A business intelligence reporting and analysis system

Abstract

The present invention belongs to the field of business analysis technology, and specifically is a business intelligence report and analysis system, including a data acquisition module, a data analysis module, an evaluation module and a decision support module; the data acquisition module is used to extract and organize front-end product information, and also includes collecting information on products to be compared related to the front-end products; the front-end product information is used to generate a front-end product marketing report; the data analysis module is used to analyze the influence relationship between the front-end product and the products to be compared, including calculating the correlation R c And according to the correlation R c Determine related products; through correlation calculation, calculate the correlation of the products to be compared that may be related to the front-end products in turn, obtain a quantitative correlation supported by long-term data, and then use the lower limit value of the first threshold interval to dynamically adjust the remaining products to be compared to perform a second screening, so as to determine the related products.

Description

Technical Field

[0001] The present invention belongs to the technical field of business analysis, and in particular to a business intelligence report and analysis system. Background Art

[0002] In the process of making product reports, companies need to measure whether there is a correlation between products and use this to develop marketing strategies to obtain greater benefits. It is understandable that, taking mobile phones as an example, mobile phones are regarded as front-end products, and correspondingly, headphones are related products. When purchasing a new mobile phone, new headphones may be purchased at the same time. Therefore, there is a certain correlation between headphones and mobile phones. It can also be said that the sales of mobile phones promote the sales of headphones. Based on this, companies need to quantify the correlation between the two products and develop marketing strategies to obtain greater benefits.

[0003] Based on the above, in the prior art, such as CN115577991A, a business intelligence data analysis system and analysis method based on big data is disclosed, by determining any front-end product, and then using the correlation between the front-end product and the product to be compared, the related product is determined;

[0004] However, since the calculated correlation is quantitative data supported by long-term data, it cannot reflect the real correlation between the front-end products and the products to be compared. It is foreseeable that with the changes in the marketing strategy of the enterprise, such as increasing the promotion intensity or increasing the promotion time, the sales data of different products will change, and the overall changes may be similar to the sales data of the front-end products. When the formula is used to calculate the similarity, a large number of related products will exist in the front-end products, which will be a great obstacle for enterprises to analyze the front-end products and related products and formulate marketing strategies based on them.

[0005] To this end, the present invention provides a business intelligence reporting and analysis system. Summary of the invention

[0006] In order to make up for the deficiencies of the prior art, at least one technical problem raised in the background technology is solved.

[0007] The technical solution adopted by the present invention to solve the technical problem is: a business intelligence reporting and analysis system described in the present invention comprises:

[0008] Data acquisition module: used to extract and organize front-end commodity information, and also includes collecting commodity information to be compared related to the front-end commodities; the front-end commodity information is used to generate front-end commodity marketing reports;

[0009] Data analysis module: used to analyze the impact relationship between the front-end products and the products to be compared, including calculating the correlation R c And according to the correlation R cIdentify related products;

[0010] The evaluation module calculates the evaluation value V and deviation value S of the related products based on the sales data of the front-end products and related products, and generates an analysis report;

[0011] A decision support module is used to compare the obtained deviation value S with the threshold value to determine whether the deviation value S is within a first threshold interval. If the deviation value S is within the first threshold interval, a decision support is provided for the marketing strategy according to the evaluation value V and the deviation value S; otherwise, it is ignored.

[0012] According to the correlation R c The method for determining related products is:

[0013] Calculate the correlation R between the product to be compared and the related products based on historical sales data c ;

[0014] Then, based on the first threshold range, irrelevant products to be compared are screened out;

[0015] Segment the marketing cycle corresponding to the historical sales data to obtain multiple evaluation cycles;

[0016] Dynamically adjust the lower limit of the first threshold interval based on multiple evaluation cycles;

[0017] Secondary screening is performed on the products to be compared that are suspected to be related in multiple evaluation cycles, and the products to be compared that meet the first condition are used as related products of the front-end products.

[0018] Preferably, the correlation R c The calculation method is:

[0019] Obtain historical sales data of front-end products and products to be compared;

[0020] Preprocess historical sales data, including cleaning, denoising, and filling missing values;

[0021] Divide the processed historical sales data into a training set and a test set;

[0022] According to the formula, the correlation R between the front-end product and the product to be compared is calculated using the training set. c ;

[0023]

[0024] Among them, R c is the correlation between product x and product y, product x is the front-end product, product y is the product to be compared, b represents the actual sales data of the front-end products in period b, represents the average marketing quantity of front-end products; yb Indicates the actual sales data of the expected comparison product in week b. Indicates the average marketing quantity of the product to be compared;

[0025] Set the first threshold interval [a,1] and calculate the correlation R c For comparison:

[0026] When the correlation R c If it is within the first threshold interval [a,1], it is considered that the product to be compared is suspected to be related to the front-end product;

[0027] Otherwise, it is determined that the product to be compared has no relevance to the front-end product;

[0028] Eliminate the unrelated products to be compared and retain the products to be compared that are suspected to be related.

[0029] Preferably, the method for obtaining the first threshold interval [a, 1] is:

[0030] Get the correlation R between multiple items to be compared and the front-end items c ;

[0031] Correlation R c Sort and take the median R med as the initial lower limit of the first threshold interval [a,1];

[0032] For the first threshold interval [R med ,1] The performance of the application on the test set is evaluated;

[0033] The evaluating includes verifying the accuracy of the first threshold interval on a test set.

[0034] Preferably, the method for dynamically adjusting the lower limit of the first threshold interval based on multiple evaluation cycles is:

[0035] The marketing cycles corresponding to the historical data are equally divided to obtain multiple evaluation cycles i;

[0036] Calculate the correlation between multiple products to be compared and the front-end products in each evaluation cycle i And use this to calculate the average correlation between all the products to be compared and the front-end products

[0037] According to the formula:

[0038]

[0039] Among them, a i is the lower limit of the initial first threshold interval of evaluation cycle i, k is the multiplier, is the standard deviation of the correlation between multiple products to be compared and the front-end products;

[0040] According to the formula, the first threshold interval [a i ,1];

[0041] According to the first threshold interval [a i ,1] Filter the products to be compared, take the products to be compared that meet the first condition as related products, sort all related products, and generate a marketing report for related products;

[0042] The first condition is that the ratio of the number of times the to-be-compared product is determined as an associated product in multiple evaluation cycles i to the total number of evaluation cycles i is greater than or equal to the preset hit rate Q ro , the hit rate Q is according to the formula:

[0043]

[0044] Wherein, Q is the hit rate, g is the number of times the to-be-compared product is determined as a related product in multiple evaluation cycles i, and n is the total number of evaluation cycles i.

[0045] Preferably, the method of equidistantly dividing the evaluation period i corresponding to the historical data is:

[0046] Get the marketing cycle corresponding to historical sales data;

[0047] Establish multiple starting time nodes in the marketing cycle, and establish the same evaluation cycle i based on the multiple starting time nodes:

[0048] Starting from the second starting time node, the midpoint of the adjacent starting time nodes is determined as the ending time node of the previous starting time node;

[0049] The period between the first start time node and the first end time node is taken as an evaluation period i, and multiple evaluation periods i are obtained by extending them one by one.

[0050] Preferably, the calculation method of the evaluation value V is:

[0051] Obtain the sales data MR of the associated products in the marketing cycle, the sales data MP of the front-end products in the marketing cycle, and the correlation R between the associated products and the front-end products c ;

[0052]

[0053] Where C(p) is the promotion intensity adjustment coefficient, C(t) is the promotion time adjustment coefficient, a is the promotion percentage of the associated product, t is the promotion time, and T is the marketing cycle time;

[0054] According to the above formula, the evaluation value V of the impact of related products on the sales volume of front-end products during the marketing cycle is calculated.

[0055] Preferably, the calculation method of the deviation value S is:

[0056] Obtain the marketing cycle to be evaluated, and divide the marketing cycle to be evaluated to obtain multiple cycles j to be evaluated;

[0057] Get the associated sales data y of the associated products in each evaluation period j j , and obtain the front-end sales data x of the front-end products in each evaluation period j j

[0058]

[0059] Where m is the total number of evaluation cycles j, is the correlation between the associated products and the front-end products in each evaluation period j;

[0060] The sales data deviation value S to be evaluated of the front-end product and the related products is calculated based on the formula.

[0061] Preferably, the decision support module further includes a marketing strategy decision unit;

[0062] The marketing strategy decision unit is used to formulate a marketing strategy according to the evaluation value V and the deviation value S of the associated product, and the method is as follows:

[0063] Obtaining a deviation value S between the front-end product and the associated product, comparing the deviation value S with a second threshold interval, and if the deviation value S is within the second threshold interval, the marketing strategy based on the associated product and the front-end product remains unchanged;

[0064] On the contrary, if the deviation value S is not within the second threshold range, based on the specific performance of the deviation value S, the marketing strategy based on the associated products and the front-end products is changed.

[0065] Preferably, the method for obtaining the second threshold interval is:

[0066] Based on the evaluation cycle i, the historical deviation value S corresponding to the impact relationship between the associated products and the front-end product sales data is calculated. i ;

[0067] Based on the multiple historical deviation values ​​S obtained i , calculate the average historical deviation value

[0068] Set the second threshold interval is multiple historical deviation values ​​S i The standard deviation of

[0069] The validity of the second threshold interval is verified, and the second threshold interval is determined according to the validity verification result.

[0070] Preferably, the decision support module further includes an inventory management unit; the inventory management unit is used to predict inventory demand according to the evaluation value V of the associated commodities and perform inventory management.

[0071] The beneficial effects of the present invention are as follows:

[0072] 1. A business intelligence report and analysis system described in the present invention calculates the correlation of the products to be compared that may be correlated with the front-end products in turn through correlation calculation, obtains a quantitative correlation supported by long-term data, and can screen the products to be compared based on a preset first threshold interval to exclude products to be compared that are basically irrelevant. For the remaining products to be compared, based on the segmented marketing cycle, the lower limit of the first threshold interval is dynamically adjusted in each evaluation cycle, so as to perform a secondary screening on the remaining products to be compared, and realizes a primary screening of multiple products to be compared from the correlation of the entire marketing cycle, and then uses the lower limit of the first threshold interval to dynamically adjust, so as to perform a secondary screening on the remaining products to be compared, so as to determine the related products;

[0073] 2. A business intelligence reporting and analysis system described in the present invention can calculate the influence relationship based on the sales data to be evaluated of the related products and the front-end products, and then subtract it from the evaluation value V calculated above to obtain a deviation value S. The deviation value S calculated here can be used to characterize the difference between the expected performance and the actual performance of the sales influence relationship between the related products and the front-end products. Based on the calculation of the deviation value S, the difference between the expected performance and the actual performance of the sales influence relationship between the related products and the front-end products can be quantified. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] The present invention will be further described below in conjunction with the accompanying drawings.

[0075] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0076] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0077] like Figure 1As shown, a business intelligence reporting and analysis system according to an embodiment of the present invention includes a data acquisition module, a data analysis module, an evaluation module and a decision support module; the data acquisition module is used to extract and organize front-end commodity information, and also includes collecting information on commodities to be compared related to the front-end commodities; the front-end commodity information is used to generate a front-end commodity marketing report; the data analysis module is used to analyze the influence relationship between the front-end commodity and the commodities to be compared, including calculating the correlation R c And according to the correlation R c Determine the associated products; the evaluation module calculates the evaluation value V and the deviation value S of the associated products based on the sales data of the front-end products and the associated products, and generates an analysis report; the decision support module is used to compare the obtained deviation value S with the threshold value, and determine whether the deviation value S is within the first threshold value interval. If the deviation value S is within the first threshold value interval, then provide decision support for the marketing strategy according to the evaluation value V and the deviation value S, otherwise, ignore it;

[0078] According to the correlation R c The method for determining related products is:

[0079] Calculate the correlation R between the product to be compared and the related products based on historical sales data c ;

[0080] Then, based on the first threshold range, irrelevant products to be compared are screened out;

[0081] Segment the marketing cycle corresponding to the historical sales data to obtain multiple evaluation cycles;

[0082] Dynamically adjust the lower limit of the first threshold interval based on multiple evaluation cycles;

[0083] Secondary screening is performed on the products to be compared that are suspected to be related in multiple evaluation cycles, and the products to be compared that meet the first condition are used as related products of the front-end products.

[0084] In one embodiment of the present invention, based on the correlation calculation, the correlation calculation is performed on the products to be compared that may be correlated with the front-end products in turn to obtain a quantitative correlation supported by long-term data. Based on the preset first threshold interval, the products to be compared can be screened once to exclude the products to be compared that are basically irrelevant. For the remaining products to be compared, based on the segmented marketing cycle, the lower limit of the first threshold interval is dynamically adjusted in each evaluation cycle, so as to perform a secondary screening on the remaining products to be compared. Since multiple evaluation cycles are set, it cannot be guaranteed that the correlation of all products to be compared in each evaluation cycle is within the first threshold interval. Therefore, it is also necessary to set a first condition so that the products to be compared that meet the first condition are retained and determined as the related products of the front-end products. Based on the above operation, using the correlation calculation formula, a primary screening of multiple products to be compared can be achieved from the correlation of the entire marketing cycle. Then, by dynamically adjusting the lower limit of the first threshold interval, the remaining products to be compared can be screened twice to determine the related products.

[0085] It is worth noting that due to promotional activities, the sales of some commodities may surge, thus creating the illusion of increased correlation with front-end commodities. However, promotional activities do not occupy a long period of the entire marketing cycle, so the illusion of increased correlation can only last for a period of time. If it is not taken into account, it may be regarded as a related commodity, thereby increasing the workload of the company's business analysis and even hindering the formulation of the company's next stage of marketing strategy. In addition, due to the different contents of promotional activities, when the company carries out a full-reduction activity, it may cause the sales volume of some of the compared commodities to converge with the front-end commodities, thus also creating the illusion of increased correlation. Taking mobile phones as an example, there are generally mobile phones, Headphones, notebooks, smart devices, etc., among which, in the historical records, the sales data of mobile phones and cameras, due to the launch of camera promotion activities, the sales volume of cameras increased, which in turn led to an increase in the correlation between cameras and mobile phones. Under normal circumstances, the correlation between products such as cameras and mobile phones is relatively low. Therefore, the correlation between cameras and mobile phones may only be high in a short period of time, and for the entire marketing cycle, the correlation between cameras and mobile phones is relatively low. By setting the first threshold interval, it is possible to screen out products such as cameras that have a higher correlation with front-end products due to promotional activities and other factors, thereby providing valuable information for corporate business analysis.

[0086] In one embodiment, the correlation R c The calculation method is:

[0087] Obtain historical sales data of front-end products and products to be compared;

[0088] Preprocess historical sales data, including cleaning, denoising, and filling missing values;

[0089] Divide the processed historical sales data into a training set and a test set;

[0090] According to the formula, the correlation R between the front-end product and the product to be compared is calculated using the training set. c ;

[0091]

[0092] Among them, R c is the correlation between product x and product y, product x is the front-end product, product y is the product to be compared, b represents the actual sales data of the front-end products in period b, represents the average marketing quantity of front-end products; y b Indicates the actual sales data of the expected comparison product in week b. Indicates the average marketing quantity of the product to be compared;

[0093] Set the first threshold interval [a,1] and calculate the correlation R c For comparison:

[0094] When the correlation R is within the first threshold interval [a, 1], it is considered that the product to be compared is suspected to be related to the front-end product;

[0095] Otherwise, it is determined that the product to be compared has no relevance to the front-end product;

[0096] Eliminate the unrelated products to be compared and retain the products to be compared that are suspected to be related.

[0097] By calculating the correlation between the front-end products and the products to be compared, it can be determined whether there is an influence relationship between the sales data of the front-end products and the sales data of the products to be compared. By calculating the correlation, it is helpful for enterprises to understand the sales influence relationship between the products, and thus formulate the next stage of marketing strategy based on this, and provide decision support for the enterprise. However, the sales data itself does not clearly show which products to be compared have correlation with the front-end products. Therefore, it is necessary to quantify this feature by calculating the correlation. By comparing the calculated correlation with the set first threshold interval, the products to be compared that have no correlation can be screened out, and the products to be compared that are suspected to have correlation can be retained. Based on this, the data calculation of the enterprise during business analysis can be reduced to ensure that the enterprise is effective in a dynamic market environment. Filter out the products to be compared that are truly related to the front-end products, so as to better support enterprises in making business decisions; it is foreseeable that, based on the above formula, the distribution of sales data of each period is used to judge the convergence of sales data of the front-end products and the products to be compared in each period, which can be used to characterize the correlation between the front-end products and the products to be compared. Based on the above, taking mobile phones as an example, when mobile phones are used as front-end products, if the products to be compared include cameras and headphones, the cameras and mobile phones are not very related at the sales level, that is, when users choose to buy a new mobile phone, they will not consider purchasing a camera at the same time. Therefore, products to be compared such as cameras will be screened out after one screening, while headphones will be retained and used as products to be compared that are suspected to be related.

[0098] In one embodiment, the method for obtaining the first threshold interval [a, 1] is:

[0099] Get the correlation R between multiple items to be compared and the front-end items c ;

[0100] Correlation R c Sort and take the median R med as the initial lower limit of the first threshold interval [a,1];

[0101] For the first threshold interval [R med ,1] The performance of the application on the test set is evaluated;

[0102] The evaluating includes verifying the accuracy of the first threshold interval on a test set.

[0103] After calculating the correlation between multiple products to be compared and the front-end products, in order to obtain related products that are truly related to the front-end products, it is necessary to set a first threshold interval to screen multiple products to be compared, reduce the amount of calculation, and effectively screen out the products to be compared that are truly related to the front-end products in a dynamic market environment for enterprises. Based on the above, the setting of the first threshold interval is particularly critical. Generally speaking, according to the above correlation calculation formula, the obtained correlation will be between [-1,1]. However, considering that the enterprise needs to obtain products to be compared that are truly related, the lower limit of the first threshold interval generally needs to be set according to the actual situation. In one embodiment of the present invention, in the absence of data support, the lower limit of the first threshold interval is generally calculated by calculating the correlation R between all products to be compared and the front-end products. c , and then sort all the obtained correlations and obtain the median R med , the median R med As the initial lower limit of the first threshold interval, it can realize one-time screening of multiple products to be compared, thereby helping enterprises to quickly screen out irrelevant products to be compared and reduce the amount of calculation;

[0104] It can be foreseen that if the lower limit of the first threshold interval is set too high, no products to be compared may be retained after comparison, that is, no related products of the front-end products, which may cause misjudgment in the formulation of corporate marketing strategies. On the contrary, if the lower limit of the first threshold interval is set too low, a large number of products to be compared will be retained, and there will be too many products to be compared that are suspected to be related to the front-end products. For the formulation of marketing strategies by enterprises, the complexity of analysis and the computational burden are increased, and it is difficult to extract useful data. Based on this, the median R med As the initial lower limit of the first threshold interval, it can effectively screen out a part of the products to be compared. For the remaining products to be compared, secondary screening can be performed by dynamically adjusting the lower limit of the first threshold interval, thereby ensuring that related products that are valuable for analyzing the front-end product sales data can be screened out, providing valuable information for enterprises to formulate marketing strategies.

[0105] In one embodiment, the method for dynamically adjusting the lower limit of the first threshold interval based on multiple evaluation cycles is:

[0106] The marketing cycles corresponding to the historical data are equally divided to obtain multiple evaluation cycles i;

[0107] Calculate the correlation between multiple products to be compared and the front-end products in each evaluation cycle i And use this to calculate the average correlation between all the products to be compared and the front-end products

[0108] According to the formula:

[0109]

[0110] Among them, a i is the lower limit of the initial first threshold interval of evaluation cycle i, k is the multiplier, is the standard deviation of the correlation between multiple products to be compared and the front-end products;

[0111] According to the formula, the first threshold interval [a i ,1];

[0112] According to the first threshold interval [a i ,1] Filter the products to be compared, take the products to be compared that meet the first condition as related products, sort all related products, and generate a marketing report for related products;

[0113] The first condition is that the ratio of the number of times the to-be-compared product is determined as an associated product in multiple evaluation cycles i to the total number of evaluation cycles i is greater than or equal to the preset hit rate Q ro , the hit rate Q is according to the formula:

[0114]

[0115] Wherein, Q is the hit rate, g is the number of times the to-be-compared product is determined as a related product in multiple evaluation cycles i, and n is the total number of evaluation cycles i.

[0116] After a screening of multiple products to be compared, the remaining products to be compared that are suspected to be associated need to be screened again. Based on the above, by segmenting the marketing cycle, the first threshold intervals corresponding to multiple evaluation cycles are dynamically adjusted. If the marketing cycle is quarterly, the evaluation cycle i here is different from the cycle in the above correlation calculation process. It can be understood that if the above cycle is monthly, the evaluation cycle i here can be weekly, or a fixed time period, such as 10 days, that is, the evaluation cycle i here is smaller than the cycle set for the above correlation calculation. That is to say, in the above calculation process, the correlation obtained is the quarterly correlation, and in this embodiment, the above cycle can be further segmented by the evaluation cycle obtained by segmentation, for example, the correlation of each month is calculated, and then the calculated correlation of each month is compared with the first threshold interval; by comparing the correlation of the product to be compared corresponding to the evaluation cycle i with the dynamic first threshold interval, it can be further determined whether the product to be compared is an associated product of the front-end product;

[0117] It is worth noting that, according to the correlation of multiple products to be compared in the evaluation cycle i, the average correlation is calculated, and the dynamic lower limit of the first threshold interval is obtained by using the average correlation and the correlation standard deviation. Based on the dynamic lower limit, the remaining products to be compared can be screened for the second time. In order to ensure that there is a real correlation between the products to be compared and the front-end products and avoid the problem of high short-term correlation due to factors such as promotional activities, a first condition can be set to measure the real correlation between the products to be compared and the front-end products. It can be understood that in the multiple evaluation cycles i after segmentation, if any product to be compared can be retained after screening according to the dynamically adjusted first threshold interval in multiple evaluation cycles i, it means that there is a real correlation between the product to be compared and the related products, that is, the product to be compared is a related product. However, in the actual calculation process, it cannot be guaranteed that all products to be compared can be retained in multiple evaluation cycles. Therefore, the products to be compared only need to meet the first condition in multiple evaluation cycles, that is, they are considered to be related products of the front-end products. In one embodiment of the present invention, the preset hit rate Q ro Set to 0.75. Taking 4 evaluation cycles as an example, if any product to be compared is retained 3 times after screening in 4 evaluation cycles, it means that the product to be compared has a strong correlation with the front-end product. Based on this, the preset hit rate Q is set manually. ro 3 / 4=0.75; through the above method, the illusion of increased correlation caused by the surge in sales of individual products due to factors such as promotional activities can be further avoided, helping companies to identify such situations and determine the related products that have a real correlation with the front-end products.

[0118] In one embodiment, the method of equidistantly dividing the evaluation period i corresponding to the historical data is:

[0119] Get the marketing cycle corresponding to historical sales data;

[0120] Establish multiple starting time nodes in the marketing cycle, and establish the same evaluation cycle i based on the multiple starting time nodes:

[0121] Starting from the second starting time node, the midpoint of the adjacent starting time nodes is determined as the ending time node of the previous starting time node;

[0122] The period between the first start time node and the first end time node is taken as an evaluation period i, and multiple evaluation periods i are obtained by extending them one by one.

[0123] In one embodiment, the calculation method of the evaluation value V is:

[0124] Obtain the sales data MR of the associated products in the marketing cycle, the sales data MP of the front-end products in the marketing cycle, and the correlation R between the associated products and the front-end products c;

[0125]

[0126] Where C(p) is the promotion intensity adjustment coefficient, C(t) is the promotion time adjustment coefficient, a is the promotion percentage of the associated product, t is the promotion time, and T is the marketing cycle time;

[0127] According to the above formula, the evaluation value V of the impact of related products on the sales volume of front-end products during the marketing cycle is calculated.

[0128] After determining the associated products of the front-end products, the correlation between the associated products and the front-end products is obtained. Since the correlation can only reflect the sales data impact relationship between the front-end products and the associated products, and cannot quantify the size of the sales data impact relationship, it is also necessary to quantify the sales data impact relationship between the front-end products and the associated products based on the correlation. In one embodiment of the present invention, the above-mentioned impact relationship is quantified based on the evaluation value. When the evaluation value V is greater than 1, it means that the sales data of the associated products is higher than the sales data of the front-end products. Based on this, it can also be characterized that the sales trend of the front-end products is saturated. Therefore, the enterprise can make marketing strategies according to the evaluation value V, such as reducing the inventory of the front-end products and replenishing the inventory of the associated products. In addition, in the prior art, only the sales data MR within the marketing cycle, the sales data MP of the front-end products within the marketing cycle and the correlation R between the associated products and the front-end products are used to determine the sales data of the associated products. c To quantify the evaluation value V, without considering the impact of promotional activities and the proportion of promotional time on the sales data of related products, the evaluation value V of the related products may be higher than the actual sales impact, which in turn affects the company's decision on marketing strategies. Based on this, by adding the promotion intensity adjustment coefficient and the promotion time adjustment coefficient, the quantification of the sales data evaluation value of the related products for the front-end products can be optimized, thereby helping companies to more accurately analyze the sales impact relationship between the related products and the front-end products.

[0129] In one embodiment, the calculation method of the deviation value S is:

[0130] Obtain the marketing cycle to be evaluated, and divide the marketing cycle to be evaluated to obtain multiple cycles j to be evaluated;

[0131] Get the associated sales data y of the associated products in each evaluation period j j , and obtain the front-end sales data x of the front-end products in each evaluation period j j

[0132]

[0133] Where m is the total number of evaluation cycles j, is the correlation between the associated products and the front-end products in each evaluation period S;

[0134] The sales data deviation value S to be evaluated of the front-end product and the related products is calculated based on the formula.

[0135] The evaluation value V calculated above is calculated based on the sales data of the associated goods and the front-end goods in the marketing cycle, and can be used to characterize the quantitative data of the sales impact relationship of the associated goods on the front-end goods. Then, based on the obtained evaluation value j, it can also be used to calculate the deviation of the to-be-evaluated sales data of the associated goods and the front-end goods. Based on the above, in one embodiment of the present invention, according to the to-be-evaluated sales data of the associated goods and the front-end goods, the impact relationship based on the to-be-evaluated sales data can be calculated, and then the deviation value S can be obtained by subtracting the evaluation value V calculated above. The deviation value S calculated here can be used to characterize the difference between the expected performance and the actual performance of the sales impact relationship between the associated goods and the front-end goods. It can be understood that: the evaluation value calculated based on the marketing cycle above is the quantitative data of the sales data impact relationship between the front-end goods and the associated goods in the marketing cycle, which can be used to characterize the expected performance of the sales data impact relationship between the associated goods and the front-end goods. Therefore, based on the calculation of the above deviation value S, the difference between the expected performance and the actual performance of the sales impact relationship between the associated goods and the front-end goods can be quantified;

[0136] It is worth noting that a positive value of the deviation S means that the evaluation value V exceeds the average value calculated based on historical data and correlation, indicating that the sales performance of the associated products is lower than expected; a negative value means that the sales performance of the associated products is higher than expected; the larger the absolute value of the deviation, the more obvious the deviation in sales performance.

[0137] In one embodiment, the decision support module further includes a marketing strategy decision unit;

[0138] The marketing strategy decision unit is used to formulate a marketing strategy according to the evaluation value V and the deviation value S of the associated product, and the method is as follows:

[0139] Obtaining a deviation value S between the front-end product and the associated product, comparing the deviation value S with a second threshold interval, and if the deviation value S is within the second threshold interval, the marketing strategy based on the associated product and the front-end product remains unchanged;

[0140] On the contrary, if the deviation value S is not within the second threshold range, based on the specific performance of the deviation value S, the marketing strategy based on the associated products and the front-end products is changed.

[0141] After obtaining the deviation value S between the associated commodity and the front-end commodity calculated based on the sales data to be evaluated, it can be used to evaluate whether the sales impact relationship between the associated commodity and the front-end commodity is in line with expectations, and make a decision based on this. In one embodiment of the present invention, the obtained deviation value S is compared with the set second threshold interval. When the deviation value S is within the second threshold interval, it means that the sales data impact relationship between the associated commodity and the front-end commodity is in a normal range, and the sales data impact relationship between the associated commodity and the front-end commodity is in line with expectations; when the deviation value S is not within the second threshold interval, it means that the sales data impact relationship between the associated commodity and the front-end commodity is not in line with expectations. At this time, the enterprise needs to adjust the marketing strategy of the front-end commodity and the associated commodity. The specific adjustment direction includes reducing inventory and replenishing inventory and updating promotional activities. Based on the above analysis, it can help enterprises quantify the sales performance of associated commodities and front-end commodities according to the sales data impact relationship between the associated commodities and the front-end commodities, so as to formulate a marketing strategy that is in line with the current time period.

[0142] In one embodiment, the method for obtaining the second threshold interval is:

[0143] Based on the evaluation cycle i, the historical deviation value S corresponding to the impact relationship between the associated products and the front-end product sales data is calculated. i ;

[0144] Based on the multiple historical deviation values ​​S obtained i , calculate the average historical deviation value

[0145] Set the second threshold interval is multiple historical deviation values ​​S i The standard deviation of

[0146] The validity of the second threshold interval is verified, and the second threshold interval is determined according to the validity verification result.

[0147] In one embodiment, the decision support module further includes an inventory management unit; the inventory management unit is used to predict inventory demand according to the evaluation value V of the associated commodities and perform inventory management.

[0148] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A business intelligence reporting and analysis system, characterized in that: include: Data acquisition module: used to extract and organize front-end product information, and also includes collecting information on products to be compared related to the front-end products; The front-end commodity information is used to generate a front-end commodity marketing report; Data analysis module: used to analyze the impact relationship between the front-end products and the products to be compared, including calculating the correlation R c And according to the correlation R c Identify related products; The evaluation module calculates the evaluation value V and deviation value S of the related products based on the sales data of the front-end products and related products, and generates an analysis report; A decision support module is used to compare the obtained deviation value S with the threshold value to determine whether the deviation value S is within a first threshold interval. If the deviation value S is within the first threshold interval, a decision support is provided for the marketing strategy according to the evaluation value V and the deviation value S; otherwise, it is ignored. According to the correlation R c The method for determining related products is: Calculate the correlation R between the product to be compared and the related products based on historical sales data c ; Then, based on the first threshold range, irrelevant products to be compared are screened out; Segment the marketing cycle corresponding to the historical sales data to obtain multiple evaluation cycles; Dynamically adjust the lower limit of the first threshold interval based on multiple evaluation cycles; Secondary screening is performed on the products to be compared that are suspected to be related in multiple evaluation cycles, and the products to be compared that meet the first condition are used as related products of the front-end products.

2. A business intelligence reporting and analysis system according to claim 1, characterized in that: The correlation R c The calculation method is: Obtain historical sales data of front-end products and products to be compared; Preprocess historical sales data, including cleaning, denoising, and filling missing values; Divide the processed historical sales data into a training set and a test set; According to the formula, the correlation R between the front-end product and the product to be compared is calculated using the training set. c ; Among them, R c is the correlation between product x and product y, product x is the front-end product, product y is the product to be compared, b represents the actual sales data of the front-end products in period b, represents the average marketing quantity of front-end products; y b Indicates the actual sales data of the expected comparison product in week b. Indicates the average marketing quantity of the product to be compared; Set the first threshold interval [a,1] and calculate the correlation R c For comparison: When the correlation R c If it is within the first threshold interval [a,1], it is considered that the product to be compared is suspected to be related to the front-end product; Otherwise, it is determined that the product to be compared has no relevance to the front-end product; Eliminate the unrelated products to be compared and retain the products to be compared that are suspected to be related.

3. A business intelligence reporting and analysis system according to claim 2, characterized in that: The method for obtaining the first threshold interval [a, 1] is: Get the correlation R between multiple items to be compared and the front-end items c ; Correlation R c Sort and take the median R med as the initial lower limit of the first threshold interval [a,1]; For the first threshold interval [R med ,1] The performance of the application on the test set is evaluated; The evaluating includes verifying the accuracy of the first threshold interval on a test set.

4. A business intelligence reporting and analysis system according to claim 1, characterized in that: The method for dynamically adjusting the lower limit of the first threshold interval based on multiple evaluation cycles is: The marketing cycles corresponding to the historical data are equally divided to obtain multiple evaluation cycles i; Calculate the correlation between multiple products to be compared and the front-end products in each evaluation cycle i And use this to calculate the average correlation between all the products to be compared and the front-end products According to the formula: Among them, a i is the lower limit of the initial first threshold interval of evaluation cycle i, k is the multiplier, is the standard deviation of the correlation between multiple products to be compared and the front-end products; According to the formula, the first threshold interval [a i ,1]; According to the first threshold interval [a i ,1] Filter the products to be compared, take the products to be compared that meet the first condition as related products, sort all related products, and generate a marketing report for related products; The first condition is that the ratio of the number of times the to-be-compared product is determined as an associated product in multiple evaluation cycles i to the total number of evaluation cycles i is greater than or equal to the preset hit rate Q ro , the hit rate Q is according to the formula: Wherein, Q is the hit rate, g is the number of times the to-be-compared product is determined as a related product in multiple evaluation cycles i, and n is the total number of evaluation cycles i.

5. A business intelligence reporting and analysis system according to claim 4, characterized in that: The method for equally dividing the evaluation period i corresponding to the historical data is: Get the marketing cycle corresponding to historical sales data; Establish multiple starting time nodes in the marketing cycle, and establish the same evaluation cycle i based on the multiple starting time nodes: Starting from the second starting time node, the midpoint of the adjacent starting time nodes is determined as the ending time node of the previous starting time node; The period between the first start time node and the first end time node is taken as an evaluation period i, and multiple evaluation periods i are obtained by extending them one by one.

6. A business intelligence reporting and analysis system according to claim 2, characterized in that: The calculation method of the evaluation value V is: Obtain the sales data MR of the associated products in the marketing cycle, the sales data MP of the front-end products in the marketing cycle, and the correlation R between the associated products and the front-end products c ; Where C(p) is the promotion intensity adjustment coefficient, C(t) is the promotion time adjustment coefficient, a is the promotion percentage of the associated product, t is the promotion time, and T is the marketing cycle time; According to the above formula, the evaluation value V of the impact of related products on the sales volume of front-end products during the marketing cycle is calculated.

7. A business intelligence reporting and analysis system according to claim 1, characterized in that: The calculation method of the deviation value S is: Obtain the marketing cycle to be evaluated, and divide the marketing cycle to be evaluated to obtain multiple cycles j to be evaluated; Get the associated sales data y of the associated products in each evaluation period j j , and obtain the front-end sales data x of the front-end products in each evaluation period j j Where m is the total number of evaluation cycles j, is the correlation between the associated products and the front-end products in each evaluation period j; The sales data deviation value S to be evaluated of the front-end product and the related products is calculated based on the formula.

8. A business intelligence reporting and analysis system according to claim 1, characterized in that: The decision support module also includes a marketing strategy decision unit; The marketing strategy decision unit is used to formulate a marketing strategy according to the evaluation value V and the deviation value S of the associated products, and the method is as follows: Obtaining a deviation value S between the front-end product and the associated product, comparing the deviation value S with a second threshold interval, and if the deviation value S is within the second threshold interval, the marketing strategy based on the associated product and the front-end product remains unchanged; On the contrary, if the deviation value S is not within the second threshold range, based on the specific performance of the deviation value S, the marketing strategy based on the associated products and the front-end products is changed.

9. A business intelligence reporting and analysis system according to claim 8, characterized in that: The method for obtaining the second threshold interval is: Based on the evaluation cycle i, the historical deviation value S corresponding to the impact relationship between the associated products and the front-end product sales data is calculated. i ; Based on the multiple historical deviation values ​​S obtained i , calculate the average historical deviation value Set the second threshold interval is multiple historical deviation values ​​S i The standard deviation of The validity of the second threshold interval is verified, and the second threshold interval is determined according to the validity verification result.

10. A business intelligence reporting and analysis system according to claim 1, characterized in that: The decision support module also includes an inventory management unit; the inventory management unit is used to predict inventory demand according to the evaluation value V of the associated commodities and perform inventory management.

Images