Supply chain demand prediction method and system based on big data analysis

Abstract

The invention relates to the technical field of supply chain demand prediction, and discloses a supply chain demand prediction method and system based on big data analysis. According to the method and the system, multi-dimensional factors such as market environment, consumer behaviors, external economic indicators and promotion information of products are fully considered, so that the demand change trend of the supply chain is reflected more comprehensively, a large amount of real-time data is cleaned and integrated, repeated and wrong information is eliminated, data vacancy is filled, and the real-time performance of the system is improved. Unification of data formats is achieved, a solid foundation is provided for subsequent analysis, actual sales data and prediction results can be compared in real time through a dynamic feedback mechanism, parameters of a prediction formula are automatically adjusted, continuous optimization of the prediction results is ensured, and the prediction efficiency is improved. The innovative method effectively overcomes the defects of the traditional demand prediction technology in the aspects of accuracy and adaptability, and provides more scientific and flexible decision support for supply chain management.

Description

Technical Field

[0001] The present invention relates to the technical field of supply chain demand forecasting, and specifically to a supply chain demand forecasting method and system based on big data analysis. Background Art

[0002] The supply chain refers to the entire process and network from the procurement of raw materials, production and manufacturing, inventory management, logistics distribution to the delivery of final products to consumers. Its main function is to ensure that products meet market demand at the appropriate time, place and quantity through reasonable resource allocation and management, thereby improving the company's operational efficiency and competitiveness. Effective supply chain management can not only reduce costs and improve service levels, but also promote cooperation between enterprises, suppliers and customers, and form a good business ecosystem.

[0003] Traditional supply chain demand forecasting technology only relies on historical sales data for forecasting during application, ignoring changes in market environment, consumer behavior and external economic factors, resulting in reduced accuracy of supply chain demand forecasting results. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the deficiencies in the prior art, the present invention provides a supply chain demand forecasting method and system based on big data analysis. The method fully considers multi-dimensional factors such as market environment, consumer behavior, external economic indicators, and product promotion information, so as to more comprehensively reflect the changing trend of supply chain demand. By cleaning and integrating a large amount of real-time data, it not only eliminates duplicate and erroneous information and fills data gaps, but also realizes the unification of data formats, providing a solid foundation for subsequent analysis. Moreover, through a dynamic feedback mechanism, it can compare actual sales data with forecast results in real time, automatically adjust the parameters of the forecast formula, and ensure continuous optimization of the forecast results. This innovative method effectively solves the shortcomings of traditional demand forecasting technology in terms of accuracy and adaptability, provides more scientific and flexible decision-making support for supply chain management, and solves the above-mentioned problems.

[0006] (II) Technical solution

[0007] To achieve the above object, the present invention provides the following technical solution: a supply chain demand forecasting method based on big data analysis, comprising the following steps:

[0008] S1. Use big data technology to capture product historical sales data, product promotion information, product inventory data, product economic market indicators, and customer feedback rating data on products;

[0009] S2: Clean and denoise the historical sales data, product promotion information, product inventory data, product economic market indicators, and customer feedback rating data captured in S1, remove duplicate and erroneous data, fill in missing data, and unify all data formats;

[0010] S3, based on the cleaned and denoised data in S2, calculate the product's historical sales volume, product seasonal adjustment index, product promotion impact factor, product inventory turnover rate, product external market index, and customer feedback index for the product, and comprehensively calculate the supply chain demand forecast value;

[0011] S4. Analyze the supply chain demand forecast value, generate future demand forecasts based on the analysis results, and provide guidance for supply chain management;

[0012] S5. Based on the future demand forecast generated in S4, a dynamic feedback mechanism is constructed to compare and analyze the actual sales data with the future demand forecast, automatically adjust the parameters and characteristics of the forecast formula according to the deviation, and generate a forecast report and store it in the demand forecast database for retrieval and reference.

[0013] Preferably, the formula for data cleaning and denoising is as follows:

[0014]

[0015] In the formula, MA t represents the moving average of the data at time t, X t-i Represents the original data point at the ti-th moment, n represents the window size, that is, the number of recent data points included when calculating the average, and i represents the index subscript.

[0016] Preferably, the calculation formula for the historical sales volume of the product is as follows:

[0017]

[0018] In the formula, Xczl represents the historical sales volume of the product, Sp k represents the sales quantity in time period k, and m represents the total number of time periods.

[0019] Preferably, the product seasonal adjustment index calculation formula is as follows:

[0020]

[0021] In the formula, Jtzs represents the seasonal adjustment index of the product. represents the average sales volume in each season, and Ztxs represents the product sales volume in all time periods.

[0022] Preferably, the calculation formula of the product promotion impact factor is as follows:

[0023]

[0024] In the formula, Cxyz represents the product promotion impact factor, Cxsl represents the total sales volume during the promotion period, and Fcxl represents the average sales volume during the non-promotion period.

[0025] Preferably, the product inventory turnover rate calculation formula is as follows:

[0026]

[0027] In the formula, Kczl represents the product inventory turnover rate, Zscb represents the total sales cost of the product, Qc represents the product's beginning inventory, and Qm represents the product's ending inventory.

[0028] Preferably, the calculation formula of the product external market index is as follows:

[0029]

[0030] In the formula, Jzgj represents the product external market index, MS i represents the market share of the ith competitor in the market, n represents the number of competitors in the market, and D represents the degree of product differentiation.

[0031] Preferably, the customer feedback index calculation formula for the product is as follows:

[0032]

[0033] In the formula, Khfk represents the customer feedback index for the product, Khhp represents the total number of positive comments from customers on the product, Khcp represents the total number of negative comments from customers on the product, and Khzp represents the total number of comments from customers on the product.

[0034] Preferably, the supply chain demand forecast value calculation formula is as follows:

[0035] Gyxc=p1*Xszl+p2*Jtzs+p3*Cxyz+p4*Kczl+p5*Jzgj+p6*Khfk

[0036] In the formula, Gyxc represents the forecast value of supply chain demand, p1, p2, p3, p4, p5, and p6 represent the weights of each calculation indicator, which are determined based on historical data analysis.

[0037] A supply chain demand forecasting system based on big data analysis, including a data capture module, a data cleaning module, a data analysis module, a supply chain demand forecasting module, and a dynamic feedback and report generation module;

[0038] The data capture module captures the product's historical sales data, product promotion information, product inventory data, product economic market indicators, and customer feedback rating data on the product through big data technology and sends them to the data cleaning module;

[0039] The data cleaning module cleans and denoises the data captured by the data capture module, removes duplicate and erroneous data, fills in missing data and unifies all data formats;

[0040] The data analysis module calculates the product historical sales volume, product seasonal adjustment index, product promotion impact factor, product inventory turnover rate, product external market index and customer feedback index of the product based on the cleaned and denoised data, and comprehensively calculates the supply chain demand forecast value and sends it to the supply chain demand forecast module;

[0041] The supply chain demand forecasting module analyzes the supply chain demand forecast value, generates a forecast of future demand based on the analysis results, and provides guidance for supply chain management;

[0042] The dynamic feedback and report generation module constructs a dynamic feedback mechanism based on the generated future demand forecast results, compares and analyzes the actual sales data with the demand forecast results, automatically adjusts the parameters and characteristics of the forecast formula according to the deviation, and generates a forecast report stored in the demand forecast database for retrieval and reference.

[0043] Compared with the prior art, the present invention provides a supply chain demand forecasting method and system based on big data analysis, which has the following beneficial effects:

[0044] The present invention fully considers multi-dimensional factors such as market environment, consumer behavior, external economic indicators and product promotion information, so as to more comprehensively reflect the changing trend of supply chain demand. By cleaning and integrating a large amount of real-time data, it not only eliminates duplicate and erroneous information and fills data gaps, but also realizes the unification of data format, providing a solid foundation for subsequent analysis. Moreover, through the dynamic feedback mechanism, it can compare actual sales data with forecast results in real time, automatically adjust the parameters of the forecast formula, and ensure the continuous optimization of the forecast results. This innovative method effectively solves the shortcomings of traditional demand forecasting technology in terms of accuracy and adaptability, and provides more scientific and flexible decision-making support for supply chain management. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the steps of the method of the present invention;

[0046] Figure 2 It is a schematic diagram of the system flow of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] In view of the problem that traditional supply chain demand forecasting technology only relies on historical sales data for forecasting during application, ignoring the changes in market environment, consumer behavior and external economic factors, which reduces the accuracy of supply chain demand forecasting results, a supply chain demand forecasting method based on big data analysis is proposed. Figure 1 , the method comprises the following steps:

[0049] S1. Use big data technology to capture product historical sales data, product promotion information, product inventory data, product economic market indicators, and customer feedback rating data on products;

[0050] Through big data technology, we can comprehensively capture and analyze the historical sales data, promotion information, inventory data, economic market indicators and customer feedback rating data of products. We can also use web crawler technology to capture the sales history and promotion information of products from e-commerce platforms and social media in real time. We can also connect with the inventory management system through API interface to obtain real-time inventory data, so as to analyze the market supply and demand of products.

[0051] S2: Clean and denoise the historical sales data, product promotion information, product inventory data, product economic market indicators, and customer feedback rating data captured in S1, remove duplicate and erroneous data, fill in missing data, and unify all data formats;

[0052] The Pandas library is used to perform preliminary data processing. The drop_duplicates() function is used to remove duplicate data to ensure the uniqueness of each record. Data validation rules and anomaly detection algorithms (such as Z-score) are used to identify and delete erroneous data to improve data quality. In terms of dealing with missing values, interpolation is used to fill in the missing values ​​according to the nature of the data to ensure data integrity and avoid bias in the analysis results. In order to unify the data format, data from different sources are standardized, including conversion of date formats, consistency of numerical units (such as converting all currencies to the same unit), and encoding of categorical variables (such as using one-hot encoding or label encoding). The data denoising formula is used Denoising the data can build a high-quality, structured database, laying a solid foundation for subsequent data analysis;

[0053] S3, based on the cleaned and denoised data in S2, calculate the product's historical sales volume, product seasonal adjustment index, product promotion impact factor, product inventory turnover rate, product external market index, and customer feedback index for the product, and comprehensively calculate the supply chain demand forecast value;

[0054] The calculation formula for the product's historical sales volume is as follows:

[0055]

[0056] Historical sales data can reveal the long-term sales trend of products and help companies identify sales changes of products in different time periods (such as quarters, months, weeks, etc.). This trend analysis helps predict future sales and thus formulate more effective production and inventory strategies. In the formula, Xszl represents the historical sales volume of the product, Sp k represents the sales quantity in time period k, and m represents the total number of time periods. Accurate historical sales data enables enterprises to better manage inventory and reduce the risk of oversupply and shortage. By predicting future demand based on historical data, enterprises can control holding costs and respond quickly to market changes.

[0057] The product seasonal adjustment index is calculated as follows:

[0058]

[0059] Seasonal adjustment enables companies to remove the interference of cyclical factors, thereby improving the accuracy of sales trend forecasts and making forecast results more targeted and credible. In the formula, Jtzs represents the product seasonal adjustment index. It represents the average sales volume in each season, and Ztxs represents the product sales volume in all time periods. By calculating the seasonal adjustment index, companies can identify seasonal changes in product demand, which is especially important for products with seasonal fluctuations (such as clothing, holiday goods, etc.), and can help companies better arrange production and inventory;

[0060] The calculation formula of product promotion impact factor is as follows:

[0061]

[0062] Calculating the promotion impact factor can help companies evaluate the actual effects of different promotional activities and analyze which promotional methods can effectively increase sales. This analysis helps companies optimize and adjust promotional strategies in future marketing activities. In the formula, Cxyz represents the product promotion impact factor, Cxsl represents the total sales volume during the promotion period, and Fcxl represents the average sales volume during the non-promotion period. By quantitatively analyzing the effectiveness of promotional activities, companies can better allocate marketing budgets and resources to ensure that every investment can generate the maximum return.

[0063] The formula for calculating product inventory turnover rate is as follows:

[0064]

[0065] Inventory turnover rate provides an intuitive assessment of inventory management efficiency. By calculating the frequency of inventory turnover, companies can quantify the effectiveness of their inventory management and formulate improvement measures. In the formula, Kczl represents the product inventory turnover rate, Zscb represents the total sales cost of the product, Qc represents the product's beginning inventory, and Qm represents the product's ending inventory. Understanding inventory turnover can also help companies quickly respond to changes in market demand, improve overall operational flexibility and response speed, and ensure that products are supplied to the market in a timely manner.

[0066] The calculation formula of the product external market index is as follows:

[0067]

[0068] External market indicators can provide companies with macroeconomic environment analysis and help them understand the health of the market and potential changes in demand. In the formula, Jzgj represents the product external market indicator, MS i represents the market share of the ith competitor in the market, n represents the number of competitors in the market, and D represents the degree of product differentiation. Understanding external market indicators can also help companies identify potential risks, conduct effective risk assessment and monitoring, and reduce operating risks;

[0069] The formula for calculating the customer feedback index for a product is as follows:

[0070]

[0071] The customer feedback index is an important indicator for evaluating customer satisfaction. It can help companies understand customers' opinions on products and changes in demand in a timely manner, so as to adjust products and services. In the formula, Khfk represents the customer feedback index for products, Khhp represents the total number of positive customer comments on products, Khcp represents the total number of negative customer comments on products, and Khzp represents the total number of customer comments on products. By regularly analyzing feedback data, companies can identify the weak links of product services and make targeted improvements or innovations to adapt to market demand.

[0072] The supply chain demand forecast value calculation formula is as follows:

[0073] Gyxc=p1*Xszl+p2*Jtzs+p3*Cxyz+p4*Kczl+p5*Jzgj+p6*Khfk

[0074] The demand forecast value calculated by combining the above data can more accurately reflect the real market demand, reduce forecast errors, and enable enterprises to reasonably arrange production and procurement to avoid waste of resources. In the formula, Gyxc represents the supply chain demand forecast value, and p1, p2, p3, p4, p5, and p6 represent the weights of each calculation indicator, which are determined based on historical data analysis. Accurate demand forecasting can improve the coordination of various links in the supply chain. All links from procurement, production to sales can make flexible adjustments according to demand, thereby maximizing operational efficiency.

[0075] S4. Analyze the supply chain demand forecast value, generate future demand forecasts based on the analysis results, and provide guidance for supply chain management;

[0076] During the analysis process, key information is extracted through data mining technology, including seasonal characteristics of demand and effects of promotional activities, to identify the main drivers of demand, and customers are segmented using cluster analysis to better understand the demand characteristics of different customer groups;

[0077] After generating future demand forecasts, we combine real-time market data and forecast results, adopt dynamic adjustment models, and timely update demand forecasts to further improve the accuracy and reliability of forecasts. This process provides valuable guidance to help supply chain management teams develop strategies to adapt to market fluctuations, such as reasonably arranging production plans and optimizing inventory levels, thereby reducing costs and improving service levels to ensure that companies maintain their competitive advantage. Such technical means can not only enhance supply chain responsiveness, but also better meet customer needs and achieve sustainable business development.

[0078] S5. Based on the future demand forecast generated in S4, a dynamic feedback mechanism is constructed to compare and analyze the actual sales data with the future demand forecast, automatically adjust the parameters and characteristics of the forecast formula according to the deviation, and generate a forecast report and store it in the demand forecast database for retrieval and reference;

[0079] Adaptive linear regression methods are used to conduct in-depth analysis of deviations, identify key factors affecting sales and quantify their impact. Based on these analysis results, the feedback mechanism can automatically adjust the parameters and feature settings of the forecasting model to optimize the accuracy of demand forecasting. The system can dynamically update key parameters such as seasonal adjustment index and promotion impact factor to reflect the latest market changes and sales trends;

[0080] In terms of report generation, automated report generation tools can be used to regularly generate detailed demand forecast reports, including forecast accuracy assessments, inferred key drivers, and recommended adjustment strategies. These reports will be stored in the demand forecast database for easy review and retrieval by different departments, which will help cross-departmental collaboration and decision support. Through this intelligent dynamic feedback mechanism, companies can improve the responsiveness and flexibility of the supply chain and further enhance their market competitiveness.

[0081] See also Figure 2 , a supply chain demand forecasting system based on big data analysis, including a data capture module, a data cleaning module, a data analysis module, a supply chain demand forecasting module, and a dynamic feedback and report generation module;

[0082] The data capture module uses big data technology to capture the product's historical sales data, product promotion information, product inventory data, product economic market indicators, and customer feedback rating data on the product and sends it to the data cleaning module;

[0083] The data cleaning module cleans and denoises the data captured by the data capture module, removes duplicate and erroneous data, fills in missing data and unifies all data formats;

[0084] The data analysis module calculates the product's historical sales volume, product seasonal adjustment index, product promotion impact factor, product inventory turnover rate, product external market index, and customer feedback index based on the cleaned and denoised data, and comprehensively calculates the supply chain demand forecast value and sends it to the supply chain demand forecast module;

[0085] The supply chain demand forecasting module analyzes the supply chain demand forecast value, generates a forecast of future demand based on the analysis results, and provides guidance for supply chain management;

[0086] The dynamic feedback and report generation module builds a dynamic feedback mechanism based on the generated future demand forecast results, compares and analyzes the actual sales data with the demand forecast results, automatically adjusts the parameters and characteristics of the forecast formula according to the deviation, and generates a forecast report stored in the demand forecast database for retrieval and reference.

[0087] By combining the above methods and systems, the shortcomings of traditional demand forecasting technology in terms of accuracy and adaptability are effectively solved, providing more scientific and flexible decision support for supply chain management.

[0088] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A supply chain demand forecasting method based on big data analysis, characterized in that: The following steps are involved: S1. Use big data technology to capture product historical sales data, product promotion information, product inventory data, product economic market indicators, and customer feedback rating data on products; S2: Clean and denoise the historical sales data, product promotion information, product inventory data, product economic market indicators, and customer feedback rating data captured in S1, remove duplicate and erroneous data, fill in missing data, and unify all data formats; S3, based on the cleaned and denoised data in S2, calculate the product's historical sales volume, product seasonal adjustment index, product promotion impact factor, product inventory turnover rate, product external market index, and customer feedback index for the product, and comprehensively calculate the supply chain demand forecast value; S4. Analyze the supply chain demand forecast value, generate future demand forecasts based on the analysis results, and provide guidance for supply chain management; S5. Based on the future demand forecast generated in S4, a dynamic feedback mechanism is constructed to compare and analyze the actual sales data with the future demand forecast, automatically adjust the parameters and characteristics of the forecast formula according to the deviation, and generate a forecast report and store it in the demand forecast database for retrieval and reference.

2. The supply chain demand forecasting method based on big data analysis according to claim 1 is characterized by: The formula for data cleaning and denoising is as follows: In the formula, MA t represents the moving average of the data at time t, X t-i Represents the original data point at the ti-th moment, n represents the window size, that is, the number of recent data points included when calculating the average, and i represents the index subscript.

3. The supply chain demand forecasting method based on big data analysis according to claim 2 is characterized by: The calculation formula for the historical sales volume of the product is as follows: In the formula, Xszl represents the historical sales volume of the product, Sp k represents the sales quantity in time period k, and m represents the total number of time periods.

4. The supply chain demand forecasting method based on big data analysis according to claim 3 is characterized by: The calculation formula of the product seasonal adjustment index is as follows: In the formula, Jtzs represents the seasonal adjustment index of the product. represents the average sales volume in each season, and Ztxs represents the product sales volume in all time periods.

5. The supply chain demand forecasting method based on big data analysis according to claim 4 is characterized by: The calculation formula of the product promotion impact factor is as follows: In the formula, Cxyz represents the product promotion impact factor, Cxsl represents the total sales volume during the promotion period, and Fcxl represents the average sales volume during the non-promotion period.

6. The supply chain demand forecasting method based on big data analysis according to claim 5 is characterized by: The product inventory turnover rate calculation formula is as follows: In the formula, Kczl represents the product inventory turnover rate, Zscb represents the total sales cost of the product, Qc represents the product's beginning inventory, and Qm represents the product's ending inventory.

7. The supply chain demand forecasting method based on big data analysis according to claim 6 is characterized by: The calculation formula of the product external market index is as follows: In the formula, Jzgj represents the product external market index, MS i represents the market share of the ith competitor in the market, n represents the number of competitors in the market, and D represents the degree of product differentiation.

8. The supply chain demand forecasting method based on big data analysis according to claim 7 is characterized by: The customer feedback index calculation formula for the product is as follows: In the formula, Khfk represents the customer feedback index for the product, Khhp represents the total number of positive comments from customers on the product, Khcp represents the total number of negative comments from customers on the product, and Khzp represents the total number of comments from customers on the product.

9. The supply chain demand forecasting method based on big data analysis according to claim 8 is characterized by: The supply chain demand forecast value calculation formula is as follows: Gyxc=p1*Xszl+p2*Jtzs+p3*Cxyz+p4*Kczl+p5*Jzgj+p6*Khfk In the formula, Gyxc represents the forecast value of supply chain demand, p1, p2, p3, p4, p5, and p6 represent the weights of each calculation indicator, which are determined based on historical data analysis.

10. A supply chain demand forecasting system based on big data analysis, characterized by: It includes data capture module, data cleaning module, data analysis module, supply chain demand forecasting module and dynamic feedback and report generation module; The data capture module captures the product's historical sales data, product promotion information, product inventory data, product economic market indicators, and customer feedback rating data on the product through big data technology and sends them to the data cleaning module; The data cleaning module cleans and denoises the data captured by the data capture module, removes duplicate and erroneous data, fills in missing data and unifies all data formats; The data analysis module calculates the product historical sales volume, product seasonal adjustment index, product promotion impact factor, product inventory turnover rate, product external market index and customer feedback index of the product based on the cleaned and denoised data, and comprehensively calculates the supply chain demand forecast value and sends it to the supply chain demand forecast module; The supply chain demand forecasting module analyzes the supply chain demand forecast value, generates a forecast of future demand based on the analysis results, and provides guidance for supply chain management; The dynamic feedback and report generation module constructs a dynamic feedback mechanism based on the generated future demand forecast results, compares and analyzes the actual sales data with the demand forecast results, automatically adjusts the parameters and characteristics of the forecast formula according to the deviation, and generates a forecast report stored in the demand forecast database for retrieval and reference.

Images