A new and popular product supervision and evaluation system based on data analysis

Abstract

The present invention relates to the technical field of new and internet celebrity product supervision, and specifically discloses a new and internet celebrity product supervision and evaluation system based on data analysis. The system comprises: a supervision and evaluation module, which screens the repeated product names collected by a product duplication identification unit, and screens the sales volume product names collected by a product sales statistics unit; compares the corresponding heat feature value of each Internet best-selling product with the feature value of non-human factors affecting it, and determines whether each Internet best-selling product is a new or internet celebrity product; the present invention obtains the heat feature value corresponding to each product name by analyzing the number of repeated appearances of the product name on the social media platform and the sales volume on the e-commerce platform after the screening, and then determines whether each product is an Internet best-selling product, thereby achieving timely and effective discovery of new and internet celebrity products, so as to achieve the purpose of effective supervision on them.

Description

Technical Field

[0001] The present invention relates to the technical field of new and internet celebrity product supervision, and specifically to a new and internet celebrity product supervision and evaluation system based on data analysis. Background Art

[0002] New and internet celebrity products refer to products that have recently become popular on the Internet. They are usually products that have received a lot of attention and praise from users. These products usually have unique designs, functions or creativity, which can arouse public interest and spread quickly on social media and online platforms. These products can be of various types, such as electronic devices, beauty products, clothing, household items, etc. At present, the popularity of many new and internet celebrity products is inseparable from the promotion of social media and the spread of user word-of-mouth. Some new and internet celebrity products may be short-lived trends, while others can maintain long-term popularity.

[0003] Existing new and internet celebrity products often see a sudden and huge increase in sales due to factors such as rapid popularity, a wide variety, and high public attention. Market supervision departments may not be able to discover new and internet celebrity products in a timely manner and supervise them effectively, which may lead to quality and safety accidents caused by new and internet celebrity products. Summary of the invention

[0004] The purpose of the present invention is to provide a new and popular product supervision and evaluation system based on data analysis to solve the following technical problems:

[0005] How to discover new and internet-famous products in a timely and effective manner in order to achieve the purpose of effective supervision over them.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A new, internet celebrity product supervision and evaluation system based on data analysis, the system comprising:

[0008] Product duplication identification unit, used to identify product names that appear repeatedly on social media platforms on the Internet and collect the corresponding number of repeated appearances;

[0009] Product sales statistics unit, used to collect the corresponding sales volume of product names appearing on e-commerce platforms on the Internet;

[0010] The supervision and assessment module screens the repeated product names collected by the product duplication identification unit, screens the sales product names collected by the product sales statistics unit, analyzes the number of repeated appearances of the screened product names on social media platforms and the sales on e-commerce platforms, obtains the corresponding heat feature values ​​of each product name, and determines whether each product is an Internet best-selling product; compares the corresponding heat feature values ​​of each Internet best-selling product with the feature values ​​of non-human factors that affect it, calculates the correlation coefficient between the heat feature values ​​of each Internet best-selling product and the feature values ​​of non-human factors that affect it, and then determines whether each Internet best-selling product is a new or Internet celebrity product.

[0011] Furthermore, the process of screening the repeated product names collected by the product duplication identification unit includes:

[0012]

[0013] The frequency characteristic value P of the collected product names appearing on social media is calculated by the above formula;

[0014] Where N is the number of social media collected by supervision, i∈[1,N], Q i is the number of times the product name appears collected on the i-th social media, ω i is the weight coefficient of the influence of the i-th social media on the product;

[0015] The product duplication identification unit is used to identify product names that appear repeatedly on social media and count the number of times each product name appears in real time;

[0016] The regulatory assessment module calculates the real-time statistical data of each product name to obtain the frequency characteristic value of each product name appearing on social media;

[0017] The curve of the frequency characteristic value of each product name changing with time is obtained by fitting through a linear regression algorithm, referred to as the frequency characteristic value curve;

[0018] Compare the frequency characteristic value curves of each product name with the preset frequency characteristic reference values ​​respectively;

[0019] If any point on the frequency characteristic value curve of a product name is greater than the preset frequency characteristic reference value, the product name and its frequency characteristic value curve are retained;

[0020] Otherwise, delete the product name and its frequency characteristic value curve.

[0021] Furthermore, the process of screening the sales product names collected by the product sales unit includes:

[0022]

[0023] The cumulative sales volume R of the collected product names on the e-commerce platform is calculated through the above formula;

[0024] Where M is the number of e-commerce platforms collected by supervision, j∈[1,M], D j is the cumulative sales volume of the product on the jth e-commerce platform;

[0025] The regulatory assessment module collects and counts the cumulative sales of each product name on the e-commerce platform in real time;

[0026] The curve of cumulative sales volume of each product name changing over time is obtained by linear regression algorithm fitting, referred to as cumulative sales volume curve;

[0027] Derivative the cumulative sales curve of each product name to obtain the sales speed curve of each product name over time;

[0028] Compare the sales speed change curve of each product name with the preset sales speed change reference value;

[0029] If the sales speed change curve of a product name is greater than the preset sales speed change reference value, the product name and its sales speed change curve are retained;

[0030] Otherwise, delete the product name and its sales speed change curve.

[0031] Furthermore, the process of determining whether each product is an Internet best-selling product includes:

[0032] S1. Mix and match the remaining product names after filtering out the repeated product names with the remaining product names after filtering out the sales volume product names, and take the product names and related data that are common in the two screening results;

[0033] S2. Analyze the frequency characteristic value curve and cumulative sales curve of each product name after the screening is completed to obtain the heat characteristic value corresponding to each product;

[0034] S3. Compare the heat characteristic value corresponding to each product with the preset heat characteristic critical value. If the heat characteristic value of a product is greater than the preset heat characteristic critical value, the product is a best-selling product on the Internet.

[0035] Furthermore, in step S2, the process of obtaining the heat characteristic value corresponding to each product includes:

[0036]

[0037] The heat characteristic value of each product within the preset time interval is calculated by the above formula;

[0038] Among them, there are a total of W product names screened in step S1, and the W product names are numbered in sequence, k∈[1,W], E k is the heat characteristic value of the kth product, [t1, t2] is the preset time interval, P k (t) is the frequency characteristic curve of the kth product, R k (t1) is the cumulative sales volume of the kth product at time point t1, R k (t2) is the cumulative sales volume of the kth product at time point t2, and τ1 and τ2 are the weight coefficients of the factors affecting the heat characteristic value.

[0039] Furthermore, the process of calculating the correlation coefficient between each Internet best-selling product and the characteristic value of the non-human factors affecting it includes:

[0040] It is determined that there are D best-selling products on the Internet, and each best-selling product on the Internet is numbered in sequence, y∈[1,D];

[0041] In the process of analyzing the heat characteristic values ​​of each product through the supervision and evaluation module, multiple heat characteristic values ​​E of the yth best-selling Internet product are taken in the preset time interval. y (t l );

[0042] The change function of the characteristic value of non-human factors affecting the yth product is preset as f y (t);

[0043] Substitute the intermediate values ​​of each preset time interval into the non-human factor characteristic value function f that affects the product y (t), obtain the f corresponding to the middle time nodes of multiple preset time intervals y (t l );

[0044]

[0045] The correlation coefficient r between the heat characteristic value of the yth Internet best-selling product and the characteristic value of the non-human factors affecting it is calculated by the above formula y ;

[0046] Among them, H is the number of collected product heat feature values ​​and their corresponding non-human factor feature values, l∈[1,H], E y (t l ) is the heat characteristic value of the yth best-selling Internet product in the lth preset time interval, is the average of the H heat feature values ​​of the yth best-selling Internet product, f y (t l) is the non-human factor characteristic value corresponding to the lth preset time interval of the yth Internet best-selling product, is the average value of the H non-human factor feature values ​​of the yth Internet best-selling product.

[0047] Furthermore, the process of obtaining the non-human factor characteristic change function affecting the y-th product includes:

[0048]

[0049] The non-human factor characteristic change function f that affects the yth product is calculated by the above formula y (t);

[0050] Among them, V y is the number of non-human factors affecting the yth product, x∈[1,V y ],g y·x (t) is the function value of the xth non-human factor affecting the yth product over time, θ y·x is the weight coefficient of the xth non-human factor affecting the yth product.

[0051] Furthermore, the process of determining whether each Internet best-selling product is a new or Internet celebrity product includes:

[0052] The correlation coefficient r corresponding to each Internet best-selling product y The correlation coefficient with the preset standard value r sth Make a comparison;

[0053] If | r y |≥r sth , indicating that there is a certain degree of correlation between the product and the characteristic values ​​of non-human factors that affect it, and it is not a new or internet-famous product;

[0054] If | r y | <r sth , indicating that there is no certain degree of correlation between the product and the characteristic values ​​of non-human factors that affect it, and it is a new and internet-famous product.

[0055] Beneficial effects of the present invention:

[0056] (1) The present invention uses a supervision and evaluation module to screen the collected fully occurring names and the collected sales product names, and then analyzes the number of repeated appearances of the product names on the social media platform and the sales volume on the e-commerce platform after the screening, and obtains the heat characteristic value corresponding to each product name, and then determines whether each product is an Internet best-selling product. Finally, the corresponding characteristic value of each Internet best-selling product is compared and analyzed with the characteristic value of the non-human factor affecting it, and the correlation coefficient between the heat characteristic value of each Internet best-selling product and the characteristic value of the non-human factor affecting it is calculated, and then it is determined whether each Internet best-selling product is a new or Internet celebrity product, so as to achieve timely and effective discovery of new and Internet celebrity products, so as to achieve the purpose of effective supervision on them. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] The present invention will be further described below in conjunction with the accompanying drawings.

[0058] Figure 1 It is a schematic block diagram of a new, Internet celebrity product supervision and evaluation system based on data analysis proposed by the present invention;

[0059] Figure 2 It is a flowchart of the steps of the method for determining whether each product is an Internet best-selling product proposed by the present invention. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] See also Figure 1 As shown, in one embodiment, a new, online celebrity product supervision and evaluation system based on data analysis is provided, the system comprising:

[0062] Product duplication identification unit, used to identify product names that appear repeatedly on social media platforms on the Internet and collect the corresponding number of repeated appearances;

[0063] Product sales statistics unit, used to collect the corresponding sales volume of product names appearing on e-commerce platforms on the Internet;

[0064] The supervision and assessment module screens the repeated product names collected by the product duplication identification unit, screens the sales product names collected by the product sales statistics unit, analyzes the number of repeated appearances of the screened product names on social media platforms and the sales on e-commerce platforms, obtains the corresponding heat feature values ​​of each product name, and determines whether each product is an Internet best-selling product; compares the corresponding heat feature values ​​of each Internet best-selling product with the feature values ​​of non-human factors that affect it, calculates the correlation coefficient between the heat feature values ​​of each Internet best-selling product and the feature values ​​of non-human factors that affect it, and then determines whether each Internet best-selling product is a new or Internet celebrity product.

[0065] Through the above technical scheme, this embodiment provides a new and Internet celebrity product supervision and evaluation system based on data analysis, wherein the product names that appear repeatedly on the social media platform on the Internet are identified through the product duplication identification unit and the corresponding number of repeated appearances are collected, and the corresponding sales volume of the product names that appear on the e-commerce platform on the Internet is collected through the product sales statistics unit, and the collected fully appeared names are screened through the supervision and evaluation module, and the collected sales volume product names are screened, and then the number of repeated appearances of the product names on the social media platform and the sales volume on the e-commerce platform after the screening are analyzed to obtain the corresponding heat feature value of each product name, and then determine whether each product is an Internet best-selling product, and finally compare and analyze the corresponding characteristic value of each Internet best-selling product with the characteristic value of non-human factors that affect it, calculate the correlation coefficient between the heat feature value of each Internet best-selling product and the characteristic value of non-human factors that affect it, and then determine whether each Internet best-selling product is a new or Internet celebrity product, so as to achieve timely and effective discovery of new and Internet celebrity products, so as to achieve the purpose of effective supervision on them.

[0066] In one embodiment, the process of screening the repeated product names collected by the product duplication identification unit includes:

[0067]

[0068] The frequency characteristic value P of the collected product names appearing on social media is calculated by the above formula;

[0069] Where N is the number of social media collected by the regulator. Existing social media can be counted and numbered, including but not limited to Weibo, WeChat, online media, Douyin, Xiaohongshu, Baidu Tieba, etc., i∈[1,N], Q i is the number of times the product name appears on the i-th social media, which can be obtained by identifying the product names appearing on social media and collecting their occurrence times through some web crawlers and text mining technologies, ω iis the weight coefficient of the influence of the i-th social media on the product, which is selected and preset according to the popularity of the social media;

[0070] The product duplication identification unit is used to identify product names that appear repeatedly on social media and count the number of times each product name appears in real time;

[0071] The supervision and evaluation module calculates the real-time statistical data of each product name to obtain the frequency characteristic value P of each product name appearing on social media;

[0072] The curve of the frequency characteristic value of each product name changing with time is obtained by fitting through a linear regression algorithm, referred to as the frequency characteristic value curve;

[0073] Compare the frequency characteristic value curves of each product name with preset frequency characteristic reference values, respectively, wherein the preset frequency characteristic reference values ​​can be preset based on empirical data;

[0074] If any point on the frequency characteristic value curve of a product name is greater than the preset frequency characteristic reference value, it means that the frequency characteristic value of the product name at a certain time point is greater than the preset frequency characteristic reference value, then the product name and its frequency characteristic value curve are retained;

[0075] Otherwise, delete the product name and its frequency characteristic value curve.

[0076] In one embodiment, the process of screening the sales product names collected by the product sales unit includes:

[0077]

[0078] The cumulative sales volume R of the collected product names on the e-commerce platform is calculated through the above formula;

[0079] Where M is the number of e-commerce platforms collected by the regulator. Existing e-commerce platforms can be counted and numbered, including but not limited to Taobao, Tmall, JD.com, Suning, Pinduoduo, Douyin, Meituan, etc., j∈[1,M], D j is the cumulative sales volume of the product on the jth e-commerce platform, obtained based on the sales volume data of the corresponding product on the corresponding e-commerce platform;

[0080] The regulatory assessment module collects and counts the cumulative sales of each product name on the e-commerce platform in real time;

[0081] The curve of cumulative sales volume of each product name changing over time is obtained by linear regression algorithm fitting, referred to as cumulative sales volume curve;

[0082] Derivative the cumulative sales curve of each product name to obtain the sales speed curve of each product name over time;

[0083] Compare the sales speed change curve of each product name with a preset sales speed change reference value, wherein the preset sales speed change reference value can be preset based on experience data;

[0084] If the sales speed change curve of a certain product name is greater than the preset sales speed change reference value, it means that the sales speed of the product name at a certain point in time is greater than the preset sales speed change reference value, then the product name and its sales speed change curve are retained;

[0085] Otherwise, delete the product name and its sales speed change curve.

[0086] In one embodiment, see Figure 2 As shown in FIG. 1 , the process of determining whether each product is an Internet best-selling product includes:

[0087] S1. Mix and match the remaining product names after filtering out the repeated product names with the remaining product names after filtering out the sales volume product names, and take the product names and related data that are common in the two screening results;

[0088] S2. Analyze the frequency characteristic value curve and cumulative sales curve of each product name after the screening is completed to obtain the heat characteristic value corresponding to each product;

[0089] S3. Compare the heat characteristic value corresponding to each product with the preset heat characteristic critical value. If the heat characteristic value of a product is greater than the preset heat characteristic critical value, the product is a best-selling product on the Internet.

[0090] In step S2, the process of obtaining the heat characteristic value corresponding to each product includes:

[0091]

[0092] The heat characteristic value of each product within the preset time interval is calculated by the above formula;

[0093] Among them, there are a total of W product names screened in step S1, and the W product names are numbered in sequence, k∈[1,W], E k is the heat characteristic value of the kth product, [t1, t2] is the preset time interval, P k (t) is the frequency characteristic curve of the kth product, R k (t1) is the cumulative sales volume of the kth product at time point t1, R k (t2) is the cumulative sales volume of the kth product at time point t2, and τ1 and τ2 are the weight coefficients of the factors affecting the heat characteristic value.

[0094] Through the above technical solution, this embodiment provides a method for determining whether each product is an Internet best-selling product. First, the remaining product names after filtering the repeated product names can be mixed and matched with the remaining product names after filtering the sales product names, and the product names and related data that exist in the two screening results are taken, and the common products are statistically numbered, and then the frequency characteristic value curve and the cumulative sales curve of the product names of each number are analyzed to obtain the heat characteristic value corresponding to each product, and finally the heat characteristic value corresponding to each product is compared with the preset heat characteristic critical value. If the heat characteristic value of a product is greater than the preset heat characteristic critical value, the product is an Internet best-selling product;

[0095] Among them, in the process of obtaining the corresponding heat characteristic value of each numbered product, the formula Calculated, where E k is the heat characteristic value of the kth product, [t1, t2] is the preset time interval, preset according to the monitoring frequency, P k (t) is the frequency characteristic curve of the kth product, R k (t1) is the cumulative sales volume of the kth product at time point t1, R k (t2) is the cumulative sales volume of the kth product at time point t2. τ1 and τ2 are weight coefficients of the factors affecting the heat characteristic value, which are selected based on empirical data.

[0096] In one embodiment, the process of calculating the correlation coefficient between each Internet best-selling product and the characteristic value of the non-human factors affecting it includes:

[0097] It is determined that there are D best-selling products on the Internet, and each best-selling product on the Internet is numbered in sequence, y∈[1,D];

[0098] In the process of analyzing the heat characteristic values ​​of each product through the supervision and evaluation module, multiple heat characteristic values ​​E of the yth best-selling Internet product are taken in the preset time interval. y (t l );

[0099] The change function of the characteristic value of non-human factors affecting the yth product is preset as f y (t);

[0100] Substitute the intermediate values ​​of each preset time interval into the non-human factor characteristic value function f that affects the product y (t), obtain the f corresponding to the middle time nodes of multiple preset time intervals y (t l );

[0101]

[0102] The correlation coefficient r between the heat characteristic value of the yth Internet best-selling product and the characteristic value of the non-human factors affecting it is calculated by the above formula y ;

[0103] Among them, H is the number of collected product heat feature values ​​and their corresponding non-human factor feature values, l∈[1,H], E y (t l ) is the heat characteristic value of the yth best-selling Internet product in the lth preset time interval, is the average of the H heat feature values ​​of the yth best-selling Internet product, f y (t l ) is the non-human factor characteristic value corresponding to the lth preset time interval of the yth Internet best-selling product, is the average value of the H non-human factor feature values ​​of the yth Internet best-selling product.

[0104] The process of obtaining the characteristic change function of non-human factors affecting the y-th product includes:

[0105]

[0106] The non-human factor characteristic change function f that affects the yth product is calculated by the above formula y (t);

[0107] Among them, V y is the number of non-human factors affecting the yth product, x∈[1,V y ],g y·x (t) is the function value of the xth non-human factor affecting the yth product over time, θ y·x is the weight coefficient of the xth non-human factor affecting the yth product.

[0108] The process of determining whether each Internet best-selling product is a new or popular product includes:

[0109] The correlation coefficient r corresponding to each Internet best-selling product y The correlation coefficient with the preset standard value r sth Make a comparison;

[0110] If | r y |≥r sth , indicating that there is a certain degree of correlation between the product and the characteristic values ​​of non-human factors that affect it, and it is not a new or internet-famous product;

[0111] If | r y | <r sth, indicating that there is no certain degree of correlation between the product and the characteristic values ​​of non-human factors that affect it, and it is a new and internet-famous product.

[0112] Through the above technical solution, this embodiment provides a method for determining whether an Internet best-selling product is a new or Internet celebrity product;

[0113] It is determined that there are D best-selling products on the Internet, and each best-selling product on the Internet is numbered in sequence, y∈[1,D];

[0114] By formula Calculate the non-human factor characteristic change function f that affects the yth product y (t), where V y is the number of non-human factors affecting the yth product, x∈[1,V y ],g y·x (t) is the function value of the xth non-human factor affecting the yth product over time, which can be obtained based on the non-human factor experiment on the yth product. The non-human factors include but are not limited to time, temperature, holidays, etc., θ y·x is the weight coefficient of the xth non-human factor affecting the yth product, which can be obtained by selecting the non-human factor experiment on the yth product;

[0115] In the process of analyzing the heat characteristic values ​​of each product through the supervision and evaluation module, multiple heat characteristic values ​​E of the yth best-selling Internet product are taken in the preset time interval. y (t l );

[0116] Substitute the intermediate values ​​of each preset time interval into the non-human factor characteristic value function f that affects the product y (t), obtain the f corresponding to the middle time nodes of multiple preset time intervals y (t l );

[0117]

[0118] The correlation coefficient r between the heat characteristic value of the yth Internet best-selling product and the characteristic value of the non-human factors affecting it is calculated by the above formula y ;

[0119] Among them, H is the number of the collected product heat characteristic values ​​and their corresponding non-human factor characteristic values, which are selected and obtained based on empirical data, l∈[1,H], E y (t l ) is the heat characteristic value of the yth Internet best-selling product in the lth preset time interval, obtained during the analysis of the heat characteristic values ​​of each product, is the average of the H heat feature values ​​of the yth best-selling Internet product, obtained by averaging all the heat feature values ​​of the product, f y (t l ) is the non-human factor feature value corresponding to the lth preset time interval of the yth best-selling Internet product. Substitute the corresponding preset time interval into the function f y (t) is obtained, is the average value of the H non-human factor characteristic values ​​of the yth Internet best-selling product, obtained by averaging all the corresponding non-human factor characteristic values ​​of the product;

[0120] Then the correlation coefficient r corresponding to each Internet best-selling product is y The correlation coefficient with the preset standard value r sth Compare and preset the standard value of correlation coefficient r sth It can be obtained based on experience.

[0121] If | r y |≥r sth , indicating that there is a certain degree of correlation between the product and the characteristic values ​​of non-human factors that affect it, and it is not a new or internet-famous product;

[0122] If | r y | <r sth , indicating that there is no certain degree of correlation between the product and the characteristic values ​​of non-human factors that affect it, and it is a new and popular product;

[0123] In this way, new and internet celebrity products can be discovered in a timely and effective manner through the supervision and evaluation module, so as to achieve the purpose of effectively supervising each new and internet celebrity product.

[0124] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A new and internet celebrity product supervision and evaluation system based on data analysis, characterized in that: The system comprises: Product duplication identification unit, used to identify product names that appear repeatedly on social media platforms on the Internet and collect the corresponding number of repeated appearances; Product sales statistics unit, used to collect the corresponding sales volume of product names appearing on e-commerce platforms on the Internet; The supervision and assessment module screens the repeated product names collected by the product duplication identification unit, screens the sales product names collected by the product sales statistics unit, analyzes the number of repeated appearances of the screened product names on social media platforms and the sales on e-commerce platforms, obtains the corresponding heat feature value of each product name, and determines whether each product is an Internet best-selling product; compares the corresponding heat feature value of each Internet best-selling product with the feature value of non-human factors that affect it, calculates the correlation coefficient between the heat feature value of each Internet best-selling product and the feature value of non-human factors that affect it, and then determines whether each Internet best-selling product is a new or Internet celebrity product; The process of screening the repeated product names collected by the product duplication identification unit includes: The frequency characteristic value P of the collected product names appearing on social media is calculated by the above formula; Where N is the number of social media collected by supervision, i∈[1,N], Q i is the number of times the product name appears collected on the i-th social media, ω i is the weight coefficient of the influence of the i-th social media on the product; The product duplication identification unit is used to identify product names that appear repeatedly on social media and count the number of times each product name appears in real time; The regulatory assessment module calculates the real-time statistical data of each product name to obtain the frequency characteristic value of each product name appearing on social media; The curve of the frequency characteristic value of each product name changing with time is obtained by fitting through a linear regression algorithm, referred to as the frequency characteristic value curve; Compare the frequency characteristic value curves of each product name with the preset frequency characteristic reference values ​​respectively; If any point on the frequency characteristic value curve of a product name is greater than the preset frequency characteristic reference value, the product name and its frequency characteristic value curve are retained; Otherwise, delete the product name and its frequency characteristic value curve; The process of screening the sales product names collected by the product sales unit includes: The cumulative sales volume R of the collected product names on the e-commerce platform is calculated through the above formula; Where M is the number of e-commerce platforms collected by supervision, j∈[1,M], D j is the cumulative sales volume of the product on the jth e-commerce platform; The regulatory assessment module collects and counts the cumulative sales of each product name on the e-commerce platform in real time; The curve of cumulative sales volume of each product name changing over time is obtained by linear regression algorithm fitting, referred to as cumulative sales volume curve; Derivative the cumulative sales curve of each product name to obtain the sales speed curve of each product name over time; Compare the sales speed change curve of each product name with the preset sales speed change reference value; If the sales speed change curve of a product name is greater than the preset sales speed change reference value, the product name and its sales speed change curve are retained; Otherwise, delete the product name and its sales speed change curve; The process of determining whether each product is an Internet bestseller includes: S1. Mix and match the remaining product names after filtering out repeated product names with the remaining product names after filtering out sales volume product names, and take the product names and related data that are common in the two screening results; S2. Analyze the frequency characteristic value curve and cumulative sales curve of each product name after the screening is completed to obtain the heat characteristic value corresponding to each product; S3, comparing the heat characteristic value corresponding to each product with the preset heat characteristic critical value. If the heat characteristic value of a product is greater than the preset heat characteristic critical value, the product is a best-selling product on the Internet; In step S2, the process of obtaining the heat characteristic value corresponding to each product includes: The heat characteristic value of each product within the preset time interval is calculated by the above formula; Among them, there are a total of W product names screened in step S1, and the W product names are numbered in sequence, k∈[1,W], E k is the heat characteristic value of the kth product, [t1, t2] is the preset time interval, P k (t) is the frequency characteristic curve of the kth product, R k (t1) is the cumulative sales volume of the kth product at time point t1, R k (t2) is the cumulative sales volume of the kth product at time point t2, τ1 and τ2 are the weight coefficients of the factors affecting the heat characteristic value; The process of calculating the correlation coefficient between each Internet best-selling product and the characteristic value of the non-human factors that affect it includes: It is determined that there are D best-selling products on the Internet, and each best-selling product on the Internet is numbered in sequence, y∈[1,D]; In the process of analyzing the heat characteristic values ​​of each product through the supervision and evaluation module, multiple heat characteristic values ​​E of the yth best-selling Internet product are taken in the preset time interval. y (t l ); The change function of the characteristic value of non-human factors affecting the yth product is preset as f y (t); Substitute the intermediate values ​​of each preset time interval into the non-human factor characteristic value function f that affects the product y (t), obtain the f corresponding to the middle time nodes of multiple preset time intervals y (t l ); The correlation coefficient r between the heat characteristic value of the yth Internet best-selling product and the characteristic value of the non-human factors affecting it is calculated by the above formula y ; Among them, H is the number of collected product heat feature values ​​and their corresponding non-human factor feature values, l∈[1,H], E y (t l ) is the heat characteristic value of the yth best-selling Internet product in the lth preset time interval, is the average of the H heat feature values ​​of the yth best-selling Internet product, f y (t l ) is the non-human factor characteristic value corresponding to the lth preset time interval of the yth Internet best-selling product, is the average value of the H non-human factor feature values ​​of the yth best-selling Internet product; The process of obtaining the characteristic change function of non-human factors affecting the y-th product includes: The non-human factor characteristic change function f that affects the yth product is calculated by the above formula y (t); Among them, V y is the number of non-human factors affecting the yth product, x∈[1,V y ],g y·x (t) is the function value of the xth non-human factor affecting the yth product over time, θ y·x is the weight coefficient of the xth non-human factor affecting the yth product; The process of determining whether each Internet best-selling product is a new or popular product includes: The correlation coefficient r corresponding to each Internet best-selling product y The correlation coefficient with the preset standard value r sth Make a comparison; If | r y |≥r sth , indicating that there is a certain degree of correlation between the product and the characteristic values ​​of non-human factors that affect it, and it is not a new or internet-famous product; If | r y | <r sth , indicating that there is no certain degree of correlation between the product and the characteristic values ​​of non-human factors that affect it, and it is a new and internet-famous product.

Images