Big data intelligent delivery system based on marketing data analysis

Abstract

This invention relates to the field of fast-moving consumer goods (FMCG) data analysis technology and discloses a big data intelligent advertising system based on marketing data analysis. The system includes a demand analysis unit, a scenario analysis unit, a fusion unit, a segmentation unit, and an advertising unit. By mining implicit demand intensity indices and combining spatiotemporal constraints with decision-making resistance coding and scenario adaptability, it generates user demand potential intensity and divides users into three circles. This accurately identifies user groups with potential consumption needs but currently weak feedback. For the main resistance factors of different circles, it adopts differentiated advertising decisions based on price sensitivity, information gaps, decision complexity, etc., and combines these with a consumption willingness value correction strategy. This effectively eliminates decision-making resistance for users in each circle, converts potential consumption potential, avoids resource misallocation, and improves the long-term return on investment for big data advertising in the home furnishing FMCG sector.

Description

Technical Field

[0001] This invention relates to the field of fast-moving consumer goods (FMCG) data analysis technology, specifically to a big data intelligent advertising system based on marketing data analysis. Background Technology

[0002] Currently, the core of intelligent coupon delivery lies in determining the target audience based on marketing data. By using machine learning algorithms to conduct in-depth analysis of users' purchase history, browsing history, favorites, and other data, more resources can be accurately delivered to user groups with strong feedback behavior, thereby ensuring the return on investment.

[0003] However, in the home furnishing and fast-moving consumer goods (FMCG) sector, big data-driven intelligent advertising has significant drawbacks. Specifically, in the intelligent advertising of coupons, the current advertising decisions rely excessively on users' explicit feedback behavior. This results in insufficient ability to identify user groups in the home furnishing and FMCG sector who have potential consumption needs but currently exhibit weak feedback behavior. Consequently, it is unable to effectively convert the consumption potential of these users, leading to a lower long-term return on investment. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a big data intelligent advertising system based on marketing data analysis, which solves the aforementioned problems.

[0005] The above-mentioned technical objective of the present invention is achieved through the following technical solution:

[0006] A big data-driven intelligent advertising system based on marketing data analysis includes:

[0007] The demand analysis unit is used to acquire marketing data for home furnishing fast-moving consumer goods, analyze the marketing data to obtain the implicit demand intensity index and the consumption intention value, and merge the implicit demand intensity index and the consumption intention value to generate the intention entropy value.

[0008] The scenario analysis unit is used to analyze the intention entropy value, generate decision-making hindrance codes with spatiotemporal constraints, extract marketing data, and obtain the user's scenario fit.

[0009] The fusion unit is used to fuse the intention entropy value, decision-making hindrance encoding, and scenario adaptability to generate the potential strength of user demand.

[0010] The segmentation unit is used to divide users into three circles based on the intensity of user demand potential and decision-making resistance coding;

[0011] The delivery unit is used to calculate the main blocking factor of the decision blocking code for each of the three types of concentric circles, and to make different delivery decisions based on the main blocking factor.

[0012] Furthermore, analysis of marketing data yields an implicit demand intensity index and a consumer willingness value, including:

[0013] Based on marketing data analysis, user focus shifts at different decision-making stages, generating behavioral entropy values;

[0014] Analyze the semantic similarity between user expressions and product attributes in marketing data to generate a semantic coupling coefficient;

[0015] The demand fluctuation coefficient is obtained by weighting and fusing the behavioral entropy value and the semantic coupling coefficient.

[0016] By integrating behavioral entropy, semantic coupling coefficient, and demand fluctuation coefficient, we obtain the implicit demand intensity index and the consumption willingness index.

[0017] Furthermore, by integrating behavioral entropy, semantic coupling coefficient, and demand fluctuation coefficient, we obtain the implicit demand intensity index and consumption intention value, including:

[0018] The behavioral entropy value, semantic coupling coefficient, and demand fluctuation coefficient are mapped to a three-dimensional tensor space to generate a composite feature matrix.

[0019] Transform the composite feature matrix to generate a four-dimensional demand potential energy field;

[0020] An analysis of the four-dimensional demand potential energy field yields the implicit demand intensity index.

[0021] The four-dimensional demand potential field is spatially divided to generate a consumption intention value, which is divided into four levels: S-level, A-level, B-level and C-level.

[0022] Furthermore, the implicit demand intensity index is integrated with the consumption intention value to generate an intention entropy value, including:

[0023] By analyzing the implicit demand intensity index and the willingness to consume, a dynamic weight matrix is ​​obtained. The dynamic weight matrix is ​​then subjected to time decay to obtain an asymmetric three-dimensional tensor.

[0024] By superimposing characteristic states on an asymmetric three-dimensional tensor, a composite quantum state vector containing the user's true intent is generated.

[0025] Constraint feature extraction is performed on the composite quantum state vector to obtain the intention entropy value with spatiotemporal characteristics.

[0026] Furthermore, the intention entropy value is analyzed to generate spatiotemporal constraint decision-making hindrance codes, including:

[0027] After mapping the intentional entropy value to a quantum state vector, a transformation is performed to generate a spatiotemporal frequency characteristic spectrum.

[0028] The spatiotemporal frequency feature spectrum is extracted to obtain the time decay coefficient, spatial thermodynamic value, and intentional divergence value;

[0029] By fusing the time decay coefficient, spatial thermodynamic value, and intentional divergence value, a multidimensional hindrance factor matrix is ​​obtained.

[0030] By analyzing and encoding the multidimensional blocking factor matrix, a decision blocking code with spatiotemporal constraints is obtained.

[0031] Furthermore, marketing data is extracted to obtain user scenario suitability, including:

[0032] Determine the spatiotemporal density of user activities based on marketing data;

[0033] By combining spatiotemporal density with the functional attributes of home furnishing fast-moving consumer goods, the field overlap coefficient is obtained;

[0034] Based on the field overlap coefficient, the evolution path of user needs is determined, and a context transition probability matrix is ​​generated;

[0035] By fusing spatiotemporal density, field overlap coefficient, and context transition probability matrix using an asymmetric tensor, the user's scenario adaptability is generated.

[0036] Furthermore, the intention entropy value, decision-making hindrance encoding, and scenario adaptability are integrated to generate the user demand potential intensity, including:

[0037] The intention entropy value and the decision-blocking code are calculated to generate a four-dimensional spatiotemporal intention field matrix;

[0038] The four-dimensional spatiotemporal intention field matrix is ​​corrected according to the scene adaptability to generate the corrected spatiotemporal intention field.

[0039] Analyze the spatiotemporal willingness field to generate a user demand potential matrix;

[0040] Extract the user demand potential energy matrix to obtain the user demand potential energy intensity.

[0041] Furthermore, based on the intensity of user demand potential and the coding of decision-making obstacles, users are divided into three circles, including:

[0042] The dynamic potential energy threshold vector is determined by analyzing the four-dimensional spatiotemporal intention field matrix.

[0043] Analysis of decision-making obstruction codes yields the resonance degree between users and the value proposition of home furnishing fast-moving consumer goods;

[0044] Based on the dynamic potential threshold vector and resonance, three types of user circles are generated in four-dimensional space: high resonance layer, steady-state resonance layer, and low resonance layer.

[0045] Furthermore, for the three types of concentric circles, the main hindering factors of the decision-making hindering coding are calculated respectively, including:

[0046] The decision-making obstruction code is extracted to obtain the coupling strength between each factor and the potential energy of user demand;

[0047] Based on the user's social circle, configure differentiated coupling strength thresholds and entropy reduction efficiency weight coefficients;

[0048] Based on the coupling strength threshold and entropy reduction efficiency weighting coefficient, the coupling strength is weighted to generate a decision resistance distribution cloud map.

[0049] Calculate the energy entropy reduction gradient of each hindering factor in the decision resistance distribution cloud map, and generate an entropy reduction efficiency ranking table.

[0050] The top three factors in the entropy reduction efficiency ranking table are extracted as candidate principal blocking factors, and the entropy reduction contribution rate of each candidate factor is calculated.

[0051] The principal blocking factor is determined based on the entropy reduction contribution rate.

[0052] Furthermore, different deployment decisions are made based on the main blocking factor, including:

[0053] When the primary blocking factor is price sensitivity, the first deployment decision is adopted;

[0054] When the primary blocking factor is information gap, the second delivery decision is adopted;

[0055] When the main blocking factor type is decision complexity, the third delivery decision is adopted.

[0056] In summary, the present invention has the following main beneficial effects:

[0057] By mining the implicit demand intensity index and combining it with spatiotemporal constraint decision-making hindrance coding and scenario adaptability, the potential energy intensity of user demand is generated. This effectively overcomes the drawback of traditional advertising that relies too much on explicit feedback. By analyzing the behavioral entropy and semantic coupling coefficient of users in the four stages of cognition, interest, consideration, and decision-making, it can accurately capture user groups with potential consumption needs but weak current feedback. This multi-dimensional integrated identification method breaks through the limitations of a single data dimension, provides a precise target group for subsequent precise advertising, and solves the problem of insufficient identification of users with implicit needs in existing solutions.

[0058] By segmenting users into three groups and developing differentiated targeting decisions based on the main bottleneck factors for each group, the big data-driven targeting can be tailored to the needs of different users. At the same time, by combining a consumption intention value correction strategy, the targeting content is ensured to be highly matched with user needs, avoiding the waste of resources from indiscriminate targeting and significantly improving the conversion efficiency of potential users, especially activating the consumption potential of users in the low resonance layer.

[0059] By activating the consumption potential of potential users and reducing resource misallocation, the long-term return on investment has been effectively improved. On the one hand, the potential value is quantified by the strength of user demand, avoiding excessive concentration of resources on users with explicit feedback. On the other hand, based on spatiotemporal dynamic analysis and scenario adaptability, the investment strategy is adjusted in real time to ensure that investment decisions are synchronized with the evolution path of user needs. This not only converts current potential users, but also optimizes long-term investment strategies by continuously mining the evolution patterns of user needs, ensuring the long-term effectiveness of big data investment in home furnishing FMCG products. Attached Figure Description

[0060] Figure 1 This is a schematic diagram of the big data intelligent advertising system based on marketing data analysis of the present invention. Detailed Implementation

[0061] 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.

[0062] refer to Figure 1 A big data-driven intelligent advertising system based on marketing data analysis includes:

[0063] The demand analysis unit is used to acquire marketing data for home furnishing fast-moving consumer goods, analyze the marketing data to obtain the implicit demand intensity index and the consumption intention value, and integrate the implicit demand intensity index and the consumption intention value to generate the intention entropy value. The marketing data includes: purchase behavior data, browsing behavior data, collection and add-to-cart data, click interaction data, evaluation and feedback data, family attribute data, as well as geographical location, environmental data and time data, etc.

[0064] The scenario analysis unit is used to analyze the intention entropy value, generate decision-making hindrance codes with spatiotemporal constraints, extract marketing data, and obtain the user's scenario fit.

[0065] The fusion unit is used to fuse the intention entropy value, decision-making hindrance encoding, and scenario adaptability to generate the potential strength of user demand.

[0066] The segmentation unit is used to divide users into three circles based on the intensity of user demand potential and decision-making resistance coding;

[0067] The delivery unit is used to calculate the main blocking factor of the decision blocking code for each of the three types of concentric circles, and to make different delivery decisions based on the main blocking factor.

[0068] By mining the implicit demand intensity index, combining it with decision-making hindrance coding under spatiotemporal constraints and scenario adaptability, the system generates user demand potential intensity and divides users into circles. This overcomes the drawbacks of over-reliance on explicit feedback, accurately identifies user groups with potential consumption needs but currently weak feedback, and adopts differentiated delivery decisions for the main hindrance factors of different circles. This can effectively convert the consumption potential of potential users and improve the long-term return on investment.

[0069] In one embodiment, marketing data is analyzed to obtain an implicit demand intensity index and a consumer willingness value, including:

[0070] Based on marketing data analysis, user focus shifts at different decision-making stages, generating behavioral entropy values. This includes identifying four stages of user decision-making: awareness (first product exposure), interest (active learning), consideration (comparison and selection), and decision (purchase). Attention metrics for each stage are extracted from marketing data. For example, in the awareness stage, this includes the number of category page visits; in the interest stage, the number of items added to cart and favorites; in the consideration stage, the frequency of competitor comparison keywords; and in the decision-making stage, the time spent on the payment page. The metrics for each stage are statistically analyzed to calculate the frequency of each attention point at that stage. For instance, in the awareness stage, if a user visits three categories of home furnishing and fast-moving consumer goods pages (Category A: 10 visits, Category B: 8 visits, Category C: 2 visits, totaling 20 visits), then the frequency of attention for Category A is calculated as follows: 10 / 20 = 0.5, Category B 0.4, Category C 0.1. Then, the entropy value of a single stage is calculated using the information entropy algorithm. Specifically, the frequency of each attention point in the stage is multiplied by its logarithm (base 2), and the negative numbers are summed to obtain the entropy value of a single stage. The number of times users shift from focusing on product price (interest stage indicator) to focusing on material (consideration stage indicator) from the interest stage to the consideration stage is counted, and divided by the total number of shifts from the interest stage to the consideration stage to obtain the shift probability of attention points between stages. Weights are assigned to the four stages (cognition 0.1, interest 0.2, consideration 0.3, decision 0.4). The entropy value of each stage is multiplied by the corresponding weight and then summed. The sum of the shift probability between each stage multiplied by the weight between stages (the weight of the shift between adjacent stages is 0.5) is then added to obtain the behavioral entropy value. The higher the behavioral entropy value, the more unstable the user's attention points are in different stages.

[0071] Analyzing the semantic similarity between user expressions and product attributes in marketing data to generate a semantic coupling coefficient involves: extracting descriptive words mentioned by users (such as durable, non-slip, etc.) from the evaluation and feedback data of marketing data; compiling a list of characteristic words for product attributes (such as sturdy material, non-slip surface, etc.); and counting the number of matches between user words and product characteristic words in the four stages of cognition, interest, consideration, and decision-making. For example, in the cognition stage, the number of user words that have the same meaning as product characteristic words is obtained. For example, if a user says "durable," and the product characteristic word is "wear-resistant material," it means that the user word and the product characteristic word have matched. The number of matches between user words and product characteristic words is divided by the total number of words to obtain the stage matching rate. The four stages are assigned weights of 0.1, 0.2, 0.3, and 0.4, respectively. The matching rates of each stage are multiplied and summed. The sum is the semantic coupling coefficient.

[0072] After weighted fusion of behavioral entropy and semantic coupling coefficient, the demand fluctuation coefficient is obtained. Specifically, the behavioral entropy and semantic coupling coefficient are assigned weights respectively. The weight of behavioral entropy is set to 0.6 and the weight of semantic coupling coefficient is set to 0.4. The behavioral entropy is multiplied by 0.6 and the semantic coupling coefficient is multiplied by 0.4 and then added together. The sum is the demand fluctuation coefficient. In home furnishing fast-moving consumer goods, behavioral entropy reflects the dynamic shift of users' focus in the four stages of cognition, interest, consideration and decision-making. It is directly related to the stability of the decision-making process. Its value can reflect the intensity of demand fluctuation in real time. It is closely related to the real-time behavioral changes that need to be captured in the decision-making process. Therefore, behavioral entropy is given a higher weight (0.6) to highlight the impact of dynamic decision-making behavior. Semantic coupling coefficient is the static matching result of user expression and product attributes. It is used to reflect the basic demand fit. Its immediate impact on demand fluctuation is weaker than dynamic behavioral changes. Therefore, semantic coupling coefficient is given a weight (0.4) as a supplement to basic matching to fit the dynamic characteristics of home furnishing fast-moving consumer goods users' decision-making.

[0073] By integrating behavioral entropy, semantic coupling coefficient, and demand fluctuation coefficient, we obtain the implicit demand intensity index and the consumption willingness index.

[0074] By analyzing the shift in user focus across four stages of decision-making to generate behavioral entropy values, and combining this with semantic coupling coefficients to obtain demand fluctuation coefficients, we can accurately capture potential consumer demand. This enhances our ability to identify users with weak feedback but potential needs. Furthermore, by integrating behavioral entropy values, semantic coupling coefficients, and demand fluctuation coefficients, we obtain implicit demand intensity indices and consumption intention values. This aligns with the dynamic decision-making characteristics of home furnishing and fast-moving consumer goods users. Based on this, our advertising decisions can reduce over-reliance on explicit feedback, improve conversion efficiency for potential users, and help increase long-term return on investment.

[0075] In one embodiment, the behavioral entropy value, semantic coupling coefficient, and demand fluctuation coefficient are integrated to obtain the implicit demand intensity index and consumption intention value, including:

[0076] The behavior entropy value, semantic coupling coefficient, and demand fluctuation coefficient are mapped to a three-dimensional tensor space to generate a composite feature matrix. Specifically, the behavior entropy value, semantic coupling coefficient, and demand fluctuation coefficient are normalized, and each coefficient value is scaled to the 0-1 interval according to its value range. A three-dimensional tensor space is constructed, and the three normalized coefficients are used as the x, y, and z axes of the space, respectively. The three normalized coefficient values ​​of a single user are mapped to a coordinate point in the space. The coordinate points of multiple users are then set to form the composite feature matrix.

[0077] The composite feature matrix is ​​transformed to generate a four-dimensional demand potential field. Specifically, this involves: sorting out the standardized data of all users' behavioral entropy values, semantic coupling coefficients, and demand fluctuation coefficients in the composite feature matrix; calculating the degree of correlation between these three coefficients; identifying the main directions that reflect these correlations; and each direction representing a set of features that integrate the key information of the three coefficients (the main directions are comprehensive feature directions that can simultaneously reflect the main change patterns of the three coefficients, such as a direction that reflects the common correlation that a high behavioral entropy value is accompanied by a high semantic coupling coefficient, and another direction that reflects the situation where the demand fluctuation coefficient changes alone). The explanatory power (the degree to which these comprehensive features reflect the overall situation of the data) of the original data is ranked according to the explanatory power of the original data, and the top four with the strongest explanatory power are retained. The key features extracted at different time points are then sorted in chronological order, and a four-dimensional space is constructed. The first three dimensions correspond to the standardized values ​​of the three coefficients, and the fourth dimension is time. Finally, the key features at each time point are mapped to specific locations in the four-dimensional space. All the locations combined form the four-dimensional demand potential field.

[0078] The analysis of the four-dimensional demand potential energy field yields an implicit demand intensity index. Specifically, this involves: extracting characteristic data for each user at different time points from the four-dimensional demand potential energy field. This data includes standardized values ​​of three coefficients and time information. Users with similar characteristics are grouped together to form multiple demand feature clusters. The density of each cluster is calculated, which is the proportion of users in the cluster to the total number of users. Simultaneously, the changing trend of each cluster in the time dimension is calculated. The weights for cluster density and time changing trend are set to 0.7 and 0.3, respectively. The density of each cluster is multiplied by the corresponding weight, and the time changing trend is multiplied by the corresponding weight to obtain the potential energy value within the cluster. The maximum and minimum potential energy values ​​in all clusters are found. The potential energy value of each user's cluster is subtracted from the minimum value, and then divided by the difference between the maximum and minimum values ​​to obtain a value between 0 and 1. This value is multiplied by 100 to obtain the implicit demand intensity index ranging from 0 to 100.

[0079] The four-dimensional demand potential energy field is spatially divided to generate consumption intention values, which are divided into four levels: S, A, B, and C. Specifically, this involves dividing all user feature points in the four-dimensional demand potential energy field into four regions, calculating the implicit demand intensity index at the center of each region, and sorting them from highest to lowest according to the implicit demand intensity index, corresponding to levels S, A, B, and C respectively. The user's region corresponds to their consumption intention value level, with level S corresponding to the implicit demand intensity index range. Grade A corresponds to the range of implicit demand intensity index. Grade B corresponds to the range of implicit demand intensity index. Grade C corresponds to the range of implicit demand intensity index. .

[0080] By constructing a three-dimensional tensor space and a four-dimensional demand potential field, user behavior and demand characteristics are analyzed in the spatiotemporal dimension, forming demand characteristic clusters and calculating potential values. This accurately extracts the implicit demand intensity index, breaking through the limitations of relying solely on explicit feedback. It can deeply explore users with weak feedback behavior but strong potential needs, and improve the ability to identify potential groups.

[0081] By dividing consumer willingness levels into four dimensions and combining them with the changing trends over time, we can dynamically capture the evolution of user needs. Based on the implicit demand intensity index and consumer willingness value, we can target and convert potential users, avoid resource misallocation, effectively activate long-term consumption potential, solve the problem of continuously declining returns on existing campaigns, and ensure the long-term effectiveness of campaigns.

[0082] In one embodiment, the implicit demand intensity index and the consumption intention value are fused to generate an intention entropy value, including:

[0083] By analyzing the implicit demand intensity index and the consumption willingness value, a dynamic weight matrix is ​​obtained. The dynamic weight matrix is ​​then decayed over time to obtain an asymmetric three-dimensional tensor. Specifically, for each consumption level under the consumption willingness value, the number of users at that level falling within each implicit demand intensity index interval is counted. Then, the number of users in each interval is divided by the total number of users at that level to obtain the correlation ratio between that level and the corresponding index interval. These ratios are arranged according to level and implicit demand intensity index interval to form a dynamic weight matrix. The collection time corresponding to each data in the dynamic weight matrix is ​​recorded. Using the current time as the benchmark, the time distance of each data point is calculated (e.g., if the data was collected 10 days ago, the distance is 10). For each weight value in the dynamic weight matrix, decay is performed according to the time distance: for every increase of 1 in distance, the weight value is multiplied by a fixed decay coefficient (0.95). Because the decay degree is different at different time points, the weight values ​​at each position in the matrix exhibit an asymmetric distribution with the time dimension, ultimately forming an asymmetric three-dimensional tensor containing the three dimensions of consumption level, index interval, and time.

[0084] The asymmetric three-dimensional tensor is superimposed with eigenstates to generate a composite quantum state vector containing the user's true intent. Specifically, this involves: defining the three dimensions of the asymmetric three-dimensional tensor as consumption level, exponential interval, and time; extracting the most representative eigenstates from each dimension, i.e., typical states reflecting the main data distribution patterns of that dimension, such as the most frequently occurring level combinations in the consumption level dimension, the range of user concentration in the exponential interval dimension, and the key time points of data change in the time dimension; calculating the weight of each eigenstate in its respective dimension, where the weight is determined by the degree of explanation of the original data, with the degree of explanation corresponding to the weight (e.g., if the degree of explanation is 60%, the weight is 0.6); adjusting all eigenstates under the three dimensions according to their respective weights; and then superimposing and merging the adjusted eigenstates of each dimension to form the final vector, which is the composite quantum state vector containing the user's true intent.

[0085] Constraint feature extraction is performed on the composite quantum state vector to obtain the intention entropy with spatiotemporal characteristics. Specifically, for the composite quantum state vector, the spatial range of time window (nearly 7 days), consumption level, and index interval are used as constraints. Within the constraint range, the fluctuation amplitude of each feature in the vector (reflecting the uncertainty of intention) is calculated, the frequency of occurrence of each fluctuation amplitude is counted, and the frequency is multiplied by its base-2 logarithm and then summed to obtain the intention entropy with spatiotemporal characteristics.

[0086] By using the time decay of a dynamic weight matrix to form an asymmetric three-dimensional tensor, and superimposing eigenstates to generate a composite quantum state vector, it is possible to deeply mine users' true intentions. Combined with spatiotemporal constraints, it extracts intention entropy, breaks through the limitations of explicit feedback, accurately captures the true intentions of users with potential consumption needs but weak feedback, and improves its recognition ability. By integrating the implicit demand intensity index and consumption intention value to generate intention entropy, it takes into account spatiotemporal characteristics and dynamic changes in user intentions, thereby enabling the effective conversion of potential user consumption potential into advertising decisions, avoiding resource misallocation, and solving the problem of low long-term advertising return rates.

[0087] In one embodiment, the intention entropy value is analyzed to generate a spatiotemporal constraint decision-making hindrance code, including:

[0088] After mapping the intention entropy value to a quantum state vector, a spatiotemporal frequency feature spectrum is generated. Specifically, this involves mapping the time window (last 7 days), consumption level, and index interval spatial range and fluctuation characteristics contained in the intention entropy value to different components of the quantum state vector. The value of each component is determined by the specific data of the corresponding element, thus completing the mapping. For the time dimension of the quantum state vector, the number of times each fluctuation feature appears in the last 7 days is counted, and the frequency of feature recurrence under different time intervals is calculated to obtain the time frequency feature. Similarly, for the spatial dimension, the distribution of each combination of consumption level and index interval is counted, and the frequency of recurrence of different combinations in the spatial range is calculated to obtain the spatial frequency feature. Finally, the time and spatial frequency features are integrated according to the corresponding spatiotemporal coordinates to form the spatiotemporal frequency feature spectrum.

[0089] The spatiotemporal frequency feature spectrum is extracted to obtain the time decay coefficient, spatial heat value, and intention divergence value. Specifically, this includes: for the time dimension data of the spatiotemporal frequency feature spectrum, determining the time frequency feature value of each day within the past 7 days, calculating the ratio of the daily feature value to the feature value of the first day, and then averaging these ratios to obtain the time decay coefficient; statistically analyzing the frequency of each consumption level and index interval combination in the spatial dimension of the spatiotemporal frequency feature spectrum, the frequency of which is the spatial heat value of that combination; calculating the frequency distribution of each feature in the spatiotemporal frequency feature spectrum, and then calculating the difference between this distribution and the average frequency distribution of all features, the difference of which is the intention divergence value.

[0090] The time decay coefficient, spatial heat value, and intention divergence value are fused to obtain a multidimensional hindering factor matrix. Specifically, the weights of the time decay coefficient, spatial heat value, and intention divergence value are set to 0.4 and 0.4 respectively. For each spatiotemporal coordinate point, the time decay coefficient, spatial heat value, and intention divergence value of that point are multiplied by their respective weights and then summed. The result is used as an element of that coordinate in the matrix. After all coordinate points are calculated, the resulting matrix is ​​the multidimensional hindering factor matrix. Among them, the time decay coefficient reflects the timeliness of the data, and user consumption decisions are more affected by recent behavior, so it is assigned a weight of 0.4. The spatial heat value reflects the mainstream user consumption pattern and has a significant impact on decision-making hindering, so it is assigned a weight of 0.4. The intention divergence value reflects the user's special needs and has a slightly weaker impact, so it is assigned a weight of 0.2.

[0091] The multidimensional hindrance factor matrix is ​​analyzed and encoded to obtain a decision hindrance code with spatiotemporal constraints. Specifically, this includes setting the range of element values ​​in the multidimensional hindrance factor matrix to 0. 1 (0 for no blockage, 1 for complete blockage), taking 0.3 and 0.7 as thresholds, and... Elements with a value of 0.3 are classified as low-impedance and assigned binary 00. Elements with a value of 0.7 are classified as medium hindrance and assigned 01. Elements with a value of 0.7 are classified as high-resistance and assigned a value of 11, according to the matrix spatiotemporal coordinates (7 days). 3) Sequentially, the binary representations of each element are concatenated and a 42-bit code is generated using a 7×3 matrix, which is the spatiotemporal constraint decision-making blocking code.

[0092] By converting the intention entropy value into a spatiotemporal frequency feature spectrum, extracting the time decay coefficient, spatial thermal value, and intention divergence value, and fusing them into a multidimensional hindering factor matrix, a decision hindering code with spatiotemporal constraints is finally generated. This code can accurately capture the hindering factors in users' consumption decisions, especially identifying the decision obstacles in the spatiotemporal dimension for user groups with strong potential needs but weak current feedback.

[0093] In-depth analysis of decision-making resistance coding can identify the main resistance factors for different user groups. Developing targeted advertising strategies for potential consumer groups can effectively reduce their decision-making resistance and activate their consumption potential. Compared with existing advertising methods that overly rely on explicit feedback, this approach can improve the conversion efficiency for users with implicit needs, thereby increasing the long-term return on investment and enhancing the accuracy and long-term effectiveness of marketing for home furnishing and fast-moving consumer goods.

[0094] In one embodiment, marketing data is extracted to obtain the user's scenario suitability, including:

[0095] Based on marketing data, determine the spatiotemporal density of user activities. Specifically, this includes: extracting user geographic location and time information from marketing data, dividing spatiotemporal units according to time span (1 hour) and spatial grid (1 square kilometer), counting the number of user activities in each spatiotemporal unit, and dividing the number of activities by the product of the time span and spatial area of ​​the corresponding unit to obtain the spatiotemporal density of user activities.

[0096] By combining spatiotemporal density with the functional attributes of home furnishing fast-moving consumer goods (FMCG), a field overlap coefficient is obtained. Specifically, this involves: determining the applicable scenarios of FMCG based on their functional attributes, such as mops corresponding to household cleaning scenarios and tableware corresponding to kitchen dining scenarios; setting corresponding time ranges (e.g., cleaning scenarios are mostly in weekday evenings) and spatial ranges (e.g., kitchen scenarios correspond to residential kitchen areas); counting the number of user activities within the time and spatial ranges corresponding to the functional attributes of each product in the spatiotemporal density of user activities; dividing the number of activities by the total number of user activities to obtain the scenario matching ratio of a single product; using the relevance between all FMCG products and users as weights; multiplying the scenario matching ratio of each product by the corresponding weights; and summing all the products to obtain the field overlap coefficient.

[0097] Based on the field overlap coefficient, the user demand evolution path is determined, and a scenario transition probability matrix is ​​generated. Specifically, the field overlap coefficients at each time point are arranged in descending order, and the scenarios corresponding to the highest values ​​are connected sequentially to form the user demand evolution path. The transition frequency of adjacent scenarios in the path is counted, and the proportion of this frequency to the total transition frequency of the initial scenario is calculated as the transition probability. The transition probabilities of all initial scenarios and subsequent scenarios are summarized into a matrix according to the correspondence between rows (initial scenarios) and columns (subsequent scenarios), which is the scenario transition probability matrix.

[0098] The spatiotemporal density, field overlap coefficient, and context transition probability matrix are fused using an asymmetric tensor to generate the user's context fit. Specifically, this involves assigning asymmetric weights to the spatiotemporal density, field overlap coefficient, and context transition probability matrix (spatiotemporal density weight is 0.3, field overlap coefficient weight is 0.2, and context transition probability matrix weight is 0.5). The spatiotemporal density is matched to the context position in the matrix according to the corresponding spatiotemporal unit and multiplied by the corresponding weight. The field overlap coefficient is directly multiplied by the corresponding weight. The probabilities of each matrix are multiplied by their corresponding weights and then summed. Finally, the three weighted values ​​are added together and normalized to 0-100, which is the user's context fit.

[0099] By extracting user spatiotemporal density, combining it with the functional attributes of home furnishing fast-moving consumer goods to generate a field overlap coefficient, and then integrating the context transition probability matrix to obtain the scene adaptability, we can accurately capture the matching degree between users and product functions in specific spatiotemporal scenarios. In particular, for potential users with weak feedback behavior but high scene adaptability, we can uncover their implicit needs, which is conducive to improving the consumption conversion ability of different users.

[0100] By using asymmetric tensors to fuse multi-dimensional data to generate scenario adaptability, taking into account both user activity patterns and product scenario attributes, it can dynamically track the evolution path of demand. Based on this, the deployment decision can be adapted to the real-world needs of users, improve the ability to identify and convert potential users, avoid resource misallocation, and effectively solve the problem of insufficient identification of users with implicit needs in existing big data deployment, thereby improving the long-term return on investment.

[0101] In one embodiment, the intention entropy value, decision-making hindrance encoding, and scenario adaptability are fused to generate the user demand potential intensity, including:

[0102] The intention entropy value and the decision-making hindrance code are calculated to generate a four-dimensional spatiotemporal intention field matrix. Specifically, this involves splitting the 42-bit binary number contained in the decision-making hindrance code into groups of two bits (corresponding to low, medium, and high hindrance levels: 00, 01, and 11), and then converting each group of binary numbers into decimal numbers (00 corresponds to 0, 01 to 1, and 11 to 3). These decimal numbers together constitute the basic hindrance value. The spatiotemporal coordinate correspondence between the intention entropy value and the basic hindrance value is then determined: the time coordinate corresponds to each day of the past 7 days, and the spatial coordinate corresponds to the 3 consumption levels. By combining the index interval with the consumption level at each time point, a unique intention entropy value and a basic resistance value can be matched with the index interval combination. Then, the specific value of the intention entropy value under the corresponding coordinate is multiplied by the basic resistance value to obtain the preliminary calculation result under that spatiotemporal coordinate. Finally, the preliminary calculation results of all spatiotemporal coordinates are arranged in chronological order and spatial combination order to form a matrix containing time and three-dimensional spatial information. Then, all the values ​​in the matrix are processed to unify the range of values ​​between 0 and 1, and the four-dimensional spatiotemporal intention field matrix can be obtained.

[0103] The four-dimensional spatiotemporal intention field matrix is ​​corrected based on the scene adaptability to generate the corrected spatiotemporal intention field. Specifically, the scene adaptability (0-100) is converted into the range of 0-1 (divided by 100). For each element in the four-dimensional spatiotemporal intention field matrix, the element value is multiplied by the converted scene adaptability. The result is the element value at the corresponding position after correction. The matrix formed by splicing all the corrected elements is the corrected spatiotemporal intention field.

[0104] The spatiotemporal willingness field is analyzed to generate a user demand potential matrix. Specifically, this includes: statistically analyzing the element values ​​of spatial coordinates at each time point (last 7 days) in the corrected spatiotemporal willingness field; using the user activity at each time point as the weight; multiplying the element values ​​of the same spatial coordinate at different time points by their corresponding weights and summing them to obtain the sum value of each spatial coordinate; arranging these sum values ​​in the order of spatial coordinates; and forming the matrix is ​​the user demand potential matrix.

[0105] Extracting the user demand potential energy matrix to obtain the user demand potential energy intensity specifically includes: calculating the sum of the values ​​of all elements in the user demand potential energy matrix, dividing the sum by the total number of elements in the matrix to obtain the average value of the matrix, normalizing the average value to the range of 0-100 (by multiplying the average value by 100), and the result is the user demand potential energy intensity.

[0106] By integrating intention entropy, decision-making hindrance coding, and scenario adaptability to generate user demand potential intensity, and constructing a four-dimensional spatiotemporal intention field matrix combined with scenario adaptability correction, it can comprehensively capture users' potential demand characteristics in the spatiotemporal dimension, accurately identify users with weak feedback behavior but high demand potential, and make up for the shortcomings of existing big data advertising in identifying users with implicit needs. Moreover, it can quantify users' real consumption potential based on the intensity of user demand potential, thereby assessing user conversion value and providing accurate basis for subsequent segmentation and big data advertising decisions. This helps to formulate effective strategies for potential users, fully convert their consumption potential, solve the problem of low long-term advertising return rate, and improve the overall efficiency of intelligent advertising for home furnishing and fast-moving consumer goods.

[0107] In one embodiment, based on the user demand potential intensity and decision-making hindrance coding, users are divided into three concentric circles, including:

[0108] The four-dimensional spatiotemporal intention field matrix is ​​analyzed to determine the dynamic potential energy threshold vector. Specifically, this includes: extracting the element values ​​of all spatiotemporal coordinates in the four-dimensional spatiotemporal intention field matrix, dividing it into daily sliding windows according to the time dimension, statistically analyzing the distribution of element values ​​for each window, calculating its 25th, 50th, and 75th percentiles as the low, medium, and high threshold components of that window, and arranging the threshold components of all windows in chronological order to form the dynamic potential energy threshold vector.

[0109] Analyzing the decision-making hindrance coding yields the resonance degree between users and the value proposition of home furnishing fast-moving consumer goods (FMCG). Specifically, this involves determining the ideal decision-making state corresponding to the value proposition of FMCG. For example, for a certain product category, the ideal low-hindrance ratio is high, corresponding to binary 00. The 42-bit binary code of the decision-making hindrance coding is divided into 21 groups of two bits each, with each group corresponding to a hindrance level in a spatiotemporal coordinate (00 for low hindrance, 01 for medium hindrance, and 11 for high hindrance). The difference between the actual hindrance level and the ideal hindrance level in each of these 21 groups is calculated. For example, if the actual level is 01 and the ideal level is 00, it is recorded as 1; if the actual level is 11 and the ideal level is 00, it is recorded as 2; if the actual level matches the ideal level, it is recorded as 0. All differences are summed to obtain the total difference. The total difference is then divided by the maximum possible difference (i.e., 42 when all 21 groups have the maximum difference of 2) to obtain the difference ratio. The difference ratio is subtracted from 1, and the result is normalized to obtain the resonance degree between users and the product's value proposition.

[0110] Based on the dynamic potential threshold vector and resonance, three types of user circles are generated in four-dimensional space, specifically including:

[0111] When the user's demand potential energy intensity exceeds the high threshold component of the dynamic potential energy threshold vector and the resonance degree 0.85, classified as a first-type layer: a high-resonance layer;

[0112] When the user demand potential energy intensity is between the middle threshold component and the high threshold component of the dynamic potential energy threshold vector, and falls within the difference between the two... Within the 15% range, with resonance between 0.6 and 0.85, it is identified as a second type of layer: a steady-state resonant layer.

[0113] When the user's demand potential energy intensity is lower than the middle threshold component of the dynamic potential energy threshold vector and the resonance degree 0.6, identified as a third type of layer: low resonance layer;

[0114] For users who do not meet the above three criteria for social circles, manual intervention is required for categorization.

[0115] By combining dynamic potential energy threshold vectors with resonance, three user segments are categorized. The dynamic potential energy threshold vector captures the spatiotemporal changes in user demand potential, while resonance reflects the alignment between users and product value propositions. The combination of these two methods can accurately identify users with high potential but weak feedback. Differentiated strategies are implemented for the three segments: high resonance segments can strengthen investment to consolidate loyalty; steady-state resonance segments need to reduce decision-making obstacles to promote conversion; and low resonance segments can tap into potential by activating latent demand. This classification method avoids over-concentrating resources on users with strong explicit feedback, while also considering the conversion of potential users. It can fully release the consumption potential of each segment, improve the long-term return on investment, and enhance the accuracy and sustainability of big data investment in home furnishing and fast-moving consumer goods.

[0116] In one embodiment, for the three types of concentric circles, the principal hindering factor of the decision hindering code is calculated respectively, including:

[0117] Extracting the decision-blocking code to obtain the coupling strength between each factor and the user demand potential energy, specifically includes: dividing the 42-bit binary of the decision-blocking code into 21 groups of two bits each, with each group corresponding to a blocking factor at a spatiotemporal coordinate; converting each group of binary bits into decimal (00 corresponds to 0, 01 corresponds to 1, and 11 corresponds to 3); finding the potential energy values ​​corresponding to each of the 21 spatiotemporal coordinates from the user demand potential energy matrix; and calculating the correlation between the decimal value of each blocking factor and the corresponding coordinate potential energy value. The correlation is the coupling strength between the blocking factor and the user demand potential energy.

[0118] Based on the user's social circle, configure differentiated coupling strength thresholds and entropy reduction efficiency weight coefficients; specifically including:

[0119] For users in the high-resonance layer, the coupling strength threshold is set to 0.7, and the entropy reduction efficiency weighting coefficient is... ;

[0120] For users of the steady-state resonant layer, the coupling strength threshold is set to 0.6, and the entropy reduction efficiency weighting coefficient is... ;

[0121] For users with low resonant layers, the coupling strength threshold is set to 0.5, and the entropy reduction efficiency weighting coefficient is... ;

[0122] Based on the coupling strength threshold and entropy reduction efficiency weighting coefficient, the coupling strength is weighted to generate a decision resistance distribution cloud map. Specifically, the following steps are taken: from all coupling strengths, values ​​that exceed the coupling strength threshold corresponding to the user's circle are selected. For each selected value, the components of each dimension (time, space, and intent) are decomposed. Each dimension component is then multiplied by the corresponding value in the entropy reduction efficiency weighting coefficient. The three products are summed to obtain the weighted value of each selected coupling strength. These weighted values ​​are mapped to the original 21 spatiotemporal coordinates, and each coordinate point is colored according to the weighted value, with larger values ​​resulting in darker colors. Finally, a decision resistance distribution cloud map is formed.

[0123] The energy entropy reduction gradient of each hindering factor is calculated in the decision resistance distribution cloud map, and an entropy reduction efficiency ranking table is generated. Specifically, this includes: extracting all spatiotemporal coordinates and their weighted values ​​corresponding to each hindering factor from the decision resistance distribution cloud map; calculating the weighted value of each hindering factor at two adjacent time points, specifically: subtracting the weighted value of the previous time point from the weighted value of the later time point to obtain the difference; if the difference is negative, it indicates that the entropy is decreasing; dividing this decrease by the interval between the two time points to obtain the entropy reduction per unit time; based on these unit time reductions, calculating the rate at which the entropy value of each hindering factor decreases the fastest, i.e., the energy entropy reduction gradient; and arranging all hindering factors in descending order of their respective energy entropy reduction gradients to form the entropy reduction efficiency ranking table.

[0124] The top three factors in the entropy reduction efficiency ranking table are extracted as candidate principal blocking factors, and the entropy reduction contribution rate of each candidate factor is calculated. Specifically, the top three blocking factors in the entropy reduction efficiency ranking table are selected as candidate principal blocking factors, the energy entropy reduction gradient of each of the three factors is calculated, the sum of the gradients of the three factors is calculated, and the gradient of each candidate factor is divided by the sum to obtain the entropy reduction contribution rate of that factor.

[0125] The principal retardation factor is determined based on the entropy reduction contribution rate, specifically including:

[0126] The contribution rates of all candidate factors are normalized to fall within the range of 0 to 1.

[0127] If the contribution rate of the candidate factor If the value is 0.6, it is directly marked as the principal inhibition factor to be verified and then verified.

[0128] If the contribution rates of all candidate factors are equal If the value is 0.6, then the two factors with the highest contribution rates are selected and merged into a joint candidate principal blocking factor, which is then verified.

[0129] The time validity value is calculated based on the time decay coefficient to determine the duration of the factor’s influence over the user’s complete decision-making cycle. Specifically, this involves: determining the total duration of the user’s complete decision-making cycle, dividing the cycle into several time periods according to time intervals, using the proportion of each time period to the total duration as a weight to multiply by the time decay coefficient corresponding to that time period, and then summing all the products. The result is the time validity value of the factor. The larger the value, the stronger the duration of influence.

[0130] Based on the scenario adaptability, the correlation density between this factor and the user's high-frequency activity scenarios is calculated to generate a spatial adaptability value. Specifically, this includes: finding the top 5 high-frequency activity scenarios with the most user occurrences, recording the specific spatial range and total number of occurrences of each scenario, calculating the proportion of the occurrences of each high-frequency scenario to the total number of occurrences of all high-frequency scenarios as the weight of that scenario, then splitting the scenario adaptability according to these high-frequency scenarios to obtain the adaptability corresponding to each scenario, multiplying the adaptability of each scenario by the weight of that scenario, and then adding all the products together to get the sum, normalizing the sum to between 0 and 1, and the normalized value is the spatial adaptability value.

[0131] The matching degree between the factor and the user's implicit needs is calculated based on the semantic coupling coefficient to generate an intent fit value. Specifically, this includes: extracting descriptive words related to the main blocking factor, compiling a list of feature words corresponding to the user's implicit needs, counting the number of matches between the factor descriptive words and the implicit need feature words, dividing the number of matches by the total number of words to obtain the basic matching rate, weighting the basic matching rate and the semantic coupling coefficient with weights of 0.6 and 0.4 respectively, and then normalizing the sum to the 0-1 range to obtain the intent fit value.

[0132] The time validity value, spatial fit value, and intention fit value are integrated to generate a comprehensive verification index. Specifically, for the high resonance layer, the weights are: time validity value 0.5, spatial fit value 0.3, and intention fit value 0.2; for the steady-state resonance layer, the weights are: time validity value 0.4, spatial fit value 0.4, and intention fit value 0.2; and for the low resonance layer, the weights are: time validity value 0.2, spatial fit value 0.3, and intention fit value 0.5. The time validity value, spatial fit value, and intention fit value are multiplied by their respective weights, summed, and then scaled to the 0-1 range to obtain the comprehensive verification index.

[0133] If the factor is the principal blocking factor to be verified and the comprehensive verification index is... If the corresponding layer thresholds (0.7 for high resonance layer, 0.6 for steady-state resonance layer, and 0.5 for low resonance layer) are met, the verification is successful, confirming it as the main hindrance factor.

[0134] If the factor is a joint candidate principal blocking factor and the comprehensive verification index If the corresponding layer thresholds (0.7 for high resonance layer, 0.6 for steady-state resonance layer, and 0.5 for low resonance layer) are met, the verification is successful, and the factor with the higher single comprehensive verification index is taken as the main hindrance factor.

[0135] If the comprehensive verification index of this factor If the corresponding layer threshold (0.7 for high resonance layer, 0.6 for steady-state resonance layer, and 0.5 for low resonance layer) is not met, the verification fails. The next candidate factor is extracted from the entropy reduction efficiency ranking table and the verification continues until the main blocking factor that meets the conditions is obtained. If the verification of three consecutive candidate factors fails, it is determined that the user has no clear main blocking factor and the deployment decision is made by manual intervention.

[0136] For three types of users, the main blocking factors of each group are accurately located by setting differentiated thresholds and multi-dimensional verification. For users in the low resonance layer, potential decision-making obstacles are explored with a lower coupling strength threshold. By combining spatiotemporal adaptation and implicit demand matching verification, the limitations of relying solely on explicit feedback are overcome, and the problem of insufficient identification of potential demand users is solved. This provides clear blocking targets for converting these users.

[0137] Differentiated targeting based on key resistance factors can eliminate decision-making resistance among users in different circles. The high-resonance layer strengthens advantages and reduces minor resistance, the steady-state resonance layer focuses on core obstacles to promote conversion, and the low-resonance layer precisely addresses key obstacles to activate potential. This precise approach can effectively convert potential users, avoid resource waste, improve the overall conversion efficiency of big data targeting, and thus increase the long-term return on investment, enhancing the effectiveness of intelligent targeting for home furnishing and fast-moving consumer goods.

[0138] In one embodiment, different deployment decisions are made based on the main blocking factor, including:

[0139] If the frequency of price-related keywords in the data associated with the main blocking factor exceeds 60% (such as words like "too expensive" or "not cost-effective"), then the type is determined to be price sensitivity.

[0140] If, in the data associated with the main blocking factor, the frequency of user reviews expressing insufficient information exceeds 60% (such as not understanding the material, wanting to know the effect of use, etc.), then the type is determined to be information missing.

[0141] If, in the data associated with the main blocking factor, the frequency of mention of competitor comparison keywords (such as which is better) exceeds 60% or the dwell time in the consideration stage exceeds the average duration of 2 minutes, it is judged as decision complexity.

[0142] When the main blocking factor is price sensitivity, the first deployment decision is adopted: deploy tiered group-buying coupons;

[0143] When the primary blocking factor is information gap, the second distribution decision is adopted: distribute trial vouchers;

[0144] When the main blocking factor is decision complexity, the third delivery decision is adopted: deliver product bundled coupons;

[0145] The first, second, and third placement decisions are adjusted based on the consumer willingness score. If the consumer willingness score is S or A:

[0146] The initial decision to distribute coupons has been revised to: distribute high-value instant discount coupons.

[0147] The second decision to distribute vouchers has been revised to: distribute unlimited-time trial vouchers;

[0148] The third decision-making step was revised to: distribute personalized bundled coupons.

[0149] If the consumer willingness level is B or C, then the original advertising decision will be maintained.

[0150] By developing differentiated targeting decisions based on different primary bottleneck factors and incorporating consumer willingness-to-pay (CWP) adjustment strategies, this approach overcomes the limitations of current over-reliance on explicit feedback. It identifies three primary bottleneck factors—price sensitivity, information gaps, and decision complexity—and matches them with precise tools such as tiered group-buying coupons, trial coupons, and product bundle coupons. This effectively eliminates decision-making barriers for users across different demographics. Furthermore, it dynamically adjusts targeting based on CWP values, upgrading offers to high-willing users to accelerate conversion, while maintaining basic strategies to nurture demand from low-willing users. This approach ensures both short-term returns and taps into long-term potential, addressing the problem of insufficient conversion of users with latent needs in existing big data-driven targeting, thereby improving the long-term return on investment.

[0151] 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 big data-driven intelligent advertising system based on marketing data analysis, characterized in that: include: The demand analysis unit is used to acquire marketing data for home furnishing fast-moving consumer goods, analyze the marketing data to obtain the implicit demand intensity index and the consumption intention value, and merge the implicit demand intensity index and the consumption intention value to generate the intention entropy value. The scenario analysis unit is used to analyze the intention entropy value and generate spatiotemporal constraint decision-making hindrance codes, including: After mapping the intentional entropy value to a quantum state vector, a transformation is performed to generate a spatiotemporal frequency characteristic spectrum. The spatiotemporal frequency feature spectrum is extracted to obtain the time decay coefficient, spatial thermodynamic value, and intentional divergence value; By fusing the time decay coefficient, spatial thermodynamic value, and intentional divergence value, a multidimensional hindrance factor matrix is ​​obtained; The multidimensional blocking factor matrix is ​​analyzed and encoded to obtain a decision blocking code with spatiotemporal constraints. Marketing data is extracted to determine user scenario suitability, including: Determine the spatiotemporal density of user activities based on marketing data; By combining spatiotemporal density with the functional attributes of home furnishing fast-moving consumer goods, the field overlap coefficient is obtained; Based on the field overlap coefficient, the evolution path of user needs is determined, and a context transition probability matrix is ​​generated; The spatiotemporal density, field overlap coefficient and context transition probability matrix are fused using an asymmetric tensor to generate the user's scenario adaptability. The fusion unit is used to fuse the intention entropy value, decision-making hindrance encoding, and scenario adaptability to generate the user demand potential intensity, including: The intention entropy value and the decision-blocking code are calculated to generate a four-dimensional spatiotemporal intention field matrix; The four-dimensional spatiotemporal intention field matrix is ​​corrected according to the scene adaptability to generate the corrected spatiotemporal intention field. Analyze the spatiotemporal willingness field to generate a user demand potential matrix; Extract the user demand potential energy matrix to obtain the user demand potential energy intensity; The segmentation unit is used to divide users into three circles based on the intensity of user demand potential and decision-making resistance coding; The delivery unit is used to calculate the main blocking factor of the decision blocking code for each of the three types of concentric circles, and to make different delivery decisions based on the main blocking factor.

2. The big data intelligent advertising system based on marketing data analysis according to claim 1, characterized in that, Analyzing marketing data yields an implicit demand intensity index and a consumer willingness index, including: Based on marketing data analysis, user focus shifts at different decision-making stages, generating behavioral entropy values; Analyze the semantic similarity between user expressions and product attributes in marketing data to generate a semantic coupling coefficient; The demand fluctuation coefficient is obtained by weighting and fusing the behavioral entropy value and the semantic coupling coefficient. By integrating behavioral entropy, semantic coupling coefficient, and demand fluctuation coefficient, we obtain the implicit demand intensity index and the consumption willingness index.

3. The big data intelligent advertising system based on marketing data analysis according to claim 2, characterized in that, By integrating behavioral entropy, semantic coupling coefficient, and demand fluctuation coefficient, we obtain the implicit demand intensity index and consumption intention value, including: The behavioral entropy value, semantic coupling coefficient, and demand fluctuation coefficient are mapped to a three-dimensional tensor space to generate a composite feature matrix. Transform the composite feature matrix to generate a four-dimensional demand potential energy field; An analysis of the four-dimensional demand potential energy field yields the implicit demand intensity index. The four-dimensional demand potential field is spatially divided to generate a consumption intention value, which is divided into four levels: S-level, A-level, B-level and C-level.

4. The big data intelligent advertising system based on marketing data analysis according to claim 3, characterized in that, The implicit demand intensity index is combined with the consumption intention value to generate an intention entropy value, which includes: By analyzing the implicit demand intensity index and the willingness to consume, a dynamic weight matrix is ​​obtained. The dynamic weight matrix is ​​then subjected to time decay to obtain an asymmetric three-dimensional tensor. By superimposing characteristic states on an asymmetric three-dimensional tensor, a composite quantum state vector containing the user's true intent is generated. Constraint feature extraction is performed on the composite quantum state vector to obtain the intention entropy value with spatiotemporal characteristics.

5. The big data intelligent advertising system based on marketing data analysis according to claim 1, characterized in that, Based on the intensity of user demand potential and the coding of decision-making obstacles, users are divided into three circles, including: The dynamic potential energy threshold vector is determined by analyzing the four-dimensional spatiotemporal intention field matrix. Analysis of decision-making obstruction codes yields the resonance degree between users and the value proposition of home furnishing fast-moving consumer goods; Based on the dynamic potential threshold vector and resonance, three types of user circles are generated in four-dimensional space: high resonance layer, steady-state resonance layer, and low resonance layer.

6. The big data intelligent advertising system based on marketing data analysis according to claim 5, characterized in that, For the three types of concentric circles, the main hindering factors of the decision-making hindering coding are calculated respectively, including: The decision-making obstruction code is extracted to obtain the coupling strength between each factor and the potential energy of user demand; Based on the user's social circle, configure differentiated coupling strength thresholds and entropy reduction efficiency weight coefficients; Based on the coupling strength threshold and entropy reduction efficiency weighting coefficient, the coupling strength is weighted to generate a decision resistance distribution cloud map. Calculate the energy entropy reduction gradient of each hindering factor in the decision resistance distribution cloud map, and generate an entropy reduction efficiency ranking table. The top three factors in the entropy reduction efficiency ranking table are extracted as candidate principal blocking factors, and the entropy reduction contribution rate of each candidate factor is calculated. The principal blocking factor is determined based on the entropy reduction contribution rate.

7. The big data intelligent advertising system based on marketing data analysis according to claim 6, characterized in that, Different deployment decisions are made based on the main blocking factor, including: When the primary blocking factor is price sensitivity, the first deployment decision is adopted; When the primary blocking factor is information gap, the second delivery decision is adopted; When the main blocking factor type is decision complexity, the third delivery decision is adopted.

Images