A human-computer interaction behavior recommendation method and system

Abstract

The present invention provides a method and system for recommending human-computer interaction behaviors, which includes: obtaining information status data and recommendation requirements of a target user; determining a set of candidate human-computer interaction behaviors of the target user based on the information status data; obtaining a preset thinking model corresponding to the recommendation requirements, parsing the preset thinking model based on the information status data, and obtaining a user thinking estimation result; calculating a decision score of each candidate human-computer interaction behavior based on the user thinking estimation result and the preset behavior value corresponding to each candidate human-computer interaction behavior in the candidate human-computer interaction behavior set; and determining the recommended human-computer interaction behavior of the target user based on the decision score of each candidate human-computer interaction behavior. The personalized recommendation service for users is realized, and the bias of the recommendation model caused by data deviation is overcome by comprehensively considering the value of the user information status data and the human-computer interaction behavior itself, and the decision score recommendation is performed in combination with the user thinking estimation result, thereby improving the user's usage experience.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and system for recommending human-computer interaction behaviors. Background Art

[0002] Intelligent decision-making of machines in human-computer interaction is one of the key technologies in the field of artificial intelligence. Specifically, the machine makes intelligent decisions and recommends to users based on the task settings in the human-computer interaction scenario and the needs of the current user. Human-computer interaction behavior decision-making technology is widely used in content recommendation systems and dialogue management systems.

[0003] At present, most human-computer interaction behavior decision-making methods adopt user data-driven statistical analysis and induction methods, that is, by using existing machine learning algorithms to perform cluster analysis on the massive user data collected, such as using matrix completion to train decision models and make human-computer interaction behavior recommendations for users. However, this attempt to cluster users and provide the same human-computer interaction behavior recommendations to each type of user does not take into account the intentions and preferences of different individual users, making it difficult to provide users with personalized recommendation services. In addition, in many application scenarios, it is also very difficult to fully collect user data, resulting in the collected user data often being unbalanced and having certain data biases. Using these data recommendation models for training will lead to model bias, making it difficult to meet the personalized human-computer interaction behavior needs of different users in different application scenarios. Summary of the invention

[0004] In view of this, an embodiment of the present invention provides a human-computer interaction behavior recommendation method and system to overcome the problem that the human-computer interaction behavior recommendation method in the prior art is difficult to meet the personalized human-computer interaction behavior needs of different users in different application scenarios.

[0005] The embodiment of the present invention provides a method for recommending human-computer interaction behaviors, including:

[0006] Acquire information status data and recommendation requirements of target users, wherein the information status data at least includes: historical human-computer interaction behavior usage data and current status data of target users;

[0007] Based on the information state data, determining a set of candidate human-machine interaction behaviors of the target user;

[0008] Obtaining a preset thinking model corresponding to the recommendation requirement, parsing the preset thinking model based on the information status data, and obtaining a user thinking estimation result;

[0009] Calculate the decision score of each candidate human-machine interaction behavior based on the user thought estimation result and the preset behavior value corresponding to each candidate human-machine interaction behavior in the candidate human-machine interaction behavior set;

[0010] According to the decision score of each candidate human-computer interaction behavior, the recommended human-computer interaction behavior for the target user is determined.

[0011] Optionally, determining the candidate human-computer interaction behavior of the target user based on the information state data includes:

[0012] Based on the historical human-computer interaction behavior usage data, a preset human-computer interaction behavior prediction model is trained to obtain a current prediction model;

[0013] The current state data is input into the current prediction model, and the candidate human-machine interaction behavior set is output.

[0014] Optionally, the analyzing the preset thinking model based on the information state data to obtain the user thinking estimation result includes:

[0015] Obtaining each thinking indicator in the preset thinking model;

[0016] Analyze the current state data according to the preset weight assignment rules to determine the indicator weight of each thinking indicator;

[0017] Based on the indicator weight of each thinking indicator, the user thinking estimation result is determined.

[0018] Optionally, the analyzing the preset thinking model based on the information state data to obtain the user thinking estimation result includes:

[0019] Based on the recommendation requirements, determining the target value of each thinking indicator in the preset thinking model;

[0020] Determine the current value of each thinking indicator in the preset thinking model based on the historical human-computer interaction behavior usage data;

[0021] Determine the indicator weight of each thinking indicator according to the relationship between the target value and the current value of each thinking indicator;

[0022] Based on the indicator weight of each thinking indicator, the user thinking estimation result is determined.

[0023] Optionally, the calculating of the decision score of each candidate human-machine interaction behavior based on the user thought estimation result and the preset behavior value corresponding to each candidate human-machine interaction behavior in the candidate human-machine interaction behavior set includes:

[0024] Obtain the preset behavior value of the current candidate human-computer interaction behavior under each thinking index from the preset behavior value evaluation database;

[0025] Based on the indicator weights of each thinking indicator and the preset behavior values ​​of the current candidate human-computer interaction behavior under each thinking indicator, the decision score of the current candidate human-computer interaction behavior is calculated.

[0026] Optionally, determining the recommended human-computer interaction behavior of the target user according to the decision score of each candidate human-computer interaction behavior includes:

[0027] Sort the decision scores of each candidate human-machine interaction behavior to obtain a sorting result;

[0028] According to the ranking result, the candidate human-computer interaction behavior with the highest decision score is determined as the recommended human-computer interaction behavior for the target user.

[0029] Optionally, the determining the current value of each thinking indicator in the preset thinking model based on the historical human-computer interaction behavior usage data includes:

[0030] Obtaining the historical human-computer interaction behavior contained in the historical human-computer interaction behavior usage data;

[0031] Obtain the preset behavior value of each historical human-computer interaction behavior under the current thinking index from the preset behavior value evaluation database;

[0032] According to the preset behavior values ​​of each historical human-computer interaction behavior under the current thinking indicator, the current value of the current thinking indicator is calculated.

[0033] The embodiment of the present invention also provides a human-computer interaction behavior recommendation system, including:

[0034] An acquisition module is used to acquire information status data and recommendation requirements of a target user, wherein the information status data at least includes: historical human-computer interaction behavior usage data and current status data of the target user;

[0035] A first processing module, configured to determine a candidate human-machine interaction behavior set of the target user based on the information state data;

[0036] The second processing module is used to obtain a preset thinking model corresponding to the recommendation requirement, and analyze the preset thinking model based on the information state data to obtain a user thinking estimation result;

[0037] A third processing module is used to calculate the decision score of each candidate human-machine interaction behavior based on the user thinking estimation result and the preset behavior value corresponding to each candidate human-machine interaction behavior in the candidate human-machine interaction behavior set;

[0038] The fourth processing module is used to determine the recommended human-computer interaction behavior of the target user according to the decision score of each candidate human-computer interaction behavior.

[0039] An embodiment of the present invention further provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the human-computer interaction behavior recommendation method provided by the embodiment of the present invention by executing the computer instructions.

[0040] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the human-computer interaction behavior recommendation method provided by the embodiment of the present invention.

[0041] The technical solution of the present invention has the following advantages:

[0042] The embodiment of the present invention provides a method and system for recommending human-computer interaction behaviors, which obtains information status data and recommendation requirements of a target user; determines a set of candidate human-computer interaction behaviors of the target user based on the information status data; obtains a preset thinking model corresponding to the recommendation requirements, analyzes the preset thinking model based on the information status data, and obtains a user thinking prediction result; calculates a decision score of each candidate human-computer interaction behavior based on the user thinking prediction result and the preset behavior value corresponding to each candidate human-computer interaction behavior in the candidate human-computer interaction behavior set; and determines the recommended human-computer interaction behavior of the target user according to the decision score of each candidate human-computer interaction behavior. Thus, by collecting the information status data of the target user itself, and using the information status data to determine the set of candidate human-computer interaction behaviors and the user thinking prediction result, by using the thinking prediction result, the value change of the human-computer interaction behavior itself is dynamically updated, and the personalized recommendation service of the user is realized. By comprehensively considering the user information status data and the value of the human-computer interaction behavior itself, the bias of the recommendation model caused by data deviation can be effectively overcome, and the value of multiple candidate human-computer interaction behaviors and the user thinking prediction result are combined to make a decision score to recommend the human-computer interaction behavior, thereby improving the user's use experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 A schematic diagram of an application scenario of the human-computer interaction behavior recommendation method in an embodiment of the present invention;

[0045] Figure 2 A flowchart of a method for recommending human-computer interaction behaviors in an embodiment of the present invention;

[0046] Figure 3 A schematic diagram of the working process of a content recommendation system in an embodiment of the present invention;

[0047] Figure 4 A schematic diagram of a user's current thinking in an embodiment of the present invention;

[0048] Figure 5 Schematic diagram of the structure of the human-computer interaction behavior recommendation system in an embodiment of the present invention;

[0049] Figure 6 Schematic diagram of the structure of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0051] The technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0052] Intelligent decision-making of machines in human-computer interaction is one of the key technologies in the field of artificial intelligence. Specifically, the machine makes intelligent decisions and recommends to users based on the task settings in the human-computer interaction scenario and the needs of the current user. Human-computer interaction behavior decision-making technology is widely used in content recommendation systems and dialogue management systems. The details are shown in Table 1.

[0053] Table 1

[0054]

[0055] The content recommendation system and dialogue management system shown in Table 1 predict the user's interest preferences based on the user's basic information (such as age, gender, region, income) and historical behavior (such as comments, favorites, likes, views, purchases, etc.), and recommend human-computer interaction behaviors (such as movie recommendations, food recommendations, personalized advertising, etc.) with users who have similar interests and preferences.

[0056] At present, most human-computer interaction behavior decision-making methods adopt user data-driven statistical analysis and induction methods, that is, by using existing machine learning algorithms to perform cluster analysis on the massive user data collected, such as using matrix completion to train decision models and make human-computer interaction behavior recommendations for users. However, this attempt to cluster users and provide the same human-computer interaction behavior recommendations to each type of user does not take into account the intentions and preferences of different individual users, making it difficult to provide users with personalized recommendation services. In addition, in many application scenarios, it is also very difficult to fully collect user data, resulting in the collected user data often being unbalanced and having certain data biases. Using these data recommendation models for training will lead to model bias, making it difficult to meet the personalized human-computer interaction behavior needs of different users in different application scenarios.

[0057] Based on the above problems, an embodiment of the present invention provides a method for recommending human-computer interaction behaviors. Figure 1 Schematic diagram of application scenarios of the human-computer interaction behavior recommendation method, which can be specifically applied to user recommendation systems such as content recommendation systems and dialogue management systems as shown in Table 1. When it is applied to content recommendation systems, Figure 2 As shown, the human-computer interaction behavior recommendation method mainly includes the following steps:

[0058] Step S101: Obtain the information status data and recommendation needs of the target user, the information status data at least includes: historical human-computer interaction behavior usage data and the current status data of the target user. Specifically, the historical human-computer interaction behavior usage data is obtained by collecting multi-dimensional and multi-scale data of the target user, and using computer vision, speech recognition and natural language understanding and other technologies to parse out data related to user behavior (including basic information of the user, the user's dialogue status, the user's historical operation information and user expressions, etc.), and according to pre-set rules, environmental related factors (such as weather, time, location and desktop), event related factors (such as system updates and system disconnection), task related factors (such as the system is in a task-free state) and context related factors (such as previously recommended human-computer interaction behaviors) are obtained from the data related to user behavior. The current status data includes the user status currently input by the user, such as: emotional state (happy, sad), entertainment state (watching movies, playing games), etc. The recommendation needs are certain functions selected by the user using the current recommendation system, such as: improving mood or planning a day's life, etc.

[0059] Step S102: Based on the information status data, determine the candidate human-computer interaction behavior set of the target user. Specifically, based on comprehensive consideration of the multi-dimensional factors in the above information status data, multiple subsequent human-computer interaction behaviors matching the target user are obtained to form a candidate human-computer interaction behavior set. For example: affected by context-related factors, as shown in Table 2, when the user's historical human-computer interaction behavior usage data uses the feeding game function, the output multiple candidate human-computer interaction behaviors are: 1) recommend taking group photos; 2) recommend interactive reading; 3) recommend card board games, etc.

[0060] Table 2

[0061]

[0062] Step S103: Obtain the preset thinking model corresponding to the recommendation demand, analyze the preset thinking model based on the information status data, and obtain the user thinking estimation result. Specifically, the recommendation system pre-establishes corresponding thinking models according to different recommendation needs. For example, when the user's recommendation demand is to arrange study life reasonably, the corresponding preset thinking model is composed of entertainment, intelligence, and emotion thinking indicators. Correspondingly, the user thinking estimation result reflects the current importance of different thinking indicators to the target user. Exemplarily, the relationship between different factors in the information status data and user thinking is shown in Table 3.

[0063] Table 3

[0064]

[0065]

[0066] Step S104: Calculate the decision score of each candidate human-computer interaction behavior based on the user's thought estimation result and the preset behavior value corresponding to each candidate human-computer interaction behavior in the candidate human-computer interaction behavior set. Specifically, the preset behavior value of different human-computer interaction behaviors can be obtained by querying from the preset behavior value evaluation database.

[0067] Step S105: Determine the recommended human-computer interaction behavior for the target user according to the decision scores of each candidate human-computer interaction behavior. Specifically, the decision scores of each candidate human-computer interaction behavior are sorted to obtain a sorting result; according to the sorting result, the candidate human-computer interaction behavior with the highest decision score is determined as the recommended human-computer interaction behavior for the target user, and the recommended human-computer interaction behavior is recommended to the target user. Figure 3 The present invention is a schematic diagram of the working process of a content recommendation system when the human-computer interaction behavior recommendation method provided by an embodiment of the present invention is applied to the content recommendation system.

[0068] Through the above steps S101 to S105, the human-computer interaction behavior recommendation method provided by the embodiment of the present invention collects the information status data of the target user itself, and uses the information status data to determine the candidate human-computer interaction behavior set and the user thinking prediction results. By using the thinking prediction results, the value changes of the human-computer interaction behavior itself are dynamically updated, thereby realizing personalized recommendation services for users. By comprehensively considering the user information status data and the value of the human-computer interaction behavior itself, the bias of the recommendation model caused by data deviation can be effectively overcome. Combined with the values ​​of multiple candidate human-computer interaction behaviors and the user thinking prediction results, decision-making scores are made to recommend human-computer interaction behaviors, thereby improving the user experience.

[0069] Specifically, in one embodiment, the above step S102 specifically includes the following steps:

[0070] Step S201: Based on the historical human-computer interaction behavior usage data, the preset human-computer interaction behavior prediction model is trained to obtain the current prediction model. Specifically, the above-mentioned multi-factor historical human-computer interaction behavior usage data and the candidate human-computer interaction behavior set are represented by multidimensional vectors, respectively represented as x and y, wherein x is a multidimensional vector, each element of which represents a factor; y is also a multidimensional vector, each element of which represents a human-computer interaction behavior, that is, it represents a set formed by multiple candidate human-computer interaction behaviors. Through the pre-constructed preset human-computer interaction behavior prediction model F, a mapping relationship between x and y is established, that is, y=F(x), and then the training samples are extracted from the historical human-computer interaction behavior usage data and input into the model F for training to obtain the current prediction model. In practical applications, the preset human-computer interaction behavior prediction model F can be not only the above-mentioned trainable model based on machine learning (for example, nonlinear regression model, periodic neural network), but also a non-learning model defined according to human experience and rules (for example, directly indicating that the information state data x does not directly correspond to the set y formed by multiple candidate human-computer interaction behaviors through calculation), and the present invention is not limited thereto.

[0071] Step S202: Input the current state data into the current prediction model and output a set of candidate human-computer interaction behaviors. Specifically, when the preset human-computer interaction behavior prediction model is a trainable model, the current state data of the target user is input into the trained model to obtain a set of candidate human-computer interaction behaviors of the target user.

[0072] Specifically, in one embodiment, the above step S103 specifically includes the following steps:

[0073] Step S301: Obtain each thinking indicator in the preset thinking model. Specifically, the preset thinking model can be a trainable model based on machine learning (such as a deep graph convolutional network), which is obtained by training using the thinking indicators and recommendation requirements of different users, or it can be a non-learning model defined according to human experience and rules (for example, directly defining the mapping relationship between recommendation requirements and thinking indicators in sequence). In practical applications, the preset thinking model can be represented in the form of an abstract graph such as a thinking analysis graph. The user's current thinking is Figure 4 As shown, the abstract graph connects subjective and objective concepts at different levels into a graph structure.

[0074] Step S302: Analyze the current state data according to the preset weight assignment rule to determine the indicator weight of each thinking indicator. Specifically, the preset weight assignment rule is set according to the recommendation needs of the target user. For example, the current user's recommendation need is to improve mood. The thinking indicators in the preset thinking model corresponding to the recommendation need include: learning, entertainment, and emotion. Since the entertainment and emotion indicators have a greater impact on improving mood, the preset weight assignment rule is to set the indicator weight of learning to 1, and the indicator weight of entertainment and emotion to 2. It should be noted that the present invention is only taken as an example and is not limited to this.

[0075] Step S303: Based on the indicator weights of each thinking indicator, determine the user thinking estimation result. In practical applications, the user thinking estimation result is the indicator weight assignment result of all thinking indicators.

[0076] In another alternative implementation, when the recommendation function selected by the user in the recommendation system is a comprehensive recommendation function, for example, arranging the user's study life for a day, when recommending human-computer interaction behaviors to the user, in order to arrange the user's behaviors more reasonably and improve the user experience, the above step S103 specifically includes the following steps:

[0077] Step S31: Based on the recommendation requirement, determine the target value of each thinking indicator in the preset thinking model. Specifically, assuming that the recommendation requirement is the above-mentioned "arrange the user's one-day study life", in order to balance learning and entertainment as well as the user's emotions and help the user to reasonably arrange the study life, the goal of the recommendation system is to make the score of each thinking indicator in the user's preset thinking model reach 100 points, that is, the target values ​​of the learning indicator, entertainment indicator and emotional indicator are all 100 points.

[0078] Step S32: Determine the current value of each thinking indicator in the preset thinking model based on the historical human-computer interaction behavior usage data. In practical applications, since the recommendation system will recommend different human-computer interaction behaviors to users at different times when "arranging the user's one-day learning life", and the human-computer interaction behaviors recommended in a day will affect each other, for example: the last human-computer interaction behavior recommended by the recommendation system to the user is "learning English", then the corresponding learning thinking indicator, entertainment thinking indicator and emotional thinking indicator all have the current value corresponding to the human-computer interaction behavior of "learning English".

[0079] Specifically, the above step S32 can be performed by obtaining the historical human-computer interaction behaviors contained in the historical human-computer interaction behavior usage data; obtaining the preset behavior value of each historical human-computer interaction behavior under the current thinking index from the preset behavior value evaluation database; and calculating the current value of the current thinking index according to the preset behavior value of each historical human-computer interaction behavior under the current thinking index. It should be noted that in actual applications, the current value of the current thinking index can also be determined by other means, such as directly calculating the index amplitude according to the user's current state data, etc., and the present invention is not limited to this.

[0080] Step S33: Determine the indicator weight of each thinking indicator according to the relationship between the target value and the current value of each thinking indicator. Specifically, according to the difference between the current value and the target value of each thinking indicator or the proportional relationship between the current value and the target value, the indicator weight of each thinking indicator is determined according to the preset weight assignment rule. Specifically, the closer the current value of the thinking indicator is to the target value, the smaller its corresponding indicator weight is, and vice versa. That is, the recommendation system needs to recommend a thinking indicator with a larger indicator weight and a greater impact on the human-computer interaction behavior for the target user to improve the score of the thinking indicator.

[0081] Step S34: Based on the indicator weights of each thinking indicator, determine the user thinking estimation result. In practical applications, the user thinking estimation result is the indicator weight assignment result of all thinking indicators.

[0082] Specifically, in one embodiment, the above step S104 specifically includes the following steps:

[0083] Step S401: Obtain the preset behavior value of the current candidate human-computer interaction behavior under each thinking indicator from the preset behavior value evaluation database. Specifically, the preset behavior value evaluation database is a preset behavior value scoring relationship database between human-computer interaction behavior and thinking indicators, and Table 4 is a schematic diagram of the preset behavior value evaluation database. In practical applications, the preset behavior value scores stored in the preset behavior value evaluation database can be scoring values ​​uniformly set by utilizing big data analysis, neural network learning, etc., or can be scoring values ​​personalized for the needs of a specific user. For example: Assuming that Table 4 is a scoring value set for all users, for a specific user A, whose purpose of using the recommendation system is to improve academic performance, the intelligent learning score value of the human-computer interaction behavior related to learning in Table 4 can be increased, and other scoring values ​​can be correspondingly reduced, etc. The present invention is not limited to this.

[0084] Table 4

[0085] entertainment wisdom emotion Take a group photo 8 0 5 Interactive Reading 6 9 9 Card board games 8 8 7

[0086] Step S402: Based on the indicator weights of each thinking indicator and the preset behavior values ​​of the current candidate human-computer interaction behavior under each thinking indicator, the decision score of the current candidate human-computer interaction behavior is calculated. Specifically, assuming that the user's recommendation requirement is happiness, and the indicator weight corresponding to the thinking indicator entertainment as shown in Table 4 is 2, and the weights of the other indicators are 1, the decision scores of the candidate human-computer interaction behaviors as shown in Table 4 are calculated as follows:

[0087] Taking a group photo = 8*2+0*1+5*1=21

[0088] Interactive reading = 6*2+9*1+9*1=30

[0089] Card board game = 8*2+8*1+7*1=31

[0090] In the end, the card board game received the highest decision score (31 points) and will be selected. The recommendation system will recommend the card board game as a human-computer interaction behavior to the user.

[0091] The human-computer interaction behavior recommendation method provided by the embodiment of the present invention will be described in detail below in conjunction with another specific application example.

[0092] When the human-computer interaction behavior recommendation method is applied to an intelligent human-computer question-answering system, the working process of the intelligent human-computer question-answering system is as follows:

[0093] 1) Users use the system and ask questions (such as "I'm in a bad mood, what can make me happy?", etc.).

[0094] 2) The system uses natural language understanding technology to parse the user's input questions and recognizes that the user's information status data is ("low mood"). At this time, similarity matching is performed on the user's information status data in the question and answer library. The specific matching method can adopt existing technologies, such as using the distance in the feature space of the trained model. The shorter the distance, the more similar. Then output the answers corresponding to multiple similar questions as recommendations for human-computer interaction behaviors (for example: a) listen to music to cheer yourself up; b) watch a movie to relax; c) make an appointment with friends to hang out).

[0095] 3) According to the relationship between information status data and user thinking model (see Table 3), the user thinking prediction results are generated based on the user's information status data. The results show that users prefer to listen to lyrical music as a channel for emotional catharsis.

[0096] 4) Comprehensively consider the user's thinking and multiple candidate human-computer interaction behaviors, and make a decision score. Finally, the human-computer interaction behavior "a) Listen to music to cheer yourself up" has the highest score, so it is selected for multimodal dialogue process generation and recommended to the user.

[0097] By executing the above steps, the human-computer interaction behavior recommendation method provided by the embodiment of the present invention collects the information status data of the target user itself, and uses the information status data to determine the candidate human-computer interaction behavior set and the user thinking prediction results. By using the thinking prediction results, the value changes of the human-computer interaction behavior itself are dynamically updated, thereby realizing the user's personalized recommendation service. By comprehensively considering the user's information status data and the value of the human-computer interaction behavior itself, the bias of the recommendation model caused by data deviation can be effectively overcome. Combined with the values ​​of multiple candidate human-computer interaction behaviors and the user's thinking prediction results, decision-making scores are made to recommend human-computer interaction behaviors, thereby improving the user's usage experience.

[0098] The embodiment of the present invention also provides a human-computer interaction behavior recommendation system, such as Figure 5 As shown, the human-computer interaction behavior recommendation system includes:

[0099] The acquisition module 101 is used to acquire the information status data and recommendation requirements of the target user, and the information status data at least includes: historical human-computer interaction behavior usage data and the current status data of the target user. For details, please refer to the relevant description of step S101 in the above method embodiment, which will not be repeated here.

[0100] The first processing module 102 is used to determine a candidate human-machine interaction behavior set of the target user based on the information state data. For details, please refer to the relevant description of step S102 in the above method embodiment, which will not be repeated here.

[0101] The second processing module 103 is used to obtain a preset thinking model corresponding to the recommendation requirement, analyze the preset thinking model based on the information state data, and obtain the user thinking estimation result. For details, please refer to the relevant description of step S103 in the above method embodiment, which will not be repeated here.

[0102] The third processing module 104 is used to calculate the decision score of each candidate human-machine interaction behavior based on the user's thought estimation result and the preset behavior value corresponding to each candidate human-machine interaction behavior in the candidate human-machine interaction behavior set. For details, please refer to the relevant description of step S104 in the above method embodiment, which will not be repeated here.

[0103] The fourth processing module 105 is used to determine the recommended human-computer interaction behavior for the target user according to the decision score of each candidate human-computer interaction behavior. For details, please refer to the relevant description of step S105 in the above method embodiment, which will not be repeated here.

[0104] Through the collaborative cooperation of the above-mentioned components, the human-computer interaction behavior recommendation system provided by the embodiment of the present invention collects the information status data of the target user himself, and uses the information status data to determine the candidate human-computer interaction behavior set and the user's thinking prediction results. By using the thinking prediction results, the value changes of the human-computer interaction behavior itself are dynamically updated, thereby realizing personalized recommendation services for users. By comprehensively considering the user's information status data and the value of the human-computer interaction behavior itself, the bias of the recommendation model caused by data deviation can be effectively overcome. Combined with the values ​​of multiple candidate human-computer interaction behaviors and the user's thinking prediction results, decision-making scores are made to recommend human-computer interaction behaviors, thereby improving the user's usage experience.

[0105] According to an embodiment of the present invention, there is also provided an electronic device, such as Figure 6 As shown, the electronic device may include a processor 901 and a memory 902, wherein the processor 901 and the memory 902 may be connected via a bus or other means. Figure 6 The example of connecting through bus is taken in the following.

[0106] The processor 901 may be a central processing unit (CPU). The processor 901 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination of the above chips.

[0107] The memory 902 is a non-transitory computer-readable storage medium that can be used to store non-transitory software programs, non-transitory computer executable programs and modules, such as program instructions / modules corresponding to the methods in the method embodiments of the present invention. The processor 901 executes various functional applications and data processing of the processor by running the non-transitory software programs, instructions and modules stored in the memory 902, that is, implementing the methods in the above method embodiments.

[0108] The memory 902 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required by at least one function; the data storage area may store data created by the processor 901, etc. In addition, the memory 902 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 902 may optionally include a memory remotely arranged relative to the processor 901, and these remote memories may be connected to the processor 901 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0109] One or more modules are stored in the memory 902 , and when executed by the processor 901 , the method in the above method embodiment is executed.

[0110] The specific details of the above electronic device can be understood by referring to the corresponding descriptions and effects in the above method embodiments, and will not be repeated here.

[0111] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above-mentioned types of memory.

[0112] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for recommending human-computer interaction behaviors, characterized in that: include: Acquire information status data and recommendation requirements of target users, wherein the information status data at least includes: historical human-computer interaction behavior usage data and current status data of target users; Based on the information state data, determining a set of candidate human-machine interaction behaviors of the target user; Obtaining a preset thinking model corresponding to the recommendation requirement, parsing the preset thinking model based on the information status data, and obtaining a user thinking estimation result; the preset thinking model includes various thinking indicators, and determining the user thinking estimation result based on the indicator weight of each thinking indicator; Based on the user thinking estimation result and the preset behavior value corresponding to each candidate human-machine interaction behavior in the candidate human-machine interaction behavior set, the decision score of each candidate human-machine interaction behavior is calculated, including: obtaining the preset behavior value of the current candidate human-machine interaction behavior under each thinking indicator from the preset behavior value evaluation database; based on the indicator weight of each thinking indicator and the preset behavior value of the current candidate human-machine interaction behavior under each thinking indicator, the decision score of the current candidate human-machine interaction behavior is calculated; According to the decision score of each candidate human-computer interaction behavior, the recommended human-computer interaction behavior for the target user is determined.

2. The method according to claim 1, characterized in that The determining, based on the information state data, the candidate human-computer interaction behavior of the target user includes: Based on the historical human-computer interaction behavior usage data, a preset human-computer interaction behavior prediction model is trained to obtain a current prediction model; The current state data is input into the current prediction model, and the candidate human-machine interaction behavior set is output.

3. The method according to claim 1, characterized in that The step of analyzing the preset thinking model based on the information state data to obtain a user thinking estimation result includes: Obtaining each thinking indicator in the preset thinking model; Analyze the current state data according to the preset weight assignment rules to determine the indicator weight of each thinking indicator; Based on the indicator weight of each thinking indicator, the user thinking estimation result is determined.

4. The method according to claim 1, characterized in that: The step of analyzing the preset thinking model based on the information state data to obtain a user thinking estimation result includes: Based on the recommendation requirements, determining the target value of each thinking indicator in the preset thinking model; Determine the current value of each thinking indicator in the preset thinking model based on the historical human-computer interaction behavior usage data; Determine the indicator weight of each thinking indicator according to the relationship between the target value and the current value of each thinking indicator; Based on the indicator weight of each thinking indicator, the user thinking estimation result is determined.

5. The method according to claim 1, characterized in that The step of determining the recommended human-computer interaction behavior for the target user according to the decision score of each candidate human-computer interaction behavior includes: Sort the decision scores of each candidate human-machine interaction behavior to obtain a sorting result; According to the ranking result, the candidate human-computer interaction behavior with the highest decision score is determined as the recommended human-computer interaction behavior for the target user.

6. The method according to claim 4, characterized in that The determining the current value of each thinking indicator in the preset thinking model based on the historical human-computer interaction behavior usage data includes: Obtaining the historical human-computer interaction behavior contained in the historical human-computer interaction behavior usage data; Obtain the preset behavior value of each historical human-computer interaction behavior under the current thinking index from the preset behavior value evaluation database; According to the preset behavior values ​​of each historical human-computer interaction behavior under the current thinking indicator, the current value of the current thinking indicator is calculated.

7. A human-computer interaction behavior recommendation system, characterized in that: include: An acquisition module is used to acquire information status data and recommendation requirements of a target user, wherein the information status data at least includes: historical human-computer interaction behavior usage data and current status data of the target user; A first processing module, configured to determine a candidate human-machine interaction behavior set of the target user based on the information state data; The second processing module is used to obtain a preset thinking model corresponding to the recommendation requirement, and analyze the preset thinking model based on the information state data to obtain a user thinking estimation result; the preset thinking model includes various thinking indicators, and the user thinking estimation result is determined based on the indicator weight of each thinking indicator; The third processing module is used to calculate the decision score of each candidate human-machine interaction behavior based on the user thinking estimation result and the preset behavior value corresponding to each candidate human-machine interaction behavior in the candidate human-machine interaction behavior set, including: obtaining the preset behavior value of the current candidate human-machine interaction behavior under each thinking indicator from the preset behavior value evaluation database; based on the indicator weight of each thinking indicator and the preset behavior value of the current candidate human-machine interaction behavior under each thinking indicator, calculating the decision score of the current candidate human-machine interaction behavior; The fourth processing module is used to determine the recommended human-computer interaction behavior of the target user according to the decision score of each candidate human-computer interaction behavior.

8. An electronic device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method according to any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the method according to any one of claims 1 to 6.

Images