A Big Data-Based Intelligent Shopping Guide Recommendation Method and System

Abstract

This invention provides a big data-based intelligent shopping guide recommendation method and system, belonging to the field of intelligent shopping guide technology. The method includes constructing a product network topology diagram; quantifying the purchasing power fluctuations of target customers, modeling decision-making patience decay curves, and analyzing shopping path tolerance to establish a three-dimensional value model for target customers; generating a dynamically optimized product penetration pool using a product association mining strategy; generating an optimal shopping guide route sequence using spatiotemporal fusion field construction technology and a full-domain dynamic path optimization algorithm; and generating dynamically adjusted shopping guide instructions based on shopping guide route deviations and shopping cart product type deviations. This invention significantly improves the efficiency of matching people, goods, and places in retail scenarios through multi-level technological collaboration, achieving an efficient and enjoyable immersive shopping experience; and constructs an intelligent shopping guide model that combines accurate recommendation capabilities, spatiotemporal resource optimization efficiency, and humanistic care attributes, providing core infrastructure support for the digital transformation of physical retail.

Description

Technical Field

[0001] This invention relates to the field of intelligent shopping guide technology, and in particular to an intelligent shopping guide recommendation method and an intelligent shopping guide recommendation system based on big data. Background Technology

[0002] With the accelerated digital transformation of the retail industry, traditional shopping malls are facing a strong impact from online e-commerce, and consumers are increasingly demanding more efficient and personalized shopping experiences. Traditional sales guide models rely on human experience, resulting in problems such as low recommendation accuracy, homogenized services, and unreasonable route planning, making it difficult to meet the diverse needs of modern consumers. For example, consumers in large shopping malls often prolong their shopping time because they cannot find their target products or receive recommendations that do not match their needs, leading to a decline in the shopping experience and impacting mall traffic and conversion rates.

[0003] To address these issues, intelligent shopping guide technology based on big data and artificial intelligence has gradually become a research hotspot. Existing technologies sometimes recommend products by analyzing users' historical consumption data, but these are often limited to single-dimensional purchase preference analysis, ignoring various dynamic factors, resulting in recommendations lacking timeliness and adaptability. Furthermore, they struggle to handle high-dimensional, large-scale product data in shopping malls, leading to insufficient mining of product relationships and weak correlation in the generated recommendation lists. In addition, shopping guide route planning often only considers physical spatial distance, easily resulting in route redundancy or deviation from the user's actual needs, reducing shopping guide efficiency.

[0004] Against this backdrop, there is an urgent need for an intelligent shopping guide recommendation method to achieve accurate recommendations and efficient shopping guidance, improve the consumer shopping experience, and enhance the core competitiveness of shopping malls. Summary of the Invention

[0005] This invention provides an intelligent shopping guide recommendation method and system based on big data to address the shortcomings of existing technologies.

[0006] On the one hand, this invention provides an intelligent shopping guide recommendation method based on big data, including:

[0007] Collect historical shopping records and product category information from shopping malls, and use a high-dimensional correlation calculation model to construct a product network topology diagram.

[0008] Obtain the target customer's historical consumption records and current shopping needs list, quantify the target customer's purchasing power fluctuations, model the decision patience decay curve, and analyze the shopping path tolerance, and establish a three-dimensional value model of the target customer.

[0009] Based on the product network topology diagram and the current shopping demand list, a dynamic optimal product penetration pool is generated using a product association mining strategy.

[0010] The system acquires the spatial coordinates of each product in the product penetration pool and a 3D map of the shopping mall. Combining this with a 3D value model, it uses spatiotemporal fusion field construction technology and a global dynamic path optimization algorithm to generate the optimal shopping guide route sequence.

[0011] It acquires real-time user location signals and dynamic changes in shopping cart items, and generates dynamically adjusted shopping guide instructions based on deviations in the shopping guide route and the types of items in the shopping cart.

[0012] According to the intelligent shopping guide recommendation method based on big data provided by the present invention, the process of constructing a product network topology diagram includes:

[0013] Data cleaning is performed on the mall's historical shopping records to extract product combination information from the complete transaction flow.

[0014] Define a product association tensor, calculate the first-order association strength between any two products, and extend it to multiple-order associations. The first-order association strength represents the geometric mean of the conditional probability and the lift index. Multiple-order associations represent higher-order collaborative purchasing patterns in product combinations.

[0015] Multi-level associations are stored in the form of three-dimensional tensors, and topological edge weights are generated through tensor decomposition.

[0016] Remove connection edges whose topological edge weights are lower than a preset association strength threshold to obtain a product network topology graph. In the graph, nodes represent products, and edges represent the association relationships between multiple product combinations.

[0017] According to the intelligent shopping guide recommendation method based on big data provided by this invention, historical consumption records include basic transaction data, product detail data, behavioral trajectory data, and return and exchange records. The current shopping needs list represents the list of planned purchases actively entered by the target customer.

[0018] According to the intelligent shopping guide recommendation method based on big data provided by the present invention, the process of quantifying the purchasing power fluctuations of target customers includes:

[0019] Based on the historical consumption records of target customers, the arithmetic mean and standard deviation of the amount spent by each target customer in a single purchase are calculated.

[0020] Identify the time-periodic characteristics of target customer consumption behavior and construct a fluctuation function incorporating a time decay factor.

[0021] Using the fluctuation function value as the center line, a dynamic fluctuation band interval is generated based on the standard deviation multiple.

[0022] By monitoring the cumulative spending amount in the target customer's current shopping needs list in real time, the fluctuation range is dynamically adjusted.

[0023] According to the intelligent shopping guide recommendation method based on big data provided by the present invention, the process of modeling the decision-making patience decay curve of target customers includes:

[0024] Analyze the target customer's historical behavior records, calculate the average time from browsing products to making a purchase decision, and map the average time to an initial patience value.

[0025] A hyperbolic decay function model is established to make the initial patience value exhibit nonlinear decay characteristics as the decision time progresses.

[0026] The decay rate coefficient is dynamically adjusted based on current environmental parameters, including mall pedestrian density and promotional activity intensity.

[0027] When the deviation between the actual decision-making time of the target customer and the predicted value exceeds the preset deviation threshold, the decay function parameters are adjusted.

[0028] According to the intelligent shopping guide recommendation method based on big data provided by the present invention, the process of conducting shopping path tolerance analysis on target customers includes:

[0029] Extract path abandonment records from the target customer's historical shopping history and fit a baseline threshold for the target customer's tolerance.

[0030] A spatial distance tolerance transformation model is established based on the benchmark threshold, and the tolerance energy function is calculated.

[0031] The system integrates the time taken for this shopping trip in real time and dynamically updates the path tolerance threshold.

[0032] According to the intelligent shopping guide recommendation method based on big data provided by the present invention, the process of generating a dynamic optimal product penetration pool using a product association mining strategy includes:

[0033] In the product network topology diagram, starting from the target customer's current shopping needs list, the penetration rate of all candidate products that can be reached through direct association paths is calculated.

[0034] Perform multi-demand collaborative correction and examine the second-order impact of the combination of items in the target customer's current shopping demand list on candidate items.

[0035] The system monitors the decay rate of product association strength within the pool in real time and filters the initial product pool based on a dynamically preset penetration rate threshold.

[0036] According to the intelligent shopping guide recommendation method based on big data provided by the present invention, the process of generating the optimal shopping guide route sequence includes:

[0037] The spatial coordinates and value weights of each product in the product penetration pool are mapped into a four-dimensional spatiotemporal field.

[0038] By superimposing the user's three-dimensional value model to form a comprehensive constraint field, a spatiotemporal optimization equation is established, resulting in a continuous path curve. The comprehensive constraint field includes the value constraint defined by the current fluctuation zone, the maximum dwell time determined by the decision patience value, and the spatial movement limit converted by the path tolerance.

[0039] The continuous path curve is discretized into a sequence of shopping guide nodes, and sorted by spatial reachability to output a set of shopping guide instructions, thus obtaining the optimal shopping guide sequence.

[0040] According to the intelligent shopping guide recommendation method based on big data provided by the present invention, the process of generating dynamically adjusted shopping guide instructions includes:

[0041] The system acquires the target customer's location coordinates and shopping cart item list in real time, and establishes a product decision entropy monitoring model based on the offset of the real-time location coordinates, shopping cart item list, optimal shopping guide route sequence, and product penetration pool.

[0042] When any deviation exceeds a preset threshold, a dynamic replanning condition is triggered. The dynamic replanning condition includes recalculating the spatiotemporal field gradient direction based on the latest data and generating a set of correction instructions, including steering commands, speed adjustments, and product replacement suggestions.

[0043] On the other hand, the present invention also provides an intelligent shopping guide recommendation system based on big data, comprising:

[0044] The product association topology construction module is used to collect historical shopping records and product category information from the mall, and to construct a product network topology diagram using a high-dimensional association calculation model.

[0045] The user 3D value modeling module is used to obtain the target customer's historical consumption records and current shopping needs list, quantify the target customer's purchasing power fluctuations, model the decision patience decay curve, and analyze the shopping path tolerance, thereby establishing a 3D value model of the target customer.

[0046] The dynamic product penetration pool generation module is used to generate a dynamic and optimal product penetration pool based on the product network topology diagram and the current shopping demand list, using a product association mining strategy.

[0047] The spatiotemporal path optimization module is used to obtain the spatial coordinates of each product in the product penetration pool and the 3D map of the shopping mall. Combined with the 3D value model, it uses spatiotemporal fusion field construction technology and global dynamic path optimization algorithm to generate the optimal shopping guide route sequence.

[0048] The dynamic instruction adjustment module is used to acquire user location signals and dynamic change data of shopping cart items in real time, and generate dynamically adjusted shopping guide instructions based on deviations in the shopping guide route and the types of items in the shopping cart.

[0049] This invention provides an intelligent shopping guide recommendation method and system based on big data. It breaks through the limitations of traditional pairwise recommendations by using advanced scenario-based association mining to achieve accurate prediction of cross-category demand. It integrates purchasing power fluctuations, decision-making patience decay, and path tolerance into a three-dimensional value assessment system, possessing adaptive capabilities to perceive users' real-time consumption capacity and behavioral tendencies. Based on spatiotemporal fusion field technology, it unifies the modeling of product value, physical movement, and time constraints into a continuous optimization problem, generating personalized shopping guide routes that maximize user value. Finally, relying on dynamic entropy change monitoring and a dual-deviation response mechanism, it achieves closed-loop instruction optimization, completely solving the pain point of recommendation failure caused by user interest drift or path deviation. Through multi-level technological collaboration, it significantly improves the efficiency of matching people, goods, and places in retail scenarios. On the user side, it reduces the cognitive burden of choosing from a massive number of products and the waste of ineffective walking, achieving an efficient and enjoyable immersive shopping experience. On the merchant side, it activates the penetration rate of long-tail products and the conversion of related sales, and optimizes inventory turnover through data-driven dynamic pricing. Ultimately, it constructs an intelligent shopping guide model that combines accurate recommendation capabilities, spatiotemporal resource optimization efficiency, and humanistic attributes, providing core infrastructure support for the digital transformation of physical retail. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0051] Figure 1 This is a flowchart illustrating an intelligent shopping guide recommendation method based on big data, provided in an embodiment of the present invention.

[0052] Figure 2 This is a schematic diagram of the structure of an intelligent shopping guide recommendation system based on big data, provided in an embodiment of the present invention. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0054] The following combination Figure 1-Figure 2 This invention describes an intelligent shopping guide recommendation method and system based on big data.

[0055] Figure 1This is a flowchart illustrating an intelligent shopping guide recommendation method based on big data, provided in an embodiment of the present invention.

[0056] like Figure 1 As shown in the figure, the present invention provides a big data-based intelligent shopping guide recommendation method and system. The executing entity can be a big data-based intelligent shopping guide recommendation method, which includes:

[0057] Collect historical shopping records and product category information from shopping malls, and construct a product network topology diagram using a high-dimensional correlation calculation model. The process includes:

[0058] Data cleaning is performed on the mall's historical shopping records to extract product combination information from complete transaction logs. A transaction matrix is ​​then constructed by grouping transactions by day, expressed by the following formula:

[0059]

[0060] Where N represents the number of valid transactions and M represents the total number of unique products.

[0061] Define product association tensor ,in This indicates the maximum correlation order.

[0062] Calculate the first-order association strength between any two items and extend it to higher-order associations. The first-order association strength represents the geometric mean of the conditional probability and the lift index, expressed by the formula:

[0063]

[0064]

[0065]

[0066] Multi-level association relationships represent higher-order collaborative purchasing patterns in product combinations, expressed by the formula:

[0067]

[0068]

[0069]

[0070] In the formula, Indicates a multi-level association relationship. Represents conditional probability. Indicates the degree of elevation. Indicates the weighting adjustment factor. This represents the smoothing constant, and m represents the correlation order. Let Q represent the set of combined goods, and let Q represent the set of all goods in the mall. Indicates that the product is included. A collection of transactions, Indicates that the product is included. A collection of transactions, Indicates product mix Support This represents the correlation strength amplification factor.

[0071] The multi-level association relationships are stored in the form of a three-dimensional tensor, and the topological edge weights are generated through tensor decomposition, as expressed by the formula:

[0072]

[0073] In the formula, This represents the product of tensors along the third dimension. The eigenvector representing the m-th order association. This represents the Frobenius norm.

[0074] Remove connection edges whose topological edge weights are lower than a preset association strength threshold to obtain the product network topology graph. V represents the set of product nodes, and E represents the set of edges.

[0075] Obtain the target customer's historical consumption records and current shopping needs list, quantify the target customer's purchasing power fluctuations, model the decision patience decay curve, and analyze the shopping path tolerance, and establish a three-dimensional value model of the target customer.

[0076] Historical purchase records include basic transaction data, product detail data, behavioral data, and return / exchange records. Basic transaction data includes transaction ID, timestamp, payment amount, and payment method. Product detail data includes product inventory information, category, unit price, quantity, and discount rate. Behavioral data includes shelf dwell time, try-on and trial records, and abandoned cart items. Return / exchange records include returned product inventory information, reason, and refund amount. The current shopping request list represents the list of planned purchases actively entered by the target customer.

[0077] The process of quantifying the purchasing power fluctuations of target customers includes:

[0078] Based on the historical consumption records of target customers, the arithmetic mean and standard deviation of the amount spent by each target customer in a single purchase are calculated.

[0079] Identify the time-periodic characteristics of target customer consumption behavior and construct a fluctuation function incorporating a time decay factor, expressed by the following formula:

[0080]

[0081] The purchasing power benchmark is expressed by the formula:

[0082]

[0083] The fluctuation range is expressed by the formula:

[0084]

[0085] In the formula, Represents the quantile function. Indicates the amount of a single transaction. This represents the set of valid transactions for the target customer, where t represents the current time. Indicates the most recent consumption timestamp. Indicates the attenuation intensity. Let f represent the time of the target customer's first purchase, f represent the purchase frequency, and B represent the total number of purchases. This represents the amount of the i-th transaction. This represents the average of all consumption.

[0086] Using the fluctuation function value as the center line, a dynamic fluctuation band interval is generated based on the standard deviation multiple, expressed by the formula:

[0087]

[0088] In the formula, s represents the confidence coefficient.

[0089] By monitoring the cumulative spending amount in the target customer's current shopping needs list in real time, the fluctuation range is dynamically adjusted.

[0090] The process of modeling the decision-making patience decay curve for target customers includes:

[0091] Analyze the target customer's historical behavior records, calculate the average time from browsing products to making a purchase decision, and map the average time to an initial patience value.

[0092] A hyperbolic decay function model is established, which makes the initial patience value exhibit nonlinear decay characteristics as the decision time progresses. The formula for the hyperbolic decay function model is expressed as:

[0093]

[0094] In the formula, This represents the initial patience value. This represents the decay rate coefficient, and t represents the current time. Indicates the start time of entering the decision-making state. This represents the nonlinear decay exponent.

[0095] The decay rate coefficient is dynamically adjusted based on current environmental parameters, including mall pedestrian density and promotional activity intensity.

[0096] When the actual decision-making time of the target customer deviates from the predicted value by more than a preset deviation threshold, the attenuation function parameters are adjusted, as expressed by the formula:

[0097]

[0098] In the formula, Indicates the learning rate. Indicates the actual time spent on decision-making. This indicates the time required for prediction and decision-making.

[0099] The process of conducting shopping path tolerance analysis for target customers includes:

[0100] Extract path abandonment records from the target customer's historical shopping history and fit a baseline threshold for the target customer's tolerance.

[0101] A spatial distance tolerance conversion model is established based on the baseline threshold, and the tolerance energy function is calculated. The formula is as follows:

[0102]

[0103] In the formula, This represents the tolerance energy integral of path Γ. Indicates a parameterized path, Let q represent the distance field function and q represent the tolerance coefficient. This represents the gradient norm of the distance field.

[0104] The system integrates the time taken for this shopping trip in real time and dynamically updates the path tolerance threshold.

[0105] Based on the product network topology diagram and the current shopping needs list, a dynamic optimal product penetration pool is generated using a product association mining strategy. The process includes:

[0106] In the product network topology diagram, starting from the target customer's current shopping needs list, the penetration rate of all candidate products reachable through direct association paths is calculated, expressed by the formula:

[0107]

[0108] In the formula, This indicates the penetration rate of candidate product C, where C represents the candidate product identifier. Indicates the strength of direct association, i.e., the products The geometric mean of the association probability with C and the lift. The logarithm of the gradient of the associated potential field represents the commodity. The direction and intensity of the impact on other commodities.

[0109] Perform multi-demand collaborative correction and examine the second-order impact of the combination of items in the target customer's current shopping demand list on candidate items.

[0110] The decay rate of product association strength within the pool is monitored in real time, and the initial product pool is selected based on a dynamically preset penetration rate threshold. The formula is as follows:

[0111]

[0112] In the formula, P represents the dynamic optimal product penetration pool. Indicates the penetration threshold. Indicates the rate of permeability decay. This indicates the lower limit of the decay rate tolerance. This indicates the conditions for inclusion in emergency response. The mixed partial derivatives of the correlation strength are represented. This represents the collaborative gain threshold.

[0113] The process involves acquiring the spatial coordinates of each product in the product penetration pool and a 3D map of the shopping mall. Combined with a 3D value model, and utilizing spatiotemporal fusion field construction technology and a global dynamic path optimization algorithm, an optimal shopping guide route sequence is generated. This process includes:

[0114] The spatial coordinates and value weights of each product in the product penetration pool are mapped into a four-dimensional spatiotemporal field.

[0115] By overlaying a three-dimensional user value model to form a comprehensive constraint field, a spatiotemporal optimization equation is established, resulting in a continuous path curve. The comprehensive constraint field includes value constraints defined by the current fluctuation zone, the maximum dwell time determined by the decision patience value, and spatial movement restrictions converted from path tolerance. The spatiotemporal optimization equation is expressed as follows:

[0116]

[0117] In the formula, This represents the total spatiotemporal loss of path Γ. This represents the current parameterized path, and t represents the current time. This represents the lower limit of integration, i.e., the moment when path planning begins. This indicates the maximum number of points you can accumulate, which is the time you plan to complete your shopping. This represents the coefficient of spatial friction, i.e., the cost of walking per unit distance. Represents the spatial-temporal field gradient. Indicates the value conversion rate. This indicates the real-time probability of purchase. This indicates dynamic purchasing power.

[0118] The variational method is used to find the continuous path curve that minimizes spatiotemporal friction loss, which includes path distance loss and value decay loss.

[0119] The continuous path curve is discretized into a sequence of shopping guide nodes, and sorted by spatial reachability to output a set of shopping guide instructions, thus obtaining the optimal shopping guide route sequence, expressed by the formula:

[0120]

[0121] In the formula, This represents the optimal shopping guide route sequence, where K represents the number of path segments. This represents the coordinates of the path node, specifically the shelf position of the k-th recommended product. This represents a time point, specifically the planned time to reach the k-th node. Let be the Lagrangian density, representing the path loss function per unit time.

[0122] The system acquires real-time user location signals and dynamic changes in shopping cart items, and generates dynamically adjusted shopping guide instructions based on deviations in the guided route and the types of items in the shopping cart. The process includes:

[0123] Real-time acquisition of target customer location coordinates and shopping cart item list; establishment of a product decision entropy monitoring model based on the offset of real-time location coordinates and shopping cart item list from the optimal shopping guide route sequence and product penetration pool; formula expressed as:

[0124]

[0125] In the formula, C represents the candidate product identifier. This represents the collection of unpurchased items. Indicates the probability of purchasing candidate items. Indicates the path deviation coupling coefficient. This represents the target customer's movement speed vector. This represents the planned path velocity vector.

[0126] When any deviation exceeds a preset threshold, a dynamic replanning condition is triggered. The dynamic replanning condition includes recalculating the spatiotemporal field gradient direction based on the latest data and generating a set of correction instructions, including steering commands, speed adjustments, and product replacement suggestions. The dynamic replanning condition formula is expressed as follows:

[0127]

[0128] In the formula, Indicates the monitoring time window. Represents real-time decision entropy. Represents the entropy of historical decisions. Represents the entropy change threshold. Represents logical operators, Indicates the real-time location of the target customer. Indicates the planned path location. Indicates the path offset distance.

[0129] In summary, this embodiment provides a big data-based intelligent shopping guide recommendation method. It breaks through the limitations of traditional pairwise recommendations by using advanced scenario-based association mining to achieve accurate prediction of cross-category demand. It integrates purchasing power fluctuations, decision-making patience decay, and path tolerance into a three-dimensional value assessment system, possessing adaptive capabilities to perceive users' real-time consumption capacity and behavioral tendencies. Based on spatiotemporal fusion field technology, it unifies the modeling of product value, physical movement, and time constraints into a continuous optimization problem, generating personalized shopping guide routes that maximize user value. Finally, relying on dynamic entropy change monitoring and a dual-deviation response mechanism, it achieves closed-loop instruction optimization, completely solving the pain point of recommendation failure caused by user interest drift or path deviation. Through multi-level technological collaboration, it significantly improves the matching efficiency of people, goods, and places in retail scenarios. On the user side, it reduces the cognitive burden of choosing from a massive number of products and the waste of ineffective walking, achieving an efficient and enjoyable immersive shopping experience. On the merchant side, it activates the penetration rate of long-tail products and the conversion of related sales, and optimizes inventory turnover through data-driven dynamic pricing. Ultimately, it constructs an intelligent shopping guide model that combines accurate recommendation capabilities, spatiotemporal resource optimization efficiency, and humanistic attributes, providing core infrastructure support for the digital transformation of physical retail.

[0130] Based on the same general inventive concept, this invention also protects a big data-based intelligent shopping guide recommendation system. The following describes a big data-based intelligent shopping guide recommendation system provided by this invention. The big data-based intelligent shopping guide recommendation system described below and the big data-based intelligent shopping guide recommendation method described above can be referred to and correspond to each other.

[0131] Figure 2 This is a schematic diagram of the structure of an intelligent shopping guide recommendation system based on big data, provided in an embodiment of the present invention.

[0132] like Figure 2 As shown, an intelligent shopping guide recommendation system based on big data includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The processor includes a product association topology construction module, a user three-dimensional value modeling module, a dynamic product penetration pool generation module, a spatiotemporal path optimization module, and a dynamic instruction adjustment module.

[0133] The product association topology construction module is used to collect historical shopping records and product category information from the mall, and uses a high-dimensional association degree calculation model to construct a product network topology diagram.

[0134] The user 3D value modeling module is used to obtain the target customer's historical consumption records and current shopping needs list, quantify the target customer's purchasing power fluctuations, model the decision patience decay curve, and analyze the shopping path tolerance, thereby establishing a 3D value model of the target customer.

[0135] The dynamic product penetration pool generation module is used to generate a dynamic and optimal product penetration pool based on the product network topology diagram and the current shopping demand list, using a product association mining strategy.

[0136] The spatiotemporal path optimization module is used to obtain the spatial coordinates of each product in the product penetration pool and the three-dimensional map of the shopping mall. Combined with the three-dimensional value model, it uses spatiotemporal fusion field construction technology and global dynamic path optimization algorithm to generate the optimal shopping guide route sequence.

[0137] The dynamic instruction adjustment module is used to acquire user location signals and dynamic change data of shopping cart items in real time, and generate dynamically adjusted shopping guide instructions based on deviations in the shopping guide route and the types of items in the shopping cart.

[0138] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0139] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart shopping guide recommendation method based on big data, characterized in that, include: Collect historical shopping records and product category information from shopping malls, and use a high-dimensional correlation calculation model to construct a product network topology diagram; Obtain the target customer’s historical consumption records and current shopping needs list, quantify the target customer’s purchasing power fluctuations, model the decision patience decay curve, and analyze the shopping path tolerance, and establish a three-dimensional value model of the target customer. The process of quantifying the purchasing power fluctuations of target customers includes: Based on the historical consumption records of target customers, calculate the arithmetic mean and standard deviation of the amount spent by each target customer in a single purchase; Identify the time-periodic characteristics of target customer consumption behavior and construct a fluctuation function incorporating a time decay factor; Using the fluctuation function value as the center line, a dynamic fluctuation band interval is generated based on the standard deviation multiple. By monitoring the cumulative spending amount in the target customer's current shopping needs list in real time, the fluctuation range is dynamically adjusted. The process of modeling the decision-making patience decay curve for target customers includes: Analyze the target customer's historical behavior records, calculate the average time from browsing products to making a purchase decision, and map the average time to an initial patience value. A hyperbolic decay function model is established so that the initial patience value exhibits nonlinear decay characteristics as the decision time progresses; The decay rate coefficient is dynamically adjusted based on current environmental parameters, including mall pedestrian density and promotional activity intensity. When the actual decision-making time of the target customer deviates from the predicted value by more than the preset deviation threshold, the decay function parameters are adjusted. The process of conducting shopping path tolerance analysis for target customers includes: Extract path abandonment records from the target customer's historical shopping history and fit a baseline threshold for the target customer's tolerance. A spatial distance tolerance transformation model is established based on the aforementioned benchmark threshold, and the tolerance energy function is calculated. The system integrates the time elapsed during this shopping trip in real time and dynamically updates the path tolerance threshold. Based on the product network topology diagram and the current shopping needs list, a dynamic optimal product penetration pool is generated using a product association mining strategy. The process includes: In the product network topology diagram, starting from the target customer's current shopping needs list, calculate the penetration rate of all candidate products that can be reached through direct association paths; Perform multi-demand collaborative correction and detect the second-order impact of the combination of items in the target customer's current shopping demand list on candidate items; The decay rate of the association strength of products in the pool is monitored in real time, and the initial product pool is selected based on the dynamically preset penetration rate threshold. The spatial coordinates of each product in the product penetration pool and the three-dimensional map of the shopping mall are obtained. Combined with the three-dimensional value model, the spatiotemporal fusion field construction technology and the global dynamic path optimization algorithm are used to generate the optimal shopping guide route sequence. It acquires real-time user location signals and dynamic changes in shopping cart items, and generates dynamically adjusted shopping guide instructions based on deviations in the shopping guide route and the types of items in the shopping cart.

2. The intelligent shopping guide recommendation method based on big data according to claim 1, characterized in that, The process of constructing a product network topology diagram includes: Data cleaning is performed on the mall's historical shopping records to extract product combination information from complete transaction records; Define a product association tensor, calculate the first-order association strength between any two products, and extend the multi-order association relationship. The first-order association strength represents the geometric mean of the conditional probability and the lift index; the multi-order association relationship represents the higher-order collaborative purchasing pattern in the product combination. The multi-level associations are stored in the form of a three-dimensional tensor, and topological edge weights are generated through tensor decomposition. Remove connection edges whose topological edge weights are lower than a preset association strength threshold to obtain a product network topology graph. In the graph, nodes represent products, and edges represent the association relationships between multiple product combinations.

3. The intelligent shopping guide recommendation method based on big data according to claim 1, characterized in that, The historical consumption records include basic transaction data, product details data, behavioral trajectory data, and return and exchange records; the current shopping needs list represents the list of planned purchases actively entered by the target customer.

4. The intelligent shopping guide recommendation method based on big data according to claim 1, characterized in that, The process of generating the optimal shopping route sequence includes: The spatial coordinates and value weights of each product in the product penetration pool are mapped into a four-dimensional spatiotemporal field. By superimposing the user's three-dimensional value model to form a comprehensive constraint field, a spatiotemporal optimization equation is established to obtain a continuous path curve. The comprehensive constraint field includes the value constraint defined by the current fluctuation zone, the maximum dwell time determined by the decision patience value, and the spatial movement limit converted by the path tolerance. The continuous path curve is discretized into a sequence of shopping guide nodes, and sorted by spatial reachability to output a set of shopping guide instructions, thus obtaining the optimal shopping guide sequence.

5. The intelligent shopping guide recommendation method based on big data according to claim 1, characterized in that, The process of generating dynamically adjusted shopping guide instructions includes: Real-time acquisition of target customer location coordinates and shopping cart item list; establishment of a product decision entropy monitoring model based on the real-time location coordinates and shopping cart item list offset from the optimal shopping guide route sequence and product penetration pool. When any deviation exceeds a preset threshold, a dynamic replanning condition is triggered. The dynamic replanning condition includes recalculating the spatiotemporal field gradient direction based on the latest data and generating a set of correction instructions, including steering commands, speed adjustments, and product replacement suggestions.

6. A big data-based intelligent shopping guide recommendation system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements a big data-based intelligent shopping guide recommendation method as described in any one of claims 1 to 5.

Images