IP scene development difficulty estimation method and system based on big data

Abstract

The invention discloses an IP scene development difficulty estimation method and system based on big data, and relates to the technical field of scene development, and the method comprises the steps: S1, determining a plurality of IP scene types, S2, constructing a first building scene, S3, carrying out deep analysis processing, and S4, generating an IP scene development estimation scheme. The method is technically characterized in that a target figure portrait is configured based on target business data, and requirements of a user customer are comprehensively known; an IP scene type and an IP theme are determined by constructing a matching association set, three-dimensional space modeling is carried out based on the IP theme, the IP scene is effectively combined with target service data and user customer requirements, and parameterized description is carried out on a building scene of space modeling; in the process, a first neural network and a second neural network are designed, a first development difficulty value and a second development difficulty value are evaluated, and an IP scene development scheme is generated based on the first development difficulty value and the second development difficulty value.

Description

Technical Field

[0001] The present invention relates to the technical field of scenario development, and in particular to a method and system for estimating the difficulty of IP scenario development based on big data. Background Art

[0002] IP is the core of cultural tourism. Characteristic IP can reflect the theme, uniqueness, storytelling, interactivity, explosiveness and extensibility of cultural tourism products, thereby attracting consumers' attention and interest, and increasing the market share and revenue of cultural tourism products. Therefore, IP scenario development requires in-depth understanding of user customer needs to ensure the success of the final project, and define IP scenarios, that is, actual application scenarios and functions, based on these needs. In recent years, with the development of big data technology, related technologies have begun to try to use neural networks to generate scenario development models; however, due to the huge amount of data, traditional IP scenario development cannot accurately capture the needs of user customers; and often only considers the characteristics corresponding to customer needs, while ignoring the characteristics corresponding to building IP scenarios, which may lead to IP scenario development not meeting customer needs, increasing the difficulty of actual IP scenario development, and to a certain extent increasing the management difficulty of IP scenario development, reducing the efficiency of actual IP scenario development. Summary of the invention

[0003] 1. Technical issues to be resolved In view of the deficiencies in the prior art, the present invention provides a method and system for estimating the difficulty of IP scenario development based on big data, which configures a target person portrait based on target business data, including age sub-portrait, gender sub-portrait, regional sub-portrait, consumption habit sub-portrait, interest sub-portrait, value sub-portrait and market hot word sub-portrait, so as to fully understand the needs of user customers; and constructs a matching association set to determine the IP scenario type and IP theme; designs a first neural network and a second neural network for joint training, performs three-dimensional space modeling based on the IP theme, evaluates a first development difficulty value and a second development difficulty value, and generates an IP scenario development plan based on the first development difficulty value and the second development difficulty value, thereby solving the problems raised in the background technology.

[0004] (II) Technical solution To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for estimating the difficulty of IP scenario development based on big data includes the following steps: Determine several IP scenario types: Acquire target business data, perform feature extraction based on the first neural network, and establish a first information feature set; configure a target persona portrait based on the target business data, and use a sub-portrait in the target persona portrait as an initial matching portrait, perform association matching between the first information feature set and the sub-portrait, generate a matching association set, assign corresponding weight values ​​according to the matching association set, and determine a number of IP scenario types and corresponding first development difficulty values; Construct the first building scene: Determine the IP theme based on the IP scene type, and determine a second information feature set of any IP theme based on the second neural network, the second information feature set includes scene components and scene attributes, and construct a first building scene based on the second information feature set; determine a deviation parameter based on the first building scene and target business data; determine a second development difficulty value based on the deviation parameter; wherein the second neural network is built into the first neural network; In-depth analysis and processing: The parameters of the second neural network are updated based on the deviation parameters, and the parameters of the first neural network are updated based on the second neural network to generate a difficulty estimation optimization model, where the difficulty estimation optimization model is an IP scenario development reference solution generated based on target business data; Generate IP scenario development estimate plan: The first development difficulty values ​​and the second development difficulty values ​​corresponding to several IP themes are counted, and the IP scenario development plans with higher priorities are determined based on the first development difficulty values ​​and the second development difficulty values.

[0005] Furthermore, the target business data includes business scale, target positioning, operation and maintenance, bidding and market acceptance; the first information feature set includes finance, time, resources, risks and opportunities; the target person portrait is a portrait of the main population established based on the target business data, including age sub-portrait, gender sub-portrait, regional sub-portrait, consumption habit sub-portrait, interest and hobby sub-portrait, value sub-portrait and market hot word sub-portrait.

[0006] Furthermore, the process of determining the first development difficulty value is: The similarity between the first information feature set and the sub-portrait is obtained, and a matching association set is constructed according to the similarity; wherein the matching association set is arranged in descending order according to the similarity, and a corresponding weight value is assigned to each sub-portrait through the matching association set, and the weight value and the similarity are kept synchronized; and the weight value of each sub-portrait is divided by its similarity, and all the results are added together to obtain a first development difficulty value; The formula for the first development difficulty value is: ; Where P1 is the first development difficulty value, Xi is the similarity corresponding to the i-th sub-portrait, φi is the weight value assigned according to the similarity, and N0 is the number of sub-portraits of the main population portrait.

[0007] Furthermore, the IP scenario type is determined based on the matching association set, and the right confirmation is carried out. When the right confirmation is passed, it is stored on the chain, and the IP theme is determined based on the IP scenario type after being stored on the chain.

[0008] Furthermore, the first neural network has a built-in second neural network, and the two networks are jointly trained. The construction process is as follows: Main network: the first neural network; Sub-network: the second neural network; Data flow: Identify the target business data flowing into the first neural network; Intermediate output: The first neural network extracts features from the target business data and flows the extracted intermediate features into the second neural network; wherein the intermediate features are the first information feature set; Sub-network processes data: The second neural network further extracts and processes features and marks the extracted features as a second information feature set; Return to the main network: The output of the second neural network is used as the input of the subsequent layers of the first neural network to make the final decision.

[0009] Furthermore, the scene components include scene structure, area allocation and device location, and the scene attributes include safety attributes, comfort attributes, theme consistency attributes and reachability attributes; the process of sub-network processing data is: Extract features of the IP theme based on the second neural network, determine a second information feature set for IP scene construction, determine scene components and scene attributes based on the second information feature set, perform three-dimensional space modeling based on the scene components and scene attributes, and determine a first building scene of the IP theme; extract the first information feature set corresponding to the target business data at the same time, and define the above extracted features as influencing factors; Based on the influencing factors of the target business data, the first building scene is corrected and calibrated to obtain the second building scene; and the deviation parameters of the first building scene and the second building scene are determined, and the parameters of the second neural network are updated based on the deviation parameters to generate a difficulty estimation optimization model; and a deviation parameter set is established, and the deviation parameter set is assigned corresponding weights, and the weights are synchronized with the deviation parameters, and the elements corresponding to the deviation parameter set are multiplied and accumulated by their weights to generate a second development difficulty value; The formula for the second development difficulty value is: ; Where P2 is the second development difficulty value, N1 is the number of elements in the deviation parameter set, Yj is the element of the deviation parameter set, and βj is the assigned weight; Based on the second neural network, the features of the second building scene are restored to obtain a corresponding third information feature set, a comparison cost parameter is determined according to the second information feature set and the third information feature set, and the parameters of the second neural network are updated based on the deviation parameter and the comparison cost parameter to generate a difficulty estimation optimization model.

[0010] Furthermore, component recognition is performed based on the second building scene to identify the geometric structures and geometric properties of different components; and the spatial relationship and geometric relationship between the components are determined to determine the corresponding scene components and scene properties to achieve feature restoration; wherein the geometric structure includes position, size and shape, and the geometric properties include function, color, texture and material.

[0011] Furthermore, the process of updating the parameters of the second neural network and updating the parameters of the first neural network according to the second neural network is as follows: One calculation: for the second neural network, determine the value of the bias parameter and calculate the gradient of the loss function to the second neural network through the back propagation algorithm; Secondary calculation, for the second neural network, determine the values ​​of the bias parameter and the comparison cost parameter, and calculate the gradient of the loss function to the second neural network through the back propagation algorithm; Three calculations, for the first neural network, pass the gradient of the second neural network to the first neural network, and determine the gradient of the loss function with respect to the first neural network; Parameter update: Use the calculated gradient to update the corresponding parameters of the first neural network and the second neural network using the gradient descent method.

[0012] Furthermore, the process of generating an IP scenario development estimate is as follows: First, obtain the first development difficulty value and the second development difficulty value under any IP theme, and assign corresponding weight coefficients respectively; then, combine the first development difficulty value and the second development difficulty value with the corresponding weight coefficients to generate a comprehensive development difficulty value; and count the comprehensive development difficulty values ​​under several IP themes, sort them in order, and prioritize them from high to low according to the sorting order; finally, extract the top N2 values ​​and obtain the corresponding IP scenario development plan.

[0013] A big data-based IP scenario development difficulty estimation system, comprising a data acquisition module, a first evaluation module, a second evaluation module and a solution generation module; Data acquisition module, to acquire target business data; A first training module determines a first information feature set according to the first neural network and the target business data, configures a target person portrait based on the first information feature set, uses a sub-portrait in the target person portrait as an initial matching portrait, performs association matching between the first information feature set and the sub-portrait, and determines a plurality of IP scenario types and corresponding first development difficulty values; The second training module determines a second information feature set based on the second neural network and the IP theme, constructs a first building scene based on the second information feature set, determines a deviation parameter based on the first building scene, establishes a deviation parameter set, and obtains a second development difficulty value; the second training module has a built-in feature restoration unit to perform feature restoration on the second building scene; The solution generation module generates a priority-ranked IP scenario development solution based on the first development difficulty value and the second development difficulty value.

[0014] (III) Beneficial effects The present invention provides a method and system for estimating the difficulty of IP scenario development based on big data, which has the following beneficial effects: 1. The present invention configures the target persona of the user customer based on the target business data. The target persona covers the main needs of the user customer, helps to fully understand the multi-dimensional characteristics and needs of the user customer, and provides a basis for subsequent analysis. Among them, by extracting the first information feature set of the target business data, performing association matching of the information feature set and the sub-portrait, and adaptively adjusting the order of matching association sets, the needs of the user customer are accurately captured to a certain extent; 2. The present invention designs a first neural network and a second neural network. The first neural network is used as the main network, and the second neural network is used as the sub-network for joint training, thereby optimizing the structure and management of the network and improving the processing efficiency and response speed. At the same time, the first development difficulty value is calculated based on the first neural network, and the second development difficulty value is calculated based on the second neural network, and the comprehensive development difficulty of IP scenario development is separated, and a high-priority IP scenario development strategy is matched, providing several opportunities for IP scenario development, thereby reducing the management difficulty of scenario development to a certain extent and improving the efficiency of IP scenario development. 3. The present invention performs three-dimensional spatial modeling on the IP theme, effectively combines the IP scenario with the target business data and user customer needs, and performs parameterized description of the architectural scene of the spatial modeling; on this basis, the parameters of the second neural network are updated by calculating the offset parameters, and the parameters of the second neural network are updated by calculating the offset parameters and comparing the cost parameters, so as to update the parameters of the first neural network and obtain a difficulty estimation optimization model; an IP scenario development estimation plan is generated through the difficulty estimation optimization model to optimize the IP scenario development process; this not only improves the efficiency of IP scenario development decision-making, but also helps to reduce the management difficulty of IP scenario development. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A flowchart of a method for estimating the difficulty of IP scenario development based on big data according to the present invention; Figure 2 This is a module schematic diagram of a system for estimating the difficulty of IP scenario development based on big data according to the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] Example 1: Please refer to Figure 1 This embodiment provides a method for estimating the difficulty of IP scenario development based on big data. The main steps of the estimation method are as follows: S1. Determine several IP scenario types: Acquire target business data, perform feature extraction based on the first neural network, and establish a first information feature set; configure a target persona portrait based on the target business data, and use a sub-portrait in the target persona portrait as an initial matching portrait, perform association matching between the first information feature set and the sub-portrait, generate a matching association set, assign corresponding weight values ​​according to the matching association set, and determine a number of IP scenario types and corresponding first development difficulty values; The target business data includes business scale, target positioning, operation and maintenance, bidding and market acceptance; the first information feature set includes finance, time, resources, risks and opportunities; the target persona is the main population portrait established based on the target business data, including age sub-portrait, gender sub-portrait, region sub-portrait, consumption habit sub-portrait, interest sub-portrait, value sub-portrait and market hot word sub-portrait; In this embodiment, the relevant terms involved in the target business data are explained as follows: Business scale: The overall size and scope of the enterprise or project, used to estimate the overall impact and resource requirements of IP scenario development; Target positioning: clarify the market objectives and target audiences of the business or project, including the characteristics of the target market, consumer demand and market positioning; used to estimate the development direction of IP scenario development and market competition strategy; Operation and maintenance: management and maintenance work during the operation of a business or project; used to evaluate the stability and sustainability of IP scenario development; Bidding: Control costs, time and quality during business or project implementation; used to estimate the efficiency of IP scenario development; Market acceptance: the degree of market acceptance of an enterprise or project, including market feedback and market satisfaction; used to estimate the success and potential of IP scenario development; The first neural network is used to extract features from the target business data to generate a first information feature set. The subset nouns involved are explained as follows: Finance: including design, construction, materials, staff salaries and other expenses, helping to evaluate budget benefits; Time: the progress required for IP scenario development, the cost of delays and accelerations, and the assessment of the economic impact on time; Resources: various resources required for IP scenario development, including manpower, equipment and materials; Risk: including the cost of risk management and contingency plans to reduce the cost of project failure; Opportunities: Potential benefits from other opportunities that were missed when selecting the corresponding IP theme for IP scenario development; By configuring the target person portrait, we can conduct demand analysis on the main population portraits based on big data. The related terms of the main population portraits are explained as follows: Age sub-portraits: portraits based on the age of the population, used to estimate the applicable population for IP scenario development, characteristics, needs, and behavioral habits of different age groups; Gender sub-portraits: portraits based on the gender of a group of people, used to estimate the needs and preferences of people of different genders in IP scenario development; Regional sub-profile: A population profile based on the geographic location of the population, used to estimate the regional distribution of IP scenario development; Consumption habit sub-profile: A crowd portrait built based on the crowd’s consumption habits, such as travel frequency, travel category, and consumption amount, used to estimate project settings for IP scenario development; Hobby sub-portraits: People portraits built based on their hobbies, such as favorite anime characters, movie roles, etc., used to estimate the scene types and themes of IP scene development; Value sub-portraits: character portraits based on people’s beliefs and life philosophies, used to estimate the resonance of values ​​among people in IP scenario development; Market buzzword sub-portraits: character portraits constructed by analyzing popular words and trends in the current market and user discussions, used to estimate market trends for IP scenario development; The process of determining the first development difficulty value is as follows: Obtaining the similarity between the first information feature set and the sub-portrait, and constructing a matching association set according to the similarity; wherein the matching association set is arranged in descending order according to the similarity; Matching is performed by calculating the similarity between the sub-portrait and the first information feature. Common methods include cosine similarity, Euclidean distance, and Manhattan distance. In this embodiment, cosine similarity is selected for calculation: the first information feature set and the sub-portrait are both converted into vector form and marked as the first feature vector A and the second feature vector B respectively; The similarity is calculated by the cosine of the angle between the two vectors; The formula based on: ; Where xs is the cosine similarity value, which ranges from [-1, 1], where 1 means complete similarity, 0 means no similarity, and -1 means complete opposite. The larger the value, the better. For example: the first information feature set includes goals, finance, time, resources, risks and opportunities, which are first constructed as the first feature vector A. Taking the second feature vector B1 constructed by the consumption habit sub-portrait as an example, the cosine similarity between the first feature vector A and the second feature vector B1 is calculated. If the value is 0, the first feature vector A-the second feature vector B1 is placed at the top of the matching association set; then taking the second feature vector B2 constructed by the interest and hobby sub-portrait as an example, the cosine similarity between the first feature vector A and the second feature vector B2 is calculated. If the value is 1, indicating complete similarity, the first feature vector A-the second feature vector B2 is arranged before the first feature vector A-the second feature vector B1, and this cycle is repeated to generate a matching association set; By matching the association set, a corresponding weight value is assigned to each sub-portrait, and the weight value and the cosine similarity value are kept synchronized. The weight value of each sub-portrait is divided by its similarity, and then all the results are added together to obtain the first development difficulty value; The formula for the first development difficulty value is: ; In the formula, P1 is the first development difficulty value, Xi is the cosine similarity value corresponding to the i-th sub-portrait, N0 is the number of sub-portraits of the main population portrait, and φi is the weight value assigned according to the cosine similarity value, for example: Example: Assume there is a matching association set, and the order of cosine similarity is: age sub-portrait > gender sub-portrait > region sub-portrait > consumption habit sub-portrait > hobby sub-portrait > value sub-portrait > market hot word sub-portrait, the corresponding cosine similarity values ​​are 1, 0.75, 0.5, 0.25, 0.15, 0.125, 0.05; Then: Age sub-portrait, weighted 1> Gender sub-portrait, weighted 2> Region sub-portrait, weighted 3> Consumption habit sub-portrait, weighted 4> Interest and hobby sub-portrait, weighted 5> Value sub-portrait, weighted 6> Market buzzword sub-portrait, weighted 7; The corresponding first development difficulty value = 1 / 1 + 2 / 0.75 + 3 / 0.5 + 4 / 0.25 + 5 / 0.15 + 6 / 0.125 + 7 / 0.05 = 247.97; the result is rounded to one decimal place. The larger the first development difficulty value is, the more difficult it is to match the association, and the more difficult it is to develop IP scenarios. Wherein, the IP scenario type is determined based on the matching association set; Example: IP scenario types include: Tourist attractions: such as museums, natural parks, historical sites, etc.; Cultural activities: such as art exhibitions, concerts, festivals, etc.; Theme parks: such as amusement parks, resorts, etc.; Urban culture: historical districts, cultural centers, etc. in the city; B&Bs and hotels: accommodations specially designed to cater to tourist experiences, etc. The rights confirmation process is then carried out. When the rights confirmation is passed, the IP is stored on the chain, and the IP theme is determined based on the IP scenario type after the chain storage. It should be noted that the rights confirmation process is a review and certification based on big data to ensure the legitimacy of the IP scenario type. There are many ways in real life, which will not be elaborated in this article. Example: For B&Bs and hotels, IP scenario topics include detailed information such as room type, price, facilities, and location; S2. Construct the first building scene: Determine the IP theme based on the IP scene type, and determine a second information feature set of any IP theme based on the second neural network, the second information feature set includes scene components and scene attributes, and construct a first building scene based on the second information feature set; determine a deviation parameter based on the first building scene and target business data; determine a second development difficulty value based on the deviation parameter; wherein the second neural network is built into the first neural network; The second neural network is built into the first neural network, and the two networks are jointly trained. The construction process is as follows: Main network: the first neural network; Sub-network: the second neural network; Data flow: Identify the target business data flowing into the first neural network; Intermediate output: The first neural network extracts features from the target business data and flows the extracted intermediate features into the second neural network; wherein the intermediate features are the first information feature set; Sub-network processes data: The second neural network further extracts and processes features and marks the extracted features as a second information feature set; Return to the main network: the output of the second neural network is used as the input of the subsequent layers of the first neural network to make the final decision; Among them, the scene components include scene structure, area allocation and device location, and the scene attributes include safety attributes, comfort attributes, theme consistency attributes and reachability attributes; the process of network processing data is: Extract features of the IP theme based on the second neural network to determine a second information feature set for the IP scene, where the second information feature set includes scene components and scene attributes; In this embodiment, the nouns involved in the second information feature set are explained as follows: Scene structure: refers to the layout rules and composition of the entire scene, including its spatial arrangement, functional areas and overall design framework; Area allocation: Divide the scene into different areas and allocate the use of these areas according to functions and needs; in the construction of IP scenes, such as display areas, dining areas, and play areas, the buildings involved include walls, windows, doors, stairs, etc. Equipment location: It involves the specific equipment installation point in the IP scenario, including cameras and commemorative coin vending machines; Security attributes: measures involving people and assets within the IP scenario, including monitoring, alarming, and access control; Comfort attributes: factors related to user experience in IP scenarios; Theme consistency attribute: involves the theme, visual and design consistency style within the IP scene, including the color and decorations of the display cabinet; Reachability attribute: involves the convenience of movement within the IP scenario; Based on the second neural network, three-dimensional spatial modeling is performed on scene components and scene attributes to determine the first building scene of the IP theme; Among them, you can use a variety of tools to carry out three-dimensional space modeling. The general steps are as follows: 1. Determine scene components and scene properties; 2. Use 3D modeling software for modeling. The modeling steps are as follows: Create basic structure: Select one or more suitable 3D world maps from the preset 3D world map library as the basic structure of the scene to form a 3D model; Add scene components: In the 3D model, fill the basic structure with scene modeling content, such as adding walls, doors, windows, floors, ceilings and other structures, and adjusting their sizes and positions; Application scene attributes: In the 3D model, further scene attributes are added to perform logical control of the scene. For example, if safety facilities are added, it is necessary to ensure that they cover all key areas and adjust the channel width and door opening and closing methods based on comfort and safety. 3. Adjust materials and lighting: Select appropriate materials, such as wood and glass, apply them to the surface of the 3D model, adjust the texture of the material, and determine the color and texture characteristics; at the same time, add light sources to the 3D model to simulate natural light and artificial light in the IP scene; 4. Rendering and display: Generate rendered images: Use rendering plug-ins to generate high-quality rendered images and animations; Create VR interactive experiences: Use Unity to create immersive virtual reality experiences; 5. Review and correction: Ensure that all building structures are arranged according to design requirements and correct problems in the model based on feedback and review results; Based on the influencing factors of the target business data, the first building scene is corrected and calibrated to obtain the second building scene; and the deviation parameters of the first building scene and the second building scene are determined, and the parameters of the second neural network are updated based on the deviation parameters to generate a difficulty estimation optimization model; and a deviation parameter set is established, and the deviation parameter set is assigned corresponding weights, and the weights are synchronized with the deviation parameters, and the elements corresponding to the deviation parameter set are multiplied and accumulated by their weights to generate a second development difficulty value; For example, the deviation parameters between the first building scene and the second building scene can be determined by mathematical comparison. The deviation parameters can help understand the difference between IP scene development and target business data, thereby optimizing IP scene development. For example: scene components and scene attributes are extracted from the second feature information set, and the first building scene is constructed by taking the wall C1 as an example, which is represented as (x1, y1, z1, d1, e1); wherein (x1, y1, z1) is the position coordinate of the wall C1, and d1 and e1 represent the color feature and texture feature of the wall C1; Similarly, scene components and scene attributes are extracted from the third feature information set, and the wall C2 corresponding to the second building scene is represented as (x2, y2, z2, d2, e2); wherein (x1, y1, z1) are the position coordinates of the wall C2, and d2 and e2 represent the color feature and texture feature of the wall C2; then, by comparing the difference between the wall C1 and the wall C2, the deviation parameter h1 between the wall C1 and the wall C2 is obtained; Similarly, obtain doors corresponding to the first building scene and the second building scene, including door J1 and door J2, door J1 is represented by (x3, y3, z3, d3, e3), and door J2 is represented by (x4, y4, z4, d4, e4); wherein (x3, y3, z3) are the position coordinates of door J1, d3 and e3 represent the color features and texture features of door J1, (x4, y4, z4) are the position coordinates of door J2, d4 and e4 represent the color features and texture features of door J2; obtain the corresponding deviation parameter h2 by comparing door J1 and door J2; Similarly, buildings corresponding to the first building scene and the second building are obtained, corresponding deviation parameters are obtained, and a deviation parameter set is constructed, H={h1, h2, ...}; The formula for the second development difficulty value is: ; Where P2 is the second development difficulty value, N1 is the number of elements in the deviation parameter set, Yj is the element of the deviation parameter set, and βj is the assigned weight; Example: If only the differences between wall C1 and wall C2 and door J1 and door J2 are considered, the deviation parameter h1 obtained between the walls is greater than the deviation parameter h2 obtained between the doors, and the corresponding values ​​are 0.75 and 0.5; Then: wall, assigned weight 2> door, assigned weight 1; Then the second development difficulty value = 0.75*2+0.5*1=2; it should be noted that the larger the second development difficulty value is, the greater the difference between the first building scene and the second building scene is, and the greater the difficulty of IP scene development is; Based on the second neural network, the second building scene is restored to obtain a corresponding third information feature set, a comparison cost parameter is determined according to the second information feature set and the third information feature set, and the second neural network is updated based on the deviation parameter and the comparison cost parameter to generate a difficulty estimation optimization model; Import the second building scene into professional software (such as BIM software, CAD tools), analyze the data structure of the corresponding 3D modeling, and identify components, such as doors and walls, and then identify the geometric structure and geometric properties of different components; the geometric structure includes position, size and shape, and the geometric properties include function, color, texture and material; determine the spatial relationship and geometric relationship between components to determine the corresponding scene components and scene properties to achieve feature restoration; S3. In-depth analysis and processing: The parameters of the second neural network are updated based on the deviation parameters, and the parameters of the first neural network are updated based on the second neural network to generate a difficulty estimation optimization model, where the difficulty estimation optimization model is an IP scenario development reference solution generated based on target business data; The process of updating the parameters of the second neural network and updating the parameters of the first neural network according to the second neural network is as follows: One calculation: for the second neural network, determine the value of the bias parameter and calculate the gradient of the loss function to the second neural network through the back propagation algorithm; For example, for the output layer, first calculate the gradient of the loss function with respect to the output (i.e., the derivative of the loss) as ; In the formula, H is the output, which is the value of the deviation parameter; Then the chain rule is used to calculate the gradient of the loss function relative to the penultimate layer: ; In the formula, z （l） is the weighted sum input of the penultimate layer, α （l） is the activation value of the second-to-last layer; Finally, the gradient of each layer is recursively calculated, back-propagating from the output layer to the input layer layer by layer; Secondary calculation, for the second neural network, determine the values ​​of the bias parameter and the comparison cost parameter, and calculate the gradient of the loss function to the second neural network through the back propagation algorithm; For example: For the output layer, first, because the output value has a bias parameter and a comparison cost parameter, the loss function L is defined using the mean square error loss sub , and then calculate the gradient of the loss function with respect to the output (i.e., the derivative of the loss), which is and ; Use the chain rule to calculate the gradient of each layer, and finally merge the gradients to get the final gradient, expressed as: ; Three calculations, for the first neural network, pass the gradient of the second neural network to the first neural network, and determine the gradient of the loss function with respect to the first neural network; In this embodiment, it is assumed that the loss function of the first neural network (main network) is L main , the loss function of the second neural network (sub-network) is L sub , then the total loss function L total is the weighted sum of the corresponding loss functions of the first neural network (main network) and the second neural network (sub-network); The formula is: ; In the formula, λ is a hyperparameter; Parameter update: Use the calculated gradient to update the parameters of the first neural network and the second neural network using the gradient descent method; The parameter update of the first neural network can be expressed as: ; The parameter update of the second neural network can be expressed as: ; In the formula, θ main Represents the update parameters of the first neural network (main network), θ sub represents the update parameters of the second neural network (sub-network), η is the learning rate, and They are the gradients of the first neural network (main network) and the second neural network (sub-network); S4. Generate IP scenario development estimate plan: Counting the first development difficulty values ​​and the second development difficulty values ​​corresponding to a number of IP themes, and determining a priority IP scenario development plan based on the first development difficulty value and the second development difficulty value; First, obtain the first development difficulty value and the second development difficulty value under any IP theme, and assign corresponding weight coefficients respectively; Then, the first development difficulty value and the second development difficulty value are combined with the corresponding weight coefficients to generate a comprehensive development difficulty value; the formula based on which is: ; Wherein, P1, P2 and P3 are the first development difficulty value, the second development difficulty value and the third development difficulty value respectively, σ1 and σ2 are weight coefficients, and σ1 and σ2 are both greater than 0, σ1+σ2=1; Example: When the first development difficulty value is 247.97 and the second development difficulty value is 2, take σ1=σ2=0.5, then the third development difficulty value = 247.97*0.5+2*0.5=124.985; It should be noted that the larger the first development difficulty value and the second development difficulty value, the larger the corresponding comprehensive development difficulty value, indicating that the IP scenario development is more difficult; Count the comprehensive development difficulty values ​​of several IP themes, sort them in ascending order, and prioritize them in descending order; Finally, extract the top N2 values ​​and obtain the corresponding IP scenario development plan.

[0018] Example 2: Please refer to Figure 2 , based on Example 1, this embodiment also provides an IP scenario development difficulty estimation system based on big data, including a data acquisition module, a first evaluation module, a second evaluation module and a solution generation module; Data acquisition module, to acquire target business data; A first training module determines a first information feature set according to the first neural network and the target business data, configures a target person portrait based on the first information feature set, uses a sub-portrait in the target person portrait as an initial matching portrait, performs association matching between the first information feature set and the sub-portrait, and determines a plurality of IP scenario types and corresponding first development difficulty values; The second training module determines a second information feature set based on the second neural network and the IP theme, constructs a first building scene based on the second information feature set, determines a deviation parameter based on the first building scene, establishes a deviation parameter set, and obtains a second development difficulty value; the second training module has a built-in feature restoration unit to perform feature restoration on the second building scene; The solution generation module generates a priority-ranked IP scenario development solution based on the first development difficulty value and the second development difficulty value.

[0019] In summary, the steps of the present invention include: S1, determining a number of IP scenario types, S2, constructing a first building scene, S3, in-depth analysis and processing, S4, generating an IP scenario development estimation plan; configuring a target persona based on target business data to fully understand the needs of user customers; determining the IP scenario type and IP theme by constructing a matching association set, and performing three-dimensional space modeling based on the IP theme, effectively combining the IP scenario with the target business data and user customer needs, and performing parameterized description of the architectural scene of the spatial modeling; in this process, designing a first neural network and a second neural network, evaluating a first development difficulty value and a second development difficulty value, and generating an IP scenario development plan based on the first development difficulty value and the second development difficulty value.

[0020] In the application, the several formulas involved are all calculated by removing the dimensions and taking their numerical values. The formula is a formula obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The formula is set by technical personnel in this field according to actual conditions.

[0021] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0022] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0023] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for estimating the difficulty of IP scenario development based on big data, characterized in that: The steps include: Determine several IP scenario types: Acquire target business data, perform feature extraction based on the first neural network, and establish a first information feature set; configure a target persona portrait based on the target business data, and use a sub-portrait in the target persona portrait as an initial matching portrait, perform association matching between the first information feature set and the sub-portrait, generate a matching association set, assign corresponding weight values ​​according to the matching association set, and determine a number of IP scenario types and corresponding first development difficulty values; Construct the first building scene: Determine the IP theme based on the IP scene type, and determine a second information feature set of any IP theme based on a second neural network, the second information feature set including scene components and scene attributes, and construct a first building scene based on the second information feature set; Determining a deviation parameter based on the first building scene and the target business data; determining a second development difficulty value based on the deviation parameter; wherein the second neural network is built into the first neural network; In-depth analysis and processing: The parameters of the second neural network are updated based on the deviation parameters, and the parameters of the first neural network are updated based on the second neural network to generate a difficulty estimation optimization model, where the difficulty estimation optimization model is an IP scenario development reference solution generated based on target business data; Generate IP scenario development estimate plan: The first development difficulty values ​​and the second development difficulty values ​​corresponding to several IP themes are counted, and the IP scenario development plans with higher priorities are determined based on the first development difficulty values ​​and the second development difficulty values.

2. The method for estimating the difficulty of IP scenario development based on big data according to claim 1, characterized in that: The target business data includes business scale, target positioning, operation and maintenance, bidding and market acceptance; the first information feature set includes finance, time, resources, risks and opportunities; the target person portrait is a portrait of the main population established based on the target business data, including age sub-portrait, gender sub-portrait, regional sub-portrait, consumption habit sub-portrait, interest and hobby sub-portrait, value sub-portrait and market hot word sub-portrait.

3. The method for estimating the difficulty of IP scenario development based on big data according to claim 2 is characterized in that: The process of determining the first development difficulty value is: The similarity between the first information feature set and the sub-portrait is obtained, and a matching association set is constructed according to the similarity; wherein the matching association set is arranged in descending order according to the similarity, and a corresponding weight value is assigned to each sub-portrait through the matching association set, and the weight value and the similarity are kept synchronized; and the weight value of each sub-portrait is divided by its similarity, and all the results are added together to obtain a first development difficulty value; The formula for the first development difficulty value is: ; Where P1 is the first development difficulty value, Xi is the similarity corresponding to the i-th sub-portrait, φi is the weight value assigned according to the similarity, and N0 is the number of sub-portraits of the main population portrait.

4. The method for estimating the difficulty of IP scenario development based on big data according to claim 3 is characterized by: The IP scenario type is determined based on the matching association set, and the rights confirmation process is carried out. When the rights confirmation is passed, it is stored on the chain, and the IP subject is determined based on the IP scenario type after being stored on the chain.

5. The method for estimating the difficulty of IP scenario development based on big data according to claim 4 is characterized in that: The first neural network has a built-in second neural network, and the two networks are jointly trained. The construction process is as follows: Main network: the first neural network; Sub-network: the second neural network; Data flow: Identify the target business data flowing into the first neural network; Intermediate output: The first neural network extracts features from the target business data and flows the extracted intermediate features into the second neural network; wherein the intermediate features are the first information feature set; Sub-network processes data: The second neural network further extracts and processes features and marks the extracted features as a second information feature set; Return to the main network: The output of the second neural network is used as the input of the subsequent layers of the first neural network to make the final decision.

6. The method for estimating the difficulty of IP scenario development based on big data according to claim 5 is characterized in that: The scene components include scene structure, area allocation and device location, and the scene attributes include safety attributes, comfort attributes, theme consistency attributes and reachability attributes. The process of sub-network processing data is as follows: Extract features of the IP theme based on the second neural network, determine the second information feature set of the IP scene, determine scene components and scene attributes based on the second information feature set, perform three-dimensional space modeling based on the scene components and scene attributes, and determine the first building scene of the IP theme; at the same time, extract the target business data corresponding to the first information feature set, and define the extracted features as influencing factors; Based on the influencing factors of the target business data, the first building scene is corrected and calibrated to obtain a second building scene; and determining deviation parameters between the first building scene and the second building scene, updating parameters of the second neural network based on the deviation parameters, and generating a difficulty estimation optimization model; A deviation parameter set is established, a corresponding weight is assigned to the deviation parameter set, and the weight is kept synchronized with the deviation parameter, and the element corresponding to the deviation parameter set is multiplied and accumulated with its weight to generate a second development difficulty value; The formula for the second development difficulty value is: ; Where P2 is the second development difficulty value, N1 is the number of elements in the deviation parameter set, Yj is the element of the deviation parameter set, and βj is the assigned weight; Based on the second neural network, the features of the second building scene are restored to obtain a corresponding third information feature set, a comparison cost parameter is determined according to the second information feature set and the third information feature set, and the parameters of the second neural network are updated based on the deviation parameter and the comparison cost parameter to generate a difficulty estimation optimization model.

7. The method for estimating the difficulty of IP scenario development based on big data according to claim 6 is characterized by: Component recognition is performed based on the second building scene to identify the geometric structures and geometric properties of different components; and the spatial relationship and geometric relationship between the components are determined to determine the corresponding scene components and scene properties to achieve feature restoration; wherein the geometric structure includes position, size and shape, and the geometric properties include function, color, texture and material.

8. The method for estimating the difficulty of IP scenario development based on big data according to claim 7 is characterized in that: The process of updating the parameters of the second neural network and updating the parameters of the first neural network according to the second neural network is as follows: One calculation: for the second neural network, determine the value of the bias parameter and calculate the gradient of the loss function to the second neural network through the back propagation algorithm; Secondary calculation, for the second neural network, determine the values ​​of the bias parameter and the comparison cost parameter, and calculate the gradient of the loss function to the second neural network through the back propagation algorithm; Three calculations, for the first neural network, pass the gradient of the second neural network to the first neural network, and determine the gradient of the loss function with respect to the first neural network; Parameter update: Use the calculated gradient to update the corresponding parameters of the first neural network and the second neural network using the gradient descent method.

9. The method for estimating the difficulty of IP scenario development based on big data according to claim 8, characterized in that: The process of generating an IP scenario development estimate is as follows: First, obtain the first development difficulty value and the second development difficulty value under any IP theme, and assign corresponding weight coefficients respectively; then, combine the first development difficulty value and the second development difficulty value with the corresponding weight coefficients to generate a comprehensive development difficulty value; and count the comprehensive development difficulty values ​​under several IP themes, sort them in order, and prioritize them from high to low according to the sorting order; finally, extract the top N2 values ​​and obtain the corresponding IP scenario development plan.

10. A system for estimating the difficulty of IP scenario development based on big data, using any method described in claims 1 to 9, characterized in that: It includes a data acquisition module, a first evaluation module, a second evaluation module and a solution generation module; Data acquisition module, to acquire target business data; A first training module determines a first information feature set according to the first neural network and the target business data, configures a target person portrait based on the first information feature set, uses a sub-portrait in the target person portrait as an initial matching portrait, performs association matching between the first information feature set and the sub-portrait, and determines a plurality of IP scenario types and corresponding first development difficulty values; The second training module determines a second information feature set based on the second neural network and the IP theme, constructs a first building scene based on the second information feature set, determines a deviation parameter based on the first building scene, establishes a deviation parameter set, and obtains a second development difficulty value; The second training module has a built-in feature restoration unit to restore the features of the second building scene; The solution generation module generates a priority-ranked IP scenario development solution based on the first development difficulty value and the second development difficulty value.

Images