CN115936757A - Method and device for predicting psychological price distribution of bidders - Google Patents
Method and device for predicting psychological price distribution of bidders Download PDFInfo
- Publication number
- CN115936757A CN115936757A CN202211684743.6A CN202211684743A CN115936757A CN 115936757 A CN115936757 A CN 115936757A CN 202211684743 A CN202211684743 A CN 202211684743A CN 115936757 A CN115936757 A CN 115936757A
- Authority
- CN
- China
- Prior art keywords
- price
- auction
- price distribution
- commodity
- historical
- 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.)
- Granted
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the invention discloses a method and a device for predicting the psychological price distribution of an auction player, wherein the method comprises the following steps: acquiring historical auction data, wherein the historical auction data comprises first commodity characteristic information and historical price distribution information; acquiring second commodity characteristic information of a commodity to be subjected to competitive shooting; and analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on the nonparametric Bayesian model to obtain price distribution prediction data. According to the invention, the historical auction data and the second commodity characteristic information can be converted into a probability distribution form by using the nonparametric Bayesian model, so that the observation problem of the psychological price is solved, and the quantitative auction participant psychological price distribution condition can be predicted through data analysis, thereby reducing the experience requirements of auction strategy makers in the auction process and enabling the auction strategy to be easier to quantify.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a method and a device for predicting the psychological price distribution of an auction player.
Background
The auction is a spot transaction mode that an auction company specially engaged in auction service accepts the entrustment of a buyer, displays the goods to be auctioned to the buyer according to a certain rule and rule at a specified time and place, publicly calls for a bid to bid, and finally the auctioneer sells the goods to the buyer with the highest bid. In an e-commerce shopping environment, network auctions can provide greater convenience to both the owners of goods and the bidders.
However, for the part of the owner or the auction platform, since the psychological price of the bidder cannot be directly observed, the owner or the auction platform can only set the auction strategy by experience, on one hand, the experience of the strategy maker is required to be high, and on the other hand, the auction strategy is difficult to quantify. Aiming at the problems that the experience requirements of auction strategy makers are high and the auction strategies are difficult to quantify in the existing auction process, further research on the technology for predicting the psychological price distribution of the auction makers needs to be carried out.
Disclosure of Invention
The embodiment of the invention provides a method and a device for predicting the psychological price distribution of an auction user, which are used for solving the problems that the experience requirement on auction strategy makers in the existing auction process is high and the auction strategy is difficult to quantify.
In one aspect, the invention provides a method for predicting the psychological price distribution of an auction player, which comprises the following steps:
acquiring historical auction data, wherein the historical auction data comprises first commodity characteristic information and historical price distribution information;
acquiring second commodity characteristic information of a commodity to be subjected to competitive shooting;
and analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on a nonparametric Bayesian model to obtain price distribution prediction data.
Further, the non-parametric bayesian model comprises a Polya Tree structure model, which is expressed as:
PT(Π,A);
wherein A represents a hyper-parameter and Π represents an interval parameter.
Further, the Polya Tree structure model encodes the interval in a binary segmentation form, the segmentation point is determined by an interval parameter pi, and pi is defined as:
Π=(B 0 ,B 1 ,B 00 ,B 01 ,...);
the Tree structure of the Polya Tree structure model is as follows:
the hyper-parameter a determines the random probability assigned to each partition set:
A=(a 0 ,a 1 ,a 00 ,a 01 ,...);
introducing a set of sets ε = ε 1 ...ε m ,ε i =0or1, random conditional probabilities are assigned to the set:
i.P(B ε0 |B ε )=C ε0 each of C ε0 Random variable, C, from a Beta distribution ε0 ~Beta(a ε0 |,a ε1 );
iiP(Bζ 1 、Bε)=Cε 1 =1-C ε0 ;
iii.C ε 0 are all independent;
get set B ε The probability of epsilon is:
further, the initial value of the hyper-parameter a is:
a ε =τ z R(B ε );
wherein, tau z Is a scalar greater than zero, z is the depth of epsilon, R is the basis measure, and R can be assumed to be a uniform distribution when the prior of V cannot be estimated.
Further, the historical price distribution information includes price distribution information of failed bidders, price distribution information of successful bidders, and price distribution information of non-bidding bidders, and the Polya Tree model includes price intervals of failed bidders, successful bidder, and non-bidding bidders, and
the price interval of the failed auction participants is as follows:
the price interval of successful bidders is:
the price interval of the bidder not bidding is as follows:
further, the step of analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on the nonparametric Bayesian model to obtain price distribution prediction data comprises the following steps:
analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on a PT-RNN recurrent neural network to obtain price distribution prediction data; the PT-RNN recurrent neural network comprises: input layer, output layer, hidden layer, training set label and loss function.
Further, the training process of the PT-RNN recurrent neural network comprises the following steps:
dividing the first commodity characteristic information and the historical auction data into a training set and a testing set;
inputting the training set into a PT-RNN cyclic neural network to be trained to train the PT-RNN cyclic neural network to be trained, so as to obtain the PT-RNN cyclic neural network to be tested;
and testing the PT-RNN recurrent neural network to be tested by using the test set, and if the test is passed, obtaining the PT-RNN recurrent neural network.
Further, after the step of analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on the non-parameter bayesian model to obtain the price distribution prediction data, the method further comprises the following steps:
obtaining the yield or the predicted sales volume of a product to be competitive;
obtaining supply chain cost data of the to-be-auction product according to the yield or the predicted sales volume;
and obtaining the optimal pricing data of the to-be-auction item based on the profit function, the supply chain cost data and the price distribution prediction data.
Further, the relationship of the profit function is:
where c is the supply chain cost of the commodity, x is commodity pricing,representing a psychological price compliance parameter being->Distribution of (2).
In order to solve the problems that the experience requirements of auction strategy makers are high and the auction strategies are difficult to quantify in the existing auction process, the invention aims to provide a method for predicting the psychological price distribution of auctioneers, which comprises the following steps: acquiring historical auction data, wherein the historical auction data comprises first commodity characteristic information and historical price distribution information; acquiring second commodity characteristic information of a commodity to be subjected to competitive shooting; and analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on a nonparametric Bayesian model to obtain price distribution prediction data.
Compared with the prior art, the invention has the following advantages:
because the historical auction process can reflect the distribution interval of the psychological price of the auction participants, the information can be converted into a probability distribution form by using a nonparametric Bayesian model, the observation problem of the psychological price is solved, and the quantitative distribution condition of the psychological price of the auction participants can be predicted through data analysis, so that the experience requirements of auction strategy makers in the auction process are reduced, and the auction strategy is easier to quantify.
In another aspect, the present invention provides an apparatus for predicting a psychological price distribution of an auctioneer, comprising:
the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring historical auction data, and the historical auction data comprises first commodity characteristic information and historical price distribution information;
the second acquisition module is used for acquiring the characteristic information of a second commodity of the article to be auctioned;
and the analysis module is used for analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on the nonparametric Bayesian model to obtain price distribution prediction data.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments of the present invention will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without inventive labor.
Fig. 1 is a flowchart of a method for predicting a psychological price distribution of an auctioneer according to an embodiment of the present invention.
Fig. 2 is a structural diagram of a Polya Tree model according to another embodiment of the present invention.
Fig. 3 is another structural schematic diagram of a Polya Tree model according to another embodiment of the present invention.
FIG. 4 is a diagram illustrating the result of the PT-RNN recurrent neural network according to still another embodiment of the present invention.
Fig. 5 is a diagram illustrating a forward propagation process of a PT-RNN recurrent neural network according to still another embodiment of the present invention.
Fig. 6 is a schematic diagram illustrating the operation of a PT-RNN recurrent neural network according to still another embodiment of the present invention.
Fig. 7 is a schematic view of an auctioneer's psychological price distribution predicting device according to yet another embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without inventive step based on the embodiments of the present invention, are within the scope of protection of the present invention.
Referring to fig. 1, a method for predicting a psychological price distribution of an auctioneer according to an embodiment of the present invention includes the following steps:
In the embodiment of the invention, the commodity characteristics comprise various parameters of commodities listed in the auction website, for example, the characteristic information of jewelry commodities comprises (production place, designer/design organization, material type, material grade/material attribute, appearance, size and weight … …).
And 102, acquiring second commodity characteristic information of a to-be-photographed commodity.
And 103, analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on the nonparametric Bayesian model to obtain price distribution prediction data.
In order to solve the problems that the experience requirements of auction strategy makers are high and the auction strategies are difficult to quantify in the existing auction process, the invention aims to provide a method for predicting the psychological price distribution of auction strategies.
Each auction participant has a judgment value for the auction products, and forms a psychological price based on the judgment value, namely the psychological price is the price subjectively given by the auction participants to one auction product. The psychological prices of all the participants who participate in a certain auction can form a psychological price distribution, but the psychological price can not be directly observed, so the method for predicting the psychological price distribution of the participants can predict the psychological price distribution of the participants according to the relevant characteristics of the products to be auctioned.
As another implementation manner of the foregoing embodiment, the non-parametric bayesian model includes a Polya Tree structure model, which is expressed as:
PT(Π,A);
wherein A represents a hyper-parameter and Π represents an interval parameter.
The psychological price of the bidder cannot be directly observed generally, but the auction process can reflect the distribution interval of the psychological price, and the information can be converted into a probability distribution form by using a PolyaTree structure, so that the problem of observing the psychological price is solved. In the embodiment of the invention, the psychological price distribution of the auction participants is described by a PolyaTree model which consists of an interval parameter Π and a hyperparameter A and is represented as PT (II, A). The Polya Tree is a process for generating a random probability measure distribution, and is an extension of the Dirichlet process and the Tailfree process. The Dirichlet process, which generates a discrete set of random probability measures with a probability of 1, tends to produce random results with poor fitness experimentally, and the Polya Tree can generate a continuous or absolutely continuous set of probability measures.
In the embodiment of the present invention, the Polya Tree structure model is a binary Tree structure, and the interval is coded in a binary partition (dynamic partition) form, as shown in fig. 2, the partition point is determined by a parameter Π, and is defined as:
Π=(B 0 ,B 1 ,B oo ,B o1 ,...)(1)
the tree structure is:
the hyper-parameter a determines the random probability assigned to each partition set, where:
A=(a 0 ,a 1 ,a 00 ,a 01 ,...)(3)
introducing a set of sets ε = ε 1 …ε m ,ε i =0or1, the polya Tree distribution assigns random conditional probabilities to such a set:
i.P(B ε0 |B ε )=C ε0 each of C ε0 The random variable from one of the Beta distributions,
C ε0 ~Beta(a ε0 ,a ε1 )
ii.P(B ε1 B ε )=C ε1 =1-C ε0
iii.C ε0 are all independent
Thus set B ε The probability of the epsilon II is as follows:
for example:
P(B 001 |A)=C 0 C 00 C 001 (5)
the iterative process of the model can be obtained according to the property of the posterior distribution of the Beta distribution:
therefore, if F is to be θ (v) Is regarded as a Polya Tree distribution, and the posteriori is also a Polya Tree distribution, and iteration can be performed by updating the hyperparameter a. It should be noted that the above process needs to be calculated once in each level interval, for example, when there is an observation result v ∈ B 10 Then, corresponding to a 1 And a10 needs to be updated to a respectively 1 +1 and a 10 +1, since the observation also satisfies v ∈ B 1 . The learning effect of the Polya Tree model is determined by the hyperparameter A, and the influence proportion of new observation is gradually reduced as the number of observation samples increases (a is larger).
Regarding the initial value of the hyper-parameter a. Let R be the prior estimate of V, and refer to R as the base measure (base measure), and when the prior estimate of V cannot be estimated, it can be assumed that R is a uniform distribution. The initial value of a can be obtained by the following equation.
a ε =τ z R(B ε ) (7)
Wherein, tau z Is a scalar quantity greater than zero and z refers to the depth of epsilon, e.g., when epsilon =001, then theta =3. The characteristics of the Polya Tree structural model are known, tau z The larger the value of (b), the more approximate V is to the shape of R, and the smaller the correction effect of the new observed value on V. Therefore, if the shape of R is certain (e.g., an expert is engaged in making an empirical estimate a priori), a larger τ can be set z The value should be as small as possible, otherwise.
The important link of the Polya Tree structure is to construct a binary Tree model, and the discrete characteristics of the discrete price model are utilized to construct the Polya Tree structure for the price sequence.
Let Π = (B) 0 ,B 1 ,B 00 ,B 01 ,..) b is divided as follows:
if the observation points of the auction bid are used as the partition points to construct the Polya Tree structure, the Tree structure needs to be updated again each time new data enters, although the complexity of the Tree is relatively simple in the initial stage, and the partition and the assignment need to be re-divided every iteration.
In the embodiment of the present invention, the historical price distribution information includes price distribution information of failed bidders, price distribution information of successful bidders, and price distribution information of non-bidding bidders, and the Polya Tree model includes price intervals of failed bidders, successful bidder, and non-bidding bidders, and
the price interval of the failed bidder is as follows:
the price interval of successful bidders is:
the price interval of the bidder not bidding is as follows:
in particular, any quote that is observed (one observation is counted for no quote)Three types of auction-type persons whose corresponding +>The interval of (a):
the third category is the non-bidding bidder, and if the data is directly deleted, the estimation of the final bidder psychological price distribution is biased. Because the part of the data contains at least one information, i.e. a part of the samples is smaller thanMay have its true v value present->At any point in between.
In practical consideration, obviously, the auction participants do not consider the value of the auction product to be 0, otherwise, the auction participants do not participate in the auction, so that the psychological price interval of the part of the population can be set asThe effect of retaining this portion of the auction data on the population is to increase the likelihood that a prediction falls ≧ or>While lowering the probability ofLow-falling on/off>This is also one of the important advantages of the model, namely to retain as much information of the data as possible (even if this data has never been observed directly).
Therefore, the psychological prices of the third type of participants are:
in combination with the above formulas (4), (9) to (11), with respect to anyThe probability of (c) can be given by:
When b is n About ∞ times
Wherein, C ε Is a random variable obeying a Beta distribution and C ε0 Are all independent and can therefore be calculated according to the following formulaNamely, it is
The posterior probability distribution of the psychological price of the first and third types of participants, i.e. how to obtain the posterior predicted distribution of new data in practical application, is described belowWhere Data represents the acquired bid Data.
When it is satisfied withThe expression of E (V | Data) can be obtained according to the nonparametric Bayes estimation formula (namely the psychological price of the first and third types of auction participants):
wherein N is ε Indicates that Data falls in the section B ε Number of samples in. Of particular note, in this model all observed samples would not fall within interval B ε0 . Let a epsilon = a ε1 +a ε。 =τR(B ε ) I.e. to simplify the model into a Dirichlet process, the above equation can be written as:
when b is n About ∞ is:
for the initial state of the model, pi in PT (pi, A) is represented by: (pi, A) 1 ) The formula is constructed and is fixed, and the only consideration is the value of A, which is defined i For parameters A, A updated after the ith round of auction 0 Is in an initial state.
The posterior probability distribution of the psychological price of the second type of auction is described below.
For each auction activity, the psychological price of one and only one auctioneer (successful auctioneer) is in the intervalAnd (4) the following steps. If N auction participants participate in the auction of a certain commodity, except for successful auction participants, the psychological price of N-1 auction participants is fallen into ^ er>Therefore, the posterior probability distribution graph should present a significant light tail distribution, and the light tail is more obvious when the N value is larger. However, the posterior probability of the above-mentioned Polya Tree model has a thick tail problem, and the probability is accumulated in ^ 4>In the interval, the probability measurement value of the interval needs to be corrected. In order to obtain +>Another Polya Tree model can be constructed. To distinguish between the previous partition Π, a new partition is defined as Π '= (B' 0 ,B′ 1 ,B′ 00 ,B′ 01 …, the binary tree structure of this model is shown in fig. 3:
in the context of figure 3, it is shown,the posterior probability of this interval can be found to be:
wherein R (B' 1 ) Is a base measure and can be considered to be a prior probability measure and is therefore a constant, the posterior probability of this interval being only inversely related to the value of N.
In one auction combining the (17) and (21), the posterior probability distribution of the psychological prices of the participants is:
from the practical application perspective, the psychological price of the auction player falls into the interval [ b L And ∞) does not affect the final result, so the tail probability distribution is processed as follows:
P(v∈[b L ,b L+1 )|v≥b)=1 (23)
as another implementation manner of the foregoing embodiment, the step of analyzing the first commodity feature information, the historical price distribution information, and the second commodity feature information based on the non-parametric bayesian model to obtain the price distribution prediction data includes:
analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on the PT-RNN recurrent neural network to obtain price distribution prediction data; the PT-RNN recurrent neural network comprises: input layer, output layer, hidden layer, training set label and loss function.
According to the embodiment of the invention, the psychological price distribution of potential demanders of a certain characteristic commodity can be obtained by analyzing the historical auction data, and then the relationship between the commodity characteristic and the psychological price distribution of the potential demanders is learned by utilizing a deep learning technology.
Referring to fig. 4, a jewelry item is taken as an example, and a ring ornament is selected by way of example. According to auction information and artificial labels provided by an auction website, selecting the following commodity characteristics: brand awareness, material, place of origin, quality, appearance pictures, design connotations, IP information, and the like. Accordingly, the relationship between the commodity characteristics and the psychological price distribution of the potential demanders is learned.
(1) Brand awareness: the brand awareness depends on the average daily search times (web search popularity) of the search engine.
(2) And (3) material quality. The materials are divided into main materials and non-main materials, and the main materials of the data are classified as follows according to the collected auction information: nephrite jade, amber, emerald, agate, crystal, tourmaline, pearl, diamond, garnet, qingjin stone, other natural jades; the non-main materials are as follows: gold, silver, other metals or materials.
(3) The place of origin. The origin is the design origin of the product, not the production origin.
(4) And (4) quality. The mass here corresponds to the mass of two types of materials.
(5) And (5) appearance pictures. Auction product picture information provided by the auction website.
(6) The design implications. The characteristics mainly refer to the connotation expressed by the ornament, and the theme of the connotation is divided into: love, family, health (peace), romance, personality and no special connotation.
(7) And IP information. The IP (Intellectual Property) information refers to other subject information included in the design of the jewelry, such as stars, animation, movies, and the like. The part is mainly judged in an artificial subjective mode, a scoring mode of 0-9 is adopted, and scoring (network searching heat degree) is carried out according to the IP influence from low to high, wherein the score of 0 does not contain any additional IP information.
The training process of the PT-RNN recurrent neural network comprises the following steps:
dividing the first commodity characteristic information and the historical auction data into a training set and a testing set;
inputting the training set into a PT-RNN recurrent neural network to be trained to train the PT-RNN recurrent neural network to be trained, so as to obtain the PT-RNN recurrent neural network to be tested;
and testing the PT-RNN recurrent neural network to be tested by using the test set, and if the test is passed, obtaining the PT-RNN recurrent neural network.
And when the test result shows that the similarity between the predicted psychological price distribution and the actual psychological price distribution exceeds the index threshold, the test is regarded as passed.
Specifically, the PT-RNN model is a recurrent neural network model of a time-based back propagation algorithm, and a sequence architecture of the poly a Tree binary segmentation is built on the basis of the recurrent neural network model. The structure and learning process of the PT-RNN model are described below.
(1) And (3) sequence structure. As shown in FIG. 5, the sequence structure adopted by PT-RNN is binary split. At the far left end of the sequence, the RNN calculates first the proportion of consumers with a psychological price greater than b1, and then this information is passed on to the next stage, i.e. the proportion of consumers with a psychological price greater than b2, among consumers with a psychological price greater than b1, and so on. When the sequence reaches bL, in order to simplify the model, b is added on the premise of not influencing pricing decision L+1 The probability measures of the subsequent intervals are all set to be 0.
( 2 ) Input and output. Input feature x (t) The composition of (1) comprises two parts, which are composed of commodity characteristics and price intervals. Merchandise features include, but are not limited to, raw material features, design features, and the like. Input feature x (t) Is represented by (f) i ,b( t) ),f i I.e. the feature vector of the commodity i, b is the price, and t represents the current stage.
Output o (t) Expressed as probabilities, therefore (f) i ,b (t) )→o (t) It can be understood that in the stage t, the price is predicted to be more than b according to the characteristics of the commodity (t) Has a conditional probability of o (t) . The price interval is divided into binary divisions, i.e. pi = (B) 0 ,B 1 ,B 00 ,B 01 ,...). Finally according to ( 13 ) The formula calculates the probability measure value of each unit price interval.
( 3 ) And hiding the layer. RNN delivers information through a hidden layer, hidden layer h (t) Indicating a price greater than b (t) This information, the letterInformation will be passed to the next layer, formulated as:
o (t) =P(v≥b (t) |v≥b (t-1) (24)
(4) And (5) training set labels. y is (t) The label is a training set label, the label is a judgment standard of the quality of an RNN output result, in this text, the true probability measures distributed in different price intervals are measured, and the result is obtained by calculation of a PT model. As shown in the training label module in fig. 5, the probability measure for the price per unit interval is given by equation (17).
(5) A loss function. The penalty function L describes the difference between the label and the model training result. As shown in fig. 6, the predicted distribution of RNNs is compared with the true distribution, and the BPTT algorithm is used to perform reverse propagation and modify parameters, thereby realizing deep learning.
The purpose of the present disclosure is to make the distribution of RNN prediction consistent with the actual distribution, so the loss function needs to represent the difference or distance between the two distributions, and the learning purpose is realized by continuously decreasing the distance. Therefore, the loss function needs to satisfy L (P, Q) ≧ 0, that is, the loss function has a minimum value, and when the value is 0, the two distributions are completely consistent.
Specifically, the loss function takes the Hellinger distance:
where P represents the predicted psychological price distribution and Q represents the auction price distribution.
As another implementation manner of the above embodiment, after the step of analyzing the first product characteristic information, the historical price distribution information, and the second product characteristic information based on the non-parametric bayesian model to obtain the price distribution prediction data, the method further includes:
obtaining the yield or the expected sales volume of a product to be competitive;
obtaining supply chain cost data of the to-be-auction product according to the yield or the predicted sales volume;
and obtaining the optimal pricing data of the to-be-auction item based on the profit function, the supply chain cost data and the price distribution prediction data.
The embodiment predicts the psychological price distribution of potential demanders by using the commodity characteristics and provides a basis for commodity pricing.
Further, the relationship of the profit function is:
where c is the supply chain cost of the commodity, x is commodity pricing,representing a psychological price obeying parameter of->Distribution of (2).
Referring to fig. 7, a device for predicting the mental price distribution of an auction player according to another embodiment of the present invention can be used to implement the corresponding steps of the above method embodiment, including:
the first acquisition module is connected with the analysis module and used for acquiring historical auction data, and the historical auction data comprises first commodity characteristic information and historical price distribution information;
the second acquisition module is connected with the analysis module and used for acquiring the second commodity characteristic information of the item to be auctioned;
and the analysis module is used for analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on the nonparametric Bayesian model to obtain price distribution prediction data.
In yet another aspect, the present invention provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described above.
In a further aspect, the invention provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, is operable to perform a method as described above.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
It can be clearly understood by those skilled in the art that, for convenience and simplicity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A method for predicting the psychological price distribution of an auction player is characterized by comprising the following steps:
acquiring historical auction data, wherein the historical auction data comprises first commodity characteristic information and historical price distribution information;
acquiring second commodity characteristic information of a to-be-photographed commodity;
analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on a nonparametric Bayesian model to obtain price distribution prediction data.
2. The method for predicting the psychological price distribution of auction participants according to claim 1, wherein said nonparametric bayes model comprises a Polya Tree structure model expressed as:
PT(Π,A);
wherein A represents a hyper-parameter and Π represents an interval parameter.
3. The method according to claim 2, wherein the Polya Tree structure model encodes an interval in a binary division form, the division point is determined by the interval parameter Π, and is defined as:
Π=(B 0 ,B 1 ,B 00 ,B 01 ,...);
the Tree structure of the Polya Tree structure model is as follows:
the hyper-parameter a determines the random probability assigned to each partition set:
A=(a 0 ,a 1 ,a 00 ,a 01 ,...);
introducing a set of sets ε = ε 1 ...ε m ,ε t =0or1, random conditional probabilities are assigned to such sets:
ⅰ.P(B ε0 |B ε )=C ε0 each of C ε0 Random variable, C, from a Beta distribution ε0 ~Beta(a ε0 ,a ε1 );;
ⅱ.P(B ε1 |B ε )=C ε1 =1-C ε0 ;
ⅲ.C ε0 Are all independent;
get set B ε The probability of epsilon is:
4. the method for predicting the mental price distribution of an auction player of claim 3, wherein the initial values of the hyper-parameter A are as follows:
a ε =τ z R(B ε );
wherein, tau z Is a scalar greater than zero, z is the depth of epsilon, R is the basis measure, and R can be assumed to be a uniform distribution when the prior of V cannot be estimated.
5. The method according to claim 4, wherein the historical price distribution information includes price distribution information of failed bidders, price distribution information of successful bidders, and price distribution information of non-bidding bidders, the Polya Tree model includes price intervals of failed bidders, successful bidders, and non-bidding bidders, and the Polya Tree model includes price intervals of failed bidders, successful bidders, and non-bidding bidders
The price interval of the failed bidder is as follows:
the price interval of the successful bidder is as follows:
the price interval of the bidder not bidding is as follows:
6. the method for predicting the psychological price distribution of the bidder according to claim 1, wherein the step of analyzing the first commodity feature information, the historical price distribution information and the second commodity feature information based on a nonparametric bayesian model to obtain price distribution prediction data comprises:
analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on a PT-RNN recurrent neural network to obtain price distribution prediction data; the PT-RNN recurrent neural network includes: input layer, output layer, hidden layer, training set label and loss function.
7. The method for predicting the mental price distribution of an auction player according to claim 6, wherein the training process of the PT-RNN recurrent neural network comprises:
dividing the first commodity feature information and the historical auction data into a training set and a testing set;
inputting the training set into a PT-RNN circulating neural network to be trained to train the PT-RNN circulating neural network to be trained, so as to obtain the PT-RNN circulating neural network to be tested;
and testing the PT-RNN recurrent neural network to be tested by using the test set, and if the test is passed, obtaining the PT-RNN recurrent neural network.
8. The method for predicting the psychological price distribution of the bidder according to claim 1, wherein after the step of analyzing the first commodity feature information, the historical price distribution information and the second commodity feature information based on the non-parametric bayesian model to obtain price distribution prediction data, the method further comprises:
obtaining the yield or the predicted sales volume of the item to be competitive-photographed;
obtaining supply chain cost data of the item to be auctioned according to the yield or the predicted sales volume;
obtaining optimal pricing data of the item to be auctioned based on a profit function, the supply chain cost data, and the price distribution forecast data.
10. A device for predicting a psychological price distribution of an auctioneer, comprising:
the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring historical auction data, and the historical auction data comprises first commodity characteristic information and historical price distribution information;
the second acquisition module is used for acquiring the characteristic information of a second commodity of the article to be auctioned;
and the analysis module is used for analyzing the first commodity characteristic information, the historical price distribution information and the second commodity characteristic information based on a nonparametric Bayesian model to obtain price distribution prediction data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211684743.6A CN115936757B (en) | 2022-12-27 | 2022-12-27 | Method and device for predicting psychological price distribution of bidders |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211684743.6A CN115936757B (en) | 2022-12-27 | 2022-12-27 | Method and device for predicting psychological price distribution of bidders |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115936757A true CN115936757A (en) | 2023-04-07 |
CN115936757B CN115936757B (en) | 2023-09-22 |
Family
ID=86650604
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211684743.6A Active CN115936757B (en) | 2022-12-27 | 2022-12-27 | Method and device for predicting psychological price distribution of bidders |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115936757B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116934438A (en) * | 2023-04-14 | 2023-10-24 | 济南明泉数字商务有限公司 | AI auction decision method and system based on chatGPT model and calculation force |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110213644A1 (en) * | 2008-05-05 | 2011-09-01 | Kaustubha Phene | Consumer goods price prediction and optimization |
CN109064212A (en) * | 2018-07-04 | 2018-12-21 | 苏州仙度网络科技有限公司 | Price forecasting of commodity method and device |
CN111275117A (en) * | 2020-01-20 | 2020-06-12 | 金华航大北斗应用技术有限公司 | Bayes-based data fluctuation prediction method and system |
CN115423576A (en) * | 2022-08-12 | 2022-12-02 | 深圳大学 | Auction parameter dynamic optimization method for digital auction platform and computing terminal |
-
2022
- 2022-12-27 CN CN202211684743.6A patent/CN115936757B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110213644A1 (en) * | 2008-05-05 | 2011-09-01 | Kaustubha Phene | Consumer goods price prediction and optimization |
CN109064212A (en) * | 2018-07-04 | 2018-12-21 | 苏州仙度网络科技有限公司 | Price forecasting of commodity method and device |
CN111275117A (en) * | 2020-01-20 | 2020-06-12 | 金华航大北斗应用技术有限公司 | Bayes-based data fluctuation prediction method and system |
CN115423576A (en) * | 2022-08-12 | 2022-12-02 | 深圳大学 | Auction parameter dynamic optimization method for digital auction platform and computing terminal |
Non-Patent Citations (1)
Title |
---|
罗英;郑灿钊;: "基于博弈机制的网上双向拍卖报价模型研究", 现代管理科学, no. 07 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116934438A (en) * | 2023-04-14 | 2023-10-24 | 济南明泉数字商务有限公司 | AI auction decision method and system based on chatGPT model and calculation force |
Also Published As
Publication number | Publication date |
---|---|
CN115936757B (en) | 2023-09-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111080413A (en) | E-commerce platform commodity recommendation method and device, server and storage medium | |
CN107657488A (en) | Advertisement putting processing method and processing device based on advertisement matching | |
CN108665311B (en) | Electric commercial user time-varying feature similarity calculation recommendation method based on deep neural network | |
CN102902691A (en) | Recommending method and recommending system | |
CN113744017A (en) | E-commerce search recommendation method and device, equipment and storage medium | |
Tallis et al. | Reacting to variations in product demand: An application for conversion rate (cr) prediction in sponsored search | |
CN115936757A (en) | Method and device for predicting psychological price distribution of bidders | |
Adams | Estimating demand from eBay prices | |
CN112967108B (en) | BP-ANN-based power retail package search ordering simulation system and method | |
CN113763095B (en) | Information recommendation method and device and model training method and device | |
Dass et al. | Maximizing bidder surplus in simultaneous online art auctions via dynamic forecasting | |
Ikonomovska et al. | Real-time bid prediction using thompson sampling-based expert selection | |
CN114493724A (en) | Multi-task learning-based bidding keyword effect estimation model construction method | |
Wu et al. | Predicting long-term returns of individual stocks with online reviews | |
CN111782926B (en) | Method and device for data interaction, storage medium and electronic equipment | |
TWM600893U (en) | Product recommendation apparatus | |
Li et al. | Pricing strategies with promotion time limitation in online social networks | |
Liu et al. | Intelligent modeling of abnormal demand forecasting for medical consumables in smart city | |
Al-Basha | Forecasting Retail Sales Using Google Trends and Machine Learning | |
Dass et al. | Dynamic price forecasting in simultaneous online art auctions | |
Öz | Empirical studies on price determinants of online auctions with machine learning applications | |
Gomrokchi et al. | Adcraft: An advanced reinforcement learning benchmark environment for search engine marketing optimization | |
CN115147156B (en) | Advertisement accurate delivery method based on neural network | |
Sharma | Performance Prediction of Amazon Ad Campaigns Through Past Performance Data Using Machine Learning | |
Wang et al. | Survey of e-commerce modeling and optimization strategies |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |