Iterative product sales prediction method based on big data

Abstract

The invention relates to the technical field of product sales prediction, and particularly discloses an iterative product sales prediction method based on big data, and the method comprises the steps: calculating a correlation degree coefficient through correlation analysis based on the obtained iterative product sales data and factor data influencing the sales, and judging the influence degree of each factor on the sales; building a first sales prediction model based on the correlation result and the sales data to obtain a first prediction result, and building a second sales prediction model according to the sales data to obtain a second prediction result; fusing the two to obtain a third prediction result; according to the method, by comprehensively considering various influence factors and sales data and combining a model fusion and optimization mechanism, the accuracy and adaptability of sales prediction are remarkably improved, an enterprise is helped to carry out sales prediction according to the accurate sales prediction result, and the sales prediction optimization index is substituted into the third prediction result for optimization. Product inventory management is optimized, and operation efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of product sales forecasting, and in particular to an iterative product sales forecasting method based on big data. Background Art

[0002] Sales forecasting is a method of comprehensively analyzing sales data to predict future sales. Effective sales forecasting can help companies manage cash flow, long- and short-term financing, production and sales staffing, equipment purchases, and inventory costs, and can inform rational investment and financing plans, as well as inventory and production schedules. Effective sales forecasting is one way companies build competitive advantage.

[0003] In today's fiercely competitive and rapidly changing market, a company's ability to accurately grasp sales trends for its iterative products plays a crucial role in its survival and development. With the rapid development of information technology and data collection methods, massive amounts of data are constantly emerging. How to efficiently mine and utilize the value contained in this data to assist in sales forecasting for iterative products has become a core issue for companies.

[0004] Traditional iterative product sales forecasting methods primarily rely on historical sales data, using single forecasting methods such as time series analysis, regression analysis, or simple statistical methods to infer future sales. These methods typically analyze past data patterns and trends to predict future sales, but often fail to fully consider the complexity and dynamic nature of external factors. While traditional iterative product sales forecasting methods can provide a certain degree of sales data forecasting, they still suffer from numerous issues, including insufficient forecast accuracy, poor adaptability to market changes, and limited ability to handle outliers and missing data. These issues limit the effectiveness of traditional methods in complex market environments.

[0005] The traditional sales forecasting method for iterative products has the following problems: on the one hand, sales forecasts are overly influenced by subjective factors and fail to consider the correlation between the various influencing factors of iterative products and the sales volume of iterative products; on the other hand, a single forecasting method has been adopted for a long time, and the forecasting system has not been adjusted according to changes in the sales environment. In addition, the traditional sales forecasting method for iterative products does not comprehensively analyze the factors affecting sales or overemphasizes the impact of certain factors, and cannot optimize and adjust the forecast results in a timely manner, making the sales forecast for iterative products not accurate enough. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an iterative product sales forecasting method based on big data to solve the problems raised in the above-mentioned background technology.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an iterative product sales forecasting method based on big data, comprising the following steps:

[0008] Step S1: Obtain sales data of the iterative product and data on factors influencing product sales, calculate the correlation coefficient by analyzing the correlation between each influencing factor of the iterative product and the sales volume of the iterative product, and determine the degree of influence of each influencing factor on the sales of the iterative product;

[0009] Step S2: establishing a first sales forecasting model based on the correlation analysis results in step S1 and the sales data of the iterative product, and calculating a first forecasting result of the iterative product sales data;

[0010] Step S3: establishing a second sales forecasting model based on the sales data of the iterative product, and calculating a second forecasting result of the sales data of the iterative product;

[0011] Step S4: performing a fusion calculation based on the first prediction result of the iterative product sales data and the second prediction result of the iterative product sales data to obtain a third prediction result of the iterative product sales data;

[0012] Step S5: Optimize the third prediction result of the iterative product sales data, calculate the sales prediction optimization index through the optimization model, substitute the sales prediction optimization index into the third prediction result of the iterative product sales data in step S4, and obtain the optimized third prediction result of the iterative product sales data.

[0013] Preferably, in step S1, the specific content of analyzing the correlation between each impact factor of the iterative product and the sales volume of the iterative product to calculate the correlation coefficient is:

[0014] Step S11: Construct a data set based on the sales data of the iterative products. The sales data set of each iterative product is: X i (n) = {x i (1),x i (2),…,x i (n)}, where i represents the number of the iterated product, i = 1, 2, ... n, and n represents the total number of samples of the iterated product sales data, that is, there are a total of n sample sales data records, each record corresponding to a specific sales volume value;

[0015] The dataset is constructed based on the obtained influencing factor data affecting iterative product sales. The datasets of each influencing factor are: Y j (m) = {y j (1),y j (2),…,y j(m)}, where j represents the number of each impact factor, j = 1, 2, ... m, and m represents the total number of impact factors;

[0016] Step S12: Standardize the sales data of the iterative product and the data of the various factors affecting the product sales. The standardized data set is still marked as X. i (n) and Y j (m):

[0017]

[0018] in, Expressed as the average value of product sales data in the i-th iteration, Expressed as the average value of the jth impact factor;

[0019] Step S13: Calculate the correlation coefficient REC between the sales data of the iterative product and the data of various influencing factors affecting the product sales. The calculation model is as follows:

[0020] Among them, X i (n) represents the sales data set of each iterative product after normalization, Y j (m) represents the dataset of each impact factor after normalization.

[0021] Preferably, in step S1, the specific content of determining the degree of influence of each influencing factor on the sales of the iterative product is:

[0022] The extracted correlation coefficient is compared and analyzed with the preset three level thresholds. If the correlation coefficient is less than or equal to the first level threshold, the j-th influencing factor is judged to be "not relevant" to the iterative product sales; if the correlation coefficient is between the first level threshold and the second level threshold, the j-th influencing factor is judged to be "weakly correlated" to the iterative product sales; if the correlation coefficient is between the second level threshold and the third level threshold, the j-th influencing factor is judged to be "strongly correlated" to the iterative product sales.

[0023] Preferably, in step S2, the specific steps of calculating and obtaining the first prediction result of iterative product sales data are:

[0024] Step S21: Based on the correlation analysis results in step S1, k influencing factors that are correlated with the sales of the iterative product are screened from the m influencing factor data sets. The correlation includes strong correlation or weak correlation, where k≤m. The screened influencing factor data set is:

[0025] Y jp (k) = {y j1 (k),y j2(k),…,y jk (k)}, where p represents the number of the influencing factors screened out that are correlated with the iterative product sales data, j represents the number of each influencing factor, p = 1, 2, ... k, k represents the total number of influencing factors screened out that are correlated with the iterative product sales, and m represents the total number of influencing factors;

[0026] Step S22: Construct a matrix Y that selects k influencing factor datasets that are correlated with the sales of iterative products from the m influencing factor datasets. The matrix Y has a dimension of n×k, with n sales data samples and k influencing factors. The specific form is as follows:

[0027]

[0028] Construct a matrix Xi of the sales data set of each iterative product, with a dimension of n×1;

[0029] Right now:

[0030] Step S23: Establish a first sales forecast model and calculate the first forecast result X of each iteration product sales data i预1 , the specific mathematical model is as follows:

[0031] X i预1 =(Y T ×Y) -1 ×Y T ×Xi×Y+ε, where X i预1 Represented as the first prediction result of each iteration product sales data, Y T It is represented as the transposed matrix of matrix Y, Xi is represented as the matrix of sales data set of each iterative product, Y is represented as the matrix of k influencing factor data sets that are correlated with iterative product sales screened from m influencing factor data sets, and ε is represented as the error factor.

[0032] Preferably, in step S3, the specific steps of calculating and obtaining the second prediction result of the iterative product sales data are:

[0033] Step S31: Arrange the iterative product sales data obtained in step S1 in chronological order to reconstruct the time series iterative product sales dataset U:

[0034] U={u1,u2,u3,…,u t ,…,u T}, where t represents the number of the time point, t = 1, 2, ... T, T represents the total number of time points, u1, u2, u3, ..., u t ,…,u T They are respectively represented as the iterative product sales data values corresponding to each time point;

[0035] Step S32: Calculate the average value F of each iteration product sales data i , the calculation model is as follows:

[0036] Among them, x i (1) is represented by the sales data value of the first sample of the iterative product corresponding to the i-th iterative product, x i (2) is represented by the sales data value of the iterative product corresponding to the second sample of the iterative product i, x i (n) represents the iterative product sales data value of the nth sample corresponding to the iterative product i, and n represents the total number of samples of iterative product sales data;

[0037] Step S33: Calculate and obtain the second prediction result X of each iteration product sales data i预2 , the calculation model is as follows:

[0038] Among them, X i预2 It represents the second prediction result of each iteration product sales data, u t Represented as the iterative product sales volume recorded at the t-th time point.

[0039] Preferably, in step S4, the calculation model for calculating the third prediction result of iterative product sales data is specifically:

[0040] X i预 =w1×X i预1 +w2×X i预2 , where X i预 It represents the third prediction result of each iteration product sales data, X i预1 Represented as the first prediction result of each iteration product sales data, X i预2 It represents the second prediction result of each iteration product sales data, w1 and w2 represent the weights of the first prediction result of each iteration product sales data and the second prediction result of each iteration product sales data, respectively, and w1+w2=1.

[0041] Preferably, in step S5, the steps for calculating the sales forecast optimization index are:

[0042] Step S51: Calculate the mean relative error MRE of the third prediction result of each iterative product i , the calculation model is as follows:

[0043] Among them, t represents the number of the time point, T represents the total number of time points, X i预 It represents the third prediction result of each iteration product sales data, Y tIt represents the actual sales data of the product at the t-th time point;

[0044] Step S52: Calculate the sales forecast optimization index OI of each iteration product i , the calculation model is as follows:

[0045] Among them, OI i It is represented by the sales forecast optimization index of each iteration product, MRE i It is expressed as the average relative error of the third prediction result of each iterative product;

[0046] Step S53: Substitute the sales forecast optimization index into the third forecast result of the iterative product sales data in step S4 to obtain the optimized third forecast result of the iterative product sales data. Mark the optimized third forecast result of each iterative product sales data as: X′ i预 , which is the final sales forecast value of each iterative product, and the specific expression is: X′ i预 =X i预 ×OI i .

[0047] Technical effects and advantages of the present invention:

[0048] 1. The present invention is based on obtaining iterative product sales data and data on factors affecting sales, calculating the correlation coefficient through correlation analysis, and judging the degree of influence of each factor on sales; then, a first sales forecasting model is constructed based on the correlation results and the sales data to obtain a first forecast result, and a second sales forecasting model is established based on the sales data to obtain a second forecast result; then, the two are integrated to obtain a third forecast result; then, a sales forecast optimization index is calculated through an optimization model, and the index is substituted into the third forecast result for optimization. By comprehensively considering multiple influencing factors and sales data, and combining model fusion and optimization mechanisms, the present invention significantly improves the accuracy and adaptability of sales forecasts, which helps enterprises optimize product inventory management and improve operational efficiency based on accurate sales forecast results;

[0049] 2. The present invention not only considers historical sales data but also comprehensively analyzes various factors that affect the sales of iterative products. By calculating the correlation coefficient, it determines the degree of influence of each influencing factor on the sales of iterative products. On this basis, two sales forecasting models are established: one based on the correlation analysis results and the sales data of iterative products, and the other based solely on the sales data of iterative products. By integrating the prediction results of these two models, the present invention obtains more accurate iterative product sales forecast data. Finally, the prediction results are optimized through the optimization model, further improving the accuracy of the prediction.

[0050] 3. By comprehensively considering the sales data of iterative products and data on multiple influencing factors, the present invention can more comprehensively analyze the driving factors of sales changes, thereby significantly improving the accuracy of sales forecasts. Compared with traditional methods that only rely on single predictions of historical sales data, the method of the present invention is more adaptable to the complexity and dynamics of the market environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The present invention is further described with reference to the accompanying drawings. However, the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without inventive effort.

[0052] Figure 1 It is a schematic diagram of the structural flow of the method of the present invention. DETAILED DESCRIPTION

[0053] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0054] See also Figure 1 As shown, the present invention provides an iterative product sales forecasting method based on big data, comprising the following steps:

[0055] Step S1: Obtain sales data of the iterative product and data on factors influencing product sales, calculate the correlation coefficient by analyzing the correlation between each influencing factor of the iterative product and the sales volume of the iterative product, and determine the degree of influence of each influencing factor on the sales of the iterative product;

[0056] In this embodiment, it is important to note that the sales data for the iterative product is a comprehensive dataset reflecting the product's sales performance. It covers multiple key indicators to comprehensively assess the product's market performance. Specifically, this data includes sales volume (the total number of products sold within a certain period); sales revenue (the total revenue generated by product sales); and sales growth rate, which measures the growth rate and trend of product sales. This data not only reflects the current sales status of the product but also provides important reference for future sales strategies.

[0057] The data on the various factors influencing iterative product sales is complex and diverse. Market demand is a key factor directly impacting product sales, determining the potential market size of the product. The competitive landscape reflects the competition among similar products in the market and is crucial for formulating sales strategies. Product performance, pricing strategy, and consumer feedback are also important factors influencing product sales. Product performance determines the product's appeal and competitiveness, pricing strategy directly influences consumer purchasing intention, and consumer feedback provides companies with valuable market information and guidance for improvement.

[0058] In this embodiment, it should be specifically explained that in step S1, the specific content of analyzing the correlation between each impact factor of the iterative product and the sales volume of the iterative product and calculating the correlation coefficient is:

[0059] Step S11: Construct a data set based on the sales data of the iterative products. The sales data set of each iterative product is: X i (n) = {x i (1),x i (2),…,x i (n)}, where i represents the number of the iterated product, i = 1, 2, ... n, and n represents the total number of samples of the iterated product sales data, that is, there are a total of n sample sales data records, each record corresponds to a specific sales volume value, for example, x1(1) is the sales volume of the first sample, x2(2) is the sales volume of the second sample, and so on;

[0060] The dataset is constructed based on the obtained influencing factor data affecting iterative product sales. The datasets of each influencing factor are: Y j (m) = {y j (1),y j (2),…,y j (m)}, where j represents the number of each impact factor, j = 1, 2, ... m, m represents the total number of impact factors, y j (1),y j (2),…,y j (m) are respectively represented as subsets in the iterative product impact factor dataset;

[0061] Step S12: Standardize the sales data of the iterative product and the data of the various factors affecting the product sales. The standardized data set is still marked as X. i (n) and Y j (m):

[0062]

[0063]

[0064] in, Expressed as the average value of product sales data in the i-th iteration, Expressed as the average value of the jth impact factor;

[0065] Step S13: Calculate the correlation coefficient REC between the sales data of the iterative product and the data of various influencing factors affecting the product sales. The calculation model is as follows:

[0066] Among them, X i (n) represents the sales data set of each iterative product after normalization, Y j (m) represents the dataset of each impact factor after normalization.

[0067] In this embodiment, it should be specifically explained that |·| represents the number of elements in the set, |X i (n)∪Y j (m)|-|X i (n)∩Y j (m)| represents the number of elements in the set consisting of the elements belonging to the sales data set of each iterative product or the data set of each influencing factor, but not belonging to both sets at the same time, that is, the number of elements in the set obtained by removing the elements in the intersection from the union;

[0068] and It is expressed as the ratio between the number of elements in the intersection of the two sets and the number of elements that belong to only one of the sets (not the intersection). It is used to measure the relative size of the common elements in the context of the elements in the different parts of the sets. If the ratio is large, it means that the common elements in the sales data set of each iterative product and the data set of each impact factor account for a large proportion of the non-overlapping elements in the sets, that is, the correlation between the two sets is higher, otherwise the proportion is smaller.

[0069] It is expressed as the multiple relationship between the number of elements in the combined sales data set of each iterative product and the number of elements in the influencing factor data set relative to the number of elements in the intersection. If this ratio is large, it means that the number of elements contained in the combined two sets is far greater than the number of elements they share, indicating that the two sets have relatively more differences; conversely, if the ratio is small, it means that the common elements of the two sets account for a large proportion in the whole, and the overlap between the sets is relatively high.

[0070] In this embodiment, it should be specifically explained that in step S1, the specific content of determining the degree of influence of each influencing factor on the sales of the iterative product is:

[0071] The extracted correlation coefficient is compared and analyzed with the preset three level thresholds respectively. If the correlation coefficient is less than or equal to the first level threshold, the jth influencing factor is judged to be "irrelevant" to the iterative product sales, and the jth influencing factor can be excluded from the factors affecting the model when subsequently establishing the sales data model; if the correlation coefficient is between the first level threshold and the second level threshold, the jth influencing factor is judged to be "weakly correlated" with the iterative product sales, which means that although the jth influencing factor has a certain correlation, its influence on driving the iterative product sales is limited, and it can be selectively included in subsequent analysis and modeling; if the correlation coefficient is between the second level threshold and the third level threshold, the jth influencing factor is judged to be "strongly correlated" with the iterative product sales, which clearly shows that the jth influencing factor plays a key positive role in promoting or restricting the iterative product sales. In the process of building sales forecast models, formulating marketing strategies, and making decisions on product improvement directions, it is necessary to focus on and deeply explore and utilize this factor to accurately grasp the quantitative relationship between it and sales volume changes.

[0072] Step S2: establishing a first sales forecasting model based on the correlation analysis results in step S1 and the sales data of the iterative product, and calculating a first forecast result of the iterative product sales data;

[0073] In this embodiment, it should be specifically explained that in step S2, the specific steps of calculating the first prediction result of the iterative product sales data are:

[0074] Step S21: Based on the correlation analysis results in step S1, k influencing factors that are correlated with the sales of the iterative product are screened from the m influencing factor data sets. The correlation includes strong correlation or weak correlation, where k≤m. The screened influencing factor data set is:

[0075] Y jp (k) = {y j1 (k),y j2 (k),…,y jk (k)}, where p represents the number of the influencing factors screened out that are correlated with the iterative product sales data, j represents the number of each influencing factor, p = 1, 2, ... k, k represents the total number of influencing factors screened out that are correlated with the iterative product sales, and m represents the total number of influencing factors;

[0076] Step S22: Construct a matrix Y that selects k influencing factor datasets that are correlated with the sales of iterative products from the m influencing factor datasets. The matrix Y has a dimension of n×k, with n sales data samples and k influencing factors. The specific form is as follows:

[0077]

[0078] The understanding of matrix Y can be expressed as follows: Assume that there are 3 sales data samples, that is, n = 3, and after the correlation analysis in step S1, 2 influencing factors are determined, that is, k = 2, which are used to build the model. The iterative product sales data are x1(1), x2(2), and x3(3). The data of the two influencing factors are:

[0079]

[0080] Construct a matrix Xi of the sales data set of each iterative product, with a dimension of n×1;

[0081] Right now:

[0082] Step S23: Establish a first sales forecast model and calculate the first forecast result X of each iteration product sales data i预1 , the specific mathematical model is as follows:

[0083] X i预1 =(Y T ×Y) -1 ×Y T ×Xi×Y+ε, where X i预1 Represented as the first prediction result of each iteration product sales data, Y T It is represented as the transposed matrix of matrix Y, Xi is represented as the matrix of sales data set of each iterative product, Y is represented as the matrix of k influencing factor data sets that are correlated with iterative product sales screened from m influencing factor data sets, and ε is represented as the error factor.

[0084] Step S3: establishing a second sales forecasting model based on the sales data of the iterative product, and calculating a second forecasting result of the sales data of the iterative product;

[0085] In this embodiment, it should be specifically explained that in step S3, the specific steps of calculating the second prediction result of the iterative product sales data are:

[0086] Step S31: Arrange the iterative product sales data obtained in step S1 in chronological order to reconstruct the time series iterative product sales dataset U:

[0087] U={u1,u2,u3,…,u t ,…,u T}, where t represents the number of the time point, t = 1, 2, ... T, T represents the total number of time points, u1, u2, u3, ..., u t ,…,u T They are respectively represented as the iterative product sales data values corresponding to each time point;

[0088] Step S32: Calculate the average value F of each iteration product sales data i , the calculation model is as follows:

[0089] Among them, x i (1) is represented by the sales data value of the first sample of the iterative product corresponding to the i-th iterative product, x i (2) is represented by the sales data value of the iterative product corresponding to the second sample of the iterative product i, x i (n) represents the iterative product sales data value of the nth sample corresponding to the iterative product i, and n represents the total number of samples of iterative product sales data;

[0090] Step S33: Calculate and obtain the second prediction result X of each iteration product sales data i预2 , the calculation model is as follows:

[0091] Among them, X i预2 It represents the second prediction result of each iteration product sales data, u t Represented as the iterative product sales volume recorded at the t-th time point.

[0092] Step S4: performing a fusion calculation based on the first prediction result of the iterative product sales data and the second prediction result of the iterative product sales data to obtain a third prediction result of the iterative product sales data;

[0093] In this embodiment, it should be specifically explained that in step S4, the calculation model for calculating the third prediction result of the iterative product sales data is specifically:

[0094] X i预 =w1×X i预1 +w2×X i预2 , where X i预 It represents the third prediction result of each iteration product sales data, X i预1 Represented as the first prediction result of each iteration product sales data, X i预2 It represents the second prediction result of each iteration product sales data, w1 and w2 represent the weights of the first prediction result of each iteration product sales data and the second prediction result of each iteration product sales data, respectively, and w1+w2=1.

[0095] Step S5: Optimize the third prediction result of the iterative product sales data, calculate the sales prediction optimization index through the optimization model, substitute the sales prediction optimization index into the third prediction result of the iterative product sales data in step S4, and obtain the optimized third prediction result of the iterative product sales data.

[0096] In this embodiment, it should be specifically explained that in step S5, the steps for calculating the sales forecast optimization index are as follows:

[0097] Step S51: Calculate the mean relative error MRE of the third prediction result of each iterative product i , the calculation model is as follows:

[0098] Among them, t represents the number of the time point, T represents the total number of time points, X i预 It represents the third prediction result of each iteration product sales data, Y t It represents the actual sales data of the product at the t-th time point;

[0099] Step S52: Calculate the sales forecast optimization index OI of each iteration product i , the calculation model is as follows:

[0100] Among them, OI i It is represented by the sales forecast optimization index of each iteration product, MRE i It is expressed as the average relative error of the third prediction result of each iterative product;

[0101] Step S53: Substitute the sales forecast optimization index into the third forecast result of the iterative product sales data in step S4 to obtain the optimized third forecast result of the iterative product sales data. Mark the optimized third forecast result of each iterative product sales data as: X′ i预 , which is the final sales forecast value of each iterative product, and the specific expression is: X′ i预 =X i预 ×OI i .

[0102] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0103] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. An iterative product sales forecasting method based on big data, characterized in that: The following steps are involved: Step S1: Obtain sales data of the iterative product and data on factors influencing product sales, calculate the correlation coefficient by analyzing the correlation between each influencing factor of the iterative product and the sales volume of the iterative product, and determine the degree of influence of each influencing factor on the sales of the iterative product; Step S2: establishing a first sales forecasting model based on the correlation analysis results in step S1 and the sales data of the iterative product, and calculating a first forecasting result of the iterative product sales data; Step S3: establishing a second sales forecasting model based on the sales data of the iterative product, and calculating a second forecasting result of the sales data of the iterative product; Step S4: performing a fusion calculation based on the first prediction result of the iterative product sales data and the second prediction result of the iterative product sales data to obtain a third prediction result of the iterative product sales data; Step S5: Optimize the third prediction result of the iterative product sales data, calculate the sales prediction optimization index through the optimization model, substitute the sales prediction optimization index into the third prediction result of the iterative product sales data in step S4, and obtain the optimized third prediction result of the iterative product sales data.

2. The iterative product sales forecasting method based on big data according to claim 1, characterized in that: In step S1, the specific content of analyzing the correlation between each impact factor of the iterative product and the sales volume of the iterative product and calculating the correlation coefficient is: Step S11: Construct a data set based on the sales data of the iterative products. The sales data set of each iterative product is: X i (n) = {x i (1),x i (2),…,x i (n)}, where i represents the number of the iterated product, i = 1, 2, ... n, and n represents the total number of samples of the iterated product sales data, that is, there are a total of n sample sales data records, each record corresponding to a specific sales volume value; The dataset is constructed based on the obtained influencing factor data affecting iterative product sales. The datasets of each influencing factor are: Y j (m) = {y j (1),y j (2),…,y j (m)}, where j represents the number of each impact factor, j = 1, 2, ... m, and m represents the total number of impact factors; Step S12: Standardize the sales data of the iterative product and the data of the various factors affecting the product sales. The standardized data set is still marked as X. i (n) and Y j (m): in, Expressed as the average value of product sales data in the i-th iteration, Expressed as the average value of the jth impact factor; Step S13: Calculate the correlation coefficient REC between the sales data of the iterative product and the data of various influencing factors affecting the product sales. The calculation model is as follows: Among them, X i (n) represents the sales data set of each iterative product after normalization, Y j (m) represents the dataset of each impact factor after normalization.

3. The iterative product sales forecasting method based on big data according to claim 1, characterized in that: In step S1, the specific content of determining the degree of influence of each influencing factor on the sales of the iterative product is: The extracted correlation coefficient is compared and analyzed with the preset three level thresholds. If the correlation coefficient is less than or equal to the first level threshold, the jth influencing factor is judged to be "not relevant" to the iterative product sales; if the correlation coefficient is between the first level threshold and the second level threshold, the jth influencing factor is judged to be "weakly correlated" to the iterative product sales; if the correlation coefficient is between the second level threshold and the third level threshold, the jth influencing factor is judged to be "strongly correlated" to the iterative product sales.

4. The iterative product sales forecasting method based on big data according to claim 1, characterized in that: In step S2, the specific steps of calculating and obtaining the first prediction result of the iterative product sales data are: Step S21: Based on the correlation analysis results in step S1, k influencing factors that are correlated with the sales of the iterative product are screened from the m influencing factor data sets. The correlation includes strong correlation or weak correlation, where k≤m. The screened influencing factor data set is: Y jp (k) = {y j1 (k),y j2 (k),…,y jk (k)}, where p represents the number of the influencing factors screened out that are correlated with the iterative product sales data, j represents the number of each influencing factor, p = 1, 2, ... k, k represents the total number of influencing factors screened out that are correlated with the iterative product sales, and m represents the total number of influencing factors; Step S22: Construct a matrix Y that selects k influencing factor datasets that are correlated with the sales of iterative products from the m influencing factor datasets. The matrix Y has a dimension of n×k, with n sales data samples and k influencing factors. The specific form is as follows: Construct a matrix Xi of the sales data set of each iterative product, with a dimension of n×1; Right now: Step S23: Establish a first sales forecast model and calculate the first forecast result X of each iteration product sales data i预1 , the specific mathematical model is as follows: X i预1 =(Y T ×Y) -1 ×Y T ×Xi×Y+ε, where X i预1 Represented as the first prediction result of each iteration product sales data, Y T It is represented as the transposed matrix of matrix Y, Xi is represented as the matrix of sales data set of each iterative product, Y is represented as the matrix of k influencing factor data sets that are correlated with iterative product sales screened from m influencing factor data sets, and ε is represented as the error factor.

5. The iterative product sales forecasting method based on big data according to claim 1, characterized in that: In step S3, the specific steps of calculating and obtaining the second prediction result of the iterative product sales data are: Step S31: Arrange the iterative product sales data obtained in step S1 in chronological order to reconstruct the time series iterative product sales dataset U: U={u1,u2,u3,…,u t ,…,u T }, where t represents the number of the time point, t = 1, 2, ... T, T represents the total number of time points, u1, u2, u3, ..., u t ,…,u T They are respectively represented as the iterative product sales data values corresponding to each time point; Step S32: Calculate the average value F of each iteration product sales data i , the calculation model is as follows: Among them, x i (1) is represented by the sales data value of the first sample of the iterative product corresponding to the i-th iterative product, x i (2) is represented by the sales data value of the iterative product corresponding to the second sample of the iterative product i, x i (n) represents the iterative product sales data value of the nth sample corresponding to the iterative product i, and n represents the total number of samples of iterative product sales data; Step S33: Calculate and obtain the second prediction result X of each iteration product sales data i预2 , the calculation model is as follows: Among them, X i预2 It represents the second prediction result of each iteration product sales data, u t Represented as the iterative product sales volume recorded at the t-th time point.

6. The iterative product sales forecasting method based on big data according to claim 1, characterized in that: In step S4, the calculation model for calculating the third prediction result of the iterative product sales data is specifically: X i预 =w1×X i预1 +w2×X i预2 , where X i预 It represents the third prediction result of each iteration product sales data, X i预1 Represented as the first prediction result of each iteration product sales data, X i预2 It represents the second prediction result of each iteration product sales data, w1 and w2 represent the weights of the first prediction result of each iteration product sales data and the second prediction result of each iteration product sales data, respectively, and w1+w2=1.

7. The iterative product sales forecasting method based on big data according to claim 1, characterized in that: In step S5, the steps for calculating the sales forecast optimization index are as follows: Step S51: Calculate the mean relative error MRE of the third prediction result of each iterative product i , the calculation model is as follows: Among them, t represents the number of the time point, T represents the total number of time points, X i预 It represents the third prediction result of each iteration product sales data, Y t It represents the actual sales data of the product at the t-th time point; Step S52: Calculate the sales forecast optimization index OI of each iteration product i , the calculation model is as follows: Among them, OI i It is represented by the sales forecast optimization index of each iteration product, MRE i It is expressed as the average relative error of the third prediction result of each iterative product; Step S53: Substitute the sales forecast optimization index into the third forecast result of the iterative product sales data in step S4 to obtain the optimized third forecast result of the iterative product sales data. Mark the optimized third forecast result of each iterative product sales data as: X′ i预 , which is the final sales forecast value of each iterative product, and the specific expression is: X′ i预 =X i预 ×OI i .

Images