Sales prediction method and store building site selection method

Abstract

The invention relates to the technical field of intelligent business analysis, and discloses a sales prediction method and a store building site selection method, which comprehensively considers various key data around candidate site selection points of a to-be-built store, including POI data, passenger flow data, consumer portrait data and store attribute data. And in combination with feature engineering processing and a sales prediction optimization model, the sales of the candidate site selection points can be comprehensively and accurately predicted. Compared with a traditional prediction method, the method can fully capture the influence of the complex environmental factors around the site selection on the sales volume, thereby remarkably improving the prediction accuracy and reliability. In addition, the method can help the enterprise to more scientifically evaluate the commercial potential of different candidate site selection points in the site selection decision process, reduce the operation risk, improve the rate of return on investment, and provide powerful support for the commercial site selection and operation decision of the enterprise.

Description

Technical Field

[0001] The present invention relates to the field of intelligent business analysis technology, and in particular to a sales forecasting method and a store site selection method. Background Art

[0002] Among the many key aspects of business operations, store location decisions play a crucial role. Accurately forecasting sales for a prospective store is a core element in this decision-making process, and its importance cannot be overstated. The accuracy of sales forecasts directly impacts a company's estimated return on investment for a store, the rationality of resource allocation, and the formulation of future business strategies.

[0003] Currently, traditional sales forecasting methods mainly rely on historical sales data and simple statistical analysis, but these methods often fail to fully consider the complex environmental factors surrounding store locations, such as POIs (Points of Interest, i.e., any non-geographically meaningful points on a map, such as stores, bars, gas stations, hospitals, train stations, etc.), customer flow, consumer profiles, etc.

[0004] Therefore, a more comprehensive and accurate sales forecasting method is needed to help companies make more scientific decisions in the site selection of stores to be built.

[0005] The above information is presented as background information only to assist with an understanding of the present disclosure and is not a determination or admission that any of the above may be applicable as prior art with respect to the present disclosure. Summary of the Invention

[0006] The present invention provides a sales forecasting method and a store site selection method to achieve accurate prediction of sales of candidate sites for store construction, providing strong support for the site selection decision of the enterprise.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides a sales forecasting method, the method comprising:

[0009] S10: Obtain POI data, customer flow data, character portrait data, and store attribute data around the candidate site of the store to be built;

[0010] S11. Perform feature engineering on the acquired data;

[0011] S12. Input the processed data into a sales forecast optimization model to predict the sales of the candidate locations of the store to be built.

[0012] Furthermore, in the sales forecasting method, in S10, the POI data includes one or more of life service places, transportation facility places, educational places, government service places, financial places, business office places, sports and fitness places, public service places, tourist attractions and accommodation places;

[0013] The character portrait data includes one or more of age, gender, spending capacity, spending preferences, and spending frequency;

[0014] The store's own attribute data includes one or more of the area, business type, floor, and whether the store is located at a corner.

[0015] Furthermore, in the sales forecasting method, S11 includes:

[0016] S111. Calculating mathematical statistical features of the acquired data; the mathematical statistical features include one or more of the mean, median, and standard deviation;

[0017] S112: Perform one-hot encoding on the geographic information corresponding to the candidate site selection points of the store to be built.

[0018] Furthermore, in the sales forecasting method, after S111, the method further includes:

[0019] S111.5. Extract time series features from the acquired data; the time series features include one or more of trend components and seasonal components.

[0020] Furthermore, in the sales forecasting method, in S12, the sales forecasting optimization model is constructed based on a machine learning model, and hyperparameters are optimized using an optimization algorithm combined with a cross-validation method.

[0021] Furthermore, the sales forecasting method further includes:

[0022] S20. Build a sales forecasting model based on a machine learning model;

[0023] S21. Optimizing hyperparameters of the sales forecast model using an optimization algorithm combined with a cross-validation method to obtain an optimized sales forecast model.

[0024] S22: Training the sales forecast optimization model to obtain the trained sales forecast optimization model.

[0025] Furthermore, in the sales forecasting method, the machine learning model is any one of a LightGBM model, an XGBoost model, a CatBoost model, or a random forest model;

[0026] The optimization algorithm is any one of a PSO algorithm, a genetic algorithm, a grid search algorithm, and a random search algorithm.

[0027] In a second aspect, the present invention provides a method for selecting a store location, the method comprising:

[0028] S100: Using the sales forecasting method provided in the first aspect above, forecast the sales of several candidate locations for store construction;

[0029] S200. Based on the prediction results, select the candidate site with the highest sales forecast value from a number of candidate sites as the final site to guide the site selection of the store to be built.

[0030] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the sales forecasting method provided in the first aspect or the store site selection method provided in the second aspect.

[0031] In a fourth aspect, the present invention provides a computer-readable storage medium having computer-executable instructions stored thereon, wherein the computer-executable instructions are executed by a computer processor to implement the sales forecasting method provided in the first aspect above, or the store site selection method provided in the second aspect above.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] The present invention provides a sales forecasting method and a store site selection method. By comprehensively considering a variety of key data around the candidate site selection point of the store to be built, including POI data, customer flow data, consumer portrait data, and the store's own attribute data, and combining feature engineering processing and sales forecast optimization models, it is possible to comprehensively and accurately predict the sales of the candidate site selection point. Compared with traditional forecasting methods, the present invention can fully capture the impact of complex environmental factors around the site selection on sales, thereby significantly improving the accuracy and reliability of the forecast. In addition, the method can also help enterprises more scientifically evaluate the commercial potential of different candidate site selection points during the site selection decision-making process, reduce operational risks, improve return on investment, and provide strong support for the company's business site selection and operational decisions.

[0034] The present invention has other features and advantages that will be apparent from or will be described in detail in the accompanying drawings and the following detailed description incorporated herein, which together serve to explain certain principles of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0036] Figure 1 This is one of the flow charts of a sales forecasting method provided in the first embodiment of the present invention;

[0037] Figure 2 This is a second flow chart of a sales forecasting method provided in the first embodiment of the present invention;

[0038] Figure 3 This is a third flow chart of a sales forecasting method provided in the first embodiment of the present invention;

[0039] Figure 4 This is a fourth flow chart of a sales forecasting method provided in the first embodiment of the present invention;

[0040] Figure 5 This is a flow chart of a store location selection method provided in the second embodiment of the present invention;

[0041] Figure 6 This is a structural diagram of a computer device provided in Example 3 of the present invention. DETAILED DESCRIPTION

[0042] In order to explain in detail the possible application scenarios, technical principles, specific solutions that can be implemented, and the purpose and effects of this application, the following is a detailed description of the specific embodiments listed in conjunction with the accompanying drawings. The embodiments described herein are only used to more clearly illustrate the technical solutions of this application and are therefore only examples and are not intended to limit the scope of protection of this application.

[0043] References to "embodiments" herein mean that the specific features, structures, or characteristics described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the word "embodiment" in various places in the specification does not necessarily refer to the same embodiment, nor does it particularly limit its independence or relevance to other embodiments. In principle, in this application, as long as there are no technical contradictions or conflicts, the various technical features mentioned in the embodiments can be combined in any manner to form a corresponding implementable technical solution.

[0044] Unless otherwise defined, the technical terms used herein have the same meanings as those generally understood by those skilled in the art to which this application belongs; the use of relevant terms herein is only for describing specific embodiments and is not intended to limit this application.

[0045] In the description of this application, the term "and / or" is used to describe a logical relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A exists, B exists, and both A and B exist. In addition, the character " / " in this document generally indicates that the objects before and after are in a logical "or" relationship.

[0046] In this application, terms such as "first" and "second" are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual quantity, priority or sequence relationship between these entities or operations.

[0047] Without further limitations, in this application, the words "include", "comprise", "have" or other similar expressions used in the sentences are intended to cover non-exclusive inclusion. These expressions do not exclude the presence of additional elements in the process, method or product including the elements, so that the process, method or product including a series of elements may include not only those defined elements, but also other elements not explicitly listed, or elements inherent to such process, method or product.

[0048] In this application, expressions such as "greater than," "less than," and "exceed" are understood to exclude the number itself; expressions such as "above," "below," and "within" are understood to include the number itself. In addition, in the description of the embodiments of this application, "multiple" means two or more (including two), and similar expressions related to "multiple" are also understood in this way, such as "multiple groups" and "multiple times," unless otherwise specifically limited.

[0049] In the description of the embodiments of the present application, the space-related expressions used, such as "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "vertical", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc., indicate the orientation or position relationship based on the orientation or position relationship shown in the specific embodiments or drawings, and are only for the convenience of describing the specific embodiments of the present application or facilitating the reader's understanding, and do not indicate or imply that the device or component referred to must have a specific position, a specific orientation, or be constructed or operated in a specific orientation. Therefore, it should not be understood as a limitation on the embodiments of the present application.

[0050] Unless otherwise expressly specified or limited, in the description of the embodiments of the present application, the terms "installed", "connected", "connected", "fixed", "set", etc. used should be understood in a broad sense. For example, the "connection" can be a fixed connection, a detachable connection, or an integrated setting; it can be a mechanical connection, an electrical connection, or a communication connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be the internal connection of two elements or the interaction relationship between two elements. For those skilled in the art of the present application, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0051] Example 1

[0052] Please refer to Figure 1 , is a flow chart of a sales forecasting method provided in Example 1 of the present invention. This method is applicable to scenarios where an enterprise predicts sales of stores to be built. This method can be implemented by software and / or hardware. The method specifically includes the following steps:

[0053] S10. Obtain POI data, customer flow data, character portrait data, and store attribute data around the candidate site selection point for the store to be built.

[0054] It should be noted that this step is the foundation of the entire sales forecasting process. By collecting multiple types of data, it provides a comprehensive and rich source of information for subsequent feature engineering processing and model prediction, ensuring that the forecast results can fully consider various factors affecting sales.

[0055] Among them, POI data: POI stands for Point of Interest, which refers to various commercial facilities and public service facilities around the candidate site. These facilities include but are not limited to shopping malls, supermarkets, restaurants, cafes, schools, hospitals, parks, banks, etc. Obtaining POI data can help us understand the richness of the surrounding business ecosystem, the competitive landscape, and the complementarity between different types of facilities. For example, if there are multiple large shopping malls and well-known restaurants around the candidate site, it may attract a large number of consumers, thereby increasing the potential customer flow of the store to be built; and the presence of nearby schools may attract students and parents, which has a positive impact on stores operating education-related businesses such as stationery and educational training.

[0056] Customer flow data: Customer flow data reflects the popularity and commercial activity of candidate locations. It includes not only customer flow statistics for different time periods (e.g., weekdays, weekends, and holidays), but also detailed information such as the source of customer flow (e.g., local residents, tourists), flow direction (e.g., the direction of entry into the area), and duration of stay. By analyzing customer flow data, we can understand peaks and troughs in customer flow during different time periods, predict potential customer flow for stores during different time periods, and assess sales trends. For example, if a candidate location experiences high customer flow on weekday evenings and weekends, then for restaurants, entertainment outlets, and other businesses, these periods may also be peak sales periods.

[0057] Profile data: Profile data depicts the characteristics of the target consumer group from multiple perspectives, including age, gender, and consumer preferences. By analyzing profile data, we can better understand the needs and expectations of target customers and predict the appeal of different candidate locations to specific consumer groups, as well as the resulting consumer behavior. For example, if the target consumer group is primarily young people, who may be more inclined to consume trendy and personalized products or services, the location selection process should consider the availability of nearby shopping venues and an atmosphere that is suitable for young people.

[0058] Store attribute data: This data covers information closely related to store operations, such as store size and business type. This data is directly related to the store's positioning and market competitiveness, and has a significant impact on sales forecasts. For example, if the store to be built is a high-end luxury store, its location may need to consider whether there is a sufficient number of high-end consumers and a good business atmosphere in the surrounding area. On the other hand, a small convenience store is more suitable for being located near a residential area to meet the daily consumption needs of residents.

[0059] In one implementation of this embodiment, the data types acquired in S10 are more carefully defined, clarifying the specific content included in POI data, character portrait data, and store attribute data. This data is an important foundation for subsequent sales forecasting. Comprehensive and accurate data collection helps improve the accuracy and reliability of the forecasting model.

[0060] The POI data covers various types of locations that are closely related to the potential customer flow, consumer demand, and business environment of the store to be built. Specifically, they may include one or more of the following:

[0061] Life service places: including restaurants, shops, leisure and entertainment venues (such as cinemas, KTVs, etc.), medical service institutions (such as hospitals, clinics), residential communities, etc.

[0062] It is understandable that the presence of restaurants and shops can attract surrounding residents and passers-by, increasing potential customer flow; leisure and entertainment venues can meet consumers' entertainment needs and attract specific consumer groups; there may be a stable population of medical seekers near medical service institutions, bringing certain consumption opportunities; the distribution and size of residential communities can reflect the surrounding population density and living conditions, which is of great significance for predicting the daily customer flow of stores.

[0063] Transportation facilities: such as bus stations, parking lots, subway stations, airports, etc.

[0064] Understandably, the convenience of transportation directly impacts consumer accessibility. Bus and subway stations typically have high foot traffic, bringing more potential customers to stores. The availability and size of parking lots are crucial for consumers arriving by car. Airport proximity can attract business travelers and tourists, creating potential value for stores operating tourism- and business-related businesses.

[0065] Educational places: kindergartens, schools, etc.

[0066] Understandably, educational venues are often surrounded by large groups of students and parents, making them attractive to stores selling stationery, educational training, and catering services. For example, stationery stores near schools often experience higher sales during the start of the school term and exam periods.

[0067] Government service places: government service centers, police stations, courts, etc.

[0068] It is understandable that the existence of these places may attract specific groups of people to conduct business. Although the direct consumption they bring may be relatively limited, a certain business atmosphere may be formed in the surrounding areas. For example, there may be some places around that provide catering and rest services for business personnel.

[0069] Financial places: banks, insurance companies, etc.

[0070] Understandably, financial venues often have a certain amount of business activity and foot traffic around them, potentially attracting some finance-related professionals and clients. For example, a coffee shop near a bank might become a venue for business negotiations, generating a certain amount of spending.

[0071] Business office places: office buildings, industrial parks, etc.

[0072] Understandably, business offices, where a large number of office workers gather, may have certain consumption needs during lunch and afternoon tea breaks, as well as after get off work. For example, restaurants and convenience stores near office buildings may experience peak customer traffic during these times.

[0073] Sports and fitness venues: gymnasiums, fitness centers, basketball courts, etc.

[0074] Understandably, these venues attract a large number of people who love sports, creating a potential market for stores selling sports equipment, health foods, sports rehabilitation, and other related businesses. At the same time, athletes may also have consumption needs such as dining and rest before and after exercise.

[0075] Public service places: parks, squares, etc.

[0076] It is understandable that parks and squares are popular places for people to relax and have fun, especially during holidays and weekends when there are large crowds of people. There may be some snack stalls, souvenir shops, etc. around them to meet the consumption needs of tourists and residents.

[0077] Tourist attractions: scenic spots, etc.

[0078] Understandably, tourist attractions attract a large number of tourists, creating huge business opportunities for stores selling tourist souvenirs, specialty restaurants, accommodation, and other related businesses. Tourists' spending behavior and spending power may be affected by factors such as the tourist season and the popularity of the attraction.

[0079] Accommodation places: B&Bs, hotels, etc.

[0080] Understandably, the presence of accommodations signifies the presence of tourists or business people, who are likely to spend money in the surrounding area, such as dining and shopping. Furthermore, the quality and scale of accommodations also influence the spending power and preferences of surrounding consumer groups.

[0081] The character portrait data depicts the characteristics of the target consumer group from multiple dimensions, which helps to more accurately predict sales. Specifically, it may include one or more of the following:

[0082] Age: Consumers of different age groups have different consumption needs and spending power. For example, young people may prefer to consume fashionable and personalized products or services, while middle-aged and elderly people may focus more on practicality and cost-effectiveness.

[0083] Gender: Men and women have different consumer preferences. For example, women may be more interested in beauty products, clothing, and accessories, while men may be more inclined towards consumer electronics and sports equipment.

[0084] Spending power: Spending power reflects a consumer's purchasing power and is typically measured through indicators such as income level and asset status. Understanding the spending power of your target consumer group helps determine a store's pricing strategy and product offerings.

[0085] Consumer preferences: Consumer preferences reflect consumers' preferences for specific products or services. For example, some consumers may prefer organic food and eco-friendly products, while others may favor high-end luxury goods. Analyzing consumer preferences can inform store product selection and marketing strategies.

[0086] Consumption frequency: Consumption frequency reflects the number of times a consumer purchases a product or service within a certain period of time. Understanding the consumption frequency of your target consumer group helps predict store traffic and sales, and optimize inventory and staffing.

[0087] The store attribute data is directly related to the store's operation and market positioning, and has a significant impact on sales forecasts. Specifically, it may include one or more of the following:

[0088] Size: A store's size determines the variety and quantity of merchandise it can accommodate, as well as the customer experience. Larger stores may offer a wider selection, but this also increases operating costs; smaller stores require a greater focus on product selection and display.

[0089] Business Type: The business type defines the range of goods or services provided by the store, such as catering, retail, and services. Different business types have different market demands and competitive situations, and their impact on sales also varies.

[0090] Floor: The store's location affects its visibility and accessibility. Generally speaking, stores on lower floors are easier for consumers to find and enter, and may have higher customer traffic; stores on higher floors may need better promotion and guidance to attract customers.

[0091] Corner location: Corner stores typically have better display and higher exposure, attracting more passersby. Compared to non-corner stores, corner stores may have higher customer traffic and sales.

[0092] Through the detailed collection and analysis of these data, comprehensive and accurate information can be provided for subsequent feature engineering processing and sales forecasting models, thereby improving the accuracy and reliability of sales forecasts and providing strong support for the site selection and operating decisions of the company's upcoming stores.

[0093] S11. Perform feature engineering on the acquired data.

[0094] It's important to note that feature engineering involves performing a series of operations on raw data to improve its quantity, quality, and usability, making it more suitable for use in subsequent sales forecasting models. Feature engineering can extract the most representative and predictive features from massive amounts of data, removing redundant and noisy information, thereby improving the model's predictive accuracy and efficiency.

[0095] S12. Input the processed data into a sales forecast optimization model to predict the sales of the candidate locations of the store to be built.

[0096] It should be noted that this step is to input the data processed by feature engineering into the pre-built sales forecast optimization model, and use the computing power and learning ability of the model to predict the sales of candidate locations for store construction, providing a basis for decision-making for the enterprise.

[0097] The sales forecast optimization model is carefully constructed and trained using advanced machine learning algorithms or deep learning techniques. During the model construction process, a large amount of historical sales data, as well as related POI data, customer flow data, character portrait data, and store attribute data, is used for training. This allows the model to automatically learn complex patterns and regularities in the data, capturing the potential relationship between complex environmental factors surrounding the selected location and sales.

[0098] The prediction process involves feeding processed data into the model, which then calculates and predicts sales for candidate locations based on its learned knowledge and patterns. The prediction result is typically a specific value or range, representing the expected sales for the candidate location over a specific timeframe.

[0099] Understandably, companies can use the forecast results to compare and evaluate different candidate locations, selecting locations with higher sales forecasts and greater commercial potential for store development. Furthermore, the forecast results can also provide a reference for the company's subsequent business strategy formulation, such as pricing and marketing strategies.

[0100] Please refer to Figure 2 In one implementation of this embodiment, S11 can be further refined to include the following sub-steps, aiming to mine more valuable information from the data through more sophisticated feature engineering processing, provide more representative features for the subsequent sales forecast optimization model, and thus improve the accuracy of the forecast:

[0101] S111. Calculate mathematical statistical features of the acquired data; the mathematical statistical features include one or more of the mean, median, and standard deviation.

[0102] It should be noted that the calculation of mathematical and statistical features is a basic and important way to process raw data. By calculating statistics such as the mean, median, and standard deviation, the type and quantity of data can be increased. It can also grasp the distribution and centralization trend of the data as a whole, reveal the basic characteristics of the data, and provide a reference for subsequent feature selection and model training.

[0103] Specific calculation and significance:

[0104] Average: Calculating the average of data can reveal the data's central tendency. For example, for customer flow data, calculating the average customer flow over different time periods (such as daily, weekly, and monthly) can provide insights into the average popularity of a candidate location within that time period. A high average customer flow at a candidate location indicates a large potential consumer base in that area, potentially positively impacting sales at the proposed store.

[0105] Median: The median is the value in the middle of a set of data after it is sorted by size. Compared to the mean, the median is less affected by extreme values ​​and more robustly reflects the central position of the data. When dealing with data that may contain outliers (such as occasional extreme peaks or valleys in passenger flow data), the median can provide a more reliable estimate of central tendency.

[0106] Standard Deviation: Standard deviation measures the degree of dispersion in data. A larger standard deviation indicates greater data volatility; a smaller standard deviation indicates greater data concentration. In sales forecasting, standard deviation can help us understand data stability. For example, calculating the standard deviation of historical sales data (if relevant reference data is available) can help us determine sales fluctuations and, in turn, assess the stability of the market environment around different candidate locations. If the standard deviation of the data around a candidate location is small, this indicates a relatively stable market environment in that area, which is conducive to the long-term operation of the store.

[0107] S112: Perform one-hot encoding on the geographic information corresponding to the candidate site selection points of the store to be built.

[0108] It's important to note that geographic information typically exists in the form of categorical variables, such as the region (e.g., city center, suburbs, commercial district, residential area) or administrative divisions where a candidate site is located. One-hot encoding is an effective method for converting categorical variables into numerical data. One-hot encoding can convert geographic information into a format that models can process, preventing the model from treating categorical variables as simply numerically related, thereby improving the model's predictive accuracy.

[0109] Specific operations and significance:

[0110] Operational steps: Suppose the geographic information for a candidate site includes a categorical variable called "region," with four possible values: "city center," "suburbs," "commercial district," and "residential district." One-hot encoding converts these four values ​​into four new binary variables, one for each value. For example, the binary variable for "city center" is [1, 0, 0, 0], the binary variable for "suburbs" is [0, 1, 0, 0], and so on.

[0111] Significance: One-hot encoding allows the model to better understand differences between geographic regions and capture the impact of geography on sales. For example, downtown areas may have higher customer traffic and spending power, while residential areas may prioritize convenience and daily consumption. Using one-hot encoded geographic information, the model can more accurately predict sales for different candidate locations.

[0112] Please refer to Figure 3 In one implementation of this embodiment, after S111, the method may be further refined to include the following sub-steps:

[0113] S111.5. Extract time series features from the acquired data; the time series features include one or more of trend components and seasonal components.

[0114] It's important to note that in real-world applications, many data types (such as customer flow and sales data) exhibit time series characteristics, meaning that the data exhibits patterns and trends over time. Extracting time series features can help us capture temporal dynamics within the data, understand its changing trends and cyclical patterns, and thus more accurately predict future sales.

[0115] Specific extraction and significance:

[0116] Trend component: The trend component reflects the overall direction of change in the data over a longer period of time. For example, by analyzing the time series trend of customer flow data around a candidate site, we can determine whether the area's customer flow is increasing, decreasing, or stable. If customer flow is increasing, it indicates that the area has great commercial development potential and may have a positive impact on the future sales of the proposed store. Conversely, if customer flow is decreasing, then the site selection should be carefully considered.

[0117] Seasonal component: Seasonal components refer to cyclical changes in data within a year or other fixed time period. For example, certain commercial areas experience significant increases in customer traffic during holidays or specific seasons (such as peak tourist season), but relatively low traffic at other times. Extracting seasonal components can help us understand this cyclical pattern and consider the impact of seasonal factors when forecasting sales. For example, for a store selling winter clothing, if customer traffic and consumer demand around a candidate location show a seasonal upward trend before winter arrives, the sales forecast for that location may increase accordingly.

[0118] Through the above further refinement of step S11, the implementation method of this embodiment can more comprehensively and deeply mine the characteristic information in the data, provide better input data for the sales forecast optimization model, thereby improving the accuracy and reliability of sales forecasts, and providing more powerful support for the location selection decisions of enterprises' stores to be built.

[0119] In one implementation of this embodiment, in S12, the sales forecast optimization model is constructed based on a machine learning model, and an optimization algorithm combined with a cross-validation method is used to optimize hyperparameters.

[0120] Please refer to Figure 4 In one implementation of this embodiment, the sales forecasting method further includes the following steps, which primarily describe the construction, optimization, and training of a sales forecast optimization model. By constructing a sales forecast model based on a machine learning model and optimizing hyperparameters using an optimization algorithm combined with cross-validation, a trained sales forecast optimization model is ultimately obtained to accurately predict sales at candidate store locations:

[0121] S20. Build a sales forecasting model based on machine learning model.

[0122] It's important to note that this step builds an initial sales forecast model, providing a foundational framework for subsequent hyperparameter optimization and model training. Choosing an appropriate machine learning model can fully leverage the data's feature information and uncover potential relationships between data and sales.

[0123] The choice of machine learning model can be any of the following:

[0124] LightGBM model: LightGBM is an efficient gradient boosting framework based on gradient boosted decision trees (GBDT). It uses a histogram-based decision tree learning algorithm, significantly reducing computational complexity and memory usage while offering fast training speed and high prediction accuracy. It excels at processing large amounts of data and is suitable for tasks such as sales forecasting that require processing diverse data types.

[0125] XGBoost model: XGBoost is a machine learning algorithm based on GBDT, which optimizes and improves upon traditional GBDT. XGBoost introduces regularization to prevent overfitting, supports parallel computing, and can quickly process large amounts of data. Furthermore, XGBoost offers a wealth of parameter tuning options for easy model optimization.

[0126] CatBoost model: CatBoost is specifically optimized for categorical features and can automatically handle categorical features without requiring additional encoding. Its symmetric tree structure reduces the risk of overfitting and is robust against noise and outliers in the data. CatBoost has advantages for sales forecasting tasks involving a high number of categorical features, such as geographic information and store type.

[0127] Random Forest Model: Random Forest is an ensemble learning algorithm that improves model accuracy and stability by building multiple decision trees and combining their predictions. It can handle high-dimensional data, has good modeling capabilities for nonlinear relationships between features, and is tolerant to missing values ​​and outliers.

[0128] S21. Optimize the hyperparameters of the sales forecast model using an optimization algorithm combined with a cross-validation method to obtain an optimized sales forecast model.

[0129] It's important to note that hyperparameters are parameters that need to be manually set before training a machine learning model. They significantly impact the model's performance. By combining optimization algorithms with cross-validation, we can automatically search for the optimal hyperparameter combination, improving the model's predictive accuracy and generalization capabilities.

[0130] The optimization algorithm can be any of the following:

[0131] Particle Swarm Optimization (PSO): A swarm intelligence-based optimization algorithm that simulates the foraging behavior of a flock of birds. In the PSO algorithm, each particle represents a potential solution (i.e., a combination of hyperparameters). The particle continuously updates its position and velocity within the search space to find the optimal solution. The PSO algorithm boasts fast convergence and strong global search capabilities, making it suitable for hyperparameter optimization problems.

[0132] Genetic Algorithm: A genetic algorithm is an optimization algorithm that simulates the process of biological evolution. It continuously generates new solutions through operations such as selection, crossover, and mutation. It evaluates the quality of these solutions based on a fitness function, gradually evolving to the optimal solution. Genetic algorithms are capable of handling complex nonlinear optimization problems and exhibit strong global search capabilities.

[0133] Grid Search Algorithm: The grid search algorithm is a simple and intuitive hyperparameter optimization method. It traverses all possible combinations in a given hyperparameter space, trains and evaluates the model for each combination, and ultimately selects the hyperparameter combination with the best performance. Although the grid search algorithm is computationally intensive, it is guaranteed to find the global optimal solution when the hyperparameter space is small.

[0134] Random search algorithm: A randomized search algorithm is a random sampling search algorithm that randomly samples a certain number of hyperparameter combinations from the hyperparameter space, then evaluates the performance of these combinations and selects the one with the best performance. Random search algorithms are relatively computationally inefficient and can find good hyperparameter combinations in a short time, but they may not necessarily find the global optimal solution.

[0135] Cross-validation: Cross-validation is an effective method for evaluating model performance and selecting hyperparameters. A common cross-validation method is k-fold cross-validation. In k-fold cross-validation, the dataset is randomly divided into k subsets. Each time, k-1 subsets are selected as the training set, and the remaining subset is used as the validation set. Training and validation are repeated k times, and the average of these k evaluation results is used as the model performance metric. Cross-validation allows for more accurate assessment of model performance on different datasets, avoiding overfitting and underfitting.

[0136] S22: Training the sales forecast optimization model to obtain the trained sales forecast optimization model.

[0137] It should be noted that this step uses a hyperparameter-optimized model and prepared training data to train the model, allowing the model to learn the relationship between the features in the data and sales, thereby obtaining a model that can accurately predict sales.

[0138] The training process specifically involves inputting feature-engineered data (including POI data, customer flow data, persona data, and store attribute data) into the sales forecast optimization model. Based on the input data and optimized hyperparameters, the model continuously adjusts its parameters through an iterative optimization algorithm (such as gradient descent) to minimize the error between the predicted and true values. The training process typically involves setting a specific number of training rounds or convergence criteria. When the model reaches convergence or the number of training rounds reaches a preset value, the training process ends, resulting in a trained sales forecast optimization model.

[0139] Through the above steps, the implementation method of this embodiment constructs a sales forecast optimization model based on a machine learning model, and optimizes and trains the model through optimization algorithms and cross-validation methods. This model can fully utilize the characteristic information in the data to accurately predict the sales of candidate locations for future store construction, providing enterprises with a scientific basis for site selection decisions. In practical applications, appropriate machine learning models and optimization algorithms can be selected based on specific data characteristics and business needs to achieve better forecasting results.

[0140] Although this application frequently uses terms such as sales and POI data, the possibility of using other terms is not excluded. These terms are used only to more conveniently describe and explain the essence of the present invention; interpreting them as any additional limitations is contrary to the spirit of the present invention.

[0141] The embodiments of the present invention provide a sales forecasting method and a store site selection method. By comprehensively considering a variety of key data around the candidate site for the store to be built, including POI data, customer flow data, consumer portrait data, and the store's own attribute data, and combining feature engineering processing and sales forecast optimization models, it is possible to comprehensively and accurately predict the sales of the candidate site. Compared with traditional forecasting methods, the embodiments of the present invention can fully capture the impact of complex environmental factors around the site on sales, thereby significantly improving the accuracy and reliability of the forecast. In addition, the method can also help enterprises more scientifically evaluate the commercial potential of different candidate sites in the site selection decision-making process, reduce operational risks, improve return on investment, and provide strong support for the company's business site selection and operational decisions.

[0142] Example 2

[0143] Please refer to Figure 5 , is a flow chart of a store location selection method provided in Example 2 of the present invention. The method specifically includes the following steps:

[0144] S100: Using the sales forecasting method provided in the first embodiment, forecast the sales of several candidate locations for store construction.

[0145] It should be noted that the core goal of this step is to quantitatively predict the future sales of each candidate site using the sales forecasting method described in Example 1. By accurately predicting the sales of each candidate site, we can provide objective, quantifiable data support for subsequent site selection decisions, avoiding the risks associated with relying solely on subjective experience when selecting a site.

[0146] For example, assuming there are three candidate sites A, B, and C, the model will output the sales forecast values ​​of these three sites in a certain period of time in the future. For example, the predicted sales of site A is X yuan, the predicted sales of site B is Y yuan, and the predicted sales of site C is Z yuan.

[0147] S200. Based on the prediction results, select the candidate site with the highest sales forecast value from a number of candidate sites as the final site to guide the site selection of the store to be built.

[0148] It should be noted that after obtaining the sales forecast values ​​of each candidate site, by comparing these forecast values, selecting the site with the highest sales forecast value as the final site can maximize the store's potential operating benefits, reduce site selection risks, and ensure that the store can achieve good sales performance after opening.

[0149] For example, among the three locations A, B, and C mentioned above, if X>Y>Z, then the sales forecast value of location A is the highest.

[0150] Then, this step will determine the site with the highest sales forecast value, that is, site A, as the final site for the store to be built.

[0151] It is understood that in actual operations, in addition to considering sales forecasts, other factors may also be considered for comprehensive evaluation, such as rental costs, surrounding competition, and policies and regulations. However, in this embodiment, the location selection decision is primarily based on sales forecasts to ensure that the store has great potential for operating efficiency.

[0152] The method for selecting a store location provided by the embodiment of the present invention has the following important significance:

[0153] Scientificity: This method makes site selection decisions based on a data-driven sales forecast model, avoiding the influence of subjective factors in traditional site selection methods and making site selection decisions more scientific and objective.

[0154] Accuracy: The sales forecasting model optimized and trained in the first embodiment can more accurately predict the sales of each candidate site, providing a reliable basis for site selection decisions and improving the accuracy of site selection.

[0155] Profitability: Selecting the location with the highest sales forecast as the final location will help improve the store's operating efficiency, reduce investment risks, and bring better economic benefits to the company.

[0156] In summary, the second embodiment of the present invention provides a scientific, accurate and effective solution for store site selection by combining the sales forecasting method provided in the first embodiment, and has important practical application value.

[0157] Example 3

[0158] Figure 6 A schematic diagram of the structure of a computer device provided in Example 3 of the present invention. Figure 6 A block diagram of an exemplary computer device 12 suitable for use in implementing embodiments of the present invention is shown. Figure 6 The computer device 12 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0159] like Figure 6 As shown, computer device 12 is implemented as a general-purpose computing device. Components of computer device 12 may include, but are not limited to, one or more processors or processing units 16, system memory 28, and a bus 18 that connects various system components (including system memory 28 and processing unit 16).

[0160] Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures. Examples of these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0161] The computer device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computer device 12, including volatile and non-volatile media, removable and non-removable media.

[0162] System memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. Computer device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 34 may be configured to read and write non-removable, non-volatile magnetic media ( Figure 6 Not shown, usually called a "hard drive"). Although Figure 6 Although not shown, a magnetic disk drive for reading and writing to a removable non-volatile magnetic disk (e.g., a "floppy disk"), as well as an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. Memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of various embodiments of the present invention.

[0163] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may include an implementation of a network environment. Program modules 42 generally implement the functions and / or methodologies of the embodiments described herein.

[0164] The computer device 12 may also communicate with one or more external devices 14 (e.g., a keyboard, a pointing device, a display 24, etc.), one or more devices that enable a user to interact with the computer device 12, and / or any device that enables the computer device 12 to communicate with one or more other computing devices (e.g., a network card, a modem, etc.). Such communication may be performed via an input / output (I / O) interface 22. Furthermore, the computer device 12 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 20. As shown, the network adapter 20 communicates with the other modules of the computer device 12 via the bus 18. It should be understood that although Figure 6 Not shown, other hardware and / or software modules may be used in conjunction with computer device 12, including but not limited to microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0165] The processing unit 16 executes various functional applications and data processing by running programs stored in the system memory 28, such as implementing the sales forecasting method provided in the first embodiment of the present invention, or the store location selection method provided in the second embodiment of the present invention.

[0166] Example 4

[0167] A fourth embodiment of the present invention provides a computer-readable storage medium having computer-executable instructions stored thereon. When the instructions are executed by a processor, the sales forecasting method provided in the first embodiment of the present invention or the store site selection method provided in the second embodiment of the present invention is implemented.

[0168] Any combination of one or more computer-readable media may be employed. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0169] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0170] Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0171] Computer program code for performing the operations of the present invention may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0172] Finally, it should be noted that although the above embodiments have been described in the specification and drawings of this application, this does not limit the scope of patent protection of this application. All technical solutions generated by replacing or modifying equivalent structures or equivalent processes based on the essential concepts of this application using the contents recorded in the specification and drawings of this application, as well as directly or indirectly implementing the technical solutions of the above embodiments in other related technical fields, are included in the scope of patent protection of this application.

Claims

1. A sales forecasting method, characterized in that: The method comprises: S10: Obtain POI data, customer flow data, character portrait data, and store attribute data around the candidate site of the store to be built; S11. Perform feature engineering on the acquired data; S12. Input the processed data into a sales forecast optimization model to predict the sales of the candidate locations of the store to be built.

2. The sales forecasting method according to claim 1, characterized in that: In S10, the POI data includes one or more of life service places, transportation facilities places, education places, government service places, financial places, business office places, sports and fitness places, public service places, tourist attractions and accommodation places; The character portrait data includes one or more of age, gender, spending capacity, spending preferences, and spending frequency; The store's own attribute data includes one or more of the area, business type, floor, and whether the store is located at a corner.

3. The sales forecasting method according to claim 1, wherein: The S11 includes: S111. Calculating mathematical statistical features of the acquired data; the mathematical statistical features include one or more of the mean, median, and standard deviation; S112: Perform one-hot encoding on the geographic information corresponding to the candidate site selection points of the store to be built.

4. The sales forecasting method according to claim 3, characterized in that: After S111, the method further includes: S111.

5. Extract time series features from the acquired data; the time series features include one or more of trend components and seasonal components.

5. The sales forecasting method according to claim 1, wherein: In S12, the sales forecast optimization model is constructed based on a machine learning model, and an optimization algorithm combined with a cross-validation method is used to optimize hyperparameters.

6. The sales forecasting method according to claim 5, characterized in that: The method further comprises: S20. Build a sales forecasting model based on a machine learning model; S21. Optimizing hyperparameters of the sales forecast model using an optimization algorithm combined with a cross-validation method to obtain an optimized sales forecast model. S22: Training the sales forecast optimization model to obtain the trained sales forecast optimization model.

7. The sales forecasting method according to claim 6, characterized in that: The machine learning model is any one of the LightGBM model, XGBoost model, CatBoost model or random forest model; The optimization algorithm is any one of a PSO algorithm, a genetic algorithm, a grid search algorithm, and a random search algorithm.

8. A method for selecting a store location, characterized in that: The method comprises: S100, using the sales forecasting method according to any one of claims 1 to 7 to forecast sales of several candidate locations for store construction; S200. Based on the prediction results, select the candidate site with the highest sales forecast value from a number of candidate sites as the final site to guide the site selection of the store to be built.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the sales forecasting method according to any one of claims 1 to 7 or the store location selection method according to claim 8 is implemented.

10. A computer-readable storage medium having computer-executable instructions stored thereon, characterized in that: The computer executable instructions are executed by a computer processor to implement the sales forecasting method according to any one of claims 1 to 7, or the store site selection method according to claim 8.

Images