A user portrait generation method and system for Yunqiwei

Abstract

The present invention relates to a user portrait generation method and system of Yunqiwei, and relates to the field of portrait generation technology. The method includes obtaining user information and real-time environmental information; determining scenario change characteristics based on real-time environmental information and historical environmental information; obtaining a portrait generation instruction in response to the scenario change characteristics being greater than a change threshold; and determining a user portrait based on user information and real-time environmental information according to the portrait generation instruction. The user portrait generation method can automatically update the user portrait according to changes in environmental information, and its user portrait generation process reduces the need for manual intervention, reduces maintenance costs, and speeds up the conversion speed from data collection to actual application, so that the generated user portrait is more in line with the user's current situation.

Description

Technical Field

[0001] The present invention relates to the technical field of portrait generation, and in particular to a method and system for generating a user portrait of Yunqiwei. Background Art

[0002] With the rapid development of information technology, user portraits, as an important tool for understanding and serving users, have been widely used in many fields, such as e-commerce, personalized recommendation systems, and social media analysis.

[0003] Yunqiwei is a system that uses cloud phones to manage WeChat for business. Cloud phones do not have a physical screen or physical SIM (Subscriber Identity Module) card, but are otherwise the same as physical phones. They run the real Android operating system and WeChat for business, and fully comply with relevant security specifications. Compared with the use of ARM (Advanced RISC Machine) servers to simulate the Android environment, they are more secure and more scalable. Each WeChat for business runs on an independent mobile phone chip, and the operating environment is the same as that of a real phone. Yunqiwei breaks the original "one person, one number" model of "person-device-account binding" and hosts WeChat for business accounts in the cloud for centralized management. Aggregate all chat sessions in the hosted WeChat for business accounts and redistribute them according to preset routing rules and operation management strategies. Set up multiple skill groups according to the team (for example: front-line agents, marketing agents, etc.), and transfer sessions seamlessly according to the classification of user questions, without the need to repeatedly add people or groups. In the chat session, Yunqiwei introduces the capabilities of intelligent customer service robots, which can automatically reply to common questions and improve the timeliness of customer response.

[0004] Traditional user portrait generation methods mainly rely on static data, such as user personal information, historical behavior records, and purchase preferences. Although this information can provide a basic understanding of the user, it often lacks consideration of the user's current state and emotional changes, which limits the effectiveness and practicality of user portraits in dynamic environments.

[0005] In addition, the portrait generation methods in related technologies lack a real-time update mechanism. Traditional user portraits are usually built based on historical data and rarely change once established. This type of user portrait cannot timely reflect the user's latest interests, needs, or emotional changes, resulting in the content provided by the service provider no longer meeting the user's current expectations and failing to match highly applicable intelligent customer service to the users in need.

[0006] Therefore, there is an urgent need for a method and system for generating user portraits for cloud enterprises and micro-enterprises. Summary of the invention

[0007] The present invention provides a method for generating a user portrait of Yunqiwei, comprising: obtaining user information and real-time environment information; determining scenario change characteristics based on the real-time environment information and historical environment information; obtaining a portrait generation instruction in response to the scenario change characteristics being greater than a change threshold; and determining a user portrait based on the user information and the real-time environment information according to the portrait generation instruction.

[0008] Furthermore, the above-mentioned real-time environmental information includes real-time environmental parameters, real-time device status and real-time character status, and the above-mentioned historical environmental information includes historical environmental parameters, historical device status and historical character status; the above-mentioned determination of scenario change characteristics based on the above-mentioned real-time environmental information and historical environmental information includes: determining natural change characteristics based on real-time environmental parameters and historical environmental parameters; determining medium change characteristics based on real-time device status and historical device status; determining character change characteristics based on real-time character status and historical character status; determining the above-mentioned scenario change characteristics based on the above-mentioned natural change characteristics, the above-mentioned medium change characteristics and the above-mentioned character change characteristics.

[0009] Furthermore, the above-mentioned real-time environmental information includes real-time location parameters, real-time time parameters and real-time climate parameters, and the above-mentioned historical environmental information includes historical location parameters, historical time parameters and historical climate parameters; the above-mentioned determination of natural change characteristics based on real-time environmental parameters and historical environmental parameters includes: determining a location change signal based on the change state of the above-mentioned real-time location parameters and the above-mentioned historical location parameters, wherein the level of the above-mentioned location change signal includes at least a first-level migration level, a second-level migration level and a third-level migration level; determining an interval time based on the above-mentioned real-time time parameters and the above-mentioned historical time parameters; determining a weather change signal based on the above-mentioned real-time climate parameters and the above-mentioned historical climate parameters, wherein the above-mentioned weather change signal includes a humidity change signal and a temperature change signal; determining an environmental judgment interval based on the above-mentioned interval time; determining natural change characteristics based on the above-mentioned location change signal and the above-mentioned weather change signal based on the above-mentioned environmental judgment interval.

[0010] Furthermore, the above-mentioned real-time device status includes real-time device type and real-time network type, and the above-mentioned historical device status includes historical device type and historical network type; the above-mentioned determination of medium change characteristics based on real-time device status and historical device status includes: determining the device change coefficient based on the difference between the real-time device type and the historical device type; determining the network change coefficient based on the difference between the real-time network type and the historical network type; determining the above-mentioned medium change characteristics based on the above-mentioned device change coefficient and the above-mentioned network change coefficient.

[0011] Furthermore, the above-mentioned real-time character status includes real-time activity status and real-time emotional status, and the above-mentioned historical device status includes historical activity status and historical emotional status; the above-mentioned determination of character change characteristics based on the real-time character status and the historical character status includes: determining the activity change coefficient based on the difference between the real-time activity status and the historical activity status; determining the emotion change coefficient based on the difference between the real-time emotional state and the historical emotional state; determining the above-mentioned character change characteristics based on the above-mentioned activity change coefficient and the above-mentioned emotion change coefficient.

[0012] Furthermore, the determining of the scene change characteristics based on the natural change characteristics, the medium change characteristics and the character change characteristics includes: ,in, For the characteristics of scene changes, It is a natural variation feature. is the medium change characteristic, For character changes, are their corresponding weight values ​​respectively.

[0013] Furthermore, according to the above-mentioned portrait generation instruction, the user portrait is determined based on the above-mentioned user information and the above-mentioned real-time environmental information, including: determining the situational trend and the emotional trend based on the above-mentioned real-time environmental information; determining the portrait bias factors based on the above-mentioned situational trend and the above-mentioned user information; determining the user portrait based on the above-mentioned emotional trend according to the above-mentioned portrait bias factors.

[0014] Furthermore, the above-mentioned determination of scenario trends and emotional trends based on the above-mentioned real-time environmental information includes: determining scenario trends based on real-time environmental parameters and real-time device status, wherein the scenario types associated with the above-mentioned scenario trends include at least office scenario types, daily scenario types and travel scenario types; determining emotional trends based on real-time character status, wherein the above-mentioned emotional trends include negative emotional types, ordinary emotional types and positive emotional types; the above-mentioned determination of portrait bias factors based on the above-mentioned scenario trends and the above-mentioned user information includes: the above-mentioned user information includes at least basic attribute information, interest preference information and social behavior information; in response to the scenario type associated with the above-mentioned scenario trend being an office scenario type, determining the above-mentioned portrait bias factors based on the above-mentioned basic attribute information; in response to the scenario type associated with the above-mentioned scenario trend being a daily scenario type, determining the above-mentioned portrait bias factors based on the above-mentioned basic attribute information and the above-mentioned interest preference information; in response to the scenario type associated with the above-mentioned scenario trend being a travel scenario type, determining the above-mentioned portrait bias factors based on the above-mentioned basic attribute information and the above-mentioned social behavior information.

[0015] Furthermore, the above-mentioned determination of emotional trends based on real-time character status includes: extracting real-time sentences based on user-generated texts according to a language processing model; marking vocabulary types in the above-mentioned real-time sentences based on the above-mentioned real-time sentences and a preset emotional dictionary; determining type proportions based on the above-mentioned vocabulary types; determining emotional trends based on the above-mentioned type proportions and a preset proportion interval.

[0016] The present invention provides a user portrait generation system for Yunqiwei, including: an acquisition module, used to acquire user information and real-time environmental information; a scenario judgment module, used to determine scenario change characteristics based on the above real-time environmental information and historical environmental information; a portrait generation module, used to obtain a portrait generation instruction in response to the above scenario change characteristics being greater than a change threshold; and also used to determine a user portrait based on the above user information and the above real-time environmental information according to the above portrait generation instruction.

[0017] Compared with the prior art, the user portrait generation method and system of Yunqiwei provided in this specification has at least the following beneficial effects:

[0018] 1. The user portrait generation method can automatically update the user portrait according to the changes in environmental information. The user portrait generation process reduces the need for manual intervention and reduces maintenance costs. At the same time, it speeds up the conversion from data collection to actual application, making the generated user portrait more in line with the user's current situation;

[0019] 2. The user portrait generation method can automatically adjust the user portrait based on the characteristics of scenario changes, so that the system can generate the most suitable user portrait for the user in different scenarios, which is conducive to providing information services to the user in the future;

[0020] 3. By comparing real-time environmental information with historical environmental information, it is possible to accurately capture subtle changes in the situation, thereby more accurately reflecting the user's current status and needs; this dynamic way of generating user portraits makes the user portrait no longer static, but can be updated with changes in time and environment, improving the practicality and accuracy of the user portrait; covering multiple dimensions such as natural change characteristics, medium change characteristics, and character change characteristics, through comprehensive analysis of these characteristics, it is possible to more comprehensively understand the user's environment and status, thereby more accurately portraying the user portrait; this multi-dimensional analysis method makes the user portrait more three-dimensional and rich; when the scenario change characteristics exceed the set threshold, the portrait generation instruction is automatically triggered, and the entire process does not require manual intervention, realizing the intelligent and automatic generation of user portraits; this not only improves efficiency, but also reduces errors caused by human factors. The generated user portrait can provide valuable user information for enterprises or platforms, helping enterprises to better understand user needs, thereby providing more personalized recommendations and services. This can not only improve user experience, but also enhance user stickiness and loyalty. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0022] Figure 1 This is a flow chart of a method for generating a user portrait of Yunqiwei shown in an embodiment of the present application;

[0023] Figure 2 It is a structural block diagram of a user portrait generation system of Yunqiwei shown in one embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly introduces the drawings required for describing the embodiments.

[0025] Please refer to Figure 1 , Figure 1It is a flow chart of a user portrait generation method of Yunqiwei shown in an embodiment of the present application. The user portrait in this embodiment is applied to the intelligent customer service scenario, specifically, to the intelligent customer service scenario of Yunqiwei. By establishing a user portrait, the intelligent customer service can generate a more matching answer. It is understandable that the existing one-to-one manual service requires a large amount of customer service manpower, and the pure manual processing efficiency is low when the consultation volume is large. The business capabilities of account managers vary, and the quality of replies varies. It is impossible to quickly transfer customers, encounter difficult problems, and it is difficult to carry out team collaborative services such as customer upgrade services and problem upgrade processing. It uses environmental information as a setting for triggering instructions, which can obtain more accurate user portraits in real time, and then in the process of intelligent customer application, it can better understand the needs of users, match or recommend more accurate solutions to users, and ultimately improve the user experience. In addition, by generating the latest user portrait (including elements such as account type, session type, customer tag, customer source, etc.), the session routing rules of intelligent customer service can also be customized, and according to different conditions such as account, session type, customer tag, customer source, etc., flexible allocation can be achieved, intelligent identification of customers, allocation of optimal seat reception, and service satisfaction can be improved.

[0026] The user portrait generation method includes at least the following steps:

[0027] Step S100, obtaining user information and real-time environment information.

[0028] It is worth noting that user information may include basic attribute information, interest preference information and social behavior information. Among them, basic attribute information may include age information, gender information, occupation information, etc. Interest preference information may include product interest preference, purchase history and content preference, etc. Social behavior information may include social media activity, social circle type and communication style, etc.

[0029] The real-time environment information in this embodiment includes real-time environment parameters, real-time device status and real-time character status. Among them, the real-time environment parameters include real-time location parameters, real-time time parameters and real-time climate parameters. The real-time device status includes real-time device type and real-time network type. The real-time character status includes real-time activity status and real-time emotional status.

[0030] It is also worth mentioning that the user portrait generation method can automatically update the user portrait according to changes in environmental information. The user portrait generation process reduces the need for manual intervention and reduces maintenance costs. At the same time, it speeds up the conversion from data collection to actual application, making the generated user portrait more in line with the user's current situation. The user portrait generation method can automatically adjust the user portrait based on the characteristics of scenario changes, so that the system can generate the user portrait that best fits the user in different situations, which is conducive to providing customer service information services to the user in the future.

[0031] Step S200, based on the above real-time environment information and historical environment information, determine the scene change characteristics. It can be understood that the historical environment information in this embodiment refers to the environment information when the portrait was last updated.

[0032] Specifically, step S200 may include steps S210 to S240.

[0033] Step S210, determining natural change characteristics based on real-time environmental parameters and historical environmental parameters.

[0034] Specifically, based on the change states of the above-mentioned real-time location parameters and the above-mentioned historical location parameters, a location change signal is determined, wherein the level to which the above-mentioned location change signal belongs includes at least a first-level migration level, a second-level migration level, and a third-level migration level. It is understandable that the user's travel status can be inferred by analyzing the user's mobility pattern (such as short distance, long distance, etc.), and then it is determined whether the user portrait needs to be updated. For example, a first-level migration indicates that the user has undergone a smaller location migration, while a third-level migration indicates that the user has undergone a larger location migration, and at this time, it is more necessary to update the user portrait. Depending on the specific implementation environment, the distance difference between the real-time location and the historical location can be determined based on the GPS signal, and then based on the value of the distance difference, it is determined whether the change state is a first-level migration level, a second-level migration level, or a third-level migration level.

[0035] Based on the above real-time time parameters and the above historical time parameters, the interval time is determined.

[0036] Based on the above-mentioned real-time climate parameters and the above-mentioned historical climate parameters, a weather change signal is determined, wherein the above-mentioned weather change signal includes a humidity change signal and a temperature change signal. It is understandable that by comparing the current humidity with the historical humidity data, humidity changes can be identified. For example, an increase in humidity is associated with rainy weather. At this time, an intelligent customer service with softer speech and communication style after big data training can be recommended. Similarly, the difference between the current temperature and the historical temperature is analyzed. For example, hot weather may be suitable for recommending intelligent customer service with cool background sounds. For another example, cold weather may be more inclined to intelligent customer service with a warmer language style. In this embodiment, the numerical expression of the weather change signal can be:

[0037] ;

[0038] in, A signal of weather change. is the temperature change signal (in this embodiment, it can be represented by the temperature difference), is the humidity change signal (in this embodiment, it can be represented by the humidity difference), is the weight of both.

[0039] According to the above time interval, the environment judgment interval is determined.

[0040] According to the above-mentioned environmental judgment interval, the natural change characteristics are determined based on the above-mentioned position change signal and the above-mentioned weather change signal.

[0041] Specifically, when the interval time is less than 1 hour, the environment judgment interval is determined to be the first preset interval. At this time, when the position change signal is the second migration level or the third migration level, and the weather change signal is greater than the weather change threshold, the natural change feature is determined to be a large change feature. At this time, the natural change feature It can be represented by a value of 10. When the interval time is less than 1 hour, if the position change signal is at the first migration level and the weather change signal is less than or equal to the weather change threshold, the natural change characteristics at this time It can be represented by a value of 1. On the contrary, when the interval is less than 1 hour, the natural variation characteristics in other cases It can be represented by the value 5.

[0042] When the interval time is 1h-3h, the environment judgment interval is determined to be the second preset interval. At this time, if the weather change signal is greater than the weather change threshold, the natural change feature is determined to be a large change feature. At this time, the natural change feature It can be represented by a value of 10. When the interval time is 1h-3h, if the weather change signal is less than or equal to the weather change threshold, and the location change signal is at the third level of migration, then the natural change characteristics at this time are determined. The value is 5. On the contrary, when the interval time is 1h-3h, the natural change characteristics in other cases It can be represented by the value 1.

[0043] When the interval is more than 3 hours, the environment judgment interval is determined to be the third preset interval. At this time, if the weather change signal is greater than the weather change threshold, or the location change signal is the second migration level or the third migration level, the natural change feature is determined to be a large change feature. At this time, the natural change feature It can be represented by a value of 10. When the interval time is more than 3 hours, if the weather change signal is less than or equal to the weather change threshold, and the location change signal is a first-level migration level, then the natural change characteristics at this time are determined. is 1. On the contrary, when the interval is more than 3 hours, the natural variation characteristics in other cases It can be represented by the value 5.

[0044] Step S220, determining medium change characteristics based on the real-time device status and the historical device status.

[0045] Specifically, based on the difference between the real-time device type and the historical device type, the device change coefficient is determined; based on the difference between the real-time network type and the historical network type, the network change coefficient is determined; based on the above-mentioned device change coefficient and the above-mentioned network change coefficient, the above-mentioned medium change characteristics are determined.

[0046] It can be understood that when the real-time device type is different from the historical device type, for example, the real-time device type is a mobile terminal and the historical device type is a host terminal, the device change coefficient is 10; and when the real-time device type is the same as the historical device type, the device change coefficient is 1. Correspondingly, when the real-time network type is different from the historical network type, for example, the real-time network type is 4G and the historical network type is WIFI, the network change coefficient is 10; and when the real-time network type is the same as the historical network type, the network change coefficient is 1. Therefore, based on the average value of the above-mentioned device change coefficient and the above-mentioned network change coefficient, the above-mentioned medium change characteristics can be determined.

[0047] ;

[0048] in, is the medium change characteristic, is the equipment variation coefficient, is the network change coefficient.

[0049] Step S230, determining the character change characteristics based on the real-time character status and the historical character status.

[0050] Specifically, the above-mentioned determination of character change characteristics based on real-time character status and historical character status includes: determining an activity change coefficient based on the difference between real-time activity status and historical activity status; determining an emotion change coefficient based on the difference between real-time emotional status and historical emotional status; determining the above-mentioned character change characteristics based on the above-mentioned activity change coefficient and the above-mentioned emotion change coefficient.

[0051] It is worth noting that the activity state can be a stationary state, a walking state, and a running state. The emotional state can include negative emotions, normal emotions, and positive emotions. At this time, the determination method of the character change characteristics can be similar to the determination method of the medium change characteristics, which will not be repeated here.

[0052] Step S240, determining the scene change characteristics based on the natural change characteristics, the medium change characteristics and the person change characteristics.

[0053] Specifically, ;

[0054] in, For the characteristics of scene changes, It is a natural variation feature. is the medium change characteristic, For the changing characteristics of characters, are their corresponding weight values ​​respectively.

[0055] Step S300, in response to the above-mentioned scene change characteristics being greater than the change threshold, obtaining a portrait generation instruction.

[0056] Step S400, determining a user portrait according to the portrait generation instruction and based on the user information and the real-time environment information.

[0057] Specifically, step S400 may include steps S410 to S430.

[0058] Step S410, determining the situation trend and the emotion trend based on the above real-time environmental information.

[0059] In this embodiment, scenario trends are determined based on real-time environmental parameters and real-time device status, wherein the scenario types associated with the above scenario trends include at least office scenario types, daily scenario types and travel scenario types; and emotional trends are determined based on real-time character status, wherein the above emotional trends include negative emotional types, ordinary emotional types and positive emotional types.

[0060] In this embodiment, real-time sentences can be extracted based on the language processing model and the text generated by the user; the vocabulary types in the real-time sentences can be marked based on the real-time sentences and a preset sentiment dictionary; the type proportions can be determined based on the vocabulary types; and the sentiment trend can be determined based on the type proportions and a preset proportion interval.

[0061] In a specific example, for example, a user posted a real-time sentence on social media: "I feel very tired working in the office today, but chatting with colleagues during lunch made me relax a lot." At this time, the real-time sentence can be extracted based on the existing language processing model: "I feel very tired working in the office today, but chatting with colleagues during lunch made me relax a lot." At this time, the following emotional words are extracted through the emotional dictionary: negative emotion: tired; positive emotion: relaxed. The marking results are: negative: 1; positive: 1. Therefore, the type proportion can be determined: the total number of words is 2 (tired, relaxed), the negative emotion accounts for 50%, and the positive emotion accounts for 50%. The system then determines the emotional trend based on the preset proportion interval. For example, the preset proportion interval is defined as: negative emotion>60% is a negative trend, positive emotion>60% is a positive trend, and the rest is a neutral trend. The following is an explanation of its formula:

[0062] ;

[0063] in, is the proportion of emotional type e, count of words of type e, is the total number of words.

[0064] Therefore, the user image generation method can capture the user's emotional state in real time by analyzing the text generated by the user, provide a dynamic user portrait, and help the system understand the user's emotional changes. Combined with the emotional trend, the user's current user portrait can be generated more accurately, and then the user can be provided with intelligent customer service that matches his or her emotional state, thereby improving the user experience.

[0065] Step S420: determining the most important factors of the portrait based on the scenario trend and the user information.

[0066] Specifically, the above-mentioned determination of the portrait bias factors based on the above-mentioned scenario trends and the above-mentioned user information includes: the above-mentioned user information at least includes basic attribute information, interest preference information and social behavior information; in response to the scenario type associated with the above-mentioned scenario trend being an office scenario type, the above-mentioned portrait bias factors are determined based on the above-mentioned basic attribute information; in response to the scenario type associated with the above-mentioned scenario trend being a daily scenario type, the above-mentioned portrait bias factors are determined based on the above-mentioned basic attribute information and the above-mentioned interest preference information; in response to the scenario type associated with the above-mentioned scenario trend being a travel scenario type, the above-mentioned portrait bias factors are determined based on the above-mentioned basic attribute information and the above-mentioned social behavior information.

[0067] It is understandable that based on the different emphasis on user information in different scenario types (office, daily life, travel), we can obtain a user portrait that better matches the current scenario, which makes it easier to provide more personalized intelligent customer service recommendations and improve user experience. In other words, it can help the subsequent recommendation system to flexibly respond to different scenario trends and automatically adjust the recommendation strategy so that users can match the appropriate intelligent customer service in different environments.

[0068] Step S430, determining the user portrait according to the portrait bias factors and based on the emotional trend.

[0069] Specifically, this step can be, for example, when the user is in an office scenario, based on basic attribute information (for example, the user's age information is 38, the living area is Sichuan, and the occupation information is programmer), determine the portrait bias factors, and the portrait bias factors determined may include: 1) Communication method: question-and-answer type) 2) Communication voice: Mandarin or Sichuan dialect; 3) Communication style: problem-solving type. If the user's emotional trend is a normal trend at this time, the user portrait can be determined according to the above-mentioned portrait bias factors. If the user's emotional trend is a positive trend, the communication style can be adaptively adjusted.

[0070] FIG2 is a structural block diagram of a user portrait generation system of Yunqiwei shown in one embodiment of the present application.

[0071] The user portrait generation system includes an acquisition module, a scenario judgment module and a portrait generation module. The acquisition module is used to acquire user information and real-time environment information; the scenario judgment module is used to determine the scenario change characteristics based on the real-time environment information and historical environment information; the portrait generation module is used to obtain a portrait generation instruction in response to the scenario change characteristics being greater than a change threshold; and is also used to determine the user portrait based on the user information and the real-time environment information according to the portrait generation instruction.

[0072] Finally, it should be understood that the above embodiments in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, as an example and not a limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly introduced and described in this specification.

Claims

1. A method for generating a user portrait of Yunqiwei, characterized in that: include: Obtain user information and real-time environment information; Determining scenario change characteristics based on the real-time environment information and the historical environment information; In response to the scenario change characteristic being greater than a change threshold, obtaining a portrait generation instruction; Determining a user portrait according to the portrait generation instruction and based on the user information and the real-time environment information; The real-time environment information includes real-time environment parameters, real-time device status and real-time character status, and the historical environment information includes historical environment parameters, historical device status and historical character status; The determining of the scenario change characteristics based on the real-time environment information and the historical environment information includes: Determine the characteristics of natural changes based on real-time and historical environmental parameters; Determine the characteristics of media changes based on real-time equipment status and historical equipment status; Determine the character's changing characteristics based on the real-time character status and historical character status; Determining the scene change feature based on the natural change feature, the medium change feature, and the character change feature; The real-time environmental parameters include real-time location parameters, real-time time parameters and real-time climate parameters, and the historical environmental parameters include historical location parameters, historical time parameters and historical climate parameters; The determining of natural change characteristics based on real-time environmental parameters and historical environmental parameters includes: Determine a position change signal based on the change states of the real-time position parameter and the historical position parameter, wherein the level of the position change signal at least includes a primary migration level, a secondary migration level, and a tertiary migration level; Determining an interval time based on the real-time time parameter and the historical time parameter; Determine a weather change signal based on the real-time climate parameter and the historical climate parameter, wherein the weather change signal includes a humidity change signal and a temperature change signal; Determining an environment judgment interval according to the interval time; Determining a natural change feature according to the environmental judgment interval and based on the location change signal and the weather change signal; in, is a weather change signal, t is a temperature change signal, ω is a humidity change signal, and q1 and q2 are weight values ​​of the two; the real-time device state includes a real-time device type and a real-time network type, and the historical device state includes a historical device type and a historical network type; The determining of the medium change characteristics based on the real-time device status and the historical device status includes: Determine the equipment change coefficient based on the difference between the real-time equipment type and the historical equipment type; Determine the network change coefficient based on the difference between the real-time network type and the historical network type; Based on the device change coefficient and the network change coefficient, determining the medium change characteristic; the real-time character status includes a real-time activity status and a real-time emotional status, and the historical device status includes a historical activity status and a historical emotional status; Determining the character change characteristics based on the real-time character status and the historical character status includes: Determine the activity change coefficient based on the difference between the real-time activity status and the historical activity status; Determine the emotion change coefficient based on the difference between the real-time emotion state and the historical emotion state; Determining the character change characteristics based on the activity change coefficient and the emotion change coefficient; The determining the scene change feature based on the natural change feature, the medium change feature and the character change feature includes: x=m·α+n·β+l·γ Wherein, x is a scene change feature, α is a natural change feature, β is a medium change feature, γ is a character change feature, and m, n, and l are their corresponding weight values ​​respectively; according to the portrait generation instruction, based on the user information and the real-time environment information, determining the user portrait includes: Determining situational trends and sentiment trends based on the real-time environmental information; Determining important factors of the portrait based on the scenario trend and the user information; Determining the user portrait based on the emotional trend according to the portrait bias factors; determining the situation trend and the emotional trend based on the real-time environmental information includes: Based on the real-time environmental parameters and the real-time device status, the scenario trend is determined, wherein the scenario types associated with the scenario trend include at least an office scenario type, a daily scenario type, and a travel scenario type; based on the real-time character status, the emotional trend is determined, wherein the emotional trend includes a negative emotional type, a normal emotional type, and a positive emotional type; Determining the important factors of the portrait based on the scenario trend and the user information includes: The user information includes at least basic attribute information, interest preference information and social behavior information; In response to the scenario type associated with the scenario trend being an office scenario type, determining the biased elements of the portrait based on the basic attribute information; In response to the scenario type associated with the scenario trend being a daily scenario type, determining the bias factors of the portrait based on the basic attribute information and the interest preference information; In response to the scenario type associated with the scenario trend being a travel scenario type, determining the important factors of the portrait based on the basic attribute information and the social behavior information; and determining the emotional trend based on the real-time character status, including: Extract real-time sentences based on user-generated text according to the language processing model; Based on the real-time sentence and a preset sentiment dictionary, marking the vocabulary type in the real-time sentence; Based on the vocabulary type, determining the type ratio; Determine the sentiment trend according to the type proportion and the preset proportion interval; Specifically, the type ratio is determined based on the following formula: Among them, R e is the proportion of emotional type e, C e Count the words of type e, C total is the total number of words.

2. A user portrait generation system for Yunqiwei, characterized in that: A method for generating a user portrait of Yunqiwei according to claim 1 is applied, comprising: Acquisition module, used to obtain user information and real-time environment information; A scenario judgment module, used to determine scenario change characteristics based on the real-time environment information and historical environment information; The portrait generation module is used to obtain a portrait generation instruction in response to the scenario change feature being greater than a change threshold; and is also used to determine a user portrait based on the user information and the real-time environment information according to the portrait generation instruction.

Images