A business push method and system based on user profiles

By constructing user profiles and adjusting chromaticity values ​​using neural networks and image chromaticity recognition models, the problem of low accuracy in pushing power service information in existing technologies has been solved, achieving more accurate service information delivery and meeting current user needs.

CN123115507APending Publication Date: 2026-11-17GUANGDONG POWER GRID CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510926367.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-07-07
Publication Date
2026-11-17

Smart Images

  • Figure CN123115507A_ABST
    Figure CN123115507A_ABST
Patent Text Reader

Abstract

This invention discloses a service push method and system based on user profiles. The method includes: acquiring business demand information of a first user; constructing different business feature tags based on the characteristic attributes of the business demand information, and constructing a first user profile using multiple business feature tags; calling a preset neural network to calculate the profile matching degree between the first user profile and several preset profiles; searching for several different business service information in a preset database based on the numerical value of the profile matching degree; filtering target business information from the several business service information and pushing the target business information to the first user. This invention constructs a profile based on the user's current business needs, and then filters business service information that meets the user's needs from the business service information of the current business cycle based on the profile, reducing matching and filtering errors and improving the accuracy of information push.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the technical field of business push processing, and in particular to a business push method and system based on user profiles. Background Technology

[0002] In the field of electricity services, in order to better plan future electricity consumption and improve the user experience, and to meet the actual electricity needs of users, business service information can be pushed to users so that they can understand different electricity business services based on the business service information and then select the electricity service that suits their individual needs.

[0003] To better meet users' actual needs and push relevant business service information, one common method is to first determine the user's historical business information (including user attributes, historical preferences, historical habits, historical behaviors, historical battery consumption, etc.), match the historical business information with multiple services related to the user's needs, and finally push the matched business services to the user for reference.

[0004] However, the commonly used business push methods have the following technical problems: as the number of users increases, the business services provided by the power company need to be constantly updated, and at the same time, users' electricity needs are also changing. Over-reliance on historical data for matching can easily lead to the push of overly simplistic and repetitive business services, resulting in a large deviation between the matched business and the user's current actual needs, low matching accuracy, and consequently, the push of business services failing to meet the user's actual needs. Summary of the Invention

[0005] This invention proposes a service push method and system based on user profiles, which can solve one or more of the above-mentioned technical problems.

[0006] A first aspect of this invention provides a service push method based on user profiles, the method comprising: Obtain the business requirements information of the first user, which includes business characteristic data from multiple dimensions; Different business feature tags are constructed based on the characteristic attributes of the business requirement information, and a first user profile is constructed using multiple business feature tags; A preset neural network is invoked to calculate the profile matching degree between the first user profile and several preset profiles, wherein each preset profile is a personal profile of the second user within the current business update cycle; Based on the numerical values ​​of several profile matching degrees, several different business service information are searched in a preset database, target business information is filtered from the several business service information, and the target business information is pushed to the first user.

[0007] In one possible implementation of the first aspect, constructing multiple different business feature tags based on the feature attributes of the business requirement information includes: The feature attributes corresponding to each business feature data are obtained from the business requirement information, and the feature attributes include numerical values ​​and text. A preset text recognition model is invoked to identify the semantics of the text for each of the aforementioned feature attributes to obtain semantic information, and the corresponding image is searched according to each of the aforementioned semantic information to obtain multiple semantic images; The chromaticity values ​​of the corresponding semantic image are adjusted according to the numerical values ​​of the feature attributes to obtain different business feature labels.

[0008] In one possible implementation of the first aspect, adjusting the chromaticity value of the corresponding semantic image based on the numerical value of the feature attribute to obtain different business feature labels includes: The semantic image is divided into multiple grids according to a preset image ratio, and the chromaticity value of one of the grids is obtained to obtain the chromaticity value to be adjusted. Determine the target chromaticity value based on the chromaticity value to be adjusted; The color of the edge grid of the semantic image is adjusted according to the numerical value of the feature attribute and the target chromaticity value to obtain the business feature label.

[0009] In one possible implementation of the first aspect, constructing the first user profile using multiple business feature tags includes: Semantic relationships determined based on the semantic information of multiple business feature tags; The first user profile is obtained by connecting multiple business feature tags based on the semantic relationships.

[0010] In one possible implementation of the first aspect, the preset neural network is an image colorimetric recognition model; The step of calling a preset neural network to calculate the matching degree between the first user profile and several preset profiles includes: The image chromaticity recognition model is invoked to identify the chromaticity value of the first user profile to obtain the tag chromaticity value, and the image chromaticity recognition model is invoked to identify the chromaticity value of each preset profile to obtain the profile chromaticity value. The tag chromaticity value is the chromaticity value of the image edge color of the business feature tag of the first user profile, and the profile chromaticity value is the chromaticity value of the image edge color of the business feature tag of the preset profile. The image matching degree is calculated based on the label chromaticity value and the image chromaticity value.

[0011] In one possible implementation of the first aspect, calculating the image matching degree based on the tag chromaticity value and the image chromaticity value includes: The label weight value is obtained by finding the corresponding weight based on the label color value, and the image weight value is obtained by finding the weight based on the image color value. The total tag value of the first user profile is calculated using several of the tag weight values, and the total profile value of each preset profile is calculated using several of the profile weight values. The image matching degree is obtained by calculating the ratio of the total tag value to the total image value.

[0012] In one possible implementation of the first aspect, after the step of constructing the first user profile using the multiple business feature tags, the method further includes: Determine the number of preset images stored in the database; If the quantity value is less than the preset quantity threshold, then multiple pieces of information to be used are obtained, wherein the information to be used is business information updated in the current business cycle or business information viewed and retained by the first user in the previous business cycle. A preset profile is constructed using the information to be used.

[0013] A second aspect of this invention provides a service push system based on user profiles, the system comprising: The acquisition module is used to acquire the business requirement information of the first user, which includes business feature data in multiple different dimensions. The construction module is used to construct different business feature tags based on the feature attributes of the business requirement information, and to construct a first user profile using multiple business feature tags; The matching module is used to call a preset neural network to calculate the matching degree between the first user profile and several preset profiles, wherein each preset profile is a personal profile of the second user in the current business update cycle. The push module is used to search for several different business service information in a preset database based on the numerical values ​​of several profile matching degrees, filter target business information from the several business service information, and push the target business information to the first user.

[0014] In one possible implementation of the second aspect, constructing multiple different business feature tags based on the feature attributes of the business requirement information includes: The feature attributes corresponding to each business feature data are obtained from the business requirement information, and the feature attributes include numerical values ​​and text. A preset text recognition model is invoked to identify the semantics of the text for each of the aforementioned feature attributes to obtain semantic information, and the corresponding image is searched according to each of the aforementioned semantic information to obtain multiple semantic images; The chromaticity values ​​of the corresponding semantic image are adjusted according to the numerical values ​​of the feature attributes to obtain different business feature labels.

[0015] In one possible implementation of the second aspect, adjusting the chromaticity value of the corresponding semantic image based on the numerical value of the feature attribute to obtain different business feature labels includes: The semantic image is divided into multiple grids according to a preset image ratio, and the chromaticity value of one of the grids is obtained to obtain the chromaticity value to be adjusted. Determine the target chromaticity value based on the chromaticity value to be adjusted; The color of the edge grid of the semantic image is adjusted according to the numerical value of the feature attribute and the target chromaticity value to obtain the business feature label.

[0016] In one possible implementation of the second aspect, constructing the first user profile using multiple business feature tags includes: Semantic relationships determined based on the semantic information of multiple business feature tags; The first user profile is obtained by connecting multiple business feature tags based on the semantic relationships.

[0017] In one possible implementation of the second aspect, the preset neural network is an image colorimetric recognition model; The step of calling a preset neural network to calculate the matching degree between the first user profile and several preset profiles includes: The image chromaticity recognition model is invoked to identify the chromaticity value of the first user profile to obtain the tag chromaticity value, and the image chromaticity recognition model is invoked to identify the chromaticity value of each preset profile to obtain the profile chromaticity value. The tag chromaticity value is the chromaticity value of the image edge color of the business feature tag of the first user profile, and the profile chromaticity value is the chromaticity value of the image edge color of the business feature tag of the preset profile. The image matching degree is calculated based on the label chromaticity value and the image chromaticity value.

[0018] In one possible implementation of the second aspect, calculating the image matching degree based on the tag chromaticity value and the image chromaticity value includes: The label weight value is obtained by finding the corresponding weight based on the label color value, and the image weight value is obtained by finding the weight based on the image color value. The total tag value of the first user profile is calculated using several of the tag weight values, and the total profile value of each preset profile is calculated using several of the profile weight values. The image matching degree is obtained by calculating the ratio of the total tag value to the total image value.

[0019] In one possible implementation of the second aspect, the system further includes: The determining module is used to determine the number of preset profiles stored in the database after the step of constructing the first user profile using multiple business feature tags; The pending information module is used to acquire multiple pending information if the quantity value is less than a preset quantity threshold, wherein the pending information is business information updated in the current business cycle or business information viewed and retained by the first user in the previous business cycle. The preset profile module is used to construct a preset profile using the information to be used.

[0020] Compared to existing technologies, the user profile-based service push method and system provided in this invention offer the following advantages: This invention can acquire the service needs information of a first user; construct different service feature tags based on the characteristic attributes of the service needs information, and construct a first user profile using multiple service feature tags; call a preset neural network to calculate the profile matching degree between the first user profile and several preset profiles; search for several different service information items in a preset database based on the numerical values ​​of the profile matching degrees, filter target service information from the several service information items, and push the target service information to the first user. This invention constructs a profile based on the user's current service needs, and then filters service information that meets the user's needs from the service information of the current business cycle based on that profile, reducing matching and filtering errors to improve the accuracy of information push. Furthermore, constructing a profile based on the user's actual characteristics reflects the user's actual needs, further improving the accuracy of matching and filtering. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating a user profile-based service push method according to an embodiment of the present invention. Figure 2 This is a schematic diagram of an existing user profile provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a service push system based on user profiles provided in an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] In the field of electricity services, in order to better plan future electricity consumption and improve the user experience, and to meet the actual electricity needs of users, business service information can be pushed to users so that they can understand different electricity business services based on the business service information and then select the electricity service that suits their individual needs.

[0024] To better meet users' actual needs and push relevant business service information, one common method is to first determine the user's historical business information (including user attributes, historical preferences, historical habits, historical behaviors, historical battery consumption, etc.), match the historical business information with multiple services related to the user's needs, and finally push the matched business services to the user for reference.

[0025] However, the commonly used business push methods have the following technical problems: as the number of users increases, the business services provided by the power company need to be constantly updated, and at the same time, users' electricity needs are also changing. Over-reliance on historical data for matching can easily lead to the push of overly simplistic and repetitive business services, resulting in a large deviation between the matched business and the user's current actual needs, low matching accuracy, and consequently, the push of business services failing to meet the user's actual needs.

[0026] To address the aforementioned issues, the following specific embodiments will provide a detailed introduction and explanation of a user profile-based service push method and system provided in this application.

[0027] Reference Figure 1 The diagram illustrates a flowchart of a user profile-based service push method according to an embodiment of the present invention.

[0028] In one embodiment, the method is applicable to an enterprise's data management platform or data processing platform. The platform can be a customer service platform that promotes the interconnection and interoperability of the company's unified user system with the power grid's online, marketing management, power grid management, integrated energy, industrial finance, digital government, and other systems. It can create a set of data operation and business operation tools such as demand management, user management, relationship management, group profiling, points operation, and channel operation, providing digital support tools for business and management personnel at all levels of the platform. This supports the marketing team in providing users with high-quality power supply services, energy value-added services, and energy ecosystem services, and promotes the development of power supply services.

[0029] The customer service platform can manage users, based on a unified user system, to collect, organize, store, modify, and retrieve user profiles; it can aggregate data from different dimensions such as basic user attributes, electricity usage behavior, and consumption habits to establish a tagging and profiling system and build a panoramic view of users; based on profiles and group characteristics, it can enhance user care services and improve user experience through points management and personalized services.

[0030] As an example, the user profile-based service push method may include: S11. Obtain the business requirement information of the first user, wherein the business requirement information includes business feature data of multiple different dimensions.

[0031] In one embodiment, when it is necessary to push business service information to the first user, the user's business demand information can be obtained first. This business demand information may be related to the personal business demand entered by the first user when inquiring about the required business service.

[0032] In order to comprehensively determine user needs, business requirement information can include multiple business characteristic data, and these multiple business characteristic data are data from different dimensions.

[0033] For example, business requirement information may include data from different dimensions such as user basic attributes, electricity consumption behavior, consumption habits, electricity consumption location, electricity consumption scenario, acceptable amount for business, and payment method.

[0034] By using data from different dimensions, we can comprehensively determine the user's actual needs. Then, we can filter and match services based on the user's actual needs, reduce the deviation between the matched services and the user's current actual needs, improve the accuracy of the matching, and thus meet the user's actual needs.

[0035] S12. Construct different business feature tags based on the characteristic attributes of the business requirement information, and construct a first user profile using multiple business feature tags.

[0036] In one embodiment, if matching is performed one by one according to different dimensions of business requirement information, the amount of data to be processed is large and the processing efficiency is low; moreover, each piece of data is independent of the others, and if matching is performed one by one according to different dimensions of data, it is difficult to determine the relationship between the data.

[0037] In order to integrate data from multiple different dimensions, in one embodiment, the feature attributes of each business feature data in the business requirement information can be obtained, and then a business feature label for each business feature data can be constructed based on the feature attributes of each business feature data, thereby obtaining multiple different business feature labels.

[0038] After obtaining multiple different business feature tags, the corresponding profile of the first user can be constructed using these multiple business feature tags, thus obtaining the first user profile.

[0039] When building a user profile, if a single text is used as a tag, the amount of text in the profile will increase as the number of tags increases. This not only makes the profile complex, but also makes it difficult to distinguish the primary and secondary tags and determine the user's needs through the profile.

[0040] In an optional embodiment, to distinguish and highlight different business feature tags, as an example, constructing multiple different business feature tags based on the feature attributes of the business requirement information may include the following sub-steps: S21. Obtain the feature attributes corresponding to each business feature data from the business requirement information. The feature attributes include numerical values ​​and text.

[0041] S22. Call the preset text recognition model to recognize the semantics of the text of each of the aforementioned feature attributes to obtain semantic information, and search for the corresponding image according to each of the aforementioned semantic information to obtain multiple semantic images.

[0042] S23. Adjust the chromaticity value of the corresponding semantic image according to the value of the feature attribute to obtain different business feature labels.

[0043] In one embodiment, feature attributes corresponding to each business feature data of the business requirement information can be read, and the feature attributes include numerical values ​​and text.

[0044] For example, business characteristic data could be electricity consumption behavior, where the text for electricity consumption behavior could be "industrial electricity consumption," and the numerical value for electricity consumption behavior could be 100kW. Another example is electricity consumption location, where the text for electricity consumption location could be "xx factory area," and the numerical value for electricity consumption location could be the address of xx factory area. Yet another example is electricity consumption scenario, where the text for electricity consumption scenario could be "daytime electricity consumption," and the numerical value for electricity consumption scenario could be 50kW.

[0045] It should be noted that some business feature data may only contain text and no numerical values. For example, the business feature data may be payment method, with the text stating "xx online" or "xx offline," without a specific numerical value. The number of characters in the text of the business feature data can be used as the numerical value of its feature attribute; in this case, the feature attribute value for payment method would be 4.

[0046] Some business characteristic data may only contain numerical values ​​without text. For example, the business characteristic data might be the acceptable amount for a business, with a numerical value of xx yuan. In this case, the text of the business characteristic data can be used as the text for its characteristic attributes, such as directly using the acceptable amount for a business as the text.

[0047] By acquiring and calling a preset text recognition model to identify the semantics of the text for each feature attribute, the semantic information of each business feature data can be determined. The preset text recognition model can be a conventional text recognition model in the field; this invention does not limit it, and it can be adjusted according to actual needs.

[0048] Finally, the corresponding image can be found for each piece of semantic information to determine the semantic image corresponding to each business feature data. Since there are multiple business feature data, multiple semantic images can be obtained in the end.

[0049] In one approach, an image-text matching model can be pre-built, and then the image-text matching model can be called to find the corresponding image based on semantic information.

[0050] The training method for the image-text matching model can be to pre-acquire image samples and text samples, and then extract the global and local representations of the image samples; and extract the global and local representations of the text samples; and train the pre-constructed matching model based on the extracted global and local representations, thereby obtaining the image-text matching model.

[0051] Specifically, the matching model maps the global representations of image samples and text samples to a specified semantic space, calculates the similarity of the global representations between heterogeneous sample pairs composed of image samples and text samples; maps the local representations of image samples and text samples to a specified semantic space, calculates the similarity of the local representations between heterogeneous sample pairs; and determines the matching degree of heterogeneous sample pairs by weighted summation based on the preset weights of the similarity of the global representations and the preset weights of the similarity of the local representations.

[0052] In one embodiment, for the input image, a local image CNN (Convolutional Neural Network) is used to extract the local representation of the image; and a global image CNN is used to extract the global representation of the image. For the input text, a local statement encoder is used to extract the local representation of the text, and a global statement encoder is used to extract the global representation of the text. After extracting the local and global representations of the image and text, the local deep matching representation maps the local representations of the image and text to a specified semantic space, and determines the similarity of the local representations between heterogeneous sample pairs composed of the image and text in the specified semantic space; similarly, the global deep matching representation maps the global representations of the image and text to a specified semantic space, and determines the similarity of the global representations between heterogeneous sample pairs composed of the image and text in the specified semantic space. Then, through model ensemble, based on the preset weights of the similarity of the global representations and the preset weights of the similarity of the local representations, a weighted summation method is used to determine the matching degree of the heterogeneous sample pairs.

[0053] In practice, after receiving the semantic information of the text, the semantic information can be used as a reference sample. Then, the global and local representations of the reference sample are extracted. The global and local representations of the semantic information are input into the image-text matching model so that the image-text matching model can calculate the matching degree between the reference sample and the corresponding material. Here, the reference sample is text, and the corresponding material is an image. The image-text matching model can determine the matching degree between the image and the text based on the global and local representations. The corresponding material with a matching degree greater than a specified matching degree is selected as the material that matches the reference sample, and thus a semantic image can be obtained.

[0054] In an optional embodiment, an image database can be pre-built, where each semantic can correspond to one or more images, and each match can be made by matching the corresponding image from the image database.

[0055] Since each business feature data has corresponding semantic information, each business feature data corresponds to a semantic image; and since each business feature data is different, the corresponding semantic images are also different from each other, thus enabling the differentiation of different business feature data through images and achieving label differentiation.

[0056] Finally, to highlight each semantic image, the chroma value of the corresponding semantic image can be adjusted according to the numerical value of the feature attribute. Finally, using the image as a label, different business feature labels can be obtained.

[0057] For example, the chroma values ​​of the semantic image of business feature data a can be adjusted using the numerical values ​​of the feature attributes of business feature data a; the chroma values ​​of the semantic image of business feature data b can be adjusted using the numerical values ​​of the feature attributes of business feature data b.

[0058] Since each semantic term corresponds to one or more images, the number of images that can be matched for each semantic term is limited. Furthermore, if different users have the same textual features, the images corresponding to their business feature tags will also be similar. By adjusting the chroma values, it's possible to quickly determine whether two users have the same or similar needs. Simultaneously, the tags can distinguish whether their required business functions are the same, facilitating subsequent matching.

[0059] Reference Figure 2 The diagram illustrates an existing user profile provided by an embodiment of the present invention.

[0060] Traditional user profile tags are usually text-based. With many tags, the resulting profile becomes complex and disorganized, making it difficult for users to distinguish between different tags.

[0061] This invention uses images to distinguish features by tags. Compared to files, images can more intuitively highlight characteristic attributes, thus facilitating subsequent business analysis and matching.

[0062] In one embodiment, different images have different colors, and the color distribution is uneven. To further highlight the image of each business feature label, as an example, adjusting the chromaticity value of the corresponding semantic image according to the value of the feature attribute to obtain different business feature labels may include the following sub-steps: S231. Divide the semantic image into multiple grids according to a preset image ratio, and obtain the chroma value of one of the grids to obtain the chroma value to be adjusted, wherein the grid corresponding to the chroma value to be adjusted is any one of the non-edge grids of the semantic image.

[0063] S232. Determine the target chromaticity value based on the chromaticity value to be adjusted.

[0064] S233. Adjust the color of the edge grid of the semantic image according to the value of the feature attribute and the target chromaticity value to obtain the business feature label.

[0065] In one embodiment, a preset image ratio value can be obtained. This preset image ratio value can be the ratio of the grid area to the image area, and the value can be 20%, 10%, or 5%, etc. The semantic image is divided into multiple grids according to the preset image ratio value. For example, if the preset image ratio value is 5%, then the ratio of the grid area to the semantic image area is 5%. The semantic image can be divided into 20 grids.

[0066] In another optional embodiment, the preset image ratio value can be a grid number value, such as 4, 16, or 64. For example, a preset image ratio value of 16 can divide the semantic image into 16 grids.

[0067] The specific division method can be adjusted according to actual needs, and different images can also be divided using different preset image ratio values. This invention does not limit this.

[0068] After dividing the grid into multiple grids, the chromaticity value of one of the grids can be obtained to get the chromaticity value to be adjusted.

[0069] In practice, a conventional pixel recognition model or color recognition model can be used to identify the color of one of the grids and determine the corresponding chromaticity value, thereby obtaining the chromaticity value to be adjusted.

[0070] When selecting a grid, you can prioritize any one of the non-edge grids in the semantic image. For example, if the semantic image is divided into 16 grids, you can choose any one of the four middle grids instead of selecting the first grid, the first grid, the last grid, and the last grid. Then, obtain the chromaticity value of that grid to get the chromaticity value to be adjusted.

[0071] For example, if a semantic image is divided into four grids, and all four grids are edges of the image, then any one of the four can be selected.

[0072] In another alternative embodiment, in order to further highlight the colors of the image, the chromaticity value of each grid color can be identified, and the area value of the color corresponding to each chromaticity value in the semantic image can be determined, thereby filtering out the chromaticity value with the largest area value to obtain the chromaticity value to be adjusted.

[0073] After determining the chromaticity value to be adjusted, the target chromaticity value can be determined based on it. In one operation mode, to improve the label display effect, the color corresponding to the chromaticity value to be adjusted can be determined, then the complementary color of that color can be found, and the chromaticity value of the complementary color can be determined to obtain the target chromaticity value.

[0074] Finally, the color of the edge grid of the semantic image can be adjusted according to the numerical value of the feature attribute and the target chromaticity value to obtain the business feature label.

[0075] In one adjustment method, the number of edge grids to be adjusted can be determined based on the value of the feature attribute. If the feature attribute value is 10 and the semantic image has 12 edge grids, 10 edge grids can be adjusted, specifically by adjusting the color of the semantic image's edge grids to the color corresponding to the target chromaticity value. If the feature attribute value is greater than the number of edge grids in the semantic image, then the color of all edge grids in the semantic image is adjusted to the color corresponding to the target chromaticity value.

[0076] In another optional adjustment method, after adjusting the color of the edge grid of the semantic image to the color corresponding to the target chromaticity value, the transparency of each edge grid can be adjusted according to the magnitude of the feature attribute value. In one operation, the comparison result of each feature attribute value with its corresponding preset threshold and the ratio of each feature attribute value to its corresponding preset threshold can be determined. If the feature attribute value is greater than the corresponding preset threshold, a positive adjustment is made; otherwise, a negative adjustment is made. The adjustment operation can be based on the ratio of each feature attribute value to its corresponding preset threshold. It should be noted that the preset transparency of each grid is 50%, i.e., a semi-transparent state. The adjustment magnitude is the product of the ratio and the adjusted interval value.

[0077] For example, if the feature attribute value is 50, the corresponding preset threshold is 10, and the adjustment range is 5, then 50 is greater than 10. Calculating the ratio, we get 50 / 10 = 5. We can adjust the mesh transparency in a positive direction, with an adjustment increment of 5 * 5 = 25, thus increasing the transparency from 50 to 75. As another example, if the feature attribute value is 50, the corresponding preset threshold is 100, and the adjustment range is 10, then 50 is less than 100. Calculating the ratio, we get 50 / 100 = 0.5. We can adjust the mesh transparency in a negative direction, with an adjustment increment of 10 * 0.5 = 5, thus decreasing the transparency from 50 to 45.

[0078] It should be noted that once the transparency is adjusted to 100% or 0%, no further adjustments will be made. All the above values ​​can be adjusted according to actual needs, and this invention does not impose any limitations on them.

[0079] Furthermore, the adjustment method can be adapted to meet the display needs of different users, and this invention does not limit it.

[0080] Since the colors corresponding to the target chromaticity value and the colors corresponding to the chromaticity value to be adjusted are complementary colors, the colors of the middle and edge positions of the business feature label can be different. Furthermore, the two colors have significant contrasting characteristics, including hue, brightness, and saturation, thus creating a strong contrast effect to highlight the entire business feature label.

[0081] Reference Figure 2 In existing user profile construction techniques, tags are generally text-based, and each tag is independent of the others. In one embodiment, to determine the association between tags, as an example, constructing a first user profile using multiple business feature tags may include the following sub-steps: S31. Semantic relationships determined based on the semantic information of multiple business feature tags.

[0082] S32. Based on the semantic relationship, connect multiple business feature tags to obtain a first user profile.

[0083] In one embodiment, the present invention can pre-establish a user's center point or a user's portrait, and connect each business feature tag to the center point to form a basic profile.

[0084] Based on this, the semantic information of each business feature tag can be determined, and the semantic relationship between each business feature tag can be determined according to the semantic information of each business feature tag. If two business feature tags have a corresponding semantic relationship, the two business feature tags can be connected. If two business feature tags do not have a corresponding semantic relationship, they can not be connected.

[0085] By connecting various business feature tags, a first user profile can be constructed.

[0086] In practice, a conventional text matching model can be called to calculate the semantic correlation between two business feature tags. If the correlation is greater than a preset value, it is determined that the two business feature tags have a corresponding semantic relationship; otherwise, it is determined that the two business feature tags do not have a corresponding semantic relationship.

[0087] S13. Call a preset neural network to calculate the matching degree between the first user profile and several preset profiles, wherein each preset profile is a personal profile of the second user within the current business update cycle.

[0088] After constructing the first user profile corresponding to the first user, a preset neural network can be invoked to calculate the profile matching degree between the first user profile and several preset profiles. Among them, the preset profiles are the personal profiles of other second users; and the preset profiles are the profiles updated and constructed within the current business update cycle.

[0089] Each service update cycle is the cycle for updating power grid service offerings, for example, one or two years. Because user electricity demand varies annually, and power supply also varies due to equipment upgrades, mergers, and overlaps, the available power services are adjusted annually. Therefore, each time a profile matching degree calculation is performed, it is best to use profiles from the same service update cycle for comparison. This ensures a closer match to the currently available services, avoiding matching discontinued or canceled services and improving the accuracy of service matching. In an optional implementation, the service update cycle can be one year.

[0090] During the calculation, the first user profile is matched with each preset profile once. Since there are several preset profiles, there are also several matching profiles to be calculated.

[0091] In one embodiment, since the first user profile is composed of multiple business feature tags, and each business feature tag is an image containing specific contrasting colors, in order to calculate the matching degree of the image, the preset neural network is, for example, an image color recognition model.

[0092] In one embodiment, the step of calling a preset neural network to calculate the matching degree between the first user profile and several preset profiles may include the following sub-steps: S131. Call the image chromaticity recognition model to identify the chromaticity value of the first user profile to obtain the tag chromaticity value and call the image chromaticity recognition model to identify the chromaticity value of each preset profile to obtain the profile chromaticity value, wherein the tag chromaticity value is the chromaticity value of the image edge color of the business feature label of the first user profile, and the profile chromaticity value is the chromaticity value of the image edge color of the business feature label of the preset profile.

[0093] S132. Calculate the image matching degree based on the label color value and the image color value.

[0094] As explained above, each profile is composed of business feature labels, and each business feature label is composed of an image corresponding to the text of the feature attribute and a specific contrast color of the image edge, in order to target different specific contrast colors.

[0095] In one embodiment, the image colorimetric recognition model can be invoked to identify the colorimetric value of the first user profile to obtain the tag colorimetric value, or the image colorimetric recognition model can be invoked to identify the colorimetric value of each preset profile to obtain the profile colorimetric value.

[0096] Specifically, the image colorimetric recognition model can be invoked to identify the image edge color of each business feature label of the first user profile, thereby obtaining several label colorimetric values. Similarly, the image colorimetric recognition model can be invoked to identify the image edge color of each business feature label of each preset profile, thereby obtaining several profile colorimetric values.

[0097] Finally, the matching degree between the first user profile and the preset profile can be calculated based on several tag color values ​​and several profile color values.

[0098] In one embodiment, calculating the image matching degree based on the tag chromaticity value and the image chromaticity value may include the following sub-steps: S1321. Obtain the label weight value by finding the corresponding weight based on the label chromaticity value, and obtain the image weight value by finding the weight based on the image chromaticity value.

[0099] S1322. The total tag value of the first user profile is calculated using several of the tag weight values, and the total profile value of each preset profile is calculated using several of the profile weight values.

[0100] S1323. Calculate the ratio of the total tag value to the total image value to obtain the image matching degree.

[0101] In one optional operation mode, the corresponding weight can be found based on the chroma value of each business feature tag to obtain the tag weight value. Correspondingly, several tag weight values ​​can be found. Similarly, the image weight value can also be calculated based on the weight found based on the image chroma value.

[0102] Since the chromaticity values ​​of an image in the RGB color space range from 0 to 255, a weight can be preset for each chromaticity value. There are 256 preset weights from 0 to 255. The corresponding weight can be found for each chromaticity value to obtain the corresponding weight value.

[0103] Next, the total tag value of the first user profile can be calculated using several tag weight values. Similarly, the total profile value of each preset profile can be calculated using several profile weight values.

[0104] In one calculation method, the total value of the labels can be calculated as follows: ; In the above formula, c is the total value of the labels, and A i Let A be the chromaticity value of the i-th business feature label. k Let A be the chromaticity value of the k-th business feature label. l P is the chromaticity value of the l-th business feature label. i Let P be the label weight value of the i-th business feature label. k P represents the label weight value of the k-th business feature label. l This is the label weight value for the l-th business feature label.

[0105] Similarly, the total value of the portrait can be calculated using the formula described above.

[0106] After calculating the total tag value and the total image value, the ratio of the total tag value to the total image value can be calculated to obtain the image matching degree.

[0107] It should be noted that the above method may only compare the chroma values ​​of each business feature tag, ignoring the text of the business feature tag and its corresponding image. However, matching by chroma value may yield a profile of a second user who is similar to the current user and has similar color characteristics. This allows the user to find the business service information of the second user for reference, enabling the user to view more diverse business service information.

[0108] Meanwhile, in order to calculate the matching degree by combining the image corresponding to the text and the color value of the image, so as to meet the actual needs of users, in an optional embodiment, several images to be matched can be selected from several preset images. These images to be matched are preset images that contain the feature attribute text corresponding to the first user image.

[0109] Based on this, the matching degree between the first user profile and several profiles to be matched is calculated according to the above operation method. Then, business filtering is performed based on the matching degree to obtain business service information that meets the actual needs of users.

[0110] S14. Based on the numerical values ​​of several profile matching degrees, search for several different business service information in a preset database, filter target business information from the several business service information, and push the target business information to the first user.

[0111] In one embodiment, after calculating several profile matching scores, the business service information of the second user corresponding to each profile matching score is retrieved. This business service information consists of the business service items selected by the second user during the current business update cycle. Target business information can be filtered from the several business service information items based on the numerical values ​​of the several profile matching scores and then pushed to the first user.

[0112] Specifically, the system can filter out business service information with the highest profile matching score, use that business service information as the target business information, and push it to the first user. Alternatively, it can filter out business service information with a profile matching score greater than a preset threshold, use that business service information as the target business information, and push it to the first user.

[0113] In an optional embodiment, the first user may be either the first user or the second user in the current service update cycle. For example, the service update date is April 1st of each year, and a service update cycle extends from April 1st of the current year to April 1st of the following year.

[0114] The first user completed their service for the previous service update cycle on March 30th. The first user needs to reselect their service for the next year on April 1st. At this point, the first user is the first to apply for this service cycle, and the number of second users used for reference is small or nonexistent. To construct a preset profile for subsequent matching, in one embodiment, after the step of constructing the first user profile using multiple service feature tags, the method may further include the following steps: S41. Determine the number of preset images stored in the database.

[0115] S42. If the quantity value is less than a preset quantity threshold, then multiple pieces of information to be used are obtained, wherein the information to be used is business information updated in the current business cycle or business information viewed and retained by the first user in the previous business cycle.

[0116] S43. Construct a preset profile using the information to be used.

[0117] Specifically, it can retrieve the number of preset profiles stored in the current database for this business cycle.

[0118] If the quantity is less than the preset threshold, it indicates that the number of preset profiles is small, and a preset profile needs to be built. Multiple pieces of information to be used can be obtained. Information to be used can be updated business information within the current business cycle, or business information viewed and retained by the first user in the previous business cycle. Specifically, it can be service items that provide business services to users.

[0119] Next, a preset profile can be constructed using the information to be used. The specific construction method can refer to the above embodiment. For example, business feature information is extracted from the information to be used and tags are constructed, and then a profile is constructed based on the tags to obtain the preset profile.

[0120] Finally, preset profiles can be stored in the database for users to match and use.

[0121] It should be noted that as the number of first users gradually increases, the first user profile of the first user can be used as a preset profile for subsequent users. However, to avoid consuming storage space and affecting matching accuracy by using the information to be used to build preset profiles, these preset profiles built from the information to be used can be deleted.

[0122] In this embodiment, the present invention provides a service push method based on user profiles. Its advantages are as follows: The present invention can obtain the service needs information of a first user; construct different service feature tags based on the characteristic attributes of the service needs information, and construct a first user profile using multiple service feature tags; call a preset neural network to calculate the profile matching degree between the first user profile and several preset profiles; search for several different service information in a preset database based on the numerical value of the profile matching degree, filter target service information from the several service information, and push the target service information to the first user. The present invention constructs a profile based on the user's current service needs, and then filters service information that meets the user's needs from the service information of the current service cycle based on the profile, reducing matching and filtering errors to improve the accuracy of information push. Moreover, constructing a profile based on the user's actual characteristics can reflect the user's actual needs, further improving the accuracy of matching and filtering.

[0123] This invention also provides a service push system based on user profiles, see [link to relevant documentation]. Figure 3 The diagram shows a schematic representation of a user profile-based service push system according to an embodiment of the present invention.

[0124] As an example, the user profile-based service push system may include: The acquisition module 301 is used to acquire the business requirement information of the first user, the business requirement information including business feature data of multiple different dimensions; The construction module 302 is used to construct different business feature tags based on the feature attributes of the business requirement information, and to construct a first user profile using multiple business feature tags; The matching module 303 is used to call a preset neural network to calculate the matching degree between the first user profile and several preset profiles, wherein each preset profile is a personal profile of the second user in the current business update cycle. The push module 304 is used to search for several different business service information in a preset database based on the numerical values ​​of several profile matching degrees, filter target business information from the several business service information, and push the target business information to the first user.

[0125] Optionally, the step of constructing multiple different business feature tags based on the feature attributes of the business requirement information includes: The feature attributes corresponding to each business feature data are obtained from the business requirement information, and the feature attributes include numerical values ​​and text. A preset text recognition model is invoked to identify the semantics of the text for each of the aforementioned feature attributes to obtain semantic information, and the corresponding image is searched according to each of the aforementioned semantic information to obtain multiple semantic images; The chromaticity values ​​of the corresponding semantic image are adjusted according to the numerical values ​​of the feature attributes to obtain different business feature labels.

[0126] Optionally, adjusting the chroma value of the corresponding semantic image based on the numerical value of the feature attribute to obtain different business feature labels includes: The semantic image is divided into multiple grids according to a preset image ratio, and the chromaticity value of one of the grids is obtained to obtain the chromaticity value to be adjusted. Determine the target chromaticity value based on the chromaticity value to be adjusted; The color of the edge grid of the semantic image is adjusted according to the numerical value of the feature attribute and the target chromaticity value to obtain the business feature label.

[0127] Optionally, constructing the first user profile using multiple business feature tags includes: Semantic relationships determined based on the semantic information of multiple business feature tags; The first user profile is obtained by connecting multiple business feature tags based on the semantic relationships.

[0128] Optionally, the preset neural network is an image colorimetric recognition model; The step of calling a preset neural network to calculate the matching degree between the first user profile and several preset profiles includes: The image chromaticity recognition model is invoked to identify the chromaticity value of the first user profile to obtain the tag chromaticity value, and the image chromaticity recognition model is invoked to identify the chromaticity value of each preset profile to obtain the profile chromaticity value. The tag chromaticity value is the chromaticity value of the image edge color of the business feature tag of the first user profile, and the profile chromaticity value is the chromaticity value of the image edge color of the business feature tag of the preset profile. The image matching degree is calculated based on the label chromaticity value and the image chromaticity value.

[0129] Optionally, calculating the image matching degree based on the tag chromaticity value and the image chromaticity value includes: The label weight value is obtained by finding the corresponding weight based on the label color value, and the image weight value is obtained by finding the weight based on the image color value. The total tag value of the first user profile is calculated using several of the tag weight values, and the total profile value of each preset profile is calculated using several of the profile weight values. The image matching degree is obtained by calculating the ratio of the total tag value to the total image value.

[0130] Optionally, the system further includes: The determining module is used to determine the number of preset profiles stored in the database after the step of constructing the first user profile using multiple business feature tags; The pending information module is used to acquire multiple pending information if the quantity value is less than a preset quantity threshold, wherein the pending information is business information updated in the current business cycle or business information viewed and retained by the first user in the previous business cycle. The preset profile module is used to construct a preset profile using the information to be used.

[0131] Those skilled in the art will understand that, for ease of description and brevity, the specific working process of the system described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0132] Furthermore, this application also provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the user profile-based service push method as described in the above embodiments.

[0133] Furthermore, embodiments of this application also provide a computer-readable storage medium storing a computer-executable program, which is used to cause a computer to execute the user profile-based service push method as described in the above embodiments.

[0134] In the description of the embodiments of the present invention, it should be noted that the terms "above," "below," etc., indicating the orientation or positional relationship are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. When an element such as a layer, region, or substrate is referred to as being "above" or "on top of" another element, it may be directly on the other element, or there may be an intermediate element. Conversely, when an element is referred to as being "directly on" or "above" another element, there is no intermediate element. It should also be understood that when an element is referred to as being "below" or "under" another element, it may be directly below or under the other element, or there may be an intermediate element. Conversely, when an element is referred to as being "directly below" or "under" another element, there is no intermediate element. Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0135] Those skilled in the art will understand that embodiments of this application may also include computer program products. Therefore, this application may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0136] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), devices, and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0137] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0138] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0139] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A service push method based on user profiling, characterized in that, The method includes: Obtain the business requirements information of the first user, which includes business characteristic data from multiple dimensions; Different business feature tags are constructed based on the characteristic attributes of the business requirement information, and a first user profile is constructed using multiple business feature tags; A preset neural network is invoked to calculate the profile matching degree between the first user profile and several preset profiles, wherein each preset profile is a personal profile of the second user within the current business update cycle; Based on the numerical values ​​of several profile matching degrees, several different business service information are searched in a preset database, target business information is filtered from the several business service information, and the target business information is pushed to the first user. 2.The user portrait-based service push method according to claim 1, wherein, The construction of multiple different business feature tags based on the feature attributes of the business requirement information includes: The feature attributes corresponding to each business feature data are obtained from the business requirement information, and the feature attributes include numerical values ​​and text. A preset text recognition model is invoked to identify the semantics of the text for each of the aforementioned feature attributes to obtain semantic information, and the corresponding image is searched according to each of the aforementioned semantic information to obtain multiple semantic images; The chromaticity values ​​of the corresponding semantic image are adjusted according to the numerical values ​​of the feature attributes to obtain different business feature labels. 3.The user portrait-based service push method according to claim 2, characterized in that, The step of adjusting the chromaticity value of the corresponding semantic image based on the numerical value of the feature attribute to obtain different business feature labels includes: The semantic image is divided into multiple grids according to a preset image ratio, and the chromaticity value of one of the grids is obtained to obtain the chromaticity value to be adjusted. Determine the target chromaticity value based on the chromaticity value to be adjusted; The color of the edge grid of the semantic image is adjusted according to the numerical value of the feature attribute and the target chromaticity value to obtain the business feature label. 4.The user portrait-based service push method according to claim 2, wherein, The construction of the first user profile using multiple business feature tags includes: Semantic relationships determined based on the semantic information of multiple business feature tags; The first user profile is obtained by connecting multiple business feature tags based on the semantic relationships. 5.The user portrait-based service push method according to claim 1, wherein, The preset neural network is an image colorimetric recognition model; The step of calling a preset neural network to calculate the matching degree between the first user profile and several preset profiles includes: The image chromaticity recognition model is invoked to identify the chromaticity value of the first user profile to obtain the tag chromaticity value, and the image chromaticity recognition model is invoked to identify the chromaticity value of each preset profile to obtain the profile chromaticity value. The tag chromaticity value is the chromaticity value of the image edge color of the business feature tag of the first user profile, and the profile chromaticity value is the chromaticity value of the image edge color of the business feature tag of the preset profile. The image matching degree is calculated based on the label chromaticity value and the image chromaticity value. 6.The user portrait-based service push method according to claim 5, wherein, The step of calculating the image matching degree based on the tag chromaticity value and the image chromaticity value includes: The label weight value is obtained by finding the corresponding weight based on the label color value, and the image weight value is obtained by finding the weight based on the image color value. The total tag value of the first user profile is calculated using several of the tag weight values, and the total profile value of each preset profile is calculated using several of the profile weight values. The image matching degree is obtained by calculating the ratio of the total tag value to the total image value.

7. The user portrait-based service push method according to any one of claims 1-6, characterized by, After the step of constructing the first user profile using multiple business feature tags, the method further includes: Determine the number of preset images stored in the database; If the quantity value is less than the preset quantity threshold, then multiple pieces of information to be used are obtained, wherein the information to be used is business information updated in the current business cycle or business information viewed and retained by the first user in the previous business cycle. A preset profile is constructed using the information to be used.

8. A service push system based on user profiling, characterized by, The system includes: The acquisition module is used to acquire the business requirement information of the first user, which includes business feature data in multiple different dimensions. The construction module is used to construct different business feature tags based on the feature attributes of the business requirement information, and to construct a first user profile using multiple business feature tags; The matching module is used to call a preset neural network to calculate the matching degree between the first user profile and several preset profiles, wherein each preset profile is a personal profile of the second user in the current business update cycle. The push module is used to search for several different business service information in a preset database based on the numerical values ​​of several profile matching degrees, filter target business information from the several business service information, and push the target business information to the first user.

9. The service push system based on user profiles according to claim 8, characterized in that, The construction of multiple different business feature tags based on the feature attributes of the business requirement information includes: The feature attributes corresponding to each business feature data are obtained from the business requirement information, and the feature attributes include numerical values ​​and text. A preset text recognition model is invoked to identify the semantics of the text for each of the aforementioned feature attributes to obtain semantic information, and the corresponding image is searched according to each of the aforementioned semantic information to obtain multiple semantic images; The chromaticity values ​​of the corresponding semantic image are adjusted according to the numerical values ​​of the feature attributes to obtain different business feature labels.

10. The user profile-based service push system according to claim 8, characterized in that, The preset neural network is an image colorimetric recognition model; The step of calling a preset neural network to calculate the matching degree between the first user profile and several preset profiles includes: The image chromaticity recognition model is invoked to identify the chromaticity value of the first user profile to obtain the tag chromaticity value, and the image chromaticity recognition model is invoked to identify the chromaticity value of each preset profile to obtain the profile chromaticity value. The tag chromaticity value is the chromaticity value of the image edge color of the business feature tag of the first user profile, and the profile chromaticity value is the chromaticity value of the image edge color of the business feature tag of the preset profile. The image matching degree is calculated based on the label chromaticity value and the image chromaticity value.