wherein the third object is an object after predicting the emotion of the push information; When the similarity is less than a second similarity threshold, a step is performed to obtain emotion prediction information of the first object with respect to the pushed information and a first weight of the emotion prediction information based on the first interaction information of the first object with respect to the pushed information.

7. The method according to any one of claims 1 to 4, characterized in that: The step of fusing the second information and the emotion prediction information based on the first weight to obtain the third information includes: Based on the first weight, determining a second weight of the second information; Based on the first weight and the second weight, the second information and the emotion prediction information are weighted and summed to obtain the third information.

8. An electronic device, characterized in that: The electronic device comprises: A memory for storing computer executable instructions; The processor is used to implement the information push method described in any one of claims 1 to 7 when executing the computer executable instructions or computer programs stored in the memory.

9. A computer-readable storage medium storing computer-executable instructions or a computer program, characterized in that: When the computer executable instructions or computer program are executed by a processor, the information push method according to any one of claims 1 to 7 is implemented.

10. A computer program product comprising computer executable instructions or a computer program, characterized in that: When the computer executable instructions or computer program are executed by a processor, the information push method according to any one of claims 1 to 7 is implemented.

Images