A data-driven decision-making method and system for pet furniture product design

Abstract

This invention discloses a data-driven method and system for decision-making on the design direction of pet furniture products, belonging to the field of information and communication technology. The method includes: constructing a multi-dimensional evaluation index system integrating static preset indicators and dynamic discovery indicators; collecting and standardizing e-commerce data based on preset product categories; calculating index weights using an objective weighting method, and calculating the comprehensive evaluation value for each category using multi-objective decision analysis; and dividing design requirement levels through clustering and outputting corresponding product design direction decision suggestions. This invention transforms market data into design requirement levels and specific design direction suggestions for pet furniture products, realizing data-driven design decision-making, significantly improving the scientific nature, accuracy, and efficiency of design decisions. Simultaneously, the dynamic indicator discovery mechanism endows the evaluation system with adaptive capabilities, enabling the design direction to be updated in real time with market changes.

Description

Technical Field

[0001] This invention relates to the field of information and communication technology, specifically to a data-driven method and system for making decisions on the design direction of pet furniture products. Background Technology

[0002] With the continued boom in the pet economy, the pet furniture market is experiencing rapid growth. As a special product connecting humans and pets, pet furniture design needs to simultaneously meet the behavioral needs of pets and the comprehensive needs of owners in terms of aesthetics, space, and cleanliness, exhibiting typical characteristics of a dual-subject relationship between humans and pets. Currently, the field of furniture design decision-making is undergoing a transformation from traditional experience-driven to data-driven approaches, with market data accumulated by e-commerce platforms providing a new data foundation for judging design directions. However, the rapid development of the pet furniture market has also brought complexity to design decisions: on the one hand, emerging design themes, such as intelligent interaction, modular structures, and environmentally friendly materials, are constantly emerging, and traditional evaluation systems based on fixed indicators struggle to capture these trends in a timely manner; on the other hand, existing analytical methods often separate market data from design needs, lacking an effective mechanism to transform data into specific design guidance.

[0003] In existing technologies, common e-commerce data analysis methods typically select indicators only from market or user dimensions, evaluating products through simple ranking or weighting to guide designers on which categories are worth developing. However, these methods have significant drawbacks: First, existing indicators only reflect market transactions or user satisfaction, failing to quantify core needs such as behavioral suitability and safety for pets as direct users, leading to a structural contradiction where designed products are liked by owners but not used by pets; second, once the evaluation indicator system is preset, it becomes static and fixed, making it difficult to adapt to the rapidly emerging design themes in the pet furniture market, resulting in a lag; third, the lack of semantic analysis of the indicator's value orientation makes it impossible to distinguish whether the indicator reflects pet needs or owner needs, leading to a lack of targeting in subsequent weight calculations and strategy generation. Therefore, there is an urgent need for a technical solution that can take into account the needs of both pets and owners, dynamically adapt to market changes, and translate data conclusions into specific design directions. Summary of the Invention

[0004] To address the aforementioned problems in the existing technology, this invention provides a targeted optimization solution for pet furniture design.

[0005] According to a first aspect of the present invention, a data-driven method for making decisions on the design direction of pet furniture products is proposed, comprising: Step S1: Select multiple evaluation indicators from market, user, and pet dimensions to construct a multi-dimensional evaluation indicator system for pet furniture; Step S2: For multiple pre-defined pet furniture product categories, obtain the raw data corresponding to the evaluation indicators from the target e-commerce platform, and clean and standardize the raw data to obtain a standardized data matrix; Step S3: Calculate the evaluation weights of each evaluation indicator using the objective weighting method, and calculate the comprehensive evaluation value of each pet furniture product category based on the evaluation weights and the standardized data matrix using the multi-objective decision analysis method. Step S4: Classify design requirements based on the comprehensive evaluation values ​​of all pet furniture product categories; Step S5: Based on the divided design requirement hierarchy, generate and output corresponding product design direction decision suggestions.

[0006] According to some embodiments, in the method of the first aspect of the present invention, step S1 of constructing a multi-dimensional evaluation index system for pet furniture specifically includes: Construct a two-dimensional coordinate space for human-pet value that includes a pet value axis and an owner value axis; Multiple initial evaluation indicators were selected from market, user and pet dimensions, and the initial evaluation indicators were mapped to their corresponding positions in a two-dimensional coordinate space. Text data on pet furniture products was collected, and unsupervised clustering algorithms were used to perform cluster analysis on the text data to discover implicit design themes. Implicit design themes are mapped to a two-dimensional coordinate space to generate dynamic evaluation indicators; By combining initial evaluation indicators and dynamic evaluation indicators, a multi-dimensional evaluation indicator system is formed.

[0007] According to some embodiments, in the method of the first aspect of the present invention, step S2, the process of determining the preset multiple pet furniture product categories specifically includes: Construct a multi-dimensional labeling system for pet furniture products. The multi-dimensional labeling system should include at least one or more of the following dimensions: applicable pet type, product function type, space utilization type, and material type. Based on the multi-dimensional labeling system, the pet furniture products to be evaluated are labeled, and products with the same label combination are grouped into the same product category.

[0008] According to some embodiments, in the method of the first aspect of the present invention, step S3 uses an objective weighting method to calculate the evaluation weights of each evaluation index, specifically including: The combined weighting method is used to calculate the comprehensive weight of each evaluation indicator, and the comprehensive weight is used as the evaluation weight; the comprehensive weight includes objective weight and subjective weight.

[0009] According to some embodiments, in the method of the first aspect of the present invention, the comprehensive weight of each evaluation index is calculated using a combined weighting method, including: The objective weights of each evaluation index are calculated using the entropy weight method; The analytic hierarchy process was used to calculate the subjective weights of each evaluation indicator based on scores from pet behavior experts and veterinarians. The objective weights and subjective weights are weighted and integrated to obtain the comprehensive weight of each evaluation indicator.

[0010] According to some embodiments, in the method of the first aspect of the present invention, step S3, calculating the comprehensive evaluation value of each type of pet furniture product using a multi-objective decision analysis method, specifically includes: The TOPSIS method is used to define the positive and negative ideal solutions for each evaluation index, calculate the distance between each evaluation index and the positive and negative ideal solutions, and use the relative proximity as the comprehensive evaluation value of the evaluation object.

[0011] According to some embodiments, in the method of the first aspect of the present invention, step S4 specifically includes: The Ward clustering method was used, with the comprehensive evaluation value of all pet furniture product categories as input, to perform hierarchical clustering analysis and generate a tree diagram structure; Based on the tree diagram structure, the samples are divided into three categories, which are defined as high design requirement level, medium design requirement level and low design requirement level respectively. The design strategy is to prioritize the development of high-level design requirements, selectively develop medium-level design requirements, and postpone the development of low-level design requirements.

[0012] According to some embodiments, in the method of the first aspect of the present invention, step S4 of dividing the design requirement hierarchy further includes: A two-dimensional cross-analysis matrix of pet type and function type is constructed. Based on the two-dimensional cross-analysis matrix, the clustering results are further refined to generate a hierarchical design requirement for specific pet type and specific function type.

[0013] According to some embodiments, in the method of the first aspect of the present invention, step S5 of generating and outputting the design strategy specifically includes: Step S51: Construct a knowledge base of pet behavior and habits, which includes information on the behavioral characteristics, spatial needs, and material preferences of different pet types; Step S52: Based on the classification of design requirements and combined with the knowledge base of pet behavior habits, generate refined design suggestions for specific pet furniture product categories. The design suggestions shall include at least one or more of the following: functional configuration suggestions, material selection suggestions, and spatial layout suggestions. Step S53: Visualize the information entropy, weights, distance between positive and negative ideal solutions, comprehensive evaluation value, and design requirement level.

[0014] According to a second aspect of the present invention, a data-driven decision-making system for pet furniture product design is proposed, comprising: The indicator system construction module is used to select multiple evaluation indicators from market, user and pet dimensions to build a multi-dimensional evaluation indicator system for pet furniture; The data acquisition and processing module is used to obtain raw data corresponding to the evaluation indicators from the target e-commerce platform for multiple preset pet furniture product categories, and to clean and standardize the raw data to obtain a standardized data matrix. The comprehensive evaluation module is used to calculate the evaluation weights of each evaluation indicator using an objective weighting method, and to calculate the comprehensive evaluation value of each pet furniture product category based on the evaluation weights and the standardized data matrix through a multi-objective decision analysis method. The requirement hierarchy module is used to classify design requirement levels based on the comprehensive evaluation value of all pet furniture product categories using a clustering algorithm; The strategy generation module is used to generate and output corresponding product design direction decision suggestions based on the divided design requirement hierarchy.

[0015] Compared with the prior art, the solution of the present invention has the following beneficial effects: 1. To address the challenges of quantifying the dual needs of humans and pets in pet furniture design decisions, and the inability to incorporate core indicators such as pet behavior compatibility and safety into the evaluation system, this invention constructs a human-pet value coordinate space that includes both pet and owner value axes. This coordinate space maps pre-defined indicators from the market, user, and pet dimensions to their corresponding positions, thereby enabling the quantitative evaluation of the dual-subject value.

[0016] 2. To address the problem that traditional evaluation index systems are static and fixed, and cannot capture emerging design themes, leading to decision-making delays, this invention introduces an unsupervised clustering algorithm to perform cluster analysis on product text data. This automatically discovers implicit design themes and maps them to the human-pet value coordinate space to generate dynamic evaluation indicators, thereby achieving adaptive expansion of the index system and forward-looking decision-making capabilities.

[0017] 3. To address the issue that purely objective weighting in pet furniture evaluation may weaken the core needs of pets, while purely subjective weighting lacks market basis, this invention combines the objective weights calculated by the entropy weighting method with the subjective weights calculated by the analytic hierarchy process based on scores from pet behavior experts and veterinarians, thereby achieving a scientific balance between market data and professional knowledge.

[0018] 4. To address the problem that simple weighting in multi-indicator evaluation cannot reflect the gap between the product and the optimal state, and that clustering and hierarchical classification lack objective basis, this invention uses the TOPSIS method to calculate the relative closeness of each product category to the positive and negative ideal solutions as a comprehensive evaluation value, and uses the Ward clustering method to automatically divide the data into three design requirement levels: high, medium, and low, based on the natural distribution of the data, thus realizing the scientific quantification of evaluation results and the objective classification of priorities.

[0019] 5. To address the problem that data analysis results are abstract and cannot be directly translated into design instructions by designers, this invention constructs a knowledge base of pet behavior and habits, and generates detailed design suggestions including functional configuration, material selection, and spatial layout by combining design requirements hierarchically. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating an embodiment 1000 of a data-driven pet furniture product design direction decision-making method according to the present invention. Figure 2 for Figure 1 A flowchart illustrating step S1 in Example 1000; Figure 3 for Figure 1 A flowchart illustrating sub-step S2A of step S2 in Example 1000; Figure 4 for Figure 1 A flowchart illustrating step S4 in Example 1000; Figure 5 for Figure 1 A flowchart illustrating step S5 in Example 1000; Figure 6 This is a schematic diagram of an embodiment 2000 of a data-driven pet furniture product design direction decision system according to the present invention; Figure 7-10 The specific interface diagram of the home furnishing product information management platform constructed according to a data-driven pet furniture product design direction decision method of the present invention is shown. Detailed Implementation

[0021] 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.

[0022] Figure 1 This is a flowchart illustrating an embodiment 1000 of the data-driven pet furniture product design direction decision-making method of the present invention. Figure 1As shown, Example 1000 includes steps S1-S5.

[0023] In step S1, a data-driven pet furniture product design direction decision-making system (hereinafter referred to as the system) selects multiple evaluation indicators from market, user, and pet dimensions to construct a multi-dimensional evaluation indicator system for pet furniture. Step S1 constructs a multi-dimensional evaluation indicator system that integrates static preset indicators and dynamic discovery indicators. This system endows the indicators with semantic meaning through the human-pet value coordinate space and achieves adaptive expansion of the indicator system through unsupervised clustering.

[0024] Optionally, the specific process of system implementation step S1 includes: constructing a two-dimensional coordinate space for human-pet value containing pet value axis and owner value axis; selecting multiple initial evaluation indicators from market dimension, user dimension and pet dimension, and mapping the initial evaluation indicators to the corresponding positions in the two-dimensional coordinate space; collecting text data of pet furniture products, using unsupervised clustering algorithm to perform cluster analysis on the text data, and discovering implicit design themes; mapping the implicit design themes to the two-dimensional coordinate space to generate dynamic evaluation indicators; and combining the initial evaluation indicators and dynamic evaluation indicators to form a multi-dimensional evaluation indicator system.

[0025] In step S2, for multiple preset pet furniture product categories, the system obtains raw data corresponding to the evaluation indicators from the target e-commerce platform, and cleans and standardizes the raw data to obtain a standardized data matrix. The standardized data matrix obtained in step S2 serves as the basis for subsequent weight calculation and comprehensive evaluation. Because the dimensions and numerical ranges of different indicators vary greatly, standardization is necessary to eliminate the influence of dimensions and ensure that the indicators are numerically comparable.

[0026] Optionally, the process of determining the multiple pet furniture product categories preset in step S2 includes: constructing a multi-dimensional labeling system for pet furniture products, wherein the multi-dimensional labeling system includes at least one or more of the following dimensions: applicable pet type dimension, product function type dimension, space utilization type dimension, and material type dimension; and labeling the pet furniture products to be evaluated according to the multi-dimensional labeling system, and grouping products with the same label combination into the same product category.

[0027] In some specific embodiments, in step S2, the system uses web crawlers or mature SaaS data tools to collect raw data corresponding to the evaluation indicators from the target e-commerce platform for each product category. The evaluation indicators include the static and dynamic indicators determined in step S1. The collected raw data may contain missing values ​​and outliers, requiring handling of these values. The main methods include: for missing values, imputation using the mean; for outliers, identifying and removing extreme values ​​caused by promotional activities, data erroneous capture, etc. Optionally, in step S2, the system uses range standardization to convert the initial values ​​of each evaluation indicator into dimensionless values, ensuring that all indicator values ​​fall within the [0,1] interval, facilitating subsequent calculations.

[0028] In step S3, the system uses the objective weighting method to calculate the evaluation weight of each evaluation indicator, and based on the evaluation weight and the standardized data matrix, calculates the comprehensive evaluation value of each pet furniture product category through a multi-objective decision analysis method.

[0029] Optionally, in step S3, the objective weighting method adopted by the system includes the combined weighting method, which determines the final evaluation weight by calculating the comprehensive weight of each evaluation index; wherein, the comprehensive weight includes objective weight and subjective weight.

[0030] In some specific embodiments, the system uses a combined weighting method to calculate the comprehensive weight of each evaluation indicator, specifically including: calculating the objective weight of each evaluation indicator according to the entropy weight method; using the analytic hierarchy process (AHP) to calculate the subjective weight of each evaluation indicator based on the scores from pet behavior experts and veterinarians; and weighting and fusing the objective weight and the subjective weight to obtain the comprehensive weight of each evaluation indicator.

[0031] In some specific embodiments, in step S3, the system uses the entropy weight method to calculate the objective weight of each evaluation index, specifically including: for the standardized data matrix, calculating the proportion of the i-th product category under the j-th index. Calculate the information entropy of the j-th indicator. Where k = 1 / ln(n), and n is the number of product categories. If If it is 0, then The value is 0. Calculate the entropy weight of the j-th indicator, i.e., the weight value. .

[0032] In some specific embodiments, in step S3, the system uses the Analytic Hierarchy Process (AHP) to calculate subjective weights. This specifically includes: inviting pet behavior experts and veterinarians to form an expert group to compare the importance of each indicator pairwise and construct a judgment matrix. The weight vectors of each indicator are calculated using the AHP method, and after consistency testing, the subjective weights are obtained. The subjective weights can be adjusted based on the position of the indicator in the coordinate space; for example, a higher weight can be given to indicators that are dominant in terms of pet value.

[0033] Optionally, in step S3, the system linearly weights the objective weights and subjective weights to obtain a comprehensive weight. Where α is an adjustment coefficient, which can be adjusted according to actual needs.

[0034] Optionally, in step S3, the system calculates the comprehensive evaluation value for each type of pet furniture product using a multi-objective decision analysis method. Specifically, this includes: using the TOPSIS method, defining the positive and negative ideal solutions for each evaluation indicator, calculating the distance between each evaluation indicator and both the positive and negative ideal solutions, and using the relative proximity as the comprehensive evaluation value for that evaluation object. In some specific embodiments, the specific implementation process of the TOPSIS method includes: (1) Constructing the weighted normalization matrix: Multiply the standardized data matrix by the weights to obtain the weighted normalization matrix. ; (2) Determine the positive ideal solution and the negative ideal solution: Positive Ideal Solution Among them, for benefit-type indicators, For cost-related indicators, In this plan, all items except "product negative review rate" are cost-based; the rest are benefit-based.

[0035] Negative ideal solution Among them: for benefit-type indicators, For cost-related indicators, .

[0036] (3) Calculate the distance Distance from the ideal solution for each product category .

[0037] Distance from the ideal solution for each product category

[0038] (4) Calculate the relative tracking schedule, i.e., the comprehensive evaluation value. , The closer the value is to 1, the better the product category.

[0039] In step S4, the system classifies design requirement levels based on the comprehensive evaluation values ​​of all pet furniture product categories. The core principle of step S4 is to automatically divide the continuous comprehensive evaluation values ​​calculated in step S3 into discrete design requirement levels with clear priorities using a hierarchical clustering algorithm.

[0040] In some specific implementations, while comprehensive evaluation values ​​can be ranked, they cannot directly answer the management decision question of "which products deserve priority resource allocation." Cluster analysis, by automatically identifying "natural groupings" based on the distribution characteristics of the data itself, transforms continuous values ​​into actionable priority levels.

[0041] In step S4, this invention employs the Ward clustering method, also known as the sum of squared deviations method. Its core idea is to select the merging method that minimizes the overall sum of squared deviations when merging two clusters at each step. This method tends to generate compact, well-balanced clusters, making it suitable for dividing comprehensive evaluation values ​​into clearly distinguishable hierarchical levels.

[0042] In some specific embodiments, the specific implementation process of step S4 includes: using the Ward clustering method, taking the comprehensive evaluation value of all pet furniture product categories as input, performing hierarchical clustering analysis, and generating a tree diagram structure; based on the tree diagram structure, determining that the samples are divided into three categories, defined as high design demand level, medium design demand level, and low design demand level respectively; assigning the high design demand level to the design strategy of priority development, the medium design demand level to the design strategy of selective development, and the low design demand level to the design strategy of postponement.

[0043] In step S5, the system generates and outputs corresponding product design direction decision suggestions based on the divided design requirement hierarchy. Step S5 also performs semantic conversion on the design requirement hierarchy divided in step S4 using a pet behavior and habit knowledge base to generate specific design instructions that designers can directly execute.

[0044] Specifically, step S5 involves building a knowledge base of pet behavior and habits, matching design requirements with pets' behavioral characteristics, spatial needs, and material preferences, and automatically generating refined design suggestions for specific product categories. Simultaneously, the system enhances the transparency and persuasiveness of decisions by visually displaying key data from the decision-making process.

[0045] In some specific embodiments, the specific implementation process of step S5 includes: constructing a pet behavior and habit knowledge base, which contains behavioral characteristics, spatial needs, and material preference information of different pet types; based on the division results of design requirement levels and combined with the pet behavior and habit knowledge base, generating refined design suggestions for specific pet furniture product categories, the design suggestions include at least one or more of the following: functional configuration suggestions, material selection suggestions, and spatial layout suggestions; and visually displaying information entropy, weight, distance between positive and negative ideal solutions, comprehensive evaluation value, and design requirement levels.

[0046] According to such Figure 1The implementation method shown in this invention involves constructing a multi-dimensional evaluation index system that integrates market, user, and pet dimensions. By combining objective weighting and multi-objective decision analysis, the original e-commerce data is transformed into comprehensive evaluation values ​​for each product category. Then, through clustering, the data is automatically divided into three design requirement levels: high, medium, and low, and corresponding priority development, selective development, and postponement development strategies are output. This forms a complete technical closed loop from data collection to design direction decision-making, thereby transforming traditional design decisions that rely on subjective experience into data-driven scientific decisions. This significantly improves the foresight, accuracy, and resource utilization efficiency of pet furniture product development.

[0047] Figure 2 for Figure 1 A flowchart illustrating step S1 in Example 1000. (See attached diagram.) Figure 2 As shown, step S1 includes steps S11-S15.

[0048] In step S11, the system constructs a two-dimensional coordinate space for the value of pets and owners, including the pet's value axis and the owner's value axis. Step S11 provides a semantic framework for subsequent indicator mapping, enabling each evaluation indicator to clearly reflect whether it serves the pet's needs or the owner's needs.

[0049] In some specific embodiments, in step S11, the system defines a two-dimensional coordinate space, including two orthogonal axes: Pet Value P-axis: Represents the degree to which a product meets a pet's behavioral, physiological, and psychological needs. The value ranges from [0,1], where 0 indicates that the product does not meet any of the pet's needs, and 1 indicates that it fully meets the pet's needs.

[0050] Owner Value O-axis: Represents the degree to which a product meets the owner's aesthetic, spatial, cleanliness, and cost-effectiveness requirements. The value range is [0,1], where 0 indicates that the product does not meet the owner's needs at all, and 1 indicates that it fully meets the owner's needs.

[0051] Any point in the coordinate space can be represented as (P, O), where P and O represent the coordinate values ​​of that point on the pet value axis and the owner value axis, respectively. For example, the point (0.9, 0.1) indicates that the indicator or theme is highly biased towards pet value, while the point (0.2, 0.8) indicates that it is highly biased towards owner value.

[0052] In step S12, the system selects multiple initial evaluation indicators from the market, user, and pet dimensions, and maps these initial evaluation indicators to their corresponding positions in a two-dimensional coordinate space. Step S12 establishes a basic evaluation framework based on expert knowledge and industry experience, and positions the preset indicators at specific locations in the coordinate space.

[0053] Optionally, the specific implementation of step S12 includes: The first step is for the system to select initial evaluation indicators from three dimensions: Market dimensions: search popularity, product supply and demand ratio, market size, brand concentration, product concentration, seller concentration, percentage of new products, percentage of new product sales, market potential, PPC bidding; User metrics: product ratings, negative review rate, product price; Pet dimensions: pet behavior fit, pet safety level.

[0054] The second step is the quantitative calculation of the indicator coordinates: (1) For pet-related and user-related metrics, an expert scoring method is used. The specific implementation process includes: 1. Establish a scoring expert group: Invite pet behavior experts and veterinarians to form a scoring group.

[0055] 2. Design scoring criteria: Define scoring scales for the pet's value axis and the owner's value axis: P-value rating criteria: 1.0, the indicator directly reflects the pet's core physiological / psychological needs; 0.7-0.9, the indicator is highly correlated with the pet's needs; 0.4-0.6, the indicator is moderately correlated with the pet's needs; 0.1-0.3, the indicator is lowly correlated with the pet's needs; 0.0, the indicator is unrelated to the pet's needs.

[0056] O-value scoring standard: Similar to the definition, it reflects the degree of correlation between the indicator and the owner's needs.

[0057] 3. Expert scoring: Each expert scores each indicator independently, including P-value and O-value.

[0058] 4. Comprehensive calculation: The final coordinates are calculated using the arithmetic mean or a weighted average based on expert authority.

[0059] (2) For market dimension indicators, semantic analysis is used. The specific implementation process includes: 1. Build a professional thesaurus: Pet behavior terminology: Contains 200-300 keywords related to pet needs, with each word assigned a pet-related value weight. For example, pet behavior keywords include climbing, scratching, hiding, jumping, resting, comfort, safety, bite resistance, scratching, activity space, and seclusion.

[0060] Homeowner Needs Keywords: Contains 200-300 keywords related to homeowner needs, with each keyword assigned a weight based on its relevance to the homeowner's needs. For example, keywords related to homeowner needs include aesthetics, simplicity, easy cleaning, space saving, cost-effectiveness, smart technology, environmental friendliness, durability, easy installation, and versatility.

[0061] 2. Metric Name Decomposition: Decompose the evaluation metric name into keyword combinations. Decomposition methods include word segmentation, synonym normalization, and removal of stop words. For example, the evaluation metric "search popularity" is decomposed into the keyword "search"; the evaluation metric "new product sales share" is decomposed into the keywords "new product" and "sales".

[0062] 3. Keyword Matching: Calculate the matching degree between the indicator name and the pet behavior terminology database; calculate the matching degree between the indicator name and the owner's demand terminology database; normalize the coordinates. Optionally, the matching degree calculation includes exact matching and fuzzy matching, where: exact matching refers to the keyword appearing directly in the terminology database; fuzzy matching calculates the similarity through word vectors, and is considered a match when the similarity exceeds a threshold.

[0063] 4. Coordinate Calculation: Calculate the P-value and O-value based on the matching results. Specifically: the pet matching degree is the sum of the weights of the matched keywords in the pet keyword database; the owner matching degree is the sum of the weights of the matched keywords in the owner keyword database; P = pet matching degree / (pet matching degree + owner matching degree), O = owner matching degree / (pet matching degree + owner matching degree).

[0064] In step S13, the system collects text data of pet furniture products and uses an unsupervised clustering algorithm to perform cluster analysis on the text data to discover implicit design themes. In some specific embodiments, semantically similar products often have the same or similar design features. By converting the text into semantic vectors and performing clustering in step S13, the system can automatically identify design patterns or emerging trends in the products that may not yet be covered in the preset initial evaluation indicators.

[0065] In some specific embodiments, in step S13, the system implementation process includes: (1) Data collection: Collect multi-source text data of pet furniture products from the target e-commerce platform, including product title, product description, attribute tags, and user review text; (2) Processing the text, including: Data cleaning: such as removing HTML tags, special characters, and stop words; For word segmentation, Chinese uses jieba segmentation, while English is segmented by spaces. Denoising and filtering text that is too short or too long, or irrelevant to the design; Text vectorization: This involves using a pre-trained language model to convert text into fixed-length semantic vectors. Commonly used models include BERT, Sentence-BERT, and RoBERTa. Each product's title, description, and tags are concatenated into a single document, which is then input into the model to obtain the product's semantic vector.

[0066] (3) Cluster analysis. Commonly used algorithms for clustering product vectors include: K-Means: Requires a preset number of clusters K, with the optimal K value determined using the elbow rule or silhouette coefficient. DBSCAN: Based on density, it automatically determines the number of clusters, can identify noise points, and is suitable for discovering clusters of arbitrary shapes. Hierarchical clustering: generates tree diagrams to facilitate observation of the hierarchical relationships between topics. The system selects an appropriate algorithm based on the data scale. After clustering, each cluster represents a latent design theme.

[0067] (4) Topic Extraction and Naming. For each cluster, calculate the TF-IDF value of all text within the cluster and extract the top N keywords as topic features. Based on the keywords and typical product descriptions, automatically generate topic names. For example, if the keyword cluster contains "smart", "APP", "automatic", and "sensor", the system will automatically name it "smart interaction".

[0068] In step S14, the system maps implicit design themes to a two-dimensional coordinate space, generating dynamic evaluation indicators. Each newly discovered implicit design theme needs to be assigned semantic meaning, that is, its position in the two-dimensional coordinate space of human-pet value needs to be determined. This is achieved through keyword matching: the keywords of the theme are matched with a pet behavior lexicon and an owner needs lexicon to calculate the theme's pet value tendency and owner value tendency. Then, the theme is transformed into a quantifiable dynamic evaluation indicator, allowing it to participate in subsequent evaluations like the initial indicator.

[0069] Specifically, the implementation process of step S14 is similar to the processing method of market dimension indicators in step S12. First, representative keywords of each theme are extracted from the clustering results, then keyword matching and coordinate calculation are performed, and finally the (P,O) coordinate values ​​of the dynamic evaluation indicators corresponding to each theme are obtained.

[0070] In step S14, the dynamism of the dynamic evaluation index is reflected in: Dynamic Topic Discovery: The system periodically re-collects the latest text data and re-performs cluster analysis. Newly emerging topics are automatically discovered, while outdated topics may be eliminated or downgraded. Dynamic Coordinates: Topic coordinates are not static. With updates to the thesaurus or semantic changes, coordinates may adjust after re-matching. Dynamic Indicators: Dynamic indicators are incremental; the system can continuously add new indicators to form an adaptive indicator system.

[0071] In some specific embodiments, in step S15, the system integrates the initial evaluation indicators and dynamic evaluation indicators to form a multi-dimensional evaluation indicator system. In step S15, the system merges the static initial evaluation indicators and dynamic evaluation indicators to form a unified multi-dimensional evaluation indicator system.

[0072] Figure 3 for Figure 1 A flowchart illustrating sub-step S2A of step S2 in Example 1000. (See attached diagram.) Figure 3 As shown, sub-step S2A is the process of determining multiple preset pet furniture product categories, specifically including steps SA1-SA2.

[0073] Step SA1: The system constructs a multi-dimensional tag system for pet furniture products. The multi-dimensional tag system includes at least one or more of the following dimensions: applicable pet type, product function type, space utilization type, and material type.

[0074] In some specific embodiments, in step SA1, the system presets multiple label dimensions, each containing several label values. This solution includes at least the four core dimensions shown in Table 1 below, which can be expanded as needed in practical applications.

[0075] Table 1. Example of a multi-dimensional labeling system for pet furniture products: .

[0076] Optionally, the labeling system in step SA1 can be constructed as follows: Expert Definition: Pet furniture industry experts and designers define the initial labeling system based on market understanding; Data-driven expansion: Through text mining, high-frequency attribute words are automatically extracted from product titles and descriptions, and then added to the tag system after filtering. For example, if "foldable" is found to be a high-frequency word, a new "foldable" tag can be added or it can be classified into "space utilization type".

[0077] In step SA2, the system labels the pet furniture products to be evaluated based on the multi-dimensional labeling system, and groups products with the same label combination into the same product category.

[0078] In some specific embodiments, in step SA2: for large-scale product processing, the label values ​​are automatically identified based on product titles, descriptions, and attribute tags through keyword matching or text classification models; for small-scale or products with ambiguous boundaries, labelers make manual judgments based on product information and images to ensure labeling accuracy.

[0079] Optionally, in step SA2, the system groups all products according to their label combinations. Specific rules: products with identical label combinations are grouped into the same product category; the label combination serves as a unique identifier for that category.

[0080] For example, the system categorizes the following products into the same category: Product A: Title "Solid Wood Floor-Standing Cat Climbing Frame", Description "Multi-level climbing, sisal posts", Tag combination: {cat, climbing, floor-standing, solid wood}; Product B: Title "Large Cat Tree", Description "Solid wood structure, sisal wrapping", Tag combination: {cat, climbing, floor-standing, solid wood}; Both have the same tag combination and are categorized as "cat-climbing-floor-mounted-solid wood".

[0081] Figure 4 for Figure 1 A flowchart illustrating step S4 in Example 1000. (See attached diagram.) Figure 4 As shown, step S4 includes steps S41-S43.

[0082] In step S41, the system uses Ward clustering, taking the comprehensive evaluation value of all pet furniture product categories as input, to perform hierarchical clustering analysis and generate a tree diagram structure. In step S41, the system takes the comprehensive evaluation value of all pet furniture product categories output in step S3 as input. Assume there are n product categories, each corresponding to a comprehensive evaluation value. (i=1,2,...,n).

[0083] In step S41, initially, each product category is a separate category, resulting in n categories. The system calculates the Ward distance between every two categories. The Ward distance is defined as the difference between the sum of squared intra-class deviations of the new category after merging two categories and the sum of squared intra-class deviations of the two categories before merging. The calculation formula is D(A,B) = ESS(AB) - [ESS(A) + ESS(B)], where: That is, the distance from each point within class A to the center of class A The sum of squared deviations is calculated as follows: ESSB is calculated similarly; ESS(AB) is the sum of squared deviations of the new class after merging. When each class contains only one sample point, ESS=0, so the initial distance is the square of the difference between the two sample points.

[0084] In step S41, after calculating the Ward distance, iterative merging of classes is performed: the two classes with the smallest Ward distance are found and merged into a new class; the Ward distance between the new class and all other classes is recalculated; this process is repeated until all samples are merged into one class. Then, the system records the class name and the distance value at the time of merging each time, generating a hierarchical clustering dendrogram. The vertical axis of the dendrogram represents the distance at the time of merging, and the horizontal axis represents each product category.

[0085] In step S42, the system determines the classification of samples into three categories based on the tree diagram structure, defined as high design requirement level, medium design requirement level, and low design requirement level, respectively. In some specific embodiments, in step S42, the system automatically determines the optimal number of clusters based on the tree diagram structure. The determination method includes: Elbow rule: Observe the jump point in the inter-class distance in the dendrogram and select the number of clusters before the jump point. When the inter-class distance suddenly increases, it means that merging two more classes will lead to a significant increase in intra-class differences, and the number of clusters at this point is optimal.

[0086] Preset Threshold: Based on business needs, the system can preset a distance threshold, stopping the merging process when the distance exceeds this threshold. In this solution, based on experimental analysis of a large amount of pet furniture data, dividing it into three levels provides the best interpretability.

[0087] (2) Define the hierarchy The system sorts the three identified clusters from highest to lowest based on their comprehensive evaluation scores, and defines them as follows: High design requirement level: Cluster with the highest comprehensive evaluation value Middle design requirement level: Clustering with a moderate overall evaluation value Lowest design requirement level: Cluster with the lowest overall evaluation value Optionally, in some specific embodiments, step S42 of the present invention further includes step S42a: the system constructs a two-dimensional cross-analysis matrix of pet type and function type, and further refines the clustering results based on the two-dimensional cross-analysis matrix to generate differentiated design requirement levels for specific pet types and specific function types.

[0088] In some specific embodiments, after obtaining three design requirement levels through Ward clustering in step S41, in step S42a, the system further refines the clustering results using a two-dimensional cross-analysis matrix of pet type and functional type. The principle behind this supplementary step is that different product categories within the same design requirement level should have differentiated design strategies due to their different pet types and functional types. For example, even within the same "high design requirement level," "cat-climbing" products should emphasize vertical space utilization and scratching functions, while "dog-sleeping" products should emphasize comfort and ease of cleaning. The two-dimensional cross-analysis matrix constructs a grid of "pet type × functional type," locating each product category to its corresponding element cell, thereby identifying the differences in design requirements across different sub-sectors.

[0089] Specifically, step S42a is implemented in the following ways: (1) Based on the multi-dimensional tag system established in step S2, the system extracts the two dimensions of applicable pet type and product function type for each product category, and constructs a two-dimensional matrix. Each element in the matrix is ​​an element cell of a subdivided field. (2) Fill the matrix and generate differentiated design requirement levels. The system locates each product category in the clustering results of step S41 to the corresponding element cell in the matrix according to its pet type and function type. For each element, the system calculates the distribution of the comprehensive evaluation value of all product categories in that sub-field and determines the comprehensive design requirement level of that element. For example: Matrix element (cat, climbing): comprehensive evaluation value 0.75-0.82, judged as high design requirement level.

[0090] The final output of step S42a is a differentiated design requirement hierarchy for specific pet types and specific functional types, such as: "Cat climbing products: high design requirement hierarchy".

[0091] In step S43, the system assigns a high design requirement level to a priority development design strategy, a medium design requirement level to a selective development design strategy, and a low design requirement level to a design strategy that is temporarily deferred. In some specific embodiments, in step S43, the system maps the three levels one-to-one with the design management strategy: High design requirements → Prioritize development: It is recommended to invest major design resources, explore design innovation, and strive to create a blockbuster product; Design requirements level → Selective development: It is recommended to allocate limited resources for targeted optimization and improvement, rather than as a key investment. Low design requirement level → Temporarily suspend development: It is recommended to temporarily suspend investment in design resources, maintain the existing product line or gradually phase it out.

[0092] According to such Figure 4 The implementation method shown in this invention proposes a solution that automatically generates a dendrogram and objectively divides the continuous comprehensive evaluation values ​​into three design requirement levels: high, medium, and low, by inputting them into the Ward clustering algorithm. This realizes the transformation from abstract numerical evaluation to executable design priorities. This process eliminates the arbitrariness of manually setting score lines, makes the decision-making basis based on the inherent distribution of data, and transforms the technical analysis results into management instructions that can directly guide the allocation of design resources.

[0093] As a preferred embodiment, this invention also refines the hierarchical segmentation through a two-dimensional cross-analysis matrix of pet type and function type. The two-dimensional matrix reveals the differences in demand among different sub-fields within the same level, providing design managers with a precise decision-making basis for "which categories to prioritize, which pets to develop, and which functions to develop".

[0094] Figure 5 for Figure 1 A flowchart illustrating step S5 in Example 1000. (See attached diagram.) Figure 5 As shown, step S5 includes steps S51-S53.

[0095] In step S51, the system constructs a pet behavior and habit knowledge base, which includes behavioral characteristics, space requirements, and material preferences for different pet types. In some specific embodiments, the pet behavior and habit knowledge base in step S51 is a structured database containing multiple knowledge entries. Each knowledge entry corresponds to a "pet type - function type" combination, recording design guidance information under that combination. The core fields of the knowledge base are shown in Table 2 below: Table 2. Example of core fields in the pet behavior and habits knowledge base: .

[0096] Alternatively, in some specific embodiments, the knowledge base can be constructed in the following ways: Expert knowledge integration: Invite pet behavior experts, veterinarians, and experienced pet furniture designers to provide professional advice and form an initial knowledge base. Literature review: Consult relevant literature on animal behavior and pet furniture design to extract general design principles. Data feedback updates: User reviews and pet behavior feedback after product launch can be used to optimize knowledge base content.

[0097] In step S52, based on the classification of design requirements and combined with the pet behavior and habit knowledge base, the system generates refined design suggestions for specific pet furniture product categories. The design suggestions include at least one or more of the following: functional configuration suggestions, material selection suggestions, and spatial layout suggestions.

[0098] In some specific embodiments, in step S52, the system inputs the output results of step S4, including: the design requirement level for each product category and the tag combination for each product category. Optionally, step S52 may also include the coordinate values ​​(P, O) of each indicator in step S1 for value orientation analysis.

[0099] Optionally, the suggested generation logic in step S52 is that the system generates design suggestions using a rule-based matching method, specifically including: retrieving matching knowledge entries from the knowledge base based on the applicable pet type and function type of the product category; if the match is successful, using the design elements, material preferences, and space requirements in the knowledge entries as basic suggestions; if multiple matches exist simultaneously, the system merges the suggestions or sorts them by confidence level; if the match fails, the system can generate suggestions based on general knowledge of pet type or function type and mark them as "to be improved".

[0100] Optionally, in step S52, the system can combine the coordinate information of the indicators in step S1 to provide differentiated suggestions for products within the same design requirement level: For pet-value-driven products (P>preset first threshold): it is recommended to focus on behavioral adaptation design, such as adding climbing structures, optimizing hiding spaces, and using wear-resistant materials; For owner-value-driven products (O>preset first threshold): it is recommended to focus on ergonomics, aesthetics and easy-to-clean design, such as simple shape, removable and washable fabric, and integration with home style; For dual-subject balanced products (preset second threshold ≤ P, O ≤ preset first threshold): it is recommended to take into account the needs of both parties, such as modular and changeable structure, and simultaneously meet the needs of pet climbing and owner storage.

[0101] In some specific embodiments, in step S52, the system outputs structured design suggestions, which include the following elements: Feature configuration suggestions: such as "It is recommended to design a 3-4 level platform with a cave-like rest area"; Material selection suggestions: such as "the main body is made of solid wood, the scratching parts are wrapped with sisal rope, and the rest area uses a removable and washable plush mat"; Space layout suggestions: such as "suitable for placement in the corner of the living room or balcony, utilizing vertical space and not taking up floor space"; Design Priority: Based on the design requirements hierarchy, prioritize development / selectively develop / delay development.

[0102] In step S53, the system visualizes the information entropy, weights, distance between positive and negative ideal solutions, comprehensive evaluation value, and design requirement hierarchy. In some specific embodiments, in step S53, the system visualizes key data from the decision-making process, including: Indicator weight chart: Bar charts or pie charts show the weight distribution of each indicator, making it easier to understand which indicators have the greatest impact on the evaluation results; Overall Evaluation Value Ranking Chart: The bar chart displays the overall evaluation value of each product category, allowing for a direct comparison of their strengths and weaknesses; Clustering dendrogram: Displays the dendrogram generated in step S4, with the dividing points of the three design requirement levels marked; Two-dimensional cross-analysis matrix: Heatmap showing the hierarchical design requirements based on pet type × functional type; Distance chart between positive and negative ideal solutions: A radar chart or scatter plot shows the distance between each product category and the positive and negative ideal solutions; According to such Figure 5In the implementation method shown, step S5 of the proposed solution of the present invention constructs a knowledge base of pet behavior habits and matches it with the design requirement hierarchy. It transforms the abstract clustering results into functional configurations, material selections and spatial layout suggestions that designers can directly execute. At the same time, it combines the value coordinate information of indicators to achieve differentiated guidance within the same level. Then, it uses data visualization to display the key parameters such as weight, distance and evaluation value of the entire decision-making process, and transforms market data into specific design instructions, so that design decisions are transformed from guesswork based on experience to technical implementation of "data-driven + professional knowledge support".

[0103] Figure 6 This is a schematic diagram of an embodiment 2000 of a data-driven pet furniture product design direction decision system according to the present invention. Figure 6 As shown, Implementation Example 2000 includes an indicator system construction module 201, a data acquisition and processing module 202, a comprehensive evaluation module 203, a demand hierarchy module 204, and a strategy generation module 205.

[0104] The indicator system construction module 201 is used to select multiple evaluation indicators from market, user, and pet dimensions to construct a multi-dimensional evaluation indicator system for pet furniture. The indicator system construction module 201 combines static indicator selection with dynamic indicator discovery to build a multi-dimensional evaluation indicator system that comprehensively reflects the needs of both humans and pets for pet furniture.

[0105] In some specific embodiments, the specific implementation process of the indicator system construction module 201 includes: (1) Initial indicator selection: Multiple initial evaluation indicators were selected from the market dimension, user dimension and pet dimension; (2) Construction of human-pet value coordinates: Define a two-dimensional coordinate space containing the pet value axis and the owner value axis, map the initial evaluation indicators to the corresponding positions in the coordinate space through expert scoring or semantic analysis, and assign value orientation labels to each indicator; (3) Dynamic indicator discovery: Collect text data of pet furniture products, use unsupervised clustering algorithm to perform cluster analysis, discover implicit design themes; match theme keywords with pet behavior lexicon and owner demand lexicon, calculate the coordinates of the theme in coordinate space, and generate dynamic evaluation indicators; (4) System integration: The static initial indicators and dynamic discovery indicators are merged to form a multi-dimensional evaluation indicator system with complete attributes such as indicator name, dimension, coordinate, type, data source, and calculation method.

[0106] The data acquisition and processing module 202 is used to acquire raw data corresponding to evaluation indicators from target e-commerce platforms for multiple preset pet furniture product categories, and to clean and standardize the raw data to obtain a standardized data matrix. The data acquisition and processing module 202 performs refined classification of pet furniture products through a multi-dimensional tagging system, and collects corresponding indicator data from e-commerce platforms, forming a data matrix with unified dimensions after standardization. Optionally, the main functions performed include: (1) Product category determination: Construct a multi-dimensional labeling system that includes multiple dimensions such as applicable pet type, product function type, space utilization type, and material type; Based on this labeling system, label the pet furniture products to be evaluated, and group products with the same label combination into the same product category. (2) Data collection: For each preset product category, raw data corresponding to the evaluation indicators are obtained from the target e-commerce platform through web crawlers or SaaS tools, including static indicator data and dynamic indicator data; (3) Data cleaning and standardization: missing values ​​are filled and outliers are removed from the original data; the range standardization method is used to convert the index data of different dimensions into dimensionless values ​​to obtain a standardized data matrix.

[0107] The comprehensive evaluation module 203 is used to calculate the evaluation weights of each evaluation indicator using an objective weighting method, and based on the evaluation weights and a standardized data matrix, to calculate the comprehensive evaluation value for each pet furniture product category using a multi-objective decision analysis method. The comprehensive evaluation module 203 determines the weights of each indicator using an objective weighting method and calculates the comprehensive evaluation value for each product category using a multi-objective decision analysis method. Its implementation process includes: (1) Weight calculation: The entropy weight method is adopted to calculate the information entropy based on the degree of dispersion of each indicator in the standardized data matrix, and then obtain the objective weight; optionally, the combined weighting method is adopted to combine the objective weight calculated by the entropy weight method with the subjective weight calculated by the analytic hierarchy process (AHP) based on the scores of pet behavior experts and veterinarians to obtain the comprehensive weight; the setting of the subjective weight can be adjusted according to the position of the indicator in the coordinate space. (2) Comprehensive evaluation: The TOPSIS method is used to multiply the standardized data matrix with the weights to obtain the weighted normalized matrix; the positive ideal solution and negative ideal solution of each indicator are defined; the Euclidean distance of each product category to the positive ideal solution and negative ideal solution is calculated, and the relative proximity is used as the comprehensive evaluation value of the product category.

[0108] The requirement hierarchy module 204 is used to divide design requirement levels based on the comprehensive evaluation values ​​of all pet furniture product categories using a clustering algorithm. The requirement hierarchy module 204 divides continuous comprehensive evaluation values ​​into discrete design requirement levels through hierarchical clustering. In a preferred embodiment, a two-dimensional cross-analysis matrix is ​​also used to refine the hierarchy.

[0109] In some specific embodiments, the requirement hierarchy module 204 completes clustering and hierarchical classification: using the Ward clustering method, with the comprehensive evaluation value of all product categories as input, hierarchical clustering analysis is performed to generate a tree diagram structure; based on the jump points of the inter-class distance in the tree diagram, the samples are automatically divided into three categories; defined as high design requirement level, medium design requirement level and low design requirement level respectively according to the comprehensive evaluation value from high to low.

[0110] In some specific embodiments, the requirement level module 204 completes the strategy mapping: the high design requirement level corresponds to the design strategy of priority development, the medium design requirement level corresponds to the design strategy of selective development, and the low design requirement level corresponds to the design strategy of postponement.

[0111] Optionally, in some specific embodiments, the requirement hierarchy module 204 constructs a two-dimensional cross-analysis matrix of pet type and function type, locates each product category to the corresponding cell, statistically analyzes the distribution of the comprehensive evaluation value of the product category in each cell, and generates a differentiated design requirement hierarchy for specific pet type and specific function type.

[0112] The strategy generation module 205 generates and outputs corresponding product design direction decision suggestions based on the divided design requirement hierarchy. The strategy generation module 205 transforms the design requirement hierarchy into specific design suggestions through a pet behavior and habit knowledge base, and enhances decision-making transparency through visualization. In some specific embodiments, the strategy generation module 205 pre-builds a pet behavior and habit knowledge base, containing information such as behavioral characteristics, spatial needs, material preferences, and design elements for different pet types; the knowledge base can be continuously updated through expert import, literature research, and data feedback.

[0113] In some specific embodiments, the strategy generation module 205, based on the hierarchical division of design requirements, retrieves matching knowledge entries from the knowledge base according to the applicable pet type and functional type of the product category, and generates refined design suggestions for specific product categories, including functional configuration suggestions, material selection suggestions, and spatial layout suggestions. Optionally, in some specific embodiments, the strategy generation module 205 combines the value orientation of indicators in coordinate space to provide differentiated suggestions for products within the same level.

[0114] In some specific embodiments, the strategy generation module 205 visualizes key data in the decision-making process, including indicator weight charts, comprehensive evaluation value ranking charts, clustering dendrograms, two-dimensional cross-analysis matrix heatmaps, and positive and negative ideal solution distance charts, supporting user interaction and report export.

[0115] Figure 7-10 A key interface diagram of a home furnishing product information management platform constructed according to a data-driven pet furniture product design direction decision-making method of the present invention.

[0116] like Figure 7 The diagram shows the interface for defining indicators. Users can add or remove indicators based on actual needs. To add an indicator, click "Add Indicator"; to delete an indicator, click "Delete". Clicking "Save" will save the defined indicator properties. New evaluations will inherit the previously saved indicator properties.

[0117] like Figure 8 The diagram shows the interface displaying the information entropy and weights of each evaluation indicator. In this invention, the entropy weight method is used to calculate the information entropy based on the dispersion of each indicator in the standardized data matrix, thereby obtaining the objective weight; the combined weighting method is used to weight and fuse the objective weights calculated by the entropy weight method with the subjective weights calculated by the Analytic Hierarchy Process (AHP) based on scores from pet behavior experts and veterinarians, to obtain the comprehensive weight.

[0118] like Figure 9 The diagram shows the interface for the distance between positive and negative ideal solutions and the overall evaluation value. In this invention, a multi-objective decision analysis method is used to calculate the overall evaluation value for each pet furniture product category.

[0119] like Figure 10 The image shown is a screenshot of the data visualization interface. Figure 10 The system provides data visualization of information entropy, weights, distance between positive and negative ideal solutions, and comprehensive evaluation, allowing users to more intuitively understand the weights of each indicator and the gaps between the evaluation results of each evaluation object.

[0120] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0121] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A data-driven pet furniture product design direction decision-making method, characterized in that, include: Step S1: Select multiple evaluation indicators from market, user, and pet dimensions to construct a multi-dimensional evaluation indicator system for pet furniture; Step S2: For multiple preset pet furniture product categories, obtain the raw data corresponding to the evaluation indicators from the target e-commerce platform, and clean and standardize the raw data to obtain a standardized data matrix; Step S3: Calculate the evaluation weights of each evaluation index using the objective weighting method, and calculate the comprehensive evaluation value of each pet furniture product category based on the evaluation weights and the standardized data matrix using a multi-objective decision analysis method. Step S4: Classify the design requirement levels based on the comprehensive evaluation values ​​of all pet furniture product categories; Step S5: Based on the defined design requirement levels, generate and output corresponding product design direction decision suggestions.

2. The method of claim 1, wherein, The construction of a multi-dimensional evaluation index system for pet furniture in step S1 specifically includes: Construct a two-dimensional coordinate space for human-pet value that includes a pet value axis and an owner value axis; Multiple initial evaluation indicators are selected from market, user and pet dimensions, and the initial evaluation indicators are mapped to their corresponding positions in the two-dimensional coordinate space. Text data on pet furniture products was collected, and unsupervised clustering algorithms were used to perform cluster analysis on the text data to discover implicit design themes. The implicit design theme is mapped to the two-dimensional coordinate space to generate dynamic evaluation indicators; The initial evaluation indicators and the dynamic evaluation indicators are combined to form the multidimensional evaluation indicator system.

3. The method of claim 1, wherein, In step S2, the process of determining the preset multiple pet furniture product categories specifically includes: Construct a multi-dimensional labeling system for pet furniture products, wherein the multi-dimensional labeling system includes at least one or more of the following dimensions: applicable pet type dimension, product function type dimension, space utilization type dimension, and material type dimension; Based on the aforementioned multi-dimensional labeling system, the pet furniture products to be evaluated are labeled, and products with the same label combination are grouped into the same product category.

4. The method of claim 1, wherein, Step S3 uses an objective weighting method to calculate the evaluation weights of each evaluation index, specifically including: The comprehensive weight of each evaluation index is calculated using a combined weighting method, and the comprehensive weight is used as the evaluation weight; the comprehensive weight includes objective weight and subjective weight.

5. The method of claim 4, wherein, The method of calculating the comprehensive weight of each evaluation index using a combined weighting method includes: The objective weights of each evaluation index are calculated using the entropy weight method. Using the analytic hierarchy process (AHP), the subjective weights of each evaluation indicator were calculated based on scores from pet behavior experts and veterinarians. The objective weights and subjective weights are weighted and fused together to obtain the comprehensive weights of each evaluation indicator.

6. The method of claim 1, wherein, In step S3, calculating the comprehensive evaluation value for each type of pet furniture product using a multi-objective decision analysis method specifically includes: The TOPSIS method is used to define the positive and negative ideal solutions for each evaluation index, calculate the distance between each evaluation index and the positive and negative ideal solutions, and use the relative proximity as the comprehensive evaluation value of the evaluation object.

7. The method of claim 1, wherein, Step S4 specifically includes: The Ward clustering method was used, with the comprehensive evaluation value of all pet furniture product categories as input, to perform hierarchical clustering analysis and generate a tree diagram structure; Based on the tree diagram structure, the samples are divided into three categories, which are defined as high design requirement level, medium design requirement level and low design requirement level, respectively. The design strategy of prioritizing development corresponds to the high design requirement level, the design strategy of selective development corresponds to the medium design requirement level, and the design strategy of postponement corresponds to the low design requirement level.

8. The method of claim 7, wherein, The step S4 of dividing the design requirement hierarchy also includes: A two-dimensional cross-analysis matrix of pet type and function type is constructed. Based on the two-dimensional cross-analysis matrix, the clustering results are further refined to generate a hierarchical design requirement for specific pet type and specific function type.

9. The method of claim 1, wherein, The specific steps for generating and outputting the design strategy in step S5 include: Step S51: Construct a pet behavior and habit knowledge base, which includes behavioral characteristics, space requirements, and material preferences of different pet types; Step S52: Based on the classification results of the design requirements hierarchy and combined with the pet behavior and habit knowledge base, generate refined design suggestions for specific pet furniture product categories. The design suggestions shall include at least one or more of the following: functional configuration suggestions, material selection suggestions, and spatial layout suggestions. Step S53: Visualize the information entropy, weight, distance between positive and negative ideal solutions, comprehensive evaluation value, and design requirement level.

10. A data-driven decision-making system for pet furniture product design, characterized in that, include: The indicator system construction module is used to select multiple evaluation indicators from market, user and pet dimensions to build a multi-dimensional evaluation indicator system for pet furniture; The data acquisition and processing module is used to acquire raw data corresponding to the evaluation indicators from the target e-commerce platform for multiple preset pet furniture product categories, and to clean and standardize the raw data to obtain a standardized data matrix. The comprehensive evaluation module is used to calculate the evaluation weight of each evaluation indicator using an objective weighting method, and to calculate the comprehensive evaluation value of each pet furniture product category based on the evaluation weight and the standardized data matrix using a multi-objective decision analysis method. The requirement hierarchy module is used to divide the design requirement levels based on the comprehensive evaluation value of all the pet furniture product categories using a clustering algorithm; The strategy generation module is used to generate and output corresponding product design direction decision suggestions based on the defined design requirement levels.

Images