CN111754262A - Pricing determination method, device, equipment and storage medium - Google Patents

Pricing determination method, device, equipment and storage medium Download PDF

Info

Publication number
CN111754262A
CN111754262A CN202010568617.9A CN202010568617A CN111754262A CN 111754262 A CN111754262 A CN 111754262A CN 202010568617 A CN202010568617 A CN 202010568617A CN 111754262 A CN111754262 A CN 111754262A
Authority
CN
China
Prior art keywords
predicted
pricing
model
term
data
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
Application number
CN202010568617.9A
Other languages
Chinese (zh)
Inventor
罗欣
李斓
雷笑雨
郭晓刚
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Concavoconvex Lexiang Suzhou Information Technology Co ltd
Original Assignee
Shanghai Xinwin Information Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shanghai Xinwin Information Technology Co ltd filed Critical Shanghai Xinwin Information Technology Co ltd
Priority to CN202010568617.9A priority Critical patent/CN111754262A/en
Publication of CN111754262A publication Critical patent/CN111754262A/en
Pending legal-status Critical Current

Links

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
    • G06Q30/0206—Price or cost determination based on market factors
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • 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
    • G06Q30/0645—Rental transactions; Leasing transactions

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Economics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Game Theory and Decision Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The embodiment of the invention discloses a pricing determining method, a pricing determining device, pricing determining equipment and a storage medium. The method comprises the following steps: acquiring data to be predicted and a target index of an object to be predicted; respectively inputting the data to be predicted into a pre-trained short-term income pricing prediction model, a short-term profit pricing prediction model and a long-term income pricing prediction model to respectively obtain the predicted pricing of the object to be predicted output by each model; and determining the target pricing of the object to be predicted based on the predicted pricing and the target index output by each model. The method has the advantages that reasonable and comprehensive pricing is carried out on the predicted objects, the individual requirements of the users are met, and the overall income is optimal.

Description

Pricing determination method, device, equipment and storage medium
Technical Field
The embodiment of the invention relates to a machine learning technology, in particular to a pricing determining method, a pricing determining device, pricing determining equipment and a storage medium.
Background
Pricing is an important link of a business process, one or combination of a cost pricing method or a competitive pricing method is generally adopted in traditional pricing to perform commodity pricing, then dynamic pricing for revenue management is evolved in the aviation and hotel industries, and price adjustment is performed according to a demand prediction and revenue maximization target.
The pricing method considers single factors, the demand and the price influence each other in the shared car renting platform, the primary goal of the shared car renting platform does not consider the single income of the product but needs to consider the income of the whole platform, and meanwhile, most of personal vehicles on the platform are independently priced, so that the environment has great uncertainty and complexity.
Disclosure of Invention
The embodiment of the invention provides a pricing determining method, a pricing determining device, equipment and a storage medium, so that reasonable comprehensive pricing is carried out on a predicted object, personalized requirements of a user are met, and the effect of optimizing the overall income is achieved.
In a first aspect, an embodiment of the present invention provides a pricing determining method, where the method includes:
acquiring data to be predicted and a target index of an object to be predicted;
respectively inputting the data to be predicted into a pre-trained short-term income pricing prediction model, a short-term profit pricing prediction model and a long-term income pricing prediction model to respectively obtain the predicted pricing of the object to be predicted output by each model;
and determining the target pricing of the object to be predicted based on the predicted pricing and the target index output by each model.
In a second aspect, an embodiment of the present invention further provides a pricing determining apparatus, where the apparatus includes:
the data acquisition module is used for acquiring data to be predicted and a target index of an object to be predicted;
the forecasting pricing obtaining module is used for respectively inputting the data to be forecasted into a pre-trained short-term income pricing forecasting model, a short-term profit pricing forecasting model and a long-term income pricing forecasting model to respectively obtain the forecasting pricing of the object to be forecasted output by each model;
and the target pricing determining module is used for determining the target pricing of the object to be predicted based on the predicted pricing and the target index output by each model.
In a third aspect, an embodiment of the present invention further provides an apparatus, where the apparatus includes:
one or more processors;
storage means for storing one or more programs;
when executed by the one or more processors, cause the one or more processors to implement the pricing determination method of any of the embodiments of the invention.
In a fourth aspect, embodiments of the present invention also provide a storage medium containing computer-executable instructions which, when executed by a computer processor, are configured to perform any of the pricing determination methods described in embodiments of the present invention.
According to the technical scheme of the embodiment of the invention, the obtained data to be predicted of the object to be predicted is respectively input into the pre-trained short-term income pricing prediction model, short-term profit pricing prediction model and long-term income pricing prediction model to respectively obtain the predicted pricing of the object to be predicted output by each model, and the target pricing of the object to be predicted can be determined according to the predicted pricing of the object to be predicted output by each model and the obtained target index of the object to be predicted, so that the reasonable comprehensive pricing of the object to be predicted can be obtained according to the user demand, and the individual demand of the user is greatly met, so that the effect of optimal overall profit is achieved.
Drawings
FIG. 1 is a flow chart of a pricing determination method in one embodiment of the invention;
FIG. 2 is a flow chart of a pricing determination method in a second embodiment of the invention;
FIG. 3 is a flow chart of a pricing determination method in a third embodiment of the invention;
FIG. 4 is a flowchart illustrating the implementation of a pricing determination method according to a third embodiment of the invention;
FIG. 5 is a schematic structural diagram of a pricing determining apparatus according to a fourth embodiment of the present invention;
fig. 6 is a schematic structural diagram of an apparatus in the fifth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Example one
Fig. 1 is a flowchart of a pricing determining method according to an embodiment of the present invention, where this embodiment is applicable to a case where a commodity to be priced is priced comprehensively in consideration of comprehensive factors, and the method may be executed by a pricing determining apparatus, where the pricing determining apparatus may be implemented by software and/or hardware, and the pricing determining apparatus may be configured on a computing device, and specifically includes the following steps:
and S110, acquiring data to be predicted and a target index of an object to be predicted.
For example, the object to be predicted may be an object to be predicted, for example, an object to be priced. The data to be predicted may be data of an object to be predicted, for example, if the object to be predicted is a money car, the data to be predicted may be access data information and/or order data information of the money car. The target index may be a target requirement of the object to be predicted, for example, the income of the object to be predicted is to reach the user requirement in a short-term time (for example, 7 days in the future), the profit of the object to be predicted is to reach the user requirement in a short-term time (for example, 7 days in the future), the income of the object to be predicted is to reach the user requirement in a long-term time (for example, 3 months in the future), and the like. The target index can be set according to the user requirement, and is not limited herein.
Therefore, the target pricing of the object to be predicted is obtained based on the acquired data to be predicted and the target index of the object to be predicted.
And S120, respectively inputting the data to be predicted into a pre-trained short-term income pricing prediction model, a pre-trained short-term profit pricing prediction model and a pre-trained long-term income pricing prediction model, and respectively obtaining the predicted pricing of the object to be predicted output by each model.
Illustratively, the short-term revenue pricing prediction model may be a reasonably priced model of the short-term revenue of the subject to be predicted derived for the data to be predicted. The short-term profit pricing prediction model may be a reasonably priced model of the short-term profit of the object to be predicted derived for the data to be predicted. The long-term revenue pricing prediction model may be a reasonably priced model of the long-term revenue of the object to be predicted derived for the data to be predicted.
The short-term revenue pricing prediction model, the short-term profit pricing prediction model, and the long-term revenue pricing prediction model may each be Sarsa reinforcement learning models or DQN deep reinforcement learning models. The short-term income pricing prediction model, the short-term profit pricing prediction model and the long-term income pricing prediction model are trained on historical training data samples. The historical training data sample can be data of the object to be predicted and other objects related to the object to be predicted, such as order data, income data, profit data, access data and the like. The historical training data sample can be obtained from a platform of the object to be predicted, for example, taking the object to be predicted as a car, the historical training data sample can be obtained from any car renting platform where the object to be predicted is located.
The forecasting pricing can be pricing output by inputting data to be forecasted into a short-term income pricing forecasting model, a short-term profit pricing forecasting model and a long-term income pricing forecasting model which are trained in advance respectively.
And respectively inputting the data to be predicted into a pre-trained short-term income pricing prediction model, a short-term profit pricing prediction model and a long-term income pricing prediction model, and respectively obtaining the predicted pricing of the object to be predicted output by each model, so that the reasonable comprehensive pricing is carried out on the object to be predicted based on the predicted pricing and the target index output by each model in the following, the personalized requirements of the user are met, and the overall income is optimized.
S130, determining the target pricing of the object to be predicted based on the predicted pricing and the target index output by each model.
Illustratively, the target pricing may be the final pricing of the object to be predicted. And determining the target pricing of the object to be predicted based on a preset pricing rule according to the predicted pricing and the target index output by each model. Therefore, reasonable comprehensive pricing can be carried out on the objects to be predicted, and the individual requirements of the users are met, so that the overall income is optimal.
Optionally, the target pricing of the object to be predicted is determined based on the predicted pricing and the target index output by each model, which may specifically be: respectively determining the weight of the predicted pricing output by each model based on the target index; and determining the target pricing of the object to be predicted based on the predicted pricing output by each model and the weight of each predicted pricing.
For example, different weights may be assigned to the predicted pricing output by each model according to the target index, and the predicted pricing output by each model and the weight of each predicted pricing may be based on a certain calculation rule, for example, the predicted pricing output by each model and the weight of each predicted pricing may be multiplied to obtain the pricing value of each predicted pricing, and then the pricing values may be superimposed to obtain the target pricing. For example, when the object to be predicted is a vehicle, the target index is a larger amount of short-term revenue, the short-term revenue pricing prediction model, the short-term profit pricing prediction model and the long-term revenue pricing prediction model respectively output 200 yuan/day, 250 yuan/day and 150 yuan/day for the vehicle, different weights are assigned to the prediction pricing output by each model according to the target index, for example, the weight assigned to the prediction pricing output by the short-term revenue pricing prediction model is 50%, the weight assigned to the prediction pricing output by the short-term profit pricing prediction model is 30%, and the weight assigned to the prediction pricing output by the long-term revenue pricing prediction model is 20%, the target pricing of the vehicle is 205 yuan/day based on a certain calculation rule, for example, based on 200 × 0.5+250 × 0.3+150 × 0.2.
Different weights are distributed to the predicted pricing output by each model based on the target indexes, and the target pricing of the object to be predicted is determined based on the predicted pricing output by each model and the weight distributed by each predicted pricing, so that more reasonable comprehensive pricing of the object to be predicted can be obtained according to user demands, personalized demands of users are met, and the overall profit is optimal.
According to the technical scheme of the embodiment of the invention, the obtained data to be predicted of the object to be predicted is respectively input into the pre-trained short-term income pricing prediction model, short-term profit pricing prediction model and long-term income pricing prediction model to respectively obtain the predicted pricing of the object to be predicted output by each model, and the target pricing of the object to be predicted can be determined according to the predicted pricing of the object to be predicted output by each model and the obtained target index of the object to be predicted, so that the reasonable comprehensive pricing of the object to be predicted can be obtained according to the user demand, and the individual demand of the user is met, so that the overall profit is optimal.
Example two
Fig. 2 is a flowchart of a pricing determining method according to a second embodiment of the present invention, which may be combined with various alternatives in the above embodiments. In this embodiment of the present invention, optionally, before the obtaining of the data to be predicted of the object to be predicted, the method further includes: determining first access record information and first order record information of the object type corresponding to the object to be predicted in a first preset time period after the current time based on historical data information and the object type corresponding to the object to be predicted; and determining the data to be predicted of the object to be predicted in a first preset time period after the current time based on the fourth access record information and the fourth ordering record information.
As shown in fig. 2, the method of the embodiment of the present invention specifically includes the following steps:
s210, determining first access record information and first order record information of the object type corresponding to the object to be predicted in a first preset time period after the current time based on historical data information and the object type corresponding to the object to be predicted.
For example, the historical data information may be information of historical data of the object to be predicted, for example, second access record information and second order record information of the object to be predicted in a second preset time period. The second preset time period here may be a preset time period, and may be, for example, the previous 2 years of the current time point. The second access record information may be data information that the object to be predicted is accessed within a second preset time period. The second order record information may be data information that the object to be predicted is placed within a second preset time period.
The object type corresponding to the object to be predicted may be other objects related to the object to be predicted and having an influence on the object to be predicted. For example, taking an object to be predicted as a vehicle, knowing that the vehicle has a style of SUV type, black color and a price of 200/day-250/day, an object type corresponding to the object to be predicted is obtained through a certain calculation rule, for example, the object type of the predicted object can be obtained based on a K-Means clustering algorithm.
The specific acquisition mode may be as follows:
optionally, historical data information of a second preset time period of the object to be predicted is obtained, where the historical data information includes: second access record information and second order record information of the object to be predicted in the second preset time period; and determining the object type corresponding to the object to be predicted based on the object to be predicted, the second access record information and the second order record information.
Illustratively, taking a vehicle with an object to be predicted being a SUV type, black and a price of 200/day to 250/day as an example, the second preset time period is 2 years before the current time, and the second access record information is data information that the object to be predicted is accessed in the second preset time period. The second order record information is, for example, data information of the order of the object to be predicted in the second preset time period.
Extracting access and/or ordering information of the object to be predicted within 2 years before the current time and attribute data of the object to be predicted, wherein the attribute data can be style, price, color and the like, for example, the object to be predicted is SUV type, black and has a price of 200/day-250/day, the extracted access and/or ordering information of the object to be predicted and the attribute data of the object to be predicted are clustered by using a K-Means clustering algorithm, the number of clusters (classes) into which the extracted access and/or ordering information of the object to be predicted and the attribute data of the object to be predicted are divided can be set according to the requirements of users, for example, the SUV type, black and the price of 200/day-250/day, the access of the previous 2 years is 9000-, a cluster may be formed. SUV type, white, price of 200/day-250/day, and can also be grouped into a cluster. The access and/or ordering information of the object to be predicted and several features in the attribute data of the object to be predicted are used as a cluster, which can be set by the user according to the requirement, and the invention is not limited herein. For example, the access information of the object to be predicted, the price of the object to be predicted, and the color may be set as one cluster, or the access information of the object to be predicted, the order information of the object to be predicted, the price of the object to be predicted, and the color may be set as one cluster. Here, the access and/or ordering information of the object to be predicted and the features in the attribute data of the object to be predicted may be freely combined.
For example, using the access and/or order information of the object to be predicted in the previous 2 years and the attribute data of the object to be predicted as a clustering center, obtaining the object type corresponding to the object to be predicted, for example, if the object to be predicted is SUV type, black, the price is 200/day-250/day, the access of the object in the previous 2 years is 9000-, Black, 200/day-250/day, 9000 + 10000 visits in the last 2 years, and 460 + 500 orders) of other vehicles, and using the obtained visit and/or order information in the last 2 years of the object to be predicted, the attribute data of the object to be predicted, other vehicles and the vehicle to be predicted as a vehicle type group to form the object type corresponding to the object to be predicted. Therefore, the object type of the prediction object meeting the user requirement can be obtained based on the K-Means clustering algorithm.
The first preset time period may be a preset time period, for example, 7 days in the future. The first access record information may be data information that the object to be predicted is accessed within a first preset time period after the current time. The first order record information may be data information that the object to be predicted is placed within a first preset time period after the current time.
Based on the determined object type and the historical data information, first access record information and first order record information of the object type corresponding to the object to be predicted in a first preset time period after the current time can be predicted based on a certain calculation rule, for example, a Prophet time sequence prediction model.
For example, the object type corresponding to the object to be predicted is SUV type, black, the price is 200/day-250/day, the access in the previous 2 years is 9000-. The first order record information is, for example, data information of an order placed in a first preset time period after the current time of the object to be predicted. And predicting access data of the object type corresponding to the object to be predicted 7 days in the future by using a Prophet time sequence prediction model for 500 times, and predicting ordering data for 100 times. The Prophet time sequence prediction model is used for predicting that the object type corresponding to the object to be predicted in the future 7 days has access data of 500 times, and the order data of 100 times means that, for example, A, B, C three vehicles exist in the object type corresponding to the object to be predicted, the access data of 500 times means the number of times that A, B, C three vehicles are accessed together in the future 7 days, for example, a vehicle a is accessed 200 times in the future 7 days, a vehicle B is accessed 180 times in the future 7 days, and a vehicle C is accessed 220 times in the future 7 days. The order data of 100 times refers to the number of times that A, B, C three vehicles are ordered together in the future 7 days, for example, it may be that A vehicle is ordered 30 times in the future 7 days, B vehicle is ordered 50 times in the future 7 days, and C vehicle is ordered 20 times in the future 7 days.
It should be noted that, a specific implementation for predicting a certain target to be predicted by using a Prophet timing prediction model belongs to the prior art, and details are not described here.
It should be noted that uncertain parameters such as holidays and seasonal parameters may also be added when the Prophet timing prediction model is trained. Therefore, uncertain parameters such as holidays, seasonal parameters and the like exist when a certain object to be predicted is predicted by the Prophet time sequence prediction model in the following process.
Therefore, according to the historical data information and the object type corresponding to the object to be predicted, the first access record information and the first order record information of the object type corresponding to the object to be predicted in the first preset time period after the current time are determined, so that the data to be predicted of the object to be predicted in the first preset time period after the current time are determined based on the first access record information and the first order record information.
S220, determining data to be predicted of the object to be predicted in a first preset time period after the current time based on the first access record information and the first order record information.
For example, according to the first access record information and the first order record information, the data to be predicted of the first preset time period after the current time of the object to be predicted can be obtained by adopting an inverse operation of a K-Means clustering algorithm. This is done in order to obtain a reasonably comprehensive target pricing of the object to be predicted based on the data to be predicted.
And S230, acquiring data to be predicted and a target index of the object to be predicted.
S240, inputting the data to be predicted into a pre-trained short-term income pricing prediction model, a pre-trained short-term profit pricing prediction model and a pre-trained long-term income pricing prediction model respectively to obtain the predicted pricing of the object to be predicted output by each model respectively.
And S250, determining the target pricing of the object to be predicted based on the predicted pricing and the target index output by each model.
According to the technical scheme of the embodiment of the invention, the first access record information and the first order record information of the object type corresponding to the object to be predicted in the first preset time period after the current time are determined according to the historical data information and the object type corresponding to the object to be predicted, and the data to be predicted in the first preset time period after the current time of the object to be predicted are obtained according to the first access record information and the first order record information. This is done in order to obtain a reasonably comprehensive target pricing of the object to be predicted based on the data to be predicted.
EXAMPLE III
Fig. 3 is a flowchart of a pricing determining method provided in the third embodiment of the present invention, and the third embodiment of the present invention may be combined with various alternatives in the foregoing embodiments. In an embodiment of the present invention, optionally, the training method of any one of the short-term revenue pricing prediction model, the short-term profit pricing prediction model and the long-term revenue pricing prediction model includes: creating initial models of at least two different structures; respectively carrying out iterative training on the initial models of the at least two different structures based on sample data of model training to obtain training models of the at least two different structures; and performing index evaluation on the trained training models with the at least two different structures based on a model index effect discriminator, and determining a target model from the training models with the at least two different structures based on the index evaluation.
As shown in fig. 3, the method of the embodiment of the present invention specifically includes the following steps:
s310, creating initial models of at least two different structures.
For example, the at least two different structured initial models may be models to be trained for any one of a short-term revenue pricing prediction model, a short-term profit pricing prediction model, and a long-term revenue pricing prediction model. For example, Sarsa reinforcement learning model and DQN deep reinforcement learning model are possible.
And creating at least two initial models with different structures for any one of the short-term income pricing prediction model, the short-term profit pricing prediction model and the long-term income pricing prediction model, so as to screen out a target model meeting the requirements based on the at least two initial models with different structures.
And S320, performing iterative training on the initial models of the at least two different structures respectively based on the sample data of the model training to obtain training models of the at least two different structures.
Illustratively, at least two initial models with different structures of any one of the short-term income pricing prediction model, the short-term profit pricing prediction model and the long-term income pricing prediction model are subjected to iterative training based on corresponding sample data, so that training models of the short-term income pricing prediction model, the short-term profit pricing prediction model and the long-term income pricing prediction model with different structures are obtained.
Taking a short-term income pricing prediction model as an example, two initial models with different structures, namely a Sarsa reinforcement learning model and a DQN depth reinforcement learning model, are created, and sample data of the short-term income pricing prediction model is used for performing iterative training on the Sarsa reinforcement learning model and the DQN depth reinforcement learning model respectively.
Optionally, the sample data of the short-term revenue pricing prediction model herein includes: the method comprises the steps of obtaining an object type corresponding to an object to be predicted, and first historical income information, third access record information and third order record information of the object type in a third preset time period before the current time.
For example, the third predetermined period may be a predetermined period, where the third predetermined period may be a short period, such as 28 days. The third prediction time period can be set according to the requirements of users. The first historical revenue information may be historical revenue data information of an object type corresponding to the object to be predicted in a third preset time period before the current time. The third access record information may be historical access data information of an object type corresponding to the object to be predicted in a third preset time period before the current time. The third order record information may be historical order data information of an object type corresponding to the object to be predicted in a third preset time period before the current time.
For example, taking an object type corresponding to the object to be predicted as a vehicle with a price of 200/day to 250/day, taking a third preset time period as 28 days as an example, the first historical income information is income data of the vehicle of the object type corresponding to the object to be predicted in 28 days before the current time and 28 days before and after the last year, the third access record information is data that the vehicle of the object type corresponding to the object to be predicted is accessed in 28 days before the current time, and the third order record information is data that the vehicle of the object type corresponding to the object to be predicted is ordered in 28 days before the current time.
It should be noted that, the third preset time period may be determined according to the first preset time period, and the third preset time period is generally set to be 3-5 times of the first preset time period.
Taking a short-term profit pricing prediction model as an example, two initial models with different structures, namely a Sarsa reinforcement learning model and a DQN depth reinforcement learning model, are created, and sample data of the short-term profit pricing prediction model is used for performing iterative training on the Sarsa reinforcement learning model and the DQN depth reinforcement learning model respectively.
Optionally, the sample data of the short-term profit pricing prediction model herein includes: the method comprises the steps of predicting the object type corresponding to an object to be predicted, and obtaining first historical profit information, third access record information and third order record information of the object type in a third preset time period before the current time.
For example, the first historical profit information may be historical profit data information of an object type corresponding to an object to be predicted at a third preset time period before the current time. For example, taking the object type corresponding to the object to be predicted as a vehicle with a price of 200/day to 250/day and the third preset time period as 28 days as an example, the first historical profit information is profit data of the vehicle of the object type corresponding to the object to be predicted 28 days before the current time.
Taking a long-term income pricing prediction model as an example, two initial models with different structures, namely a Sarsa reinforcement learning model and a DQN depth reinforcement learning model, are created, and sample data of the long-term income pricing prediction model is used for performing iterative training on the Sarsa reinforcement learning model and the DQN depth reinforcement learning model respectively.
Optionally, the sample data of the long-term revenue pricing prediction model herein includes: the object type corresponding to the object to be predicted, and the second historical income information, the fourth access record information and the fourth order record information of the object type in a fourth preset time period before the current time.
For example, referring to fig. 4, an execution flow diagram of the pricing determining method is described, in fig. 4, a first preset time period is exemplified by 7 days, a second preset time period is exemplified by 2 years, a third preset time period is exemplified by 30 days, and a fourth preset time period is exemplified by 1 year.
For example, the fourth preset time period may be a preset time period, where the fourth preset time period may be a long time period, for example, 30 days or 90 days. The fourth prediction time period can be set according to the requirements of users. The first historical revenue information may be historical revenue data information of an object type corresponding to the object to be predicted at a fourth preset time period before the current time. The fourth access record information may be historical access data information of an object type corresponding to the object to be predicted in a fourth preset time period before the current time. The fourth order record information may be historical order data information of an object type corresponding to the object to be predicted in a fourth preset time period before the current time.
For example, if the object type corresponding to the object to be predicted is a vehicle with a price of 200/day to 250/day, and the predicted pricing of the object to be predicted is predicted for 90 days in the future, the fourth preset time period may be 1 year, the second historical income information is 1 year before the current time, income data of the vehicle of the object type corresponding to the object to be predicted, the fourth access record information is 1 year before the current time, accessed data of the vehicle of the object type corresponding to the object to be predicted, and the fourth order record information is data of the order of the vehicle of the object type corresponding to the object to be predicted, which is 1 year before the current time.
It should be noted that the fourth preset time period is usually 3 to 5 times the first preset time period.
And carrying out iterative training on the short-term income pricing prediction model, the short-term profit pricing prediction model and the long-term income pricing prediction model of at least two different structures based on different sample data, and obtaining the training models of at least two different structures for any one of the short-term income pricing prediction model, the short-term profit pricing prediction model and the long-term income pricing prediction model, so that the actual short-term income pricing prediction model, the short-term profit pricing prediction model and the long-term income pricing prediction model can be applied to be screened out on the basis of a certain rule in the subsequent process.
S330, performing index evaluation on the trained training models with the at least two different structures based on a model index effect discriminator, and determining a target model from the training models with the at least two different structures based on the index evaluation.
For example, the model index effect discriminator may be a device for evaluating the performance of the training model, and an evaluation index for evaluating the performance of the training model is set therein, for example, the evaluation index of the model index effect discriminator may include: at least one of an iteration speed, a price offset, and a logarithmic loss function.
Respectively evaluating indexes of at least two training models with different structures in the trained short-term income pricing prediction model, short-term profit pricing prediction model and long-term income pricing prediction model according to a model index effect discriminator, screening out a model with an output result closer to an actual result from the training models with different structures of each model as a target model, for example, taking the short-term income pricing prediction model as an example, the model comprises two trained models of a Sarsa reinforcement learning model and a DQN depth reinforcement learning model, wherein the output result of the Sarsa reinforcement learning model is 100 yuan/day, the output result of the DQN depth reinforcement learning model is 150 yuan/day, and if the short-term income reality of an object type corresponding to an object to be predicted is known to be 98 yuan/day, the output result of the Sarsa reinforcement learning model is closer to the actual result, the Sarsa reinforcement learning model is determined to be the final short term revenue pricing prediction model.
It should be noted that, taking the short-term revenue pricing prediction model as an example, after sample data is input, the short-term revenue pricing model outputs the predicted pricing of the object to be predicted on the basis of the condition that the overall revenue of the object type corresponding to the object to be predicted in the first preset time period is maximum.
The short-term profit pricing prediction model and the long-term revenue pricing prediction model are determined in a manner consistent with that of the short-term revenue pricing prediction model and will not be described in detail herein.
And determining a target model from at least two training models with different structures by using a model index effect discriminator, so that a more accurate short-term income pricing prediction model, a more accurate short-term profit pricing prediction model and a more accurate long-term income pricing prediction model can be obtained.
S340, determining first access record information and first order record information of the object type corresponding to the object to be predicted in a first preset time period after the current time based on historical data information and the object type corresponding to the object to be predicted.
And S350, determining data to be predicted of the object to be predicted in a first preset time period after the current time based on the first access record information and the first order record information.
And S360, acquiring data to be predicted and a target index of the object to be predicted.
And S370, inputting the data to be predicted into a pre-trained short-term income pricing prediction model, a pre-trained short-term profit pricing prediction model and a pre-trained long-term income pricing prediction model respectively to obtain the predicted pricing of the object to be predicted output by each model respectively.
And S380, determining the target pricing of the object to be predicted based on the predicted pricing and the target index output by each model.
According to the technical scheme of the embodiment of the invention, the created initial models with at least two different structures are respectively subjected to iterative training by sample data based on model training to obtain training models with at least two different structures, the trained training models with at least two different structures are subjected to index evaluation based on a model index effect discriminator, and a target model is determined from the training models with at least two different structures based on the index evaluation, so that a more accurate short-term income pricing prediction model, a short-term profit pricing prediction model and a long-term income pricing prediction model can be obtained.
Example four
Fig. 5 is a schematic structural diagram of a pricing determining apparatus according to a fourth embodiment of the present invention, as shown in fig. 5, the apparatus includes: a data acquisition module 31, a predicted pricing acquisition module 32, and a target pricing determination module 33.
The data acquiring module 31 is configured to acquire data to be predicted and a target index of an object to be predicted;
the forecast pricing obtaining module 32 is configured to input the data to be forecasted into a pre-trained short-term income pricing forecasting model, a short-term profit pricing forecasting model and a long-term income pricing forecasting model respectively, and obtain forecast pricing of the object to be forecasted output by each model respectively;
and the target pricing determining module 33 is used for determining the target pricing of the object to be predicted based on the predicted pricing and the target index output by each model.
On the basis of the technical solution of the above embodiment, the target pricing determining module 33 includes:
the weight determining unit is used for respectively determining the weight of the predicted pricing output by each model based on the target index;
and the target pricing determining unit is used for determining the target pricing of the object to be predicted based on the predicted pricing output by each model and the weight of each predicted pricing.
On the basis of the technical scheme of the embodiment, the device further comprises:
the information determining module is used for determining first access record information and first order record information of the object type corresponding to the object to be predicted in a first preset time period after the current time based on historical data information and the object type corresponding to the object to be predicted;
and the data to be predicted determining module is used for determining the data to be predicted of the object to be predicted in a first preset time period after the current time based on the first access record information and the first order record information.
On the basis of the technical scheme of the embodiment, the device further comprises:
the historical data information acquiring module is used for acquiring the historical data information of the object to be predicted in a second preset time period, wherein the historical data information comprises: second access record information and second order record information of the object to be predicted in the second preset time period;
and the object type determining module is used for determining the object type corresponding to the object to be predicted based on the object to be predicted, the second access record information and the second order record information.
On the basis of the technical scheme of the embodiment, the device further comprises:
the initial model creating module is used for creating initial models of at least two different structures;
the model training module is used for respectively carrying out iterative training on the initial models of the at least two different structures based on sample data of model training to obtain training models of the at least two different structures;
and the target model determining module is used for performing index evaluation on the trained training models with the at least two different structures based on a model index effect discriminator and determining a target model from the training models with the at least two different structures based on the index evaluation.
Optionally, the sample data of the short-term revenue pricing prediction model includes: the method comprises the steps that an object type corresponding to an object to be predicted, and first historical income information, third access record information and third order record information of the object type in a third preset time period before the current time; sample data for the short-term profit pricing prediction model includes: the method comprises the steps that an object type corresponding to an object to be predicted, and first historical profit information, third access record information and third order record information of the object type in a third preset time period before the current time; sample data for the long-term revenue pricing prediction model includes: the object type corresponding to the object to be predicted, and the second historical income information, the fourth access record information and the fourth order record information of the object type in a fourth preset time period before the current time.
Optionally, the initial models of the at least two different structures at least include: a Sarsa reinforcement learning model and a DQN deep reinforcement learning model; the evaluation indexes of the model index effect discriminator include: at least one of an iteration speed, a price offset, and a logarithmic loss function.
The pricing determining device provided by the embodiment of the invention can execute the pricing determining method provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects of the executing method.
EXAMPLE five
Fig. 6 is a schematic structural diagram of an apparatus according to a fifth embodiment of the present invention, as shown in fig. 6, the apparatus includes a processor 70, a memory 71, an input device 72, and an output device 73; the number of processors 70 in the device may be one or more, and one processor 70 is taken as an example in fig. 6; the processor 70, the memory 71, the input device 72 and the output device 73 of the apparatus may be connected by a bus or other means, as exemplified by the bus connection in fig. 6.
The memory 71, as a computer-readable storage medium, may be used for storing software programs, computer-executable programs, and modules, such as program instructions/modules (e.g., the data acquisition module 31, the predicted pricing acquisition module 32, and the target pricing determination module 33) corresponding to the pricing determination method in the embodiments of the present invention. The processor 70 executes various functional applications of the device/terminal/server and data processing by running software programs, instructions and modules stored in the memory 71, i.e. implements the pricing determination method described above.
The memory 71 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the terminal, and the like. Further, the memory 71 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, the memory 71 may further include memory located remotely from the processor 70, which may be connected to the device over a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The input device 72 may be used to receive entered numeric or character information and to generate key signal inputs relating to user settings and function controls of the apparatus. The output device 73 may include a display device such as a display screen.
EXAMPLE six
An embodiment of the present invention also provides a storage medium containing computer-executable instructions that, when executed by a computer processor, perform a pricing determination method.
Of course, the storage medium provided by the embodiment of the present invention contains computer-executable instructions, and the computer-executable instructions are not limited to the operations of the method described above, and may also perform related operations in the pricing determination method provided by any embodiment of the present invention.
From the above description of the embodiments, it is obvious for those skilled in the art that the present invention can be implemented by software and necessary general hardware, and certainly, can also be implemented by hardware, but the former is a better embodiment in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which may be stored in a computer-readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a FLASH Memory (FLASH), a hard disk or an optical disk of a computer, and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device) to execute the methods according to the embodiments of the present invention.
It should be noted that, in the embodiment of the pricing determining apparatus, the included units and modules are merely divided according to functional logic, but are not limited to the above division as long as the corresponding functions can be implemented; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims (10)

1. A pricing determination method, comprising:
acquiring data to be predicted and a target index of an object to be predicted;
respectively inputting the data to be predicted into a pre-trained short-term income pricing prediction model, a short-term profit pricing prediction model and a long-term income pricing prediction model to respectively obtain the predicted pricing of the object to be predicted output by each model;
and determining the target pricing of the object to be predicted based on the predicted pricing and the target index output by each model.
2. The method of claim 1, wherein determining the target pricing of the object to be predicted based on the predicted pricing and the target indicator output by each model comprises:
respectively determining the weight of the predicted pricing output by each model based on the target index;
and determining the target pricing of the object to be predicted based on the predicted pricing output by each model and the weight of each predicted pricing.
3. The method according to claim 2, wherein before the obtaining the data to be predicted of the object to be predicted, the method further comprises:
determining first access record information and first order record information of the object type corresponding to the object to be predicted in a first preset time period after the current time based on historical data information and the object type corresponding to the object to be predicted;
and determining data to be predicted of the object to be predicted in a first preset time period after the current time based on the first access record information and the first order record information.
4. The method according to claim 3, before determining, based on the historical data information and the type corresponding to the object to be predicted, first access record information and first order record information of the object type corresponding to the object to be predicted in a preset time period after the current time, further comprising:
acquiring historical data information of a second preset time period of the object to be predicted, wherein the historical data information comprises: second access record information and second order record information of the object to be predicted in the second preset time period;
and determining the object type corresponding to the object to be predicted based on the object to be predicted, the second access record information and the second order record information.
5. The method of claim 1, wherein the method of training any of the short term revenue pricing prediction model, the short term profit pricing prediction model, and the long term revenue pricing prediction model comprises:
creating initial models of at least two different structures;
respectively carrying out iterative training on the initial models of the at least two different structures based on sample data of model training to obtain training models of the at least two different structures;
and performing index evaluation on the trained training models with the at least two different structures based on a model index effect discriminator, and determining a target model from the training models with the at least two different structures based on the index evaluation.
6. The method of claim 5, wherein the sample data for the short-term revenue pricing prediction model comprises: the method comprises the steps that an object type corresponding to an object to be predicted, and first historical income information, third access record information and third order record information of the object type in a third preset time period before the current time;
sample data for the short-term profit pricing prediction model includes: the method comprises the steps that an object type corresponding to an object to be predicted, and first historical profit information, third access record information and third order record information of the object type in a third preset time period before the current time;
sample data for the long-term revenue pricing prediction model includes: the object type corresponding to the object to be predicted, and the second historical income information, the fourth access record information and the fourth order record information of the object type in a fourth preset time period before the current time.
7. The method of claim 5, wherein the initial models of the at least two different structures comprise at least: a Sarsa reinforcement learning model and a DQN deep reinforcement learning model;
the evaluation indexes of the model index effect discriminator include: at least one of an iteration speed, a price offset, and a logarithmic loss function.
8. A pricing determination device, comprising:
the data acquisition module is used for acquiring data to be predicted and a target index of an object to be predicted;
the forecasting pricing obtaining module is used for respectively inputting the data to be forecasted into a pre-trained short-term income pricing forecasting model, a short-term profit pricing forecasting model and a long-term income pricing forecasting model to respectively obtain the forecasting pricing of the object to be forecasted output by each model;
and the target pricing determining module is used for determining the target pricing of the object to be predicted based on the predicted pricing and the target index output by each model.
9. An apparatus, characterized in that the apparatus comprises:
one or more processors;
storage means for storing one or more programs;
when executed by the one or more processors, cause the one or more processors to implement a pricing determination method according to any of claims 1-7.
10. A storage medium containing computer-executable instructions for performing the pricing determination method of any of claims 1-7 when executed by a computer processor.
CN202010568617.9A 2020-06-19 2020-06-19 Pricing determination method, device, equipment and storage medium Pending CN111754262A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010568617.9A CN111754262A (en) 2020-06-19 2020-06-19 Pricing determination method, device, equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010568617.9A CN111754262A (en) 2020-06-19 2020-06-19 Pricing determination method, device, equipment and storage medium

Publications (1)

Publication Number Publication Date
CN111754262A true CN111754262A (en) 2020-10-09

Family

ID=72675726

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010568617.9A Pending CN111754262A (en) 2020-06-19 2020-06-19 Pricing determination method, device, equipment and storage medium

Country Status (1)

Country Link
CN (1) CN111754262A (en)

Similar Documents

Publication Publication Date Title
WO2021081962A1 (en) Recommendation model training method, recommendation method, device, and computer-readable medium
CN107818344A (en) The method and system that user behavior is classified and predicted
US20210103858A1 (en) Method and system for model auto-selection using an ensemble of machine learning models
CN106251174A (en) Information recommendation method and device
CN110264270B (en) Behavior prediction method, behavior prediction device, behavior prediction equipment and storage medium
CN109523117A (en) Risk Forecast Method, device, computer equipment and storage medium
CN111881158A (en) Management report data processing method and device, computer system and readable storage medium
CN111797320A (en) Data processing method, device, equipment and storage medium
CN111179055A (en) Credit limit adjusting method and device and electronic equipment
CN113850669A (en) User grouping method and device, computer equipment and computer readable storage medium
CN114692889A (en) Meta-feature training model for machine learning algorithm
CN111754261B (en) Method and device for evaluating taxi willingness and terminal equipment
CN110458664B (en) User travel information prediction method, device, equipment and storage medium
CN110910241B (en) Cash flow evaluation method, apparatus, server device and storage medium
CN116501979A (en) Information recommendation method, information recommendation device, computer equipment and computer readable storage medium
CN111368195A (en) Model training method, device, equipment and storage medium
CN108830302B (en) Image classification method, training method, classification prediction method and related device
CN111275062A (en) Model training method, device, server and computer readable storage medium
CN111754262A (en) Pricing determination method, device, equipment and storage medium
CN111507471B (en) Model training method, device, equipment and storage medium
CN114925919A (en) Service resource processing method and device, computer equipment and storage medium
CN114022045A (en) Regional economic development level determining method and device and terminal equipment
CN115796984A (en) Training method of item recommendation model, storage medium and related equipment
CN111861227A (en) User recommendation method and device based on membership analysis, electronic equipment and medium
CN110083517A (en) A kind of optimization method and device of user's portrait confidence level

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
CB02 Change of applicant information

Address after: Room 587, building 3, 333 Hongqiao Road, Xuhui District, Shanghai 200030

Applicant after: Shanghai Lexiang Sijin Technology Co.,Ltd.

Address before: Room 587, building 3, 333 Hongqiao Road, Xuhui District, Shanghai 200030

Applicant before: Shanghai xinwin Information Technology Co.,Ltd.

CB02 Change of applicant information
TA01 Transfer of patent application right

Effective date of registration: 20230329

Address after: 2406, Building 1, Tongji Science and Technology Plaza, No. 2 Hushan Road, Changshu High and New Technology Industrial Development Zone, Suzhou City, Jiangsu Province, 215500

Applicant after: Concavoconvex Lexiang (Suzhou) Information Technology Co.,Ltd.

Address before: Room 587, building 3, 333 Hongqiao Road, Xuhui District, Shanghai 200030

Applicant before: Shanghai Lexiang Sijin Technology Co.,Ltd.

TA01 Transfer of patent application right