Service user demand prediction system and method based on big data

Abstract

The invention discloses a service user demand prediction system and method based on big data, relates to the field of user prediction, and solves the problem that the service user demand prediction system is poor in prediction pertinence. Comprising a basic information module which is used for obtaining a first consumer label, a second consumer label and a third consumer label corresponding to a target consumer to carry out to-be-predicted commodity analysis so as to obtain demand prediction object data, the commodity analysis module is used for respectively acquiring a plurality of commodity user preference indexes, a plurality of commodity consumption price ratios and a plurality of historical purchase indexes of commodities of the same type by analyzing the demand prediction object data to obtain commodity analysis data; and the demand prediction module is used for carrying out commodity demand prediction on each to-be-predicted commodity according to the commodity analysis data. According to the invention, the comprehensiveness and the accuracy of a consumption prediction result can be improved.

Description

Technical Field

[0001] The present invention belongs to the field of user prediction and relates to big data technology, and specifically to a service user demand prediction system and method based on big data. Background Art

[0002] The existing user demand forecasting system has the following specific defects when forecasting user commodity demand:

[0003] 1. The existing demand forecasting system does not perform label screening on the predicted products according to the user's age, gender and income when forecasting commodity demand, which makes the forecasting process lack of pertinence and causes a certain waste of resources;

[0004] 2. When the existing demand forecasting system forecasts commodity demand, the indicators used in the forecasting process are single, which leads to the lack of accuracy of the forecast results;

[0005] To this end, we propose a service user demand prediction system and method based on big data. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention aims to provide a service user demand prediction system and method based on big data. The present invention analyzes the commodities to be predicted based on respectively obtaining the first consumer tag, the second consumer tag and the third consumer tag corresponding to the target consumer to obtain demand prediction object data, performs commodity preference analysis on the target consumer according to the demand prediction object data to obtain the commodity user preference index, performs consumption level analysis on the target consumer to obtain the commodity consumption price ratio, performs historical consumption analysis on the target consumer to obtain the historical purchase index of the same type of commodities to obtain commodity analysis data, obtains the demand assessment coefficient corresponding to each commodity to be predicted according to the analysis of the commodity analysis data, obtains multiple demand assessment coefficients, obtains a demand assessment coefficient threshold, numerically compares the multiple demand assessment coefficients with the demand assessment coefficient threshold, and predicts the commodity demand for the target consumer according to the comparison result.

[0007] In order to achieve the above purpose, the present invention adopts the following technical solution: a service user demand prediction system based on big data, and the specific working process of each module is as follows:

[0008] Basic information module: used to obtain the first consumer tag, the second consumer tag and the third consumer tag corresponding to the target consumer to analyze the commodity to be predicted and obtain the demand prediction object data;

[0009] Product analysis module: used to analyze the product preferences of target consumers according to the demand forecast object data to obtain the product user preference index, analyze the consumption level of target consumers to obtain the product consumption price ratio, and analyze the historical consumption of target consumers to obtain the historical purchase index of the same type of products to obtain product analysis data;

[0010] Demand forecasting module: used to obtain the demand assessment coefficient corresponding to each commodity to be predicted based on the analysis of commodity analysis data, obtain multiple demand assessment coefficients, obtain the demand assessment coefficient threshold, compare the multiple demand assessment coefficients with the demand assessment coefficient threshold, and forecast commodity demand for target consumers based on the comparison results.

[0011] Furthermore, the basic information module acquires demand forecast object data as follows:

[0012] Obtain the age value corresponding to the target consumer to obtain the consumer age value, set a first consumer age threshold and a second consumer age threshold to compare the consumer age value, and when the consumer age value is less than the first consumer age threshold, classify the target consumer as a young consumer; when the consumer age value is greater than or equal to the first consumer age threshold and less than or equal to the second consumer age threshold, classify the target consumer as a middle-aged consumer; when the consumer age value is greater than the second consumer age threshold, classify the target consumer as an elderly consumer, and obtain a first consumer label;

[0013] Obtain the gender classification corresponding to the target consumer and obtain the second consumer label;

[0014] Conduct occupational salary analysis on target consumers to obtain third-party consumer labels;

[0015] Using the big data crawler technology, the first consumer tag, the second consumer tag, and the third consumer tag are used as keywords to crawl multiple marketing products to obtain multiple products to be predicted;

[0016] A plurality of commodities to be predicted are defined as demand prediction target data.

[0017] Furthermore, the basic information module obtains the third consumer tag as follows:

[0018] Obtain the occupation type corresponding to the target consumer, obtain a number of consumers with the same occupation type as the target consumer through big data technology, obtain multiple occupational sample consumers, obtain the monthly salary corresponding to each occupational sample consumer, and obtain multiple monthly salaries;

[0019] The average of the multiple monthly salaries obtained was calculated to obtain the average monthly salary of the occupation;

[0020] The variance of the obtained monthly salaries is calculated to obtain the monthly salary variance of the occupation;

[0021] The first monthly salary critical value corresponding to the target consumer is obtained by calculating the monthly average salary of the occupation and the monthly salary variance of the occupation;

[0022] The critical value of the first monthly salary is calculated using the following formula:

[0023] Xlj1=Xzp×(1-Xfc);

[0024] Among them, Xl j1 is the critical value of the first monthly salary, Xzp is the monthly average salary of the occupation, and Xfc is the monthly salary variance of the occupation;

[0025] The second monthly salary critical value corresponding to the target consumer is obtained by calculating the monthly average salary of the occupation and the monthly salary variance of the occupation;

[0026] The second monthly salary critical value is calculated using the following formula:

[0027] Xlj2=Xzp×(1-Xfc);

[0028] Among them, Xl j2 is the second monthly salary critical value, Xzp is the monthly average salary of the occupation, and Xfc is the monthly salary variance of the occupation;

[0029] The numerical range between the first monthly salary critical value and the second monthly salary critical value is marked as the target consumer salary range to obtain a second consumer label.

[0030] Furthermore, the commodity analysis module acquires commodity analysis data as follows:

[0031] Obtain demand forecast object data, and obtain multiple commodities to be forecasted according to the demand forecast object data;

[0032] Randomly select a product from multiple products to be predicted as a feature prediction product;

[0033] The commodity type corresponding to the characteristic predicted commodity is obtained to obtain the characteristic commodity type, and a number of commodities of the same type as the characteristic commodity type are marked on the shopping platform to obtain multiple sample characteristic commodities;

[0034] Conduct product preference analysis on target consumers and obtain the product user preference index corresponding to the feature-predicted product;

[0035] Repeat the process of obtaining the product user preference index corresponding to the feature predicted product, and obtain multiple product user preference indexes for each product to be predicted;

[0036] Analyze the consumption level of target consumers and obtain the commodity consumption price ratio corresponding to the characteristic predicted commodity;

[0037] Repeat the process of obtaining the commodity consumption price ratio corresponding to the feature predicted commodity, respectively obtain the commodity consumption price ratio corresponding to each commodity to be predicted, and obtain multiple commodity consumption price ratios;

[0038] Conduct historical consumption analysis on target consumers and obtain historical purchase indexes of similar products corresponding to the feature-predicted products;

[0039] Repeat the process of obtaining the historical purchase index of the same type of goods corresponding to the feature prediction goods, respectively obtain the historical purchase index of the same type of goods corresponding to each product to be predicted, and obtain multiple historical purchase indexes of the same type of goods;

[0040] The user preference indexes of multiple commodities, the consumption price ratios of multiple commodities and the historical purchase indexes of multiple commodities of the same type are defined as commodity analysis data.

[0041] A method for predicting service user demand based on big data, comprising the following specific steps:

[0042] Step S1: respectively obtaining the first consumer tag, the second consumer tag, and the third consumer tag corresponding to the target consumer to analyze the commodity to be predicted, and obtaining demand prediction object data;

[0043] Step S2: According to the demand forecast object data, a product preference analysis is performed on the target consumer to obtain a product user preference index, a consumption level analysis is performed on the target consumer to obtain a product consumption price ratio, and a historical consumption analysis is performed on the target consumer to obtain a historical purchase index of the same type of products, thereby obtaining product analysis data;

[0044] Step S3: Obtain the demand assessment coefficient corresponding to each commodity to be predicted based on the analyzed commodity analysis data, obtain multiple demand assessment coefficients, obtain the demand assessment coefficient threshold, perform numerical comparison between the multiple demand assessment coefficients and the demand assessment coefficient threshold, and predict commodity demand for target consumers based on the comparison results.

[0045] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0046] 1. The present invention performs label screening on the products to be predicted according to the user's age, gender and income, making the prediction process more targeted and realizing efficient use of prediction resources;

[0047] 2. The present invention can achieve comprehensiveness and accuracy of consumption prediction results by predicting consumption of target consumers from aspects such as product preference, consumption level and historical consumption records. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0049] Figure 1 is the overall system block diagram of the present invention;

[0050] Figure 2 It is a diagram of the implementation steps of the present invention;

[0051] Figure 3 It is the homepage image of the e-commerce platform of the present invention. DETAILED DESCRIPTION

[0052] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. 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.

[0053] Embodiment 1

[0054] See also Figure 1 , the present invention provides a technical solution: a service user demand prediction system based on big data, comprising a basic information module, a commodity analysis module, a demand prediction module and a server, wherein the basic information module, the commodity analysis module and the demand prediction module are respectively connected to the server, and the server controls the basic information module, the commodity analysis module and the demand prediction module respectively;

[0055] The basic information module obtains the first consumer tag, the second consumer tag, and the third consumer tag corresponding to the target consumer to analyze the commodity to be predicted and obtain the demand prediction object data;

[0056] Obtain the age value corresponding to the target consumer to obtain the consumer age value, set a first consumer age threshold and a second consumer age threshold to compare the consumer age value, and when the consumer age value is less than the first consumer age threshold, classify the target consumer as a young consumer; when the consumer age value is greater than or equal to the first consumer age threshold and less than or equal to the second consumer age threshold, classify the target consumer as a middle-aged consumer; when the consumer age value is greater than the second consumer age threshold, classify the target consumer as an elderly consumer, and obtain a first consumer label;

[0057] It should be noted here that:

[0058] In this application, the first consumption age threshold is set at 35 years old, and the second consumption age threshold is set at 60;

[0059] Obtain the gender classification corresponding to the target consumer and obtain the second consumer label;

[0060] Conduct occupational salary analysis on target consumers to obtain third-party consumer labels;

[0061] The details are as follows:

[0062] Obtain the occupation type corresponding to the target consumer, obtain a number of consumers with the same occupation type as the target consumer through big data technology, obtain multiple occupational sample consumers, obtain the monthly salary corresponding to each occupational sample consumer, and obtain multiple monthly salaries;

[0063] The average of the multiple monthly salaries obtained was calculated to obtain the average monthly salary of the occupation;

[0064] The variance of the obtained monthly salaries is calculated to obtain the monthly salary variance of the occupation;

[0065] The first monthly salary critical value corresponding to the target consumer is obtained by calculating the monthly average salary of the occupation and the monthly salary variance of the occupation;

[0066] The critical value of the first monthly salary is calculated using the following formula:

[0067] Xlj1=Xzp×(1-Xfc);

[0068] Among them, Xl j1 is the critical value of the first monthly salary, Xzp is the monthly average salary of the occupation, and Xfc is the monthly salary variance of the occupation;

[0069] The second monthly salary critical value corresponding to the target consumer is obtained by calculating the monthly average salary of the occupation and the monthly salary variance of the occupation;

[0070] The second monthly salary critical value is calculated using the following formula:

[0071] Xlj2=Xzp×(1-Xfc);

[0072] Among them, Xl j2 is the second monthly salary critical value, Xzp is the monthly average salary of the occupation, and Xfc is the monthly salary variance of the occupation;

[0073] Mark the numerical range between the first monthly salary critical value and the second monthly salary critical value as the target consumer salary range to obtain a second consumer label;

[0074] Using the big data crawler technology, the first consumer tag, the second consumer tag, and the third consumer tag are used as keywords to crawl multiple marketing products to obtain multiple products to be predicted;

[0075] It should be noted here that:

[0076] The products to be predicted here are recommended products on the homepage of the e-commerce platform that match the first consumer tag, the second consumer tag, and the third consumer tag;

[0077] For example, by adding the first consumer tag, the second consumer tag, and the third consumer tag respectively through the existing e-commerce platform, the products recommended on the homepage of the e-commerce platform are the products to be predicted;

[0078] A plurality of commodities to be predicted are defined as demand prediction object data;

[0079] The product analysis module performs product preference analysis on the target consumers based on the demand forecast object data to obtain the product user preference index, performs consumption level analysis on the target consumers to obtain the product consumption price ratio, performs historical consumption analysis on the target consumers to obtain the historical purchase index of the same type of products, and obtains product analysis data;

[0080] Obtain demand forecast object data, and obtain multiple commodities to be forecasted according to the demand forecast object data;

[0081] Randomly select a product from multiple products to be predicted as a feature prediction product;

[0082] The commodity type corresponding to the characteristic predicted commodity is obtained to obtain the characteristic commodity type, and a number of commodities of the same type as the characteristic commodity type are marked on the shopping platform to obtain multiple sample characteristic commodities;

[0083] It should be noted here that:

[0084] In this application, when the feature prediction product is Mengniu pure milk, the sample feature products can be Yili pure milk, Telunsu pure milk and Jindian pure milk;

[0085] Conduct product preference analysis on target consumers and obtain the product user preference index corresponding to the feature-predicted product;

[0086] The details are as follows:

[0087] During the period of time when the target consumer's product preference analysis is performed, it is marked as a user preference monitoring cycle with a duration equal to the preference feature duration;

[0088] In the user preference monitoring cycle, a number of feature monitoring time points are randomly selected, and the quantity values ​​corresponding to the feature monitoring time points are obtained to obtain the quantity values ​​of the feature time points;

[0089] See also Figure 3, respectively obtain the e-commerce platform homepage image corresponding to each feature monitoring time point, obtain multiple e-commerce platform homepage images, obtain the cumulative number of times the feature-predicted product is recommended in the multiple e-commerce platform homepage images, obtain the feature-predicted product recommendation number, calculate the ratio of the feature-predicted product recommendation number to the feature time point quantity value, and obtain the first product recommendation ratio;

[0090] It should be noted here that:

[0091] In this application, the e-commerce platform involved here is specifically an online shopping platform, and the e-commerce platform involved here can be JD.com, Pinduoduo and Taobao;

[0092] Obtain the cumulative number of times that multiple sample feature products are recommended in the homepage images of multiple e-commerce platforms to obtain the number of times the multiple sample feature products are recommended, and average the obtained number of times the multiple sample feature products are recommended to obtain the average number of times the sample feature products are recommended, and calculate the ratio of the average number of times the sample feature products are recommended to the number of feature time points to obtain the second product recommendation ratio;

[0093] The first product recommendation ratio and the second product recommendation ratio are calculated to obtain a product user preference index corresponding to the feature prediction product;

[0094] The user preference index of the product corresponding to the feature prediction product is calculated as follows:

[0095]

[0096] Among them, Xhs is the product user preference index corresponding to the feature prediction product, Tb1 is the first product recommendation ratio, and Tb2 is the second product recommendation ratio;

[0097] Repeat the process of obtaining the product user preference index corresponding to the feature predicted product, and obtain multiple product user preference indexes for each product to be predicted;

[0098] Analyze the consumption level of target consumers and obtain the commodity consumption price ratio corresponding to the characteristic predicted commodity;

[0099] In the process of analyzing the consumption level of target consumers, the time value corresponding to the current moment is marked as the first characteristic time point, the characteristic consumption duration before the first characteristic time point is marked as the second characteristic time point, and the period between the first characteristic time point and the second characteristic time point is marked as the consumption matching analysis period;

[0100] Obtain the cumulative amount of consumption of target consumers on the e-commerce platform during the consumption matching analysis period to obtain the online consumption amount during the period;

[0101] Select several price monitoring time points in the consumption matching degree analysis cycle, obtain the commodity price value corresponding to each price detection time point of the feature prediction commodity, obtain multiple commodity price values, and average the obtained multiple commodity price values ​​to obtain the commodity period average price value;

[0102] Get the ratio of the average price value of the commodity period to the online consumption amount of the period, and get the commodity consumption price ratio corresponding to the feature prediction commodity;

[0103] Repeat the process of obtaining the commodity consumption price ratio corresponding to the feature predicted commodity, respectively obtain the commodity consumption price ratio corresponding to each commodity to be predicted, and obtain multiple commodity consumption price ratios;

[0104] Conduct historical consumption analysis on target consumers and obtain historical purchase indexes of similar products corresponding to the feature-predicted products;

[0105] The details are as follows:

[0106] In the process of historical consumption analysis of target consumers, the time value corresponding to the current moment is marked as the first consumption analysis time point, the characteristic consumption duration before the first consumption analysis time point is marked as the second consumption analysis time point, and the period between the first consumption analysis time point and the second consumption analysis time point is marked as the historical consumption analysis cycle;

[0107] And obtain the numerical value of the time length corresponding to the historical consumption analysis cycle to obtain the consumption cycle duration value;

[0108] Obtain the historical consumption records of the target consumer in the historical consumption analysis period, obtain the number of times the target consumer consumes each sample characteristic product in the historical consumption analysis period according to the historical consumption records, obtain the number of times the target consumer consumes multiple characteristic products, and calculate the average of the obtained number of times the consumer consumes the multiple characteristic products to obtain the cumulative number of times the characteristic products are consumed;

[0109] Calculate the ratio of the cumulative consumption times of the feature product to the consumption cycle duration value, and obtain the historical purchase index of the same type of products corresponding to the feature prediction product;

[0110] Repeat the process of obtaining the historical purchase index of the same type of goods corresponding to the feature prediction goods, respectively obtain the historical purchase index of the same type of goods corresponding to each product to be predicted, and obtain multiple historical purchase indexes of the same type of goods;

[0111] The user preference indexes of multiple commodities, the consumption price ratios of multiple commodities, and the historical purchase indexes of multiple commodities of the same type are defined as commodity analysis data;

[0112] The commodity analysis module acquires commodity analysis data and transmits it to the demand forecasting module;

[0113] The demand forecasting module obtains the demand assessment coefficient corresponding to each commodity to be forecasted based on the commodity analysis data, obtains multiple demand assessment coefficients, obtains the demand assessment coefficient threshold, compares the multiple demand assessment coefficients with the demand assessment coefficient threshold, and forecasts the commodity demand for the target consumers based on the comparison results;

[0114] Obtain product analysis data, and obtain the product user preference index, product consumption price ratio, and historical purchase index of the same type of products corresponding to each product to be predicted based on the product analysis data;

[0115] The demand assessment coefficient corresponding to the commodity to be predicted is obtained by calculating the commodity user preference index, commodity consumption price ratio and the historical purchase index of the same type of commodity;

[0116] Calculate the demand assessment coefficient corresponding to the predicted product. The specific formula is as follows:

[0117]

[0118] Among them, Xpx is the demand evaluation coefficient corresponding to the product to be predicted, Xhs is the product user preference index, Gmz is the product consumption price ratio, and Jgb is the historical purchase index of the same type of product;

[0119] Obtaining the demand assessment coefficient corresponding to each commodity to be predicted respectively, and obtaining multiple demand assessment coefficients;

[0120] Obtaining a demand assessment coefficient threshold, comparing multiple demand assessment coefficients with the demand assessment coefficient threshold, and performing demand forecasting on the commodity to be forecasted based on the numerical comparison result;

[0121] The details are as follows:

[0122] Obtain the user preference index threshold of the product, the price ratio threshold of the product, and the historical purchase index threshold of the same type of products respectively;

[0123] It should be noted here that:

[0124] In this application, the commodity user preference index threshold, commodity consumption price ratio threshold, and the same type commodity historical purchase index threshold involved here are respectively the maximum commodity user preference index, average commodity consumption price ratio, and maximum same type commodity historical purchase index corresponding to weak demand commodities;

[0125] The threshold of the demand assessment coefficient is obtained by calculating the threshold of the commodity user preference index, the threshold of the commodity consumption price ratio, and the threshold of the historical purchase index of the same type of commodities;

[0126] The demand assessment coefficient threshold is calculated as follows:

[0127]

[0128] Among them, Xpxy is the demand evaluation coefficient threshold corresponding to the product to be predicted, Xhsy is the product user preference index threshold, Gmzy is the product consumption price ratio threshold, and Jgby is the historical purchase index threshold of the same type of product;

[0129] When the demand assessment coefficient is greater than or equal to the demand assessment coefficient threshold, the corresponding commodity to be predicted is predicted as a commodity with strong demand, and the target consumer is automatically recommended to purchase the commodity;

[0130] When the demand assessment coefficient is less than the demand assessment coefficient threshold, the corresponding commodity to be predicted is predicted as a weak demand commodity, and the target consumer is automatically dissuaded from purchasing the commodity.

[0131] In this application, if corresponding calculation formulas appear, the above calculation formulas are all dimensionless and take their numerical calculations. The weight coefficients, proportional coefficients and other coefficients in the formulas are set to a result value obtained by quantifying each parameter. The size of the weight coefficient and the proportional coefficient can be determined as long as it does not affect the proportional relationship between the parameter and the result value.

[0132] Embodiment 2

[0133] Based on another concept of the same invention, a service user demand prediction method based on big data is now proposed, comprising the following steps:

[0134] See also Figure 2 , Step S1: respectively obtain the first consumer tag, the second consumer tag and the third consumer tag corresponding to the target consumer to analyze the commodity to be predicted, and obtain demand prediction object data;

[0135] Step S11: Obtain the age value corresponding to the target consumer to obtain the consumer age value, set a first consumer age threshold and a second consumer age threshold to compare the consumer age value, and when the consumer age value is less than the first consumer age threshold, classify the target consumer as a young consumer; when the consumer age value is greater than or equal to the first consumer age threshold and less than or equal to the second consumer age threshold, classify the target consumer as a middle-aged consumer; when the consumer age value is greater than the second consumer age threshold, classify the target consumer as an elderly consumer, and obtain a first consumer label;

[0136] Step S12: Obtain the gender classification corresponding to the target consumer to obtain a second consumer label;

[0137] Step S13: Performing occupational salary analysis on the target consumer to obtain a third consumer label;

[0138] The details are as follows:

[0139] Step S131: Obtain the occupation type corresponding to the target consumer, obtain a number of consumers with the same occupation type as the target consumer through big data technology, obtain a plurality of occupational sample consumers, obtain the monthly salary corresponding to each occupational sample consumer, and obtain a plurality of monthly salaries;

[0140] Step S132: Calculate the average of the multiple monthly salaries to obtain the monthly average salary of the occupation;

[0141] Step S133: performing variance calculation on the obtained multiple monthly salaries to obtain the monthly salary variance of the occupation;

[0142] Step S134: Calculate the occupational monthly average salary and the occupational monthly salary variance to obtain the first monthly salary critical value corresponding to the target consumer;

[0143] The critical value of the first monthly salary is calculated using the following formula:

[0144] Xlj1=Xzp×(1-Xfc);

[0145] Among them, Xl j1 is the critical value of the first monthly salary, Xzp is the monthly average salary of the occupation, and Xfc is the monthly salary variance of the occupation;

[0146] Step S135: Calculate the occupational monthly average salary and the occupational monthly salary variance to obtain the second monthly salary critical value corresponding to the target consumer;

[0147] The second monthly salary critical value is calculated using the following formula:

[0148] Xlj2=Xzp×(1-Xfc);

[0149] Among them, Xl j2 is the second monthly salary critical value, Xzp is the monthly average salary of the occupation, and Xfc is the monthly salary variance of the occupation;

[0150] Step S136: marking the numerical range between the first monthly salary threshold value and the second monthly salary threshold value as the target consumer salary range, and obtaining a second consumer label;

[0151] Step S14: using the big data crawler technology to crawl multiple marketing products using the first consumer tag, the second consumer tag, and the third consumer tag as keywords to obtain multiple products to be predicted;

[0152] Step S15: defining a plurality of commodities to be predicted as demand prediction object data;

[0153] Step S2: According to the demand forecast object data, a product preference analysis is performed on the target consumer to obtain a product user preference index, a consumption level analysis is performed on the target consumer to obtain a product consumption price ratio, and a historical consumption analysis is performed on the target consumer to obtain a historical purchase index of the same type of products, thereby obtaining product analysis data;

[0154] Step S21: Obtain demand forecast object data, and obtain a plurality of commodities to be forecasted according to the demand forecast object data;

[0155] Step S22: randomly selecting a product from multiple products to be predicted as a feature prediction product;

[0156] Step S23: Acquire the commodity type corresponding to the characteristic predicted commodity to obtain the characteristic commodity type, and mark a number of commodities of the same type as the characteristic commodity type on the shopping platform to obtain a plurality of sample characteristic commodities;

[0157] Step S24: Analyze the product preferences of target consumers and obtain the product user preference index corresponding to the feature-predicted product;

[0158] The details are as follows:

[0159] Step S241: marking a user preference monitoring period with a duration equal to the preference feature duration within the period of time during which the target consumer's product preference analysis is performed;

[0160] Step S242: randomly selecting a number of characteristic monitoring time points in the user preference monitoring cycle, and acquiring the quantity values ​​corresponding to the characteristic monitoring time points to obtain the quantity values ​​of the characteristic time points;

[0161] Step S243: respectively obtaining the e-commerce platform homepage image corresponding to each feature monitoring time point, obtaining multiple e-commerce platform homepage images, obtaining the cumulative number of times the feature-predicted product is recommended in the multiple e-commerce platform homepage images, obtaining the number of feature-predicted product recommendations, calculating the ratio of the number of feature-predicted product recommendations to the number of feature time points, and obtaining the first product recommendation ratio;

[0162] Step S244: Obtain the cumulative number of times that multiple sample feature products are recommended in the homepage images of multiple e-commerce platforms to obtain the number of times that multiple sample feature products are recommended, and average the obtained number of times that multiple sample feature products are recommended to obtain the average number of times that the sample feature products are recommended, and calculate the ratio of the average number of times that the sample feature products are recommended to the number of feature time points to obtain the second product recommendation ratio;

[0163] Step S245: Calculate the first product recommendation ratio and the second product recommendation ratio to obtain a product user preference index corresponding to the feature prediction product;

[0164] The user preference index of the product corresponding to the feature prediction product is calculated as follows:

[0165]

[0166] Among them, Xhs is the product user preference index corresponding to the feature prediction product, Tb1 is the first product recommendation ratio, and Tb2 is the second product recommendation ratio;

[0167] Step S25: respectively obtain a plurality of commodity user preference indexes for each commodity to be predicted corresponding to the commodity user preference index;

[0168] Step S26: Analyze the consumption level of the target consumers and obtain the commodity consumption price ratio corresponding to the characteristic predicted commodity;

[0169] The details are as follows:

[0170] Step S261: in the process of analyzing the consumption level of the target consumer, the time value corresponding to the current moment is marked as the first characteristic time point, the characteristic consumption duration before the first characteristic time point is marked as the second characteristic time point, and the period between the first characteristic time point and the second characteristic time point is marked as the consumption matching degree analysis period;

[0171] Step S262: Obtain the cumulative amount of consumption of the target consumer on the e-commerce platform during the consumption matching analysis period to obtain the periodic online consumption amount;

[0172] Step S263: selecting a number of price monitoring time points in the consumption matching degree analysis period, respectively obtaining the commodity price value corresponding to each price detection time point of the feature prediction commodity, obtaining a plurality of commodity price values, and averaging the obtained plurality of commodity price values ​​to obtain an average price value of the commodity period;

[0173] Step S264: Obtain the ratio of the average price value of the commodity period to the online consumption amount of the period, and obtain the commodity consumption price ratio corresponding to the feature prediction commodity;

[0174] Step S265: acquiring the commodity consumption price ratio corresponding to each commodity to be predicted, and obtaining a plurality of commodity consumption price ratios;

[0175] Step S27: Analyze the historical consumption of the target consumer and obtain the historical purchase index of the same type of goods corresponding to the feature prediction goods;

[0176] The details are as follows:

[0177] Step S271: in the process of performing historical consumption analysis on the target consumer, the time value corresponding to the current moment is marked as the first consumption analysis time point, the characteristic consumption duration before the first consumption analysis time point is marked as the second consumption analysis time point, and the period between the first consumption analysis time point and the second consumption analysis time point is marked as the historical consumption analysis period;

[0178] Step S272: obtaining a numerical value of the time length corresponding to the historical consumption analysis cycle to obtain a consumption cycle duration value;

[0179] Step S273: Obtain the historical consumption records of the target consumer in the historical consumption analysis period, obtain the number of times the target consumer consumes each sample characteristic commodity in the historical consumption analysis period according to the historical consumption records, obtain multiple characteristic commodity consumption times, and average the obtained multiple characteristic commodity consumption times to obtain the cumulative characteristic commodity consumption times;

[0180] Step S274: Calculate the ratio of the cumulative number of consumption times of the characteristic commodity to the consumption cycle duration value, and obtain the historical purchase index of the same type of commodity corresponding to the characteristic prediction commodity;

[0181] Step S275: acquiring the historical purchase index of the same type of goods corresponding to each of the goods to be predicted, and obtaining a plurality of historical purchase indexes of the same type of goods;

[0182] Step S28: defining multiple commodity user preference indexes, multiple commodity consumption price ratios, and multiple historical purchase indexes of the same type of commodities as commodity analysis data;

[0183] Step S3: Obtain the demand assessment coefficient corresponding to each commodity to be predicted according to the analyzed commodity analysis data, obtain multiple demand assessment coefficients, obtain the demand assessment coefficient threshold, compare the multiple demand assessment coefficients with the demand assessment coefficient threshold, and predict the commodity demand for target consumers according to the comparison results;

[0184] Step S31: Obtain commodity analysis data, and obtain commodity user preference index, commodity consumption price ratio and historical purchase index of commodities of the same type corresponding to each commodity to be predicted according to the commodity analysis data;

[0185] Step S32: Calculate the commodity user preference index, commodity consumption price ratio and the historical purchase index of the same type of commodity to obtain the demand assessment coefficient corresponding to the commodity to be predicted;

[0186] Calculate the demand assessment coefficient corresponding to the predicted product. The specific formula is as follows:

[0187]

[0188] Among them, Xpx is the demand evaluation coefficient corresponding to the product to be predicted, Xhs is the product user preference index, Gmz is the product consumption price ratio, and Jgb is the historical purchase index of the same type of product;

[0189] Step S33: acquiring the demand assessment coefficient corresponding to each commodity to be predicted, and obtaining a plurality of demand assessment coefficients;

[0190] Step S34: obtaining a demand assessment coefficient threshold, comparing multiple demand assessment coefficients with the demand assessment coefficient threshold, and performing demand forecasting on the commodity to be forecasted according to the numerical comparison result;

[0191] The details are as follows:

[0192] Step S341: respectively obtaining a commodity user preference index threshold, a commodity consumption price ratio threshold, and a historical purchase index threshold for commodities of the same type;

[0193] Step S342: Calculate the demand assessment coefficient threshold by using the commodity user preference index threshold, the commodity consumption price ratio threshold, and the historical purchase index threshold of the same type of commodity;

[0194] The demand assessment coefficient threshold is calculated as follows:

[0195] The demand assessment coefficient threshold corresponding to the predicted product is calculated. The specific formula is as follows:

[0196]

[0197] Among them, Xpxy is the demand evaluation coefficient threshold corresponding to the product to be predicted, Xhsy is the product user preference index threshold, Gmzy is the product consumption price ratio threshold, and Jgby is the historical purchase index threshold of the same type of product;

[0198] Step S343: When the demand evaluation coefficient is greater than or equal to the demand evaluation coefficient threshold, the corresponding commodity to be predicted is predicted as a commodity with strong demand, and the target consumer is automatically recommended to purchase the commodity;

[0199] Step S344: When the demand assessment coefficient is less than the demand assessment coefficient threshold, the corresponding commodity to be predicted is predicted as a weak-demand commodity, and the target consumer is automatically dissuaded from purchasing the commodity.

[0200] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A service user demand prediction system based on big data, characterized in that: include: Basic information module: used to obtain the first consumer tag, the second consumer tag and the third consumer tag corresponding to the target consumer to analyze the commodity to be predicted and obtain the demand prediction object data; Product analysis module: used to analyze the product preferences of target consumers according to the demand forecast object data to obtain the product user preference index, analyze the consumption level of target consumers to obtain the product consumption price ratio, and analyze the historical consumption of target consumers to obtain the historical purchase index of the same type of products to obtain product analysis data; Demand forecasting module: used to obtain the demand assessment coefficient corresponding to each commodity to be predicted based on the analysis of commodity analysis data, obtain multiple demand assessment coefficients, obtain the demand assessment coefficient threshold, compare the multiple demand assessment coefficients with the demand assessment coefficient threshold, and forecast commodity demand for target consumers based on the comparison results.

2. A service user demand prediction system based on big data according to claim 1, characterized in that: The basic information module acquires demand forecast object data as follows: The age value corresponding to the target consumer is obtained to obtain the consumer age value. The first consumer age threshold and the second consumer age threshold are set to compare the values ​​with the consumer age value. When the consumer age value is less than the first consumer age threshold, the target consumer is classified as a young consumer. When the consumer age value is greater than or equal to the first consumer age threshold and less than or equal to the second consumer age threshold, the target consumer is classified as a middle-aged consumer. When the consumer age value is greater than the second consumer age threshold, the target consumer is classified as an elderly consumer to obtain the first consumer label.

3. The service user demand prediction system based on big data according to claim 2 is characterized in that: The basic information module acquires the demand forecast object data, specifically as follows: acquiring the gender classification corresponding to the target consumer to obtain the second consumer label; Conduct occupational salary analysis on target consumers to obtain third-party consumer labels; Using the big data crawler technology, the first consumer tag, the second consumer tag, and the third consumer tag are used as keywords to crawl multiple marketing products to obtain multiple products to be predicted; A plurality of commodities to be predicted are defined as demand prediction target data.

4. The service user demand prediction system based on big data according to claim 3 is characterized in that: The basic information module obtains the third consumer tag as follows: Obtain the occupation type corresponding to the target consumer, obtain several consumers with the same occupation type as the target consumer, obtain multiple occupational sample consumers, obtain the monthly salary corresponding to each occupational sample consumer, and obtain multiple monthly salaries.

5. The service user demand prediction system based on big data according to claim 4 is characterized in that: The basic information module obtains the third consumer tag as follows: The average of the multiple monthly salaries obtained was calculated to obtain the average monthly salary of the occupation; The variance of the obtained monthly salaries is calculated to obtain the monthly salary variance of the occupation; The first monthly salary critical value corresponding to the target consumer is obtained by calculating the average monthly salary of the occupation and the monthly salary variance of the occupation.

6. The service user demand prediction system based on big data according to claim 5 is characterized in that: The basic information module obtains the third consumer tag as follows: Calculate the first monthly salary threshold; The second monthly salary critical value corresponding to the target consumer is obtained by calculating the monthly average salary of the occupation and the monthly salary variance of the occupation; Calculate the second monthly salary threshold; The numerical range between the first monthly salary critical value and the second monthly salary critical value is marked as the target consumer salary range to obtain a second consumer label.

7. The service user demand prediction system based on big data according to claim 1 is characterized in that: The commodity analysis module acquires commodity analysis data, as follows: Obtain demand forecast object data, and obtain multiple commodities to be forecasted according to the demand forecast object data; Randomly select a product from multiple products to be predicted as a feature prediction product; The commodity type corresponding to the feature predicted commodity is obtained to obtain the feature commodity type, and a number of commodities of the same type as the feature commodity type are marked on the shopping platform to obtain multiple sample feature commodities.

8. The service user demand prediction system based on big data according to claim 7 is characterized in that: The commodity analysis module acquires commodity analysis data, specifically as follows: performs commodity preference analysis on target consumers, and acquires commodity user preference indexes corresponding to feature-predicted commodities; The user preference index of each commodity to be predicted is respectively calculated to obtain multiple user preference indexes of the commodity; Analyze the consumption level of target consumers and obtain the commodity consumption price ratio corresponding to the characteristic predicted commodity; Obtain the commodity consumption price ratio corresponding to each commodity to be predicted respectively, and obtain multiple commodity consumption price ratios; Conduct historical consumption analysis on target consumers and obtain historical purchase indexes of similar products corresponding to the feature-predicted products.

9. The service user demand prediction system based on big data according to claim 8, characterized in that: The commodity analysis module acquires commodity analysis data, specifically as follows: acquiring the historical purchase index of commodities of the same type corresponding to each commodity to be predicted, and obtaining multiple historical purchase indexes of commodities of the same type; The user preference indexes of multiple commodities, the consumption price ratios of multiple commodities and the historical purchase indexes of multiple commodities of the same type are defined as commodity analysis data.

10. A method for predicting service user demand based on big data, applicable to a system for predicting service user demand based on big data as claimed in any one of claims 1 to 9, characterized in that: The demand forecasting method comprises the following specific steps: Step S1: respectively obtaining the first consumer tag, the second consumer tag, and the third consumer tag corresponding to the target consumer to analyze the commodity to be predicted, and obtaining demand prediction object data; Step S2: According to the demand forecast object data, a product preference analysis is performed on the target consumer to obtain a product user preference index, a consumption level analysis is performed on the target consumer to obtain a product consumption price ratio, and a historical consumption analysis is performed on the target consumer to obtain a historical purchase index of the same type of products, thereby obtaining product analysis data; Step S3: Obtain the demand assessment coefficient corresponding to each commodity to be predicted based on the analyzed commodity analysis data, obtain multiple demand assessment coefficients, obtain the demand assessment coefficient threshold, perform numerical comparison between the multiple demand assessment coefficients and the demand assessment coefficient threshold, and predict commodity demand for target consumers based on the comparison results.

Images