CN114693368A - Behavior data-based customer maintenance method and device and storage medium - Google Patents
Behavior data-based customer maintenance method and device and storage medium Download PDFInfo
- Publication number
- CN114693368A CN114693368A CN202210395910.9A CN202210395910A CN114693368A CN 114693368 A CN114693368 A CN 114693368A CN 202210395910 A CN202210395910 A CN 202210395910A CN 114693368 A CN114693368 A CN 114693368A
- Authority
- CN
- China
- Prior art keywords
- price
- historical
- time period
- prediction
- function
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000012423 maintenance Methods 0.000 title claims abstract description 44
- 238000003860 storage Methods 0.000 title claims abstract description 11
- 230000006870 function Effects 0.000 claims abstract description 177
- 238000009499 grossing Methods 0.000 claims description 115
- 230000006399 behavior Effects 0.000 claims description 52
- 230000001419 dependent effect Effects 0.000 claims description 16
- 238000012417 linear regression Methods 0.000 claims description 16
- 238000005457 optimization Methods 0.000 claims description 12
- 238000010276 construction Methods 0.000 claims description 11
- 230000003542 behavioural effect Effects 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 8
- 238000012886 linear function Methods 0.000 claims description 6
- 238000004891 communication Methods 0.000 claims description 3
- 230000009193 crawling Effects 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 238000013461 design Methods 0.000 abstract description 11
- 230000000875 corresponding effect Effects 0.000 description 18
- 230000008569 process Effects 0.000 description 13
- 241001061264 Astragalus Species 0.000 description 7
- 239000009636 Huang Qi Substances 0.000 description 7
- 235000006533 astragalus Nutrition 0.000 description 7
- 229940107666 astragalus root Drugs 0.000 description 7
- 238000012545 processing Methods 0.000 description 7
- 210000004233 talus Anatomy 0.000 description 7
- 230000000694 effects Effects 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 230000008531 maintenance mechanism Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 230000002596 correlated effect Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0206—Price or cost determination based on market factors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0207—Discounts or incentives, e.g. coupons or rebates
- G06Q30/0222—During e-commerce, i.e. online transactions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0207—Discounts or incentives, e.g. coupons or rebates
- G06Q30/0239—Online discounts or incentives
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0611—Request for offers or quotes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Engineering & Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a customer maintenance method, a device and a storage medium based on behavior data, firstly, the preference commodity of a user is determined through the behavior data of the user, then, the historical selling price and the market capacity data of the preference commodity in a historical time period are crawled, so that two price prediction functions are established based on the historical selling price and the market capacity data, and then, a price prediction model is established based on the two price prediction functions to realize the prediction of the price of the preference commodity in a future time period; finally, generating first pre-purchase recommendation information based on the predicted selling price, and sending the first pre-purchase recommendation information to the user terminal; through the design, the price of the preferred commodity of the user can be predicted, and the purchase recommendation information is generated based on the predicted price and sent to the user, so that the active maintenance of the user is realized, the platform is not limited, and the timeliness and effectiveness of the maintenance are improved.
Description
Technical Field
The invention belongs to the technical field of electronic commerce, and particularly relates to a customer maintenance method and device based on behavior data and a storage medium.
Background
With the rapid development of the internet, the e-commerce platform gradually becomes one of the mainstream shopping modes of people, and with the continuous improvement of the e-commerce platform and the payment means, the electronic commodity transaction amount in China is steadily increasing, and the innovation of the operation mode is increasingly active, and the electronic commerce platform is rapidly developing towards a multi-level and diversified direction.
At present, most of user maintenance mechanisms of e-commerce platforms display user preference commodities on a platform home page, and the basic working logic of the user maintenance mechanisms is as follows: collecting browsing data of a user on a platform for commodities, taking the browsed commodities as preference commodities of the user, and displaying the preference commodities to the user when the user enters the platform next time; the aforementioned user maintenance mechanism has the following disadvantages: the system is a passive maintenance mechanism and is limited to the platform, so that the maintenance purpose cannot be achieved when a user does not enter the platform, and the problems of untimely maintenance and poor maintenance effect exist; therefore, it is urgent to provide a customer maintenance method with timely maintenance and good maintenance effect.
Disclosure of Invention
The invention aims to provide a customer maintenance method, a customer maintenance device and a storage medium based on behavior data, and aims to solve the problems of untimely maintenance and poor maintenance effect of the existing maintenance mechanism.
In order to achieve the purpose, the invention adopts the following technical scheme:
in a first aspect, the present invention provides a method for customer maintenance based on behavior data, including:
acquiring behavior data of a user on a shopping platform;
determining the preference commodity of the user according to the behavior data;
obtaining historical selling price and market capacity data of the preferred commodity in a historical time period;
constructing a first price prediction function according to the historical selling price and market capacity data in the historical time period, and constructing a second price prediction function according to the historical selling price in the historical time period;
constructing a price prediction model based on the first price prediction function and the second price prediction function;
obtaining the selling prices of the preference commodities in different prediction time periods by utilizing the price prediction model;
generating first pre-purchase recommendation information according to the selling prices of the preference commodities in different prediction time periods;
and sending the first pre-purchase recommendation information to a user terminal so that a user purchases the preference commodity according to the first pre-purchase recommendation information.
Based on the disclosure, firstly, the preferred commodity of the user is determined through the behavior data of the user, then the historical selling price and the market capacity data of the preferred commodity in a historical time period are crawled, so that two price prediction functions are established based on the historical selling price and the market capacity data, and then a price prediction model is established based on the two price prediction functions to realize the prediction of the price of the preferred commodity in a future time period; finally, generating first pre-purchase recommendation information based on the predicted selling price, and sending the first pre-purchase recommendation information to the user terminal; through the design, the price of the preferred commodity of the user can be predicted, and the purchase recommendation information is generated based on the predicted price and sent to the user, so that the active maintenance of the user is realized, the platform is not limited, and the timeliness and effectiveness of the maintenance are improved.
In one possible design, the market capacity data includes: the method comprises the steps that the yield of preferred commodities in a historical time period, the stocking amount of target merchants for the preferred commodities in the historical time period and the inventory amount of target manufacturers for the preferred commodities in the historical time period are obtained, wherein the target merchants are selling merchants of the preferred commodities, and the target manufacturers are manufacturers supplying the target merchants with the preferred commodities;
correspondingly, according to the historical selling price and the market capacity data in the historical time period, a first price prediction function is constructed, and the method comprises the following steps:
taking the input goods quantity, the inventory quantity and the output as price influence factors, and respectively calculating the correlation value of each price influence factor and the historical selling price;
establishing a multiple linear regression function by taking the price influence factor with the correlation value larger than the preset correlation value as an independent variable and the selling price of the preference commodity as a dependent variable;
determining each regression coefficient in the multiple linear regression function by using a target influence factor and the historical sale price in the historical time period, wherein the target influence factor is a price influence factor of which the correlation value is greater than a preset correlation value;
and obtaining the first price prediction function based on the multiple linear regression function and the regression coefficient.
Based on the above disclosure, the invention discloses a specific construction process of a first price prediction function, namely, the correlation between each data in market capacity data and the selling price is judged, the data with high correlation is taken as a price influence factor, then the price influence factor is taken as an independent variable, the selling price of a preferred commodity is taken as a dependent variable, a multiple linear regression function is established, a regression coefficient is determined based on the data corresponding to the independent variable and the historical selling price, and finally, the regression coefficient is substituted into the multiple linear regression function, so that the first price prediction function can be obtained.
In one possible design, constructing a second price prediction function according to the historical selling prices in the historical time period includes:
according to the historical selling price in the historical time period, calculating to obtain a primary exponential smooth value and a secondary exponential smooth value of the preference commodity in the historical time period;
establishing an exponential smoothing function based on the primary exponential smoothing value and the secondary exponential smoothing value, and with a prediction time period as an independent variable and the selling price of the preferred commodity as a dependent variable;
and carrying out iterative computation on the exponential smoothing function to obtain an optimal exponential smoothing function so as to take the optimal exponential smoothing function as the second price prediction function.
Based on the above disclosure, the invention discloses a specific construction process of the second price prediction function, namely, a first exponential smoothing value and a second exponential smoothing value of a preference commodity are calculated by utilizing historical selling prices, then an exponential smoothing function is constructed based on the two smoothing values, and finally, the exponential smoothing function is continuously iterated to obtain a function with the optimal fitting degree, and the function is used as the optimal exponential smoothing function, so that the optimal exponential smoothing function is used as the second price prediction function.
In one possible design, calculating a first exponential smoothing value and a second exponential smoothing value of the preferred product in the historical time period according to the historical selling price in the historical time period includes:
obtaining a primary smoothing initial value and a secondary smoothing initial value according to the historical selling price in the historical time period;
calculating to obtain a primary exponential smoothing value and a secondary exponential smoothing value based on the primary smoothing initial value and the secondary smoothing initial value according to the following formula (1) and formula (2);
in the above-mentioned formula (1) and formula (2), S1Representing a first exponential smoothing value, S2Represents a quadratic exponential smoothing value, a represents a smoothing weight, and takes a value of [ 0.1%]，ytIndicating the historical selling price during the t time period within the historical time period,
the initial value of the first smoothing is represented,
representing the initial value of the secondary smoothing, wherein n is the total number of time periods;
correspondingly, based on the first exponential smoothing value and the second exponential smoothing value, and with a prediction time period as an independent variable and a selling price of the preferred commodity as a dependent variable, establishing an exponential smoothing function, including:
calculating a first smoothing coefficient K and a second smoothing coefficient b according to the first exponential smoothing value and the second exponential smoothing value and the following formula (3) and formula (4);
b＝2S1-S2 (3)
and establishing to obtain the exponential smoothing function by taking the first smoothing coefficient K as a linear term coefficient of a linear function, the second smoothing coefficient b as a constant term of the linear function, and a prediction time period as an independent variable and the selling price of the preferred commodity as a dependent variable.
In one possible design, constructing a price prediction model based on the first price prediction function and the second price prediction function includes:
calculating to obtain the average absolute percentage error of the first price prediction function by using the historical selling price and the market capacity data in the historical time period;
calculating to obtain the average absolute percentage error of the second price prediction function by using the historical selling prices in the historical time period;
obtaining a prediction weight of the first price prediction function and a prediction weight of the second price prediction function based on the average absolute percentage error of the first price prediction function and the average percentage error of the second price prediction function;
calculating the product of the prediction weight of the first price prediction function and the first price prediction function to obtain a first optimization function; calculating the product of the prediction weight of the second price prediction function and the second price prediction function to obtain a second optimization function;
and summing the first optimization function and the second optimization function to obtain the price prediction model.
Based on the above disclosure, the invention discloses a specific construction process of a price prediction model, namely, the average absolute percentage error of two prediction functions is calculated by utilizing historical selling price and market capacity data, then, the prediction weight of each prediction function is calculated based on the average absolute percentage error of the two prediction functions, and finally, the prediction functions are multiplied by the respective prediction weights and summed to obtain the price prediction model.
In one possible design, the prediction weight of the first price prediction function and the prediction weight of the second price prediction function are obtained according to the following formula (5);
in the above formula (5), ωiRepresenting the prediction weight, w, of the ith price prediction functioniThe average absolute percentage error of the ith price prediction function is expressed, and m represents the total number of the price prediction functions.
In one possible design, after deriving the selling prices of the preferred item over different predicted time periods, the method further comprises:
it is determined whether discount information exists for the preferred goods,
if so, obtaining the discount remaining time limit of the preference commodity based on the discount information;
screening out the selling prices in target prediction time periods from the selling prices in different prediction time periods, wherein each target prediction time period in the target prediction time periods is a time period of the preferred commodity after the discount information is invalid;
generating second pre-purchase recommendation information based on the discount remaining term of the preferred commodity and the selling price in the target prediction time period;
and sending the second pre-purchase recommendation information to a user terminal so that the user purchases the preference commodity according to the second pre-purchase recommendation information.
Based on the above disclosure, the present invention may further combine the discount information of the preferred product to generate the second pre-purchase recommendation information, which is substantially as follows: reminding the user of the discount service life of the preferred goods by using the discount remaining life; and comparing the price in the discount with the selling price after the discount is invalid, thereby recommending the optimal purchasing time for the user.
In a second aspect, the present invention provides a customer maintenance device based on user behavior, including:
the behavior data crawling unit is used for acquiring behavior data of a user on the shopping platform;
the commodity determining unit is used for determining the preference commodity of the user according to the behavior data;
a commodity data acquisition unit for acquiring historical selling prices and market capacity data of the preferred commodities in a historical time period;
the model building unit is used for building a first price prediction function according to the historical selling price and the market capacity data in the historical time period, and building a second price prediction function according to the historical selling price in the historical time period;
the model construction unit is also used for constructing a price prediction model based on the first price prediction function and the second price prediction function;
the forecasting unit is used for obtaining the selling prices of the preference commodities in different forecasting time periods by utilizing a price forecasting model;
the price table generating unit is used for generating first pre-purchase recommendation information according to the selling prices of the preference commodities in different prediction time periods;
and the sending unit is used for sending the first pre-purchase recommendation information to a user terminal so that a user can purchase the preference commodity according to the first pre-purchase recommendation information.
In a third aspect, the present invention provides another customer maintenance apparatus based on behavior data, taking an apparatus as an electronic device as an example, including a memory, a processor and a transceiver, which are sequentially connected in communication, where the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the customer maintenance method based on behavior data as described in the first aspect or any one of the possible designs of the first aspect.
In a fourth aspect, the present invention provides a storage medium having stored thereon instructions for executing the behavioral data-based customer care method according to the first aspect or any one of the possible designs of the first aspect when the instructions are run on a computer.
In a fifth aspect, the present invention provides a computer program product comprising instructions which, when run on a computer, cause the computer to perform the behavioural data based customer care method as claimed in the first aspect or any one of the first aspects.
Drawings
FIG. 1 is a schematic diagram of a system architecture of a customer care system based on behavior data according to the present invention;
FIG. 2 is a flow chart illustrating steps of a method for customer maintenance based on behavior data according to the present invention;
FIG. 3 is a schematic structural diagram of a customer care device based on behavior data according to the present invention;
fig. 4 is a schematic structural diagram of an electronic device provided in the present invention.
Detailed Description
The invention is further described with reference to the following figures and specific examples. It should be noted that the description of the embodiments is provided to help understanding of the present invention, but the present invention is not limited thereto. Specific structural and functional details disclosed herein are merely representative for purposes of describing example embodiments of the present invention. This invention may, however, be embodied in many alternate forms and should not be construed as limited to the embodiments set forth herein.
It will be understood that, although the terms first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element, without departing from the scope of example embodiments of the present invention.
It should be understood that, for the term "and/or" as may appear herein, it is merely an associative relationship that describes an associated object, meaning that three relationships may exist, e.g., a and/or B may mean: a exists independently, B exists independently, and A and B exist simultaneously; for the term "/and" as may appear herein, which describes another associative object relationship, it means that two relationships may exist, e.g., a/and B, may mean: a exists independently, and A and B exist independently; in addition, for the character "/" that may appear herein, it generally means that the former and latter associated objects are in an "or" relationship.
Examples
Referring to fig. 1, a system architecture is provided for the present application, where the system includes a server and a plurality of user terminals, where the server is communicatively connected to each user terminal to obtain behavior data of a user on a shopping platform on each user terminal, and after obtaining the behavior data of the user corresponding to each user terminal, the server may use the method provided in this embodiment to determine a preferred commodity of the user and a selling price of the preferred commodity in a future time period (i.e., a predicted time period), and finally, based on the predicted selling price, may generate first pre-purchase recommendation information and send the first pre-purchase recommendation information to the user terminal; through the design, the system can realize active maintenance of the user based on the behavior data of the user, so that the system is not limited to the platform, and the timeliness and effectiveness of maintenance are improved.
Referring to fig. 2, in the customer maintenance method based on behavior data provided in the first aspect of this embodiment, after a user browses a shopping platform each time, active maintenance of the user may be implemented, where the maintenance method may be, but is not limited to, run on a server side, and of course, the execution subject does not constitute a limitation on the embodiment of this application; specifically, the foregoing maintenance method may include, but is not limited to, the following steps S1 to S8.
S1, acquiring behavior data of a user on a shopping platform; in a specific application, the behavior data of the user on the shopping platform may be captured when the user logs in the shopping platform each time, optionally, the behavior data may include, but is not limited to: the system comprises commodity browsing data, commodity purchasing data (used for representing that commodities are added into a shopping cart), commodity searching data and commodity collection data (used for representing that commodities are added into a collection), wherein the commodity browsing data comprises browsing duration.
Specifically, after the behavior data of the user is obtained, the preferred commodity of the user can be determined based on the behavior data, as shown in the following step S2.
S2, determining the preference commodities of the user according to the behavior data; the determination of the preferred commodities, when applied specifically, is as shown in step S21, step S22, and step S23 described below.
S21, when the behavior data are commodity purchase adding data, commodity searching data and/or commodity collection data, taking the corresponding commodities of the behavior data, the commodity purchase adding data, the commodity searching data and/or the commodity collection data as the preference commodities; that is, the behavior of the user on the commodity is shopping, searching and/or collecting, which can indicate that the purchase willingness degree of the user on the commodity is large, so that the commodity corresponding to the behavior data can be directly used as a preference commodity.
Meanwhile, the user may make a purchase on the platform, which may cause a problem of misoperation, and thus, for the commodity browsing data, it is necessary to determine whether the user is interested in the commodity according to the browsing duration of the user, as shown in steps S22 and S23 below.
And S22, when the behavior data is the commodity browsing data, judging whether the browsing duration is greater than a preset threshold value.
S23, if yes, taking the commodity corresponding to the commodity browsing data as the preference commodity; specifically, when the browsing duration of a certain commodity by the user is longer than the preset browsing duration, the user is explained to check the introduction of the user in detail, so that the commodity can be regarded as the preferred commodity of the user when the browsing duration is longer than the preset browsing duration; if the preset browsing time is 1 minute, the client browses the astragalus commodity detail page, and the browsing time is 2 minutes, so that astragalus can be used as a preferred commodity of the user.
After the preferred commodity of the user is determined based on the behavior data, the price of the preferred commodity can be predicted, namely the selling price of the preferred commodity in a future time period is predicted, so that a first pre-purchase recommendation message is generated and sent to the user based on the predicted selling price, and active maintenance of the user is achieved.
Specifically, the future time period of this embodiment refers to: the time period after the user behavior data is generated, if the behavior data is 2022-3-10,10 o' clock, the future time period is the time period after the moment, such as the selling price of 3 month and 17 days in 2022, the selling price of 3 month and 24 days in 2022, and the selling price of 3 month and 31 days in 2022; of course, the specific time period may be set according to actual use.
Optionally, the present embodiment predicts the selling price of the preferred commodity in the future time period by constructing a price prediction model, as shown in steps S3 to S6.
S3, obtaining historical selling price and market capacity data of the preferred commodities in a historical time period; in specific application, the historical time period may be, but is not limited to, the first 6 months of the time when the user behavior data is generated, for example, on the basis of the foregoing example, the historical time period is the first 6 months of 3 months and 10 days in 2022; of course, the history period may be set specifically at the time of actual use, and is not limited thereto.
Alternatively, example market capacity data may include, but is not limited to: the method comprises the steps that the yield of preferred commodities in a historical time period, the stocking amount (equivalent to the demand amount of a target merchant) of target merchants for the preferred commodities in the historical time period and the inventory amount of target manufacturers for the preferred commodities in the historical time period are obtained, wherein the target merchants are selling merchants of the preferred commodities, and the target manufacturers are manufacturers supplying the target merchants with the preferred commodities; in specific application, a market capacity database is arranged in the server, and the yield of each commodity per month, the commodity quantity of each commodity to each commodity and the stock quantity of each commodity in the platform are stored in the server (for example, the commodity quantity is uploaded to the server by a merchant terminal in real time); thus, the data can be searched in the market capacity database based on the names of the preferred commodities.
After obtaining the historical selling price and the market capacity data of the preferred commodity, the price prediction model can be constructed, in this embodiment, two prediction functions are firstly established, and then the price prediction model is constructed based on the two prediction functions, wherein the construction process of the two price prediction functions is as shown in the following step S4.
S4, constructing a first price prediction function according to the historical selling price and market capacity data in the historical time period, and constructing a second price prediction function according to the historical selling price in the historical time period; in specific application, for example, the first price prediction function is a multiple linear regression function, and the second price prediction function is an exponential smoothing function, so that a combined price prediction model is constructed by 2 different types of prediction functions, the problem of low precision of a single prediction function is solved, and the accuracy of price prediction is improved.
Alternatively, the construction process of the first price prediction function is exemplified as shown in the following steps S41 to S44.
S41, taking the input goods quantity, the inventory quantity and the output as price influence factors, and respectively calculating a correlation value of each price influence factor and the historical selling price; specifically, the correlation value refers to a pearson correlation coefficient, which can be obtained according to a calculation formula corresponding to the pearson correlation coefficient, and is a common technique for determining the correlation of the factors.
Specifically, when the correlation value is more than or equal to 0 and less than or equal to 0.3, the price influencing factor is weakly correlated with the historical selling price, and when the correlation value is more than 0.3 and less than or equal to 0.5, the price influencing factor is weakly correlated with the historical selling price; when the relevance value is more than 0.5 and less than or equal to 0.8, the price influence factor is obviously related to the historical sale price; when the relevance value is more than 0.8 and less than or equal to 1, the price influence factor is highly relevant to the historical sale price; in this embodiment, the exemplary prediction threshold is 0.7.
In addition, in this embodiment, the correlation values of the input quantity, the inventory quantity and the output with the historical selling price are all greater than 0.7, and therefore, all the correlation values are independent variables, that is, the output, the demand quantity and the inventory quantity all affect the price of the preferred commodity, if the output is reduced, the price of the preferred commodity is increased, otherwise, the price is reduced, and if the demand quantity of the target merchant is increased, it indicates that the higher the selling value of the preferred commodity is, the corresponding price should be increased; for another example, the smaller the inventory of the supplier, the more intense the preferred commodity resource, and the corresponding price will rise.
After the correlation value between the price influencing factors and the historical selling prices is obtained, the price influencing factors with the correlation value larger than the preset threshold value can be used as independent variables to construct a multiple linear regression function with the selling prices of the preferred commodities as dependent variables, as shown in step S42.
S42, establishing a multiple linear regression function by taking the price influence factor with the correlation value larger than the preset correlation value as an independent variable and the selling price of the preference commodity as a dependent variable; in specific application, a multiple regression function can be constructed according to the following formula:
yT＝b0+β1XT1+β2XT2+β3XT3
in the above formula, yTIndicating the selling price of the preference commodity predicted by the first price prediction function in the prediction time period T; b0，β1，β2And beta3Respectively represent the regression coefficients, XT1Represents the production amount, X, of the preferred product in the prediction time period TT2Indicating the amount of preferred merchandise in the forecast time period TT3Indicating the stock quantity of the preferred commodities in the prediction time period T.
As can be seen from the above equation, the selling price of the preferred commodity is predicted using the predicted yield, the predicted stocking amount, and the predicted inventory amount within the prediction time period T; more specifically, the estimated quantity of the target merchant in the predicted time period T, the estimated stock of the target manufacturer in the predicted time period T, and the estimated yield are equivalent; optionally, the target merchant may send the inventory amounts corresponding to different forecast times, the target manufacturer may send the inventory amounts corresponding to different forecast times, and the manufacturer of the preferred goods may send the production amounts corresponding to different forecast times to the server.
After the multiple linear regression function is constructed, each regression coefficient in the multiple linear regression function needs to be determined, and specifically, the regression coefficient may be obtained according to the data corresponding to each independent variable and the historical selling price, as shown in step S43 below.
S43, determining each regression coefficient in the multiple linear regression function by using a target influence factor and the historical sale price in the historical time period, wherein the target influence factor is a price influence factor of which the correlation value is greater than a preset correlation value; specifically, the data corresponding to the target influence factor (i.e., the yield, the inventory, and the stock in the historical time period) and the historical selling price may be input into an SPSS (Statistical Product and Service Solutions) 24 software to calculate each regression coefficient, which is a commonly used coefficient calculation software in the multiple regression field and may be used directly.
After each regression coefficient is obtained, the regression coefficients may be substituted into a multiple linear regression function to obtain a first cost function, as shown in step S44 below.
And S44, obtaining the first price prediction function based on the multiple linear regression function and the regression coefficient.
Similarly, the second price prediction function is constructed as shown in the following steps S45 to S47.
S45, calculating to obtain a primary exponential smoothing value and a secondary exponential smoothing value of the preference commodity in the historical time period according to the historical sale price in the historical time period; for specific applications, the calculation process of the first exponential smoothing value and the second exponential smoothing value is as shown in step S45a and step S45b.
S45a, obtaining a primary smooth initial value and a secondary smooth initial value according to the historical selling price in the historical time period; specifically, the historical time period may be divided into n time periods, for example, the historical time period may be divided into 6 months, and the time period may be divided into 9 weeks (less than one week, and calculated by one week), and then the average value of the historical selling prices of the previous 3 weeks may be taken as the primary smoothing initial value and the secondary smoothing initial value; if there is a different historical selling price in any one week of the previous 3 weeks, the historical selling price in any one week is the average selling price of 7 days in any one week.
After the primary and secondary smoothing initial values are obtained, the primary and secondary exponential smoothing values may be calculated, as shown in coefficient step S45b.
S45b, calculating to obtain a primary exponential smoothing value and a secondary exponential smoothing value based on the primary smoothing initial value and the secondary smoothing initial value according to the following formula (1) and formula (2);
in the above-mentioned formula (1) and formula (2), S1Representing a first exponential smoothing value, S2Represents a quadratic exponential smoothing value, a represents a smoothing weight, and takes a value of [ 0.1%]，ytIndicating the historical selling price during the t time period within the historical time period,
the initial value of the first smoothing is represented,
represents the initial value of the second smoothing, and n is the total number of time segments.
On the basis of the foregoing example, the foregoing formula (1) is set forth, and the foregoing n is 9, then the calculation process of the first exponential smoothing value of the preferred commodity in the historical time period is: respectively calculating the first exponential smoothing values corresponding to 9 time periods (namely, each week in 9 weeks), then calculating the first exponential smoothing values of 9 time periods, and then averaging to obtain S1(ii) a In the same way, S2Obtained byThe process is consistent with the foregoing example, and is not described herein again.
After the first exponential smoothing value and the second exponential smoothing value of the preferred commodity in the historical time period are obtained, the construction of the exponential smoothing function can be performed, as shown in the following step S46.
S46, establishing an exponential smoothing function based on the primary exponential smoothing value and the secondary exponential smoothing value and by taking a prediction time period as an independent variable and the selling price of the preferred commodity as a dependent variable; when applied specifically, step S46 includes the following step S46a and step S46b.
S46a, calculating a first smoothing coefficient K and a second smoothing coefficient b according to the primary exponential smoothing value and the secondary exponential smoothing value and the following formula (3) and formula (4).
b＝2S1-S2 (3)
S46b, establishing and obtaining the exponential smoothing function by taking the first smoothing coefficient K as a first-order coefficient of a linear function and the second smoothing coefficient b as a constant item of the linear function, and taking a prediction time period as an independent variable and the selling price of the preferred commodity as a dependent variable; in particular applications, the exponential smoothing function can be expressed by the following formula:
YT＝KT+b
in the above formula, T represents a prediction time period, YTRepresenting the sales price obtained using an exponential smoothing function over the prediction period.
Since the smoothing weight a exists in the first smoothing coefficient K and the second smoothing coefficient b, an optimal smoothing weight needs to be determined to obtain the most accurate prediction result, and thus, an exponential smoothing function needs to be iteratively calculated as shown in step S47 below.
S47, performing iterative computation on the exponential smoothing function to obtain an optimal exponential smoothing function so as to take the optimal exponential smoothing function as the second price prediction function; in specific application, in an iteration process, a range of preset time T is selected, then smoothing weights with different values are substituted to obtain different predicted values, then linear fitting degree is obtained based on a least square method, and an exponential smoothing function with the optimal linear fitting degree is used as a second price prediction function.
In this embodiment, the aforementioned predictions of the preferred commodity yield, the goods quantity and the stock quantity may also be performed by using an exponential smoothing function, that is, the historical yield, the historical goods quantity and the historical stock quantity of the preferred commodity in the historical time period are obtained; then, an exponential smoothing function with the prediction time period as an independent variable and the prediction yield as a dependent variable can be constructed by using the steps S45-S47; an exponential smoothing function which takes the forecasting time period as an independent variable and forecasts the cargo quantity as a dependent variable, and an exponential smoothing function which takes the forecasting time period as an independent variable and forecasts the stock quantity as a dependent variable; finally, the prediction of the yield, the goods quantity and the inventory quantity in the prediction time period can be realized based on respective exponential smoothing functions.
After the first price prediction function and the second price prediction function are obtained, the two prediction functions are combined to construct a combined prediction model as a price prediction model, as shown in the following step S5.
S5, constructing a price prediction model based on the first price prediction function and the second price prediction function; when the method is specifically applied, the weight value of each price prediction function is calculated, then the weight value is multiplied by the corresponding prediction function, and the sum is obtained to obtain the price prediction model, wherein the construction process is shown in the following steps S51-S55.
And S51, calculating to obtain the average absolute percentage error of the first price prediction function by using the historical sale price and the market capacity data in the historical time period.
S52, calculating to obtain the average absolute percentage error of the second price prediction function by utilizing the historical selling price in the historical time period; when the method is applied specifically, the same physical quantity is measured for multiple times, absolute values of absolute errors of the measurements are taken, and then the absolute values are averaged, so that average absolute errors can be obtained; therefore, market capacity data can be substituted into the first price prediction function, then prediction is carried out for multiple times to obtain multiple predicted values, the predicted values are compared with actual values (namely historical selling prices) of the predicted values to obtain absolute values of absolute errors, and then the average value of the results is taken to obtain the average absolute percentage error of the first price prediction function; similarly, the calculation principle of the average absolute percentage error of the second price prediction function is the same as the foregoing example, and is not described herein again.
After the average absolute percentage error of the two price prediction functions is obtained, the calculation of the prediction weights may be performed, as shown in step S53 described below.
S53, obtaining a prediction weight of the first price prediction function and a prediction weight of the second price prediction function based on the average absolute percentage error of the first price prediction function and the average percentage error of the second price prediction function; in specific application, the prediction weight of each prediction function can be calculated according to the following formula:
in the above formula (5), ωiRepresenting the prediction weight, w, of the ith price prediction functioniThe average absolute percentage error of the ith price prediction function is expressed, and m represents the total number of the price prediction functions.
After the prediction weights of the respective price prediction functions are obtained from the above equation (5), a price prediction model can be constructed as shown in the following steps S54 and S55.
S54, calculating the product of the prediction weight of the first price prediction function and the first price prediction function to obtain a first optimization function; and calculating the product of the prediction weight of the second price prediction function and the second price prediction function to obtain a second optimization function.
And S55, summing the first optimization function and the second optimization function to obtain the price prediction model.
The foregoing steps S54 and S55 are summarized in one formula as follows:
PT＝ω1yT+ω2YT (6)
in the above formula (6), PTIndicating the selling price, omega, over a predicted time period T1A prediction weight, ω, representing a first price prediction function2Representing the prediction weights of the second lattice prediction function.
After the price prediction model is obtained, the price of the preferred commodity is predicted as shown in the following step S6.
S6, obtaining the selling prices of the preference commodities in different prediction time periods by utilizing the price prediction model; in specific applications, the division of the prediction time period is the same as the division of the historical time period, for example, the historical time period is divided by weeks, so the prediction time period is also in units of weeks; such as a sale price in the first week after the generation of the behavior data, a sale price in the second week, etc.
After the selling prices of the few products in different prediction time periods are obtained, the first pre-purchase recommendation information can be generated, as shown in the following step S7.
S7, generating first pre-purchase recommendation information according to the selling prices of the preference commodities in different prediction time periods; in specific application, the selling prices corresponding to different prediction time periods can be compared with the actual selling price of the preference commodity (namely, the corresponding selling price when the behavior data is generated), so that a difference value is obtained, and finally, the prediction time period corresponding to the optimal price is used as the recommended purchasing time period.
For example, in the first week 10 o' clock after 3 months 10 d 2022, the selling price of astragalus root is: 10 yuan each kilogram.
The selling price of the astragalus root in the second week 10 o' clock after 3 months and 10 days in 2022 is as follows: 9.8 yuan per kilogram.
The selling price of the astragalus root in the third week 10 o' clock after 3 months and 10 days in 2022 is as follows: 9 yuan each jin.
In the fourth week 10 o' clock after 3 months and 10 days in 2022, the selling price of radix astragali is: 10.5 yuan.
In the fifth week 10 o' clock after 3 months and 10 days in 2022, the selling price of radix astragali is: 10.3 yuan.
At 10 o 10/3/2022, the selling price of radix astragali is: 9.7 jin, therefore, the first pre-purchase recommendation information may be, but is not limited to:
in the first week 10 o' clock after 3 months 10 d in 2022, the selling price of radix astragali is: 10 yuan per jin, and 0.3 yuan more is expected to be spent than the current one; the selling price of the astragalus root in the second week 10 o' clock after 3 months and 10 days in 2022 is as follows: 9.8 yuan per kilogram, 0.1 yuan more is estimated to be spent than the current one kilogram; the selling price of the astragalus root in the third week 10 o' clock after 3 months and 10 days in 2022 is as follows: 9 yuan jin, expect that jin is current to be saved 0.7 yuan than, in 2022 year 3 month 10 day 10 later fourth week, the selling price of astragalus is: 10.5 yuan per jin, and 0.8 yuan is expected to be spent more than the current jin; in the fifth week 10 o' clock after 3 months and 10 days in 2022, the selling price of radix astragali is: 10.3 yuan per jin, and 0.6 yuan is expected to be spent more than the current jin; xxx users are advised to make a purchase of astragalus the third week 10 o' clock after 3 months 10 of 2022.
After the first pre-purchase recommendation information is obtained, the recommendation information can be sent to the user terminal so as to implement active maintenance of the user, as shown in step S7 below.
And S8, sending the first pre-purchase recommendation information to a user terminal so that a user can purchase the preference commodity according to the first pre-purchase recommendation information.
In addition, the present embodiment may also generate recommendation information in combination with the preference information of the preferred merchandise, as shown in steps S9 to S13 below.
S9, judging whether the preference commodities have discount information or not; in a specific application, the discount information comprises shop discount information and/or platform discount information.
And S10, if so, obtaining the discount remaining time limit of the preference commodity based on the discount information.
S11, screening out the selling prices in target prediction time periods from the selling prices in different prediction time periods, wherein each target prediction time period in the target prediction time periods is a time period of the preferred commodity after discount information is invalid; in specific application, the selling price after the discount service life is screened out from different prediction time periods.
The explanation is made on the basis of the above examples: for example, the discount information of the preferred commodity astragalus is shop discount information, and the service life is: 24 o ' clock at 3/2022 to 24 o ' clock at 3/2022, the discount remaining period is 6 days and 13 hours, and the target prediction time period is 2 nd to 5 th weeks after 10 o ' clock at 3/2022.
After the discount remaining term and the selling price within the target prediction time period are obtained, the second pre-purchase recommendation information may be generated, as shown in step S12 described below.
S12, generating second pre-purchase recommendation information based on the discount remaining time limit of the preference commodity and the selling price in the target prediction time period; when the method is applied specifically, the second pre-purchase recommendation information is divided into two parts, and discount remaining time limit reminding is carried out; the best time to purchase is compared to the price within the discounted lifespan.
For example, assuming that the discount information is 9 discount per jin, the selling price of astragalus root in the discount service life is: 8.73 yuan/jin, therefore, the second pre-purchase recommendation information may be, but is not limited to:
the astragalus discount time remains for 6 days and 13 hours, please note the discount time for the xxx user;
in the second week 10 o' clock after 3 months and 10 days in 2022, the selling price of radix astragali is: 9.8 yuan per kilogram, and the cost of each kilogram is estimated to be 1.07 yuan more than that of the current kilogram; the selling price of the astragalus root in the third week 10 o' clock after 3 months and 10 days in 2022 is as follows: 9 yuan jin, expect that jin is than present saving 0.27 yuan, in 2022 year 3 month 10 day 10 after the fourth week, the selling price of astragalus is: 10.5 yuan per jin, and the estimated cost of each jin is 1.77 yuan more than the current cost; in the fifth week 10 o' clock after 3 months and 10 days in 2022, the selling price of radix astragali is: 10.3 yuan per jin, and the estimated cost of each jin is 1.57 yuan more than the current cost; the xxx users are advised to make purchases within the discount expiry date.
After the second recommended pre-purchase information is obtained, the information can be sent to the user terminal so as to achieve the active maintenance of the user, as shown in the following step S13.
And S13, sending the second pre-purchase recommendation information to a user terminal so that the user can purchase the preference commodity according to the second pre-purchase recommendation information.
In this embodiment, the pushing manner of the first pre-purchase recommendation information and the second pre-purchase recommendation information may be, but is not limited to: and pushing the short message and/or the platform message bar.
Therefore, through the detailed explanation of the customer maintenance method based on the behavior data, the price of the preference commodity of the user can be predicted, and the purchase recommendation information is generated based on the predicted price and sent to the user, so that the active maintenance of the user is realized, the platform is not limited, and the timeliness and the effectiveness of the maintenance are improved.
As shown in fig. 3, a second aspect of the present embodiment provides a hardware device for implementing the behavioral data-based customer maintenance method described in the first aspect of the embodiment, including:
and the behavior data crawling unit is used for acquiring the behavior data of the user on the shopping platform.
And the commodity determining unit is used for determining the preference commodity of the user according to the behavior data.
And the commodity data acquisition unit is used for acquiring the historical selling price and the market capacity data of the preferred commodity in a historical time period.
And the model construction unit is used for constructing a first price prediction function according to the historical sales price and the market capacity data in the historical time period, and constructing a second price prediction function according to the historical sales price in the historical time period.
And the model construction unit is also used for constructing a price prediction model based on the first price prediction function and the second price prediction function.
And the predicting unit is used for obtaining the selling prices of the preference commodities in different predicting time periods by utilizing the price predicting model.
And the price table generating unit is used for generating first pre-purchase recommendation information according to the selling prices of the preference commodities in different prediction time periods.
And the sending unit is used for sending the first pre-purchase recommendation information to a user terminal so that a user can purchase the preference commodity according to the first pre-purchase recommendation information.
For the working process, the working details, and the technical effects of the hardware apparatus provided in this embodiment, reference may be made to the first aspect of the embodiment, which is not described herein again.
As shown in fig. 4, a third aspect of this embodiment provides another customer maintenance apparatus based on behavior data, taking an apparatus as an electronic device as an example, including: a memory, a processor and a transceiver, which are sequentially connected in communication, wherein the memory is used for storing a computer program, the transceiver is used for transceiving a message, and the processor is used for reading the computer program and executing the customer maintenance method based on behavior data according to the first aspect of the embodiment.
For example, the Memory may include, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Flash Memory (Flash Memory), a First In First Out (FIFO), a First In Last Out (FILO), and/or a First In Last Out (FILO); in particular, the processor may include one or more processing cores, such as a 4-core processor, an 8-core processor, and so on. The processor may be implemented in at least one hardware form of a DSP (Digital Signal Processing), an FPGA (Field Programmable Gate Array), and a PLA (Programmable Logic Array), and may also include a main processor and a coprocessor, where the main processor is a processor for Processing data in an awake state and is also referred to as a Central Processing Unit (CPU); a coprocessor is a low power processor for processing data in a standby state.
In some embodiments, the processor may be integrated with a GPU (Graphics Processing Unit) which is responsible for rendering and drawing contents required to be displayed on the display screen, for example, the processor may not be limited to a processor adopting a model STM32F105 series microprocessor, a Reduced Instruction Set Computer (RISC) microprocessor, an architecture processor such as X86, or a processor integrating an embedded neural Network Processing Unit (NPU); the transceiver may be, but is not limited to, a wireless fidelity (WIFI) wireless transceiver, a bluetooth wireless transceiver, a General Packet Radio Service (GPRS) wireless transceiver, a ZigBee wireless transceiver (ieee802.15.4 standard-based low power local area network protocol), a 3G transceiver, a 4G transceiver, and/or a 5G transceiver, etc. In addition, the device may also include, but is not limited to, a power module, a display screen, and other necessary components.
For the working process, the working details, and the technical effects of the electronic device provided in this embodiment, reference may be made to the first aspect of the embodiment, which is not described herein again.
A fourth aspect of the present embodiment provides a storage medium storing instructions including the behavior data based customer maintenance method according to the first aspect of the present embodiment, that is, the storage medium stores instructions that, when executed on a computer, perform the behavior data based customer maintenance method according to the first aspect.
The storage medium refers to a carrier for storing data, and may include, but is not limited to, a floppy disk, an optical disk, a hard disk, a flash Memory, a flash disk and/or a Memory Stick (Memory Stick), etc., and the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
For the working process, the working details, and the technical effects of the storage medium provided in this embodiment, reference may be made to the first aspect of the embodiment, which is not described herein again.
A fifth aspect of the present embodiments provides a computer program product comprising instructions which, when run on a computer, are adapted to cause the computer to perform the method for behavioral data-based customer care according to the first aspect of the embodiments, wherein the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable apparatus.
Finally, it should be noted that: the above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A method for customer maintenance based on behavioral data, comprising:
acquiring behavior data of a user on a shopping platform;
determining the preference commodity of the user according to the behavior data;
acquiring historical selling price and market capacity data of the preferred commodity in a historical time period;
constructing a first price prediction function according to the historical selling price and market capacity data in the historical time period, and constructing a second price prediction function according to the historical selling price in the historical time period;
constructing a price prediction model based on the first price prediction function and the second price prediction function;
obtaining the selling prices of the preference commodities in different prediction time periods by utilizing the price prediction model;
generating first pre-purchase recommendation information according to the selling prices of the preference commodities in different prediction time periods;
and sending the first pre-purchase recommendation information to a user terminal so that a user purchases the preference commodity according to the first pre-purchase recommendation information.
2. The method of claim 1, wherein the market capacity data comprises: the method comprises the steps that the yield of preferred commodities in a historical time period, the stocking amount of target merchants for the preferred commodities in the historical time period and the inventory amount of target manufacturers for the preferred commodities in the historical time period are obtained, wherein the target merchants are selling merchants of the preferred commodities, and the target manufacturers are manufacturers supplying the target merchants with the preferred commodities;
correspondingly, according to the historical selling price and the market capacity data in the historical time period, a first price prediction function is constructed, and the method comprises the following steps:
taking the input goods amount, the inventory amount and the output as price influence factors, and respectively calculating a correlation value of each price influence factor and the historical sale price;
establishing a multiple linear regression function by taking the price influence factor with the correlation value larger than the preset correlation value as an independent variable and the selling price of the preference commodity as a dependent variable;
determining each regression coefficient in the multiple linear regression function by using a target influence factor and the historical sale price in the historical time period, wherein the target influence factor is a price influence factor of which the correlation value is greater than a preset correlation value;
and obtaining the first price prediction function based on the multiple linear regression function and the regression coefficient.
3. The method of claim 1, wherein constructing a second price prediction function based on historical sales prices over the historical time period comprises:
according to the historical selling price in the historical time period, calculating to obtain a primary exponential smooth value and a secondary exponential smooth value of the preference commodity in the historical time period;
establishing an exponential smoothing function based on the primary exponential smoothing value and the secondary exponential smoothing value, and with a prediction time period as an independent variable and the selling price of the preferred commodity as a dependent variable;
and carrying out iterative computation on the exponential smoothing function to obtain an optimal exponential smoothing function so as to take the optimal exponential smoothing function as the second price prediction function.
4. The method of claim 3, wherein calculating a first exponential smoothing value and a second exponential smoothing value for the preferred item over the historical time period based on the historical selling prices over the historical time period comprises:
obtaining a primary smoothing initial value and a secondary smoothing initial value according to the historical selling price in the historical time period;
calculating to obtain a primary exponential smoothing value and a secondary exponential smoothing value based on the primary smoothing initial value and the secondary smoothing initial value according to the following formula (1) and formula (2);
in the above-mentioned formula (1) and formula (2), S1Represents a first order exponentially smoothed value, S2Represents a quadratic exponential smoothing value, a represents a smoothing weight, and takes a value of [ 0.1%]，ytIndicating the historical selling price during the t time period within the historical time period,
it is shown that the initial value of the first smoothing,
representing the initial value of the secondary smoothing, wherein n is the total number of time periods;
correspondingly, based on the first exponential smoothing value and the second exponential smoothing value, and with a prediction time period as an independent variable and a selling price of the preferred commodity as a dependent variable, establishing an exponential smoothing function, including:
calculating a first smoothing coefficient K and a second smoothing coefficient b according to the first exponential smoothing value and the second exponential smoothing value and the following formula (3) and formula (4);
b＝2S1-S2 (3)
and establishing to obtain the exponential smoothing function by taking the first smoothing coefficient K as a linear term coefficient of a linear function, the second smoothing coefficient b as a constant term of the linear function, and a prediction time period as an independent variable and the selling price of the preferred commodity as a dependent variable.
5. The method of claim 1, wherein constructing a price prediction model based on the first price prediction function and the second price prediction function comprises:
calculating to obtain the average absolute percentage error of the first price prediction function by using the historical selling price and the market capacity data in the historical time period;
calculating to obtain the average absolute percentage error of the second price prediction function by using the historical selling prices in the historical time period;
obtaining a prediction weight of the first price prediction function and a prediction weight of the second price prediction function based on the average absolute percentage error of the first price prediction function and the average percentage error of the second price prediction function;
calculating the product of the prediction weight of the first price prediction function and the first price prediction function to obtain a first optimization function; calculating the product of the prediction weight of the second price prediction function and the second price prediction function to obtain a second optimization function;
and summing the first optimization function and the second optimization function to obtain the price prediction model.
6. The method according to claim 5, wherein the prediction weight of the first price prediction function and the prediction weight of the second price prediction function are obtained according to the following formula (5);
in the above formula (5), ωiRepresenting the prediction weight, w, of the ith price prediction functioniThe average absolute percentage error of the ith price prediction function is expressed, and m represents the total number of the price prediction functions.
7. The method of claim 1, wherein after deriving the selling prices of the preferred good for different predicted time periods, the method further comprises:
determining whether discount information exists for the preferred goods,
if so, obtaining the discount remaining time limit of the preference commodity based on the discount information;
screening out the selling prices in target prediction time periods from the selling prices in different prediction time periods, wherein each target prediction time period in the target prediction time periods is a time period of the preferred commodity after the discount information is invalid;
generating second pre-purchase recommendation information based on the discount remaining term of the preferred commodity and the selling price in the target prediction time period;
and sending the second pre-purchase recommendation information to a user terminal so that the user purchases the preference commodity according to the second pre-purchase recommendation information.
8. A customer care appliance based on behavioral data, comprising:
the behavior data crawling unit is used for acquiring behavior data of a user on the shopping platform;
the commodity determining unit is used for determining the preference commodity of the user according to the behavior data;
a commodity data acquisition unit for acquiring historical selling prices and market capacity data of the preferred commodities in a historical time period;
the model building unit is used for building a first price prediction function according to the historical selling price and market capacity data in the historical time period and building a second price prediction function according to the historical selling price in the historical time period;
the model construction unit is also used for constructing a price prediction model based on the first price prediction function and the second price prediction function;
the forecasting unit is used for obtaining the selling prices of the preference commodities in different forecasting time periods by utilizing a price forecasting model;
the price table generating unit is used for generating first pre-purchase recommendation information according to the selling prices of the preference commodities in different prediction time periods;
and the sending unit is used for sending the first pre-purchase recommendation information to a user terminal so that a user can purchase the preference commodity according to the first pre-purchase recommendation information.
9. A customer care appliance based on behavioral data, comprising: the client maintenance method based on the behavior data comprises a memory, a processor and a transceiver which are sequentially connected in a communication mode, wherein the memory is used for storing a computer program, the transceiver is used for transmitting and receiving messages, and the processor is used for reading the computer program and executing the client maintenance method based on the behavior data according to any one of claims 1 to 7.
10. A storage medium having stored thereon instructions for performing the method of behavioral data-based customer maintenance according to any one of claims 1 to 7 when the instructions are run on a computer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210395910.9A CN114693368A (en) | 2022-04-14 | 2022-04-14 | Behavior data-based customer maintenance method and device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210395910.9A CN114693368A (en) | 2022-04-14 | 2022-04-14 | Behavior data-based customer maintenance method and device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114693368A true CN114693368A (en) | 2022-07-01 |
Family
ID=82143042
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210395910.9A Pending CN114693368A (en) | 2022-04-14 | 2022-04-14 | Behavior data-based customer maintenance method and device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114693368A (en) |
Citations (32)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050004819A1 (en) * | 2003-03-27 | 2005-01-06 | Oren Etzioni | Performing predictive pricing based on historical data |
| US20050240539A1 (en) * | 2004-04-22 | 2005-10-27 | Thomas Olavson | Method and system for forecasting commodity prices using capacity utilization data |
| US7437323B1 (en) * | 2003-06-25 | 2008-10-14 | Pros Revenue Management; L.P. | Method and system for spot pricing via clustering based demand estimation |
| US20110040656A1 (en) * | 2009-08-12 | 2011-02-17 | Groetzinger Jon D | System and method for generating predictions of price and availability of event tickets on secondary markets |
| JP2012003677A (en) * | 2010-06-21 | 2012-01-05 | Nippon Telegr & Teleph Corp <Ntt> | Commodity recommendation device, commodity recommendation method and commodity recommendation program |
| CN102324082A (en) * | 2011-09-08 | 2012-01-18 | 上海烟草集团有限责任公司 | Cigarette market price prediction method based on multiple linear regression |
| CN104217355A (en) * | 2014-09-11 | 2014-12-17 | 北京京东尚科信息技术有限公司 | Method and device for predicting sales volume of promotion items |
| JP2015041121A (en) * | 2013-08-20 | 2015-03-02 | 株式会社日立製作所 | Sales forecast system and sales forecast method |
| CN104866474A (en) * | 2014-02-20 | 2015-08-26 | 阿里巴巴集团控股有限公司 | Personalized data searching method and device |
| CN105243563A (en) * | 2015-10-21 | 2016-01-13 | 百度在线网络技术（北京）有限公司 | Commodity price predicting method and apparatus |
| JP2016133816A (en) * | 2015-01-15 | 2016-07-25 | 株式会社そごう・西武 | Commodity demand prediction system |
| CN105869019A (en) * | 2016-03-31 | 2016-08-17 | 金蝶软件（中国）有限公司 | Method and apparatus for predicting goods price |
| CN109214893A (en) * | 2018-08-31 | 2019-01-15 | 深圳春沐源控股有限公司 | Method of Commodity Recommendation, recommender system and computer installation |
| JP2019028816A (en) * | 2017-08-01 | 2019-02-21 | 株式会社オープンソース活用研究所 | Prospective customer information price determination system, price determination method, server device, and program |
| CN110276652A (en) * | 2018-03-14 | 2019-09-24 | 北京京东尚科信息技术有限公司 | Method and apparatus for pushed information |
| CN110276495A (en) * | 2019-06-26 | 2019-09-24 | 东莞市盟大塑化科技有限公司 | Commercial articles ordering method, apparatus, computer and storage medium based on artificial intelligence |
| CN110517072A (en) * | 2019-08-14 | 2019-11-29 | 平安科技（深圳）有限公司 | Method for pushing, device, equipment and the computer readable storage medium of information of vehicles |
| CN111127149A (en) * | 2019-12-23 | 2020-05-08 | 卓尔智联(武汉)研究院有限公司 | Bulk commodity supply information recommendation method and device and computer equipment |
| CN111178972A (en) * | 2019-12-30 | 2020-05-19 | 北京每日优鲜电子商务有限公司 | Message pushing method and device, storage medium and equipment |
| CN111681099A (en) * | 2020-06-03 | 2020-09-18 | 中国银行股份有限公司 | Product information pushing method and device, computer equipment and readable storage medium |
| CN111680221A (en) * | 2020-08-11 | 2020-09-18 | 腾讯科技（深圳）有限公司 | Information recommendation method, device, equipment and computer readable storage medium |
| CN111709813A (en) * | 2020-06-19 | 2020-09-25 | 深圳市赛宇景观设计工程有限公司 | Commodity recommendation method based on big data line |
| CN111899063A (en) * | 2020-06-17 | 2020-11-06 | 东南大学 | Fresh agricultural product online recommendation method considering customer consumption behaviors and preference |
| CN112132618A (en) * | 2020-09-23 | 2020-12-25 | 杭州拼便宜网络科技有限公司 | Commodity price determining method, device and equipment and readable storage medium |
| CN113157752A (en) * | 2021-03-12 | 2021-07-23 | 北京航空航天大学 | Scientific and technological resource recommendation method and system based on user portrait and situation |
| CN113269618A (en) * | 2021-06-10 | 2021-08-17 | 广州富港生活智能科技有限公司 | Optimal price display method and system based on selection preference and readable storage medium |
| CN113298550A (en) * | 2020-08-21 | 2021-08-24 | 阿里巴巴集团控股有限公司 | Commodity price adjusting method and device and server |
| CN113379516A (en) * | 2021-08-12 | 2021-09-10 | 永正信息技术(南京)有限公司 | Recommended product determination method and device |
| KR20220022544A (en) * | 2020-08-19 | 2022-02-28 | 이니그마(주) | Price adjustment system based on customer demand |
| CN114240556A (en) * | 2021-12-17 | 2022-03-25 | 支付宝(杭州)信息技术有限公司 | Information pushing method and device and electronic equipment |
| CN114331594A (en) * | 2021-12-14 | 2022-04-12 | 芸豆数字科技有限公司 | Commodity recommendation method |
| CN115170253A (en) * | 2022-09-07 | 2022-10-11 | 国连科技(浙江)有限公司 | Method and device for pushing commodity information in sales promotion activities |
-
2022
- 2022-04-14 CN CN202210395910.9A patent/CN114693368A/en active Pending
Patent Citations (32)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050004819A1 (en) * | 2003-03-27 | 2005-01-06 | Oren Etzioni | Performing predictive pricing based on historical data |
| US7437323B1 (en) * | 2003-06-25 | 2008-10-14 | Pros Revenue Management; L.P. | Method and system for spot pricing via clustering based demand estimation |
| US20050240539A1 (en) * | 2004-04-22 | 2005-10-27 | Thomas Olavson | Method and system for forecasting commodity prices using capacity utilization data |
| US20110040656A1 (en) * | 2009-08-12 | 2011-02-17 | Groetzinger Jon D | System and method for generating predictions of price and availability of event tickets on secondary markets |
| JP2012003677A (en) * | 2010-06-21 | 2012-01-05 | Nippon Telegr & Teleph Corp <Ntt> | Commodity recommendation device, commodity recommendation method and commodity recommendation program |
| CN102324082A (en) * | 2011-09-08 | 2012-01-18 | 上海烟草集团有限责任公司 | Cigarette market price prediction method based on multiple linear regression |
| JP2015041121A (en) * | 2013-08-20 | 2015-03-02 | 株式会社日立製作所 | Sales forecast system and sales forecast method |
| CN104866474A (en) * | 2014-02-20 | 2015-08-26 | 阿里巴巴集团控股有限公司 | Personalized data searching method and device |
| CN104217355A (en) * | 2014-09-11 | 2014-12-17 | 北京京东尚科信息技术有限公司 | Method and device for predicting sales volume of promotion items |
| JP2016133816A (en) * | 2015-01-15 | 2016-07-25 | 株式会社そごう・西武 | Commodity demand prediction system |
| CN105243563A (en) * | 2015-10-21 | 2016-01-13 | 百度在线网络技术（北京）有限公司 | Commodity price predicting method and apparatus |
| CN105869019A (en) * | 2016-03-31 | 2016-08-17 | 金蝶软件（中国）有限公司 | Method and apparatus for predicting goods price |
| JP2019028816A (en) * | 2017-08-01 | 2019-02-21 | 株式会社オープンソース活用研究所 | Prospective customer information price determination system, price determination method, server device, and program |
| CN110276652A (en) * | 2018-03-14 | 2019-09-24 | 北京京东尚科信息技术有限公司 | Method and apparatus for pushed information |
| CN109214893A (en) * | 2018-08-31 | 2019-01-15 | 深圳春沐源控股有限公司 | Method of Commodity Recommendation, recommender system and computer installation |
| CN110276495A (en) * | 2019-06-26 | 2019-09-24 | 东莞市盟大塑化科技有限公司 | Commercial articles ordering method, apparatus, computer and storage medium based on artificial intelligence |
| CN110517072A (en) * | 2019-08-14 | 2019-11-29 | 平安科技（深圳）有限公司 | Method for pushing, device, equipment and the computer readable storage medium of information of vehicles |
| CN111127149A (en) * | 2019-12-23 | 2020-05-08 | 卓尔智联(武汉)研究院有限公司 | Bulk commodity supply information recommendation method and device and computer equipment |
| CN111178972A (en) * | 2019-12-30 | 2020-05-19 | 北京每日优鲜电子商务有限公司 | Message pushing method and device, storage medium and equipment |
| CN111681099A (en) * | 2020-06-03 | 2020-09-18 | 中国银行股份有限公司 | Product information pushing method and device, computer equipment and readable storage medium |
| CN111899063A (en) * | 2020-06-17 | 2020-11-06 | 东南大学 | Fresh agricultural product online recommendation method considering customer consumption behaviors and preference |
| CN111709813A (en) * | 2020-06-19 | 2020-09-25 | 深圳市赛宇景观设计工程有限公司 | Commodity recommendation method based on big data line |
| CN111680221A (en) * | 2020-08-11 | 2020-09-18 | 腾讯科技（深圳）有限公司 | Information recommendation method, device, equipment and computer readable storage medium |
| KR20220022544A (en) * | 2020-08-19 | 2022-02-28 | 이니그마(주) | Price adjustment system based on customer demand |
| CN113298550A (en) * | 2020-08-21 | 2021-08-24 | 阿里巴巴集团控股有限公司 | Commodity price adjusting method and device and server |
| CN112132618A (en) * | 2020-09-23 | 2020-12-25 | 杭州拼便宜网络科技有限公司 | Commodity price determining method, device and equipment and readable storage medium |
| CN113157752A (en) * | 2021-03-12 | 2021-07-23 | 北京航空航天大学 | Scientific and technological resource recommendation method and system based on user portrait and situation |
| CN113269618A (en) * | 2021-06-10 | 2021-08-17 | 广州富港生活智能科技有限公司 | Optimal price display method and system based on selection preference and readable storage medium |
| CN113379516A (en) * | 2021-08-12 | 2021-09-10 | 永正信息技术(南京)有限公司 | Recommended product determination method and device |
| CN114331594A (en) * | 2021-12-14 | 2022-04-12 | 芸豆数字科技有限公司 | Commodity recommendation method |
| CN114240556A (en) * | 2021-12-17 | 2022-03-25 | 支付宝(杭州)信息技术有限公司 | Information pushing method and device and electronic equipment |
| CN115170253A (en) * | 2022-09-07 | 2022-10-11 | 国连科技(浙江)有限公司 | Method and device for pushing commodity information in sales promotion activities |
Non-Patent Citations (10)
| Title |
|---|
| 人人都是产品经理社区: "分享每一种值：什么值得买产品分析", Retrieved from the Internet <URL:https://www.163.com/dy/article/G7V5Q63O0511805E.html> * |
| 孙文存: "基于搜索关键词关注度的中国股票市场波动研究", 万方学位论文, 25 December 2012 (2012-12-25) * |
| 尹秋菊;闵立辉;金丹;: "面向客户行为的C2C市场价格离散演化模型", 北京理工大学学报, no. 09, 15 September 2011 (2011-09-15) * |
| 张鹏翼;王丹雪;焦凡;陈秀雨;王军;: "基于用户浏览日志的移动购买预测研究", 数据分析与知识发现, no. 01, 25 January 2018 (2018-01-25) * |
| 李东进 等: "《管理学原理》", 31 August 2014, 中国发展出版社, pages: 175 - 183 * |
| 李家华: "基于大数据的人工智能跨境电商导购平台信息个性化推荐算法", 科学技术与工程, no. 14, 5 August 2019 (2019-08-05) * |
| 潘文婧;赵安平;王大山;王晓东;肖金科;: "组合预测模型在畜产品价格预测中的应用研究", 中国食物与营养, no. 11, 28 November 2016 (2016-11-28) * |
| 猫哥清华新传考研: "【每日干货】新媒体研究方法", pages 10 - 12, Retrieved from the Internet <URL:https://zhuanlan.zhihu.com/p/271111693> * |
| 王茜;喻继军;: "基于价格消费习惯的商品多样性推荐", ***工程, no. 03, 28 March 2018 (2018-03-28) * |
| 谢意 等: "基于浏览偏好挖掘的实时商品推荐方法", 计算机应用, vol. 31, no. 01, 31 January 2011 (2011-01-31) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8117061B2 (en) | System and method of using demand model to generate forecast and confidence interval for control of commerce system | |
| US20210158404A1 (en) | Method, Apparatus, And Computer Program Product For Facilitating Dynamic Pricing | |
| Berry et al. | Probabilistic forecasting of heterogeneous consumer transaction–sales time series | |
| US8447664B1 (en) | Method and system for managing inventory by expected profitability | |
| US20230410135A1 (en) | System and method for selecting promotional products for retail | |
| US8224688B2 (en) | System and method for disaggregating weekly forecast into daily components | |
| US7881986B1 (en) | Method and system for event-driven inventory disposition | |
| US8027897B2 (en) | System and method for optimizing financial performance generated by marketing investments under budget constraints | |
| US20020046128A1 (en) | Automatic pricing method and device | |
| JP6885356B2 (en) | Recommended information identification device, recommended information identification system, recommended information identification method, and program | |
| US20100106555A1 (en) | System and Method for Hierarchical Weighting of Model Parameters | |
| US9715702B2 (en) | System and method for determining retail-business-rule coefficients from current prices | |
| US20130325554A1 (en) | Commerce System and Method of Optimizing Profit for Retailer from Price Elasticity of Other Retailers | |
| US20140172502A1 (en) | Consumer walker reports | |
| JP2018139036A (en) | Analysis device | |
| WO2022247741A1 (en) | Commodity object price information processing method and apparatus, and electronic device | |
| Joo | Rational inattention as an empirical framework: Application to the welfare effects of new-product introduction | |
| CN111768243A (en) | Sales prediction method, prediction model construction method, device, equipment and medium | |
| JP2005084970A (en) | Method and system for managing retail price | |
| CN115147144A (en) | Data processing method and electronic equipment | |
| Salvietti et al. | A profit-maximizing economic lot scheduling problem with price optimization | |
| KR102499689B1 (en) | Method for automatic screening of search keywords | |
| CN114693368A (en) | Behavior data-based customer maintenance method and device and storage medium | |
| CN114757703A (en) | Method and device for replenishing traditional Chinese medicine commodities, electronic equipment and storage medium | |
| JP2007503651A (en) | Manufacture of item units according to the demand of items expected from page view data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |