CN115271826A - Logistics line price interval prediction method and device - Google Patents

Logistics line price interval prediction method and device Download PDF

Info

Publication number
CN115271826A
CN115271826A CN202210976114.4A CN202210976114A CN115271826A CN 115271826 A CN115271826 A CN 115271826A CN 202210976114 A CN202210976114 A CN 202210976114A CN 115271826 A CN115271826 A CN 115271826A
Authority
CN
China
Prior art keywords
bargaining
price
prediction
feature data
classification model
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
CN202210976114.4A
Other languages
Chinese (zh)
Inventor
王松松
陈兰欢
王译堃
刘雪
杜毅
贺思琦
庞云冰
赵俊哲
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.)
Beijing Jingdong Zhenshi Information Technology Co Ltd
Original Assignee
Beijing Jingdong Zhenshi 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 Beijing Jingdong Zhenshi Information Technology Co Ltd filed Critical Beijing Jingdong Zhenshi Information Technology Co Ltd
Priority to CN202210976114.4A priority Critical patent/CN115271826A/en
Publication of CN115271826A publication Critical patent/CN115271826A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201Market modelling; Market analysis; Collecting market data
    • G06Q30/0206Price or cost determination based on market factors
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q10/00Administration; Management
    • G06Q10/08Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/083Shipping

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Development Economics (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Accounting & Taxation (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Finance (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Game Theory and Decision Science (AREA)
  • Data Mining & Analysis (AREA)
  • Human Resources & Organizations (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The disclosure provides a method and a device for predicting a price interval of a logistics route. Wherein, the method comprises the following steps: acquiring characteristic data corresponding to a flow line of an analyte; inputting the characteristic data into a preset bargaining forecast mixed classification model for bargaining amplitude reduction forecast analysis, and outputting forecast probability values corresponding to different bargaining intervals respectively; the bargaining forecasting mixed classification model comprises a bargaining gradient classification model for extracting high-dimensional combination characteristics and a bargaining amplitude reduction forecasting model for performing characteristic data correlation analysis; and outputting a corresponding bargaining prediction result based on the size of the prediction probability value. According to the price interval prediction method for the logistics route, accurate bargaining analysis can be carried out on the logistics route through the bargaining prediction mixed classification model, the bargaining analysis result is output in time to guide the bargaining process of the logistics route, the accuracy and timeliness of bargaining amplitude reduction prediction of the logistics route are effectively improved, and the transportation cost of the logistics route is reduced.

Description

Logistics line price interval prediction method and device
Technical Field
The disclosure relates to the technical field of big data processing, in particular to a method and a device for predicting price intervals of logistics lines. In addition, an electronic device and a processor-readable storage medium are also related.
Background
In recent years, electronic commerce has been rapidly developed along with rapid popularization of internet technology. Currently, related logistics transportation management bargaining in the field of electronic commerce is mainly classified into bargaining in a process and bargaining in daily line loss. In the prior art, operators need to issue bargained line information to business personnel according to current requirements, or business personnel spontaneously carry out line sorting at monthly granularity, and the bargained line information is screened mainly by comparing the cost with the cost of a single kilometer in the same flow direction. However, as the number of the logistics lines and the threshold-level dimensions is increased, the energy of business personnel for combing negotiable lines is limited, the negotiable lines are arranged monthly, the lines with higher cost are difficult to find in time, higher hysteresis is achieved, meanwhile, the dimension for judging whether the lines can be negotiable is single, the limitation is higher, and the accuracy and the timeliness of negotiable price reduction prediction of the logistics lines are poor. Therefore, how to design a more effective price interval prediction scheme of the logistics route becomes a difficult problem to be solved urgently.
Disclosure of Invention
Therefore, the method and the device for predicting the price interval of the logistics route are provided, and the defects that the price interval prediction scheme of the logistics route in the prior art is high in limitation, the bargaining price reduction prediction accuracy and timeliness of the logistics route are poor and the like are overcome.
The present disclosure provides a method for predicting a price interval of a logistics route, including:
acquiring characteristic data corresponding to a flow line of an analyte;
inputting the characteristic data into a preset bargaining forecast mixed classification model for bargaining amplitude reduction forecast analysis, and outputting forecast probability values corresponding to different bargaining intervals respectively; the bargaining forecasting mixed classification model comprises a bargaining gradient classification model for extracting high-dimensional combination features and a bargaining amplitude reduction estimation model for performing feature data association analysis;
and outputting a corresponding bargaining prediction result based on the size of the prediction probability value.
Further, the feature data are input into a preset bargaining forecast mixed classification model for bargaining price reduction forecast analysis, and forecast probability values respectively corresponding to different bargaining intervals are output, which specifically includes:
inputting the feature data into the bargaining gradient classification model for high-dimensional combined feature extraction to obtain high-dimensional combined feature data output by the bargaining gradient classification model; the high-dimensional combined feature data is a high-dimensional feature vector represented based on index values of leaf nodes of a decision tree structure in the bargained gradient classification model;
inputting the high-dimensional combination feature data and the feature data into the bargaining price reducing estimation model for feature data association analysis to obtain a prediction probability value corresponding to each bargaining interval output by the bargaining price reducing estimation model; wherein the feature data is low-dimensional discrete feature data associated with the analyte flow line;
the bargaining gradient classification model and the bargaining amplitude reduction estimation model are obtained by training based on sample characteristic data corresponding to a sample logistics line and a bargaining prediction result corresponding to the sample characteristic data.
Further, inputting the high-dimensional combined feature data and the feature data into the bargaining price reduction estimation model for feature data association analysis, and obtaining a prediction probability value corresponding to each bargaining interval output by the bargaining price reduction estimation model, specifically comprising:
based on the bargaining price reduction pre-estimation model, performing feature combination on the high-dimensional combined feature data and the feature data according to different feature data types to realize that the feature data of the same type are classified into the same domain, and determining corresponding hidden vectors aiming at the domain; and determining a prediction probability value corresponding to each bargaining interval based on the hidden vector.
Further, outputting a corresponding bargaining prediction result based on the size of the prediction probability value specifically includes: and normalizing the predicted probability values, determining the target probability value corresponding to each bargaining interval, determining the bargaining interval corresponding to the maximum target probability value as a target bargaining interval, and taking the target bargaining interval as first bargaining guide information in the bargaining prediction result.
Further, outputting a corresponding bargaining prediction result based on the size of the prediction probability value, specifically comprising: and determining an abnormal freight rate logistics route in the logistics route to be analyzed based on the prediction probability value, and taking attribute information corresponding to the abnormal freight rate logistics route as second bargaining guide information in the bargaining prediction result.
Further, the method for predicting the price interval of the logistics route, after outputting the corresponding bargaining prediction result based on the prediction probability value, further comprises:
attributing and analyzing a plurality of characteristic data types corresponding to the bargaining prediction result, and determining a sharp value corresponding to each characteristic data type;
and based on the value of the Charprili value, carrying out influence sequencing on a plurality of characteristic data types corresponding to the bargaining prediction result, and outputting a corresponding characteristic data type influence sequencing result.
The present disclosure also provides a price interval prediction apparatus for a logistics route, including:
the characteristic data acquisition unit is used for acquiring characteristic data corresponding to a flow line of the analyte;
the bargaining price reducing amplitude prediction analysis unit is used for inputting the characteristic data into a preset bargaining price prediction mixed classification model to carry out bargaining price reducing amplitude prediction analysis and outputting prediction probability values respectively corresponding to different bargaining intervals; the bargaining forecasting mixed classification model comprises a bargaining gradient classification model for extracting high-dimensional combination features and a bargaining amplitude reduction estimation model for performing feature data association analysis;
and the bargaining prediction result output unit is used for outputting a corresponding bargaining prediction result based on the prediction probability value.
Further, the bargained price reduction prediction analysis unit is specifically configured to:
inputting the feature data into the bargaining gradient classification model for high-dimensional combined feature extraction to obtain high-dimensional combined feature data output by the bargaining gradient classification model; the high-dimensional combined feature data are high-dimensional feature vectors represented by index values of leaf nodes of a decision tree structure in the bargained gradient classification model;
inputting the high-dimensional combination feature data and the feature data into the bargaining price reducing estimation model for feature data association analysis to obtain a prediction probability value corresponding to each bargaining interval output by the bargaining price reducing estimation model; wherein the feature data is low-dimensional discrete feature data associated with the analyte flow line;
the bargaining gradient classification model and the bargaining amplitude reduction estimation model are obtained by training based on sample characteristic data corresponding to a sample logistics line and a bargaining prediction result corresponding to the sample characteristic data.
Further, inputting the high-dimensional combined feature data and the feature data into the bargaining price reduction estimation model for feature data association analysis, and obtaining a prediction probability value corresponding to each bargaining interval output by the bargaining price reduction estimation model, specifically comprising:
based on the bargaining price reduction pre-estimation model, performing feature combination on the high-dimensional combined feature data and the feature data according to different feature data types to realize that the feature data of the same type are classified into the same domain, and determining corresponding hidden vectors aiming at the domain; and determining a prediction probability value corresponding to each bargaining interval based on the hidden vector.
Further, the bargaining prediction result output unit is specifically configured to: and normalizing the prediction probability values, determining the target probability values corresponding to the bargaining intervals, determining the bargaining interval corresponding to the maximum target probability value as a target bargaining interval, and taking the target bargaining interval as first bargaining guide information in the bargaining prediction result.
Further, the bargaining prediction result output unit is specifically further configured to: and determining an abnormal freight rate logistics route in the logistics route to be analyzed based on the prediction probability value, and taking attribute information corresponding to the abnormal freight rate logistics route as second bargaining guide information in the bargaining prediction result.
Further, the device for predicting the price interval of the logistics route, after outputting the corresponding bargained prediction result based on the prediction probability value, further comprises:
the attribution analysis unit is used for attributing and analyzing a plurality of characteristic data types corresponding to the bargaining prediction result and determining a sharp value corresponding to each characteristic data type;
and the characteristic influence sequencing unit is used for sequencing the influence of a plurality of characteristic data types corresponding to the bargained forecast result based on the magnitude of the Charpy value and outputting a corresponding characteristic data type influence sequencing result.
The present disclosure also provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the method for predicting the price interval of a logistics route as described in any one of the above items when executing the computer program.
The present disclosure also provides a processor-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method for predicting a price interval of a logistics route as described in any one of the above.
The method for predicting the price interval of the logistics route comprises the steps of obtaining characteristic data corresponding to a logistics route to be analyzed, inputting the characteristic data into a preset bargaining prediction mixed classification model for bargaining amplitude reduction prediction analysis, and outputting prediction probability values corresponding to different bargaining intervals; the bargaining forecasting mixed classification model comprises a bargaining gradient classification model for extracting high-dimensional combination characteristics and a bargaining amplitude reduction forecasting model for performing characteristic data correlation analysis; and finally outputting a corresponding bargaining prediction result based on the size of the prediction probability value. By utilizing the bargaining forecasting mixed classification model, accurate bargaining analysis can be carried out on the logistics route, and the bargaining analysis result is output in time to guide the bargaining process of the logistics route, so that the accuracy and timeliness of bargaining amplitude reduction forecasting of the logistics route are effectively improved, and the transportation cost of the logistics route is reduced.
Drawings
In order to clearly illustrate the embodiments or technical solutions of the present disclosure, the drawings used in the embodiments or technical solutions of the present disclosure will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present disclosure, and those skilled in the art can obtain other drawings without creative efforts.
Fig. 1 is a schematic flow chart of a method for predicting a price interval of a logistics route according to an embodiment of the present disclosure;
fig. 2 is a schematic processing flow diagram of a bargaining forecasting hybrid classification model in a price interval forecasting method for a logistics route according to an embodiment of the present disclosure;
fig. 3 is a schematic diagram of a specific application of the method for predicting the price interval of a logistics route according to the embodiment of the present disclosure;
fig. 4 is a complete flow diagram of a method for predicting a price interval of a logistics route according to an embodiment of the present disclosure;
fig. 5 is a schematic structural diagram of a price interval prediction device of a logistics route according to an embodiment of the disclosure;
fig. 6 is a schematic physical structure diagram of an electronic device provided in an embodiment of the present disclosure.
Detailed Description
In order to make the technical solutions and advantages of the embodiments of the present disclosure more clear, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the drawings in the embodiments of the present disclosure. All other embodiments, which can be derived by a person skilled in the art from the embodiments disclosed herein without making any creative effort, shall fall within the protection scope of the present disclosure.
The following describes an embodiment in detail based on a method for predicting a price interval of a logistics route according to the present disclosure. As shown in fig. 1, which is a schematic flow chart of a method for predicting a price interval of a logistics route according to an embodiment of the present disclosure, a specific implementation process includes the following steps:
step 101: and acquiring characteristic data corresponding to the flow line of the analyte.
In the embodiment of the present disclosure, effective model training needs to be performed in advance before executing this step, so as to obtain a bargaining prediction hybrid classification model satisfying a preset condition. It should be noted that the bargaining prediction hybrid classification model disclosed in the present disclosure includes a bargaining Gradient classification model (e.g., a Light Gradient Boosting Machine, lightGBM model) for extracting high-dimensional combination features and a bargaining downscaling prediction model (e.g., a Field-aware Factorization Machines, FFM model) for performing feature data association analysis.
In the model training process, firstly, data preparation is needed, and attribute data associated with the logistics line is determined. The attribute data includes line side data, carrier side data, and external association data. The line side data comprises historical carrying conditions of the logistics line, such as the cargo volume carried each time, contract price, start and stop cities, business types and vehicle types of the line, and the like; the carrier side data comprises the grade of a carrier, the scale of a driver of the carrier, the total scale of vehicles, the proportion of each vehicle, the type of main service and the like; the external associated data comprises current light and busy season conditions, bargaining time, friend single kilometer price, current oil price, high-speed general policy and the like. And after the data preparation is finished, performing data integration and updating. Specifically, data are obtained from each data source, preliminary cleaning and integration are carried out, the data of each data source are updated in time based on the upper edition bargaining and reducing and the carrier will feedback conditions, and the current data on the logistics line side and the data on the carrier side are updated synchronously.
In the model feature determination process, feature data such as corresponding line-side features, carrier-side features, external associated features of the current line carrier at the current time and the like can be respectively extracted based on the line-side data, the carrier-side data and the external associated data, and the feature data is processed through data preprocessing, feature selection, dimension reduction and the like. Wherein the line side features may include: carrying price on each bank level at the last time, friend line quotation crawler data, start and stop city regions, provinces of start and stop cities, historical start and stop city/region inlet and outlet quantity in the last 3 months, historical daily average quantity in the last 3 months, predicted future daily average quantity of a logistics line and each bank level proportion of the line in the last 3 months; the future daily average cargo volume of the logistics route can use an Autoregressive model (AR model) and two submodels of the same-proportion historical cargo volume and the LSTM (Long Short-Term Memory) recurrent neural network to perform result fusion, and the cargo volume reference values of 3 months, 6 months, 9 months and 12 months in the future are predicted in an average weighting mode, for example, and are finally obtained and used as the input of a training bargaining prediction mixed classification model. The carrier side characteristics can include the number of bidding carriers in the area, the negotiation success rate (success times/negotiation total times) of all current carrier historical routes, negotiation amplitude reduction statistics (median, maximum amplitude reduction and minimum amplitude reduction) of all current carrier routes, negotiation amplitude reduction statistics (median, maximum amplitude reduction and minimum amplitude reduction) of current carrier routes in the same flow direction with the current routes. The external association features may include: the current waning season, bargaining time, average settlement price per kilometer of each threshold level of each carrier between provinces, average settlement price per kilometer of each threshold level of each carrier between regions, and average settlement price per kilometer of each threshold level of each market carrier (for example, three average settlement prices take the last three months or half a year). Meanwhile, the method is the same as the prediction of the future daily average cargo quantity of the logistics route, the future oil price plays a key role in the bargained price reduction of the route, and oil price references of 3 months, 6 months, 9 months and 12 months in the future can be respectively obtained by using the same method and used as the input of a training bargained prediction mixed classification model. Further, a corresponding training sample and a corresponding prediction sample need to be constructed respectively. Specifically, the training sample may be constructed according to price changes of a route history, for example, a logistics route in a city a-B, where price changes of 3 times occur in the history of a first price, a second price, and a third price, and the first price and the second price form a training sample, and the second price and the third price form a training sample; when building the label, the difference (amplitude reduction) between the front price and the back price is used, then the amplitude reduction of all the training samples is performed with barrel division, for example, the label can be divided into 7 barrels according to the actual experience of the business, and the training samples in each barrel are guaranteed to be the same in quantity. The prediction sample may take the current shipping threshold of the current shipping route in the last month, and its threshold at the shipping price. The bank level is each charging interval with different settlement prices when the logistics line is recruited.
Further, in the model building process, the feature data may be classified into identification feature data of a carrier corresponding to a logistics route (i.e., ID class feature), cross feature data between the carrier and the logistics route (i.e., cross feature), current negotiated time feature data (i.e., time feature), and cost reference factor feature data (i.e., statistical feature). Wherein the identification characteristic data of the carrier comprises a carrier ID, a vehicle model ID and the like. The LightGBM model adopts a Many vs Many segmentation mode, so that the optimal segmentation of the category characteristics is realized, and the carrier ID, the vehicle model ID and the like can be directly input. Specifically, firstly, the characteristics to be combined are subjected to One-Hot coding (One-Hot Encoding), for example, the region where the carrier resources are located is represented as [0,1, 0], the circuit region is represented as [1, 0] and the cross multiplication is 25-dimensional coding, so that 24 codes of 0,1 and 1 are obtained. The statistical characteristics include a single kilometer price mean of the logistics route.
And continuously learning and constructing a decision tree through a Boosting algorithm of a core in a bargaining gradient classification model (namely, a LightGBM model) based on the feature data. Specifically, a corresponding decision tree is generated by continuously performing feature splitting, and a decision tree is added every time, namely a new objective function is learned to fit a residual error of a last bargained prediction result.
Specifically, the square loss function of n samples and the bargaining gradient classification model needing to be fitted are subjected to sigma summation to obtain a result
Figure BDA0003798424770000091
Regular term omega (f) corresponding to complexity of decision tree t ) And the sum of the preset constants constitutes an objective function Obj (t). By analyzing and optimizing the objective function Obj (t), f is obtained t (x i )。
Wherein, f t (x i ) The method is a new model needing to be added at this time, namely a bargaining gradient classification model needing to be fitted; x is a radical of a fluorine atom i Representing a sample; y is i For the true value of the loss function,
Figure BDA0003798424770000092
expressing a predicted value of a loss function, the loss function being a relationship between the predicted value and a true value; omega (f) t ) Representing the complexity of the bargaining gradient classification model, wherein if the new model is a decision tree model, indexes such as the depth of the decision tree, the number of leaf nodes and the like can reflect the complexity of the decision tree model; constant is a preset constant.
In embodiments of the present disclosure, high-dimensional combination features may be able to be extracted based on the LightGBM model. As shown in fig. 2, if the training sample traverses the decision tree structure model in the LightGBM model 201 and reaches a bold path with a bold leaf node, the feature data in the path is the high-dimensional combined feature 202 extracted from the training sample, the high-dimensional combined feature 202 can be represented by an index value of the leaf node, such as [0,1, 0], where each index value corresponds to different feature data, an index value of "1" indicates that the corresponding feature data has an influence on the bargained price reduction prediction result or the bargained price prediction result, and an index value of "0" indicates that the corresponding feature data has no influence on the bargained price reduction prediction result or the bargained price prediction result. Further, the LightGBM model extracts high-dimensional combination features (index values of leaf nodes of each decision tree) and low-order discrete features, and the high-dimensional combination features and the low-order discrete features are simultaneously used as input of a bargaining price reduction prediction model (namely, the FFM model 203) to predict bargaining price reduction, and corresponding parameters are set.
In particular, the features xi and xj may be combined two by two, and such combined features cannot be expressed by the linear combination of xi and xj. The resulting product xi xj becomes a new feature. In order not to miss any one of the possible useful combination features, in the implementation process, by exhaustively exhausting all the combinations of i and j, xixj,1 ≦ i ≦ n, and i < j ≦ n are added to the features, it needs to be noted that even if some xi are not one-hot features or some xixj are not useful features, the model will train the weighting coefficients of those useless features to 0 through the training of a large number of samples. The features xi and xj may refer to feature data randomly acquired from the high-dimensional combined feature and the low-dimensional discrete feature data. The FFM model corresponds to the following specific contents:
global bias term w 0 Characteristic xi and weight coefficient w thereof i Operation result of sigma summation after multiplication
Figure BDA0003798424770000101
New characteristic xi xj and weight coefficient w thereof ij Result of sigma summation operation after multiplication
Figure BDA0003798424770000102
Adding to obtain a predicted value
Figure BDA0003798424770000103
Wherein, w 0 Is a global bias term (constant term); w is a i A weight coefficient (contribution degree) of the ith feature; w is a ij The weight coefficient of the ith characteristic; the training samples x are n-dimensional vectors, x i Is a value in the ith dimension; x is a radical of a fluorine atom j Is the value in the jth dimension. w is a ij Is x i x j The corresponding hidden vector.
In the embodiment of the disclosure, the hidden vector is further refined based on the FFM model, and a field (domain) concept is introduced, and the information of different fields where the features are located is also taken into consideration. By introducing the concept of field, the FFM model is based on the FFM model to attribute the features with the same property to the same field, the hidden vector of each feature is not only one, but an independent hidden vector is learned for each field, and mutual influence is prevented. In addition, based on the LightGBM model, the FFM model and the corresponding function formula, softMax can be selected as an activation function, cross entropy is a loss function, each training sample is input, the gradient of the loss function is calculated until iteration is finished or the loss value of the training sample is not reduced, and the bargaining prediction hybrid classification model meeting the preset conditions is obtained.
In this step, the acquired feature data corresponding to the analyte flow line includes: identification characteristic data of a carrier corresponding to the analyte flow line, cross characteristic data between the carrier and the analyte flow line, current bargaining time characteristic data, cost reference factor characteristic data and the like.
Step 102: inputting the characteristic data into a preset bargaining forecasting mixed classification model for bargaining price reduction forecasting analysis, and outputting forecasting probability values corresponding to different bargaining intervals respectively; the bargaining price forecasting mixed classification model comprises a bargaining price gradient classification model used for extracting high-dimensional combination characteristics and a bargaining price reducing forecast model used for carrying out association analysis on characteristic data.
As shown in fig. 2, in a specific implementation process, firstly, the feature data is input into the negotiated price gradient classification model 201 obtained by training to perform high-dimensional combined feature extraction, so as to obtain high-dimensional combined feature data 202 output by the negotiated price gradient classification model; the high-dimensional combined feature data 202 is a high-dimensional feature vector represented based on index values of leaf nodes of a decision tree structure in the bargained gradient classification model 201. The feature data is low-dimensional discrete feature data associated with the analyte flow line, such as ID class features, cross features, temporal features, and statistical features corresponding to the analyte flow line. Then, the high-dimensional combined feature data 202 and the ID features and the cross features in the feature data are input into the bargaining price reducing estimation model for feature data association analysis, so as to obtain the prediction probability values corresponding to each bargaining interval (such as the bargaining price reducing interval) output by the bargaining price reducing estimation model 203. The bargaining gradient classification model and the bargaining reduced-amplitude estimation model are obtained by training based on sample characteristic data corresponding to a sample logistics route and a bargaining prediction result corresponding to the sample characteristic data, and the specific training process can refer to the above description and is not repeated herein.
The high-dimensional combined feature data and the feature data are input into the bargaining price reducing estimation model to perform feature data association analysis (for example, association analysis is performed according to the property and the type of the feature data), so as to obtain a prediction probability value corresponding to each bargaining price interval output by the bargaining price reducing estimation model, and the corresponding specific implementation includes: based on the bargaining price reduction pre-estimation model, performing feature combination on the high-dimensional combined feature data and the feature data according to different feature data types to realize that the feature data of the same type are classified into the same domain, and determining corresponding hidden vectors aiming at the domain; and determining a prediction probability value corresponding to each bargaining interval based on the hidden vector. For example, the features xi and xj may be combined in pairs, the combined feature cannot be represented by the linear combination of xi and xj, and the obtained product xixj becomes a new feature. In order to obtain effective combination characteristics, in the specific implementation process, through exhaustively exhausting all i, j combinations, xixj,1 is more than or equal to i and less than or equal to n, and i is more than or equal to j and less than or equal to n is added into the characteristics.
Step 103: and outputting a corresponding bargaining prediction result based on the size of the prediction probability value.
In the embodiment of the disclosure, the prediction probability value may be normalized based on a SoftMax logistic regression model, the size of a target probability value corresponding to each bargaining interval is determined, the bargaining interval corresponding to the maximum target probability value is determined as a target bargaining interval, and the target bargaining interval is sent to the user side as first bargaining guide information in the bargaining prediction result. In addition, an abnormal freight rate logistics line in the logistics line to be analyzed can be determined based on the prediction probability value, and attribute information corresponding to the abnormal freight rate logistics line is sent to the user side as second bargaining guide information in the bargaining prediction result. Furthermore, in order to provide an additional reference for bargaining to the user side, attribution analysis can be performed on a plurality of characteristic data types corresponding to the bargaining prediction result, and a sharp value corresponding to each characteristic data type is determined; and based on the magnitude of the Charpy value, carrying out influence sequencing on a plurality of characteristic data types corresponding to the bargained forecast result, and outputting a corresponding characteristic data type influence sequencing result.
In the process of predicting the price reduction of each current carrying logistics route, the bargaining interval corresponding to the new and old contracts can be equally divided into 10 categories, namely 10 labels, [ (-0.252, -0.182), (-0.182, -0.132), (-0.132, -0.0901), (-0.0901, -0.06), (-0.06, -0.0435), (-0.0435, -0.0233), (-0.0233, -0.00605), (-0.00605, 0.0), (0.0, 0.0400, 0401), (0.0401, 0.15) ]; wherein, the calculation mode of price amplitude reduction is (back freight rate-front freight rate)/front freight rate. For example, the negotiated price prediction mixed classification model obtained through training predicts price reduction ranges of all the current routes at the current threshold level in the last month, normalizes the prediction probability values based on a SoftMax logistic regression model, determines the target probability value corresponding to each negotiated price interval, and selects the price reduction range interval corresponding to the category with the highest probability (namely, the negotiated price interval corresponding to the maximum target probability value) as first negotiated price guidance information.
In the attribution analysis process of the target bargaining interval, a Shapley value can be used to generate a numerical value for each feature of each analyte flow line to represent the importance of the feature to the bargaining prediction result.
Assuming that the ith analyte flow line is xi, the jth characteristic of the ith analyte flow line is xi, j, the prediction probability value of the bargaining prediction mixed classification model to the ith analyte flow line is yi, and the baseline of the whole bargaining prediction mixed classification model (generally, the baseline is the target variable of all analyte flow lines)Value) of y base Then SHAP value obeys: predicting the mean value y of the target variables of all the analyzed flow lines corresponding to the whole bargaining mixed classification model base The method comprises the steps of obtaining a final prediction probability value yi of a current line of the ith analyte, summing the contribution value f (xi, 1) of the 1 st feature to the final prediction probability value yi in the current line of the ith analyte, the contribution value f (xi, 2) of the 2 nd feature to the final prediction probability value yi in the current line of the ith analyte and the contribution value f (xi, k) of the kth feature to the final prediction probability value yi in the current line of the ith analyte, and obtaining the prediction probability value yi of the current line of the ith analyte.
Wherein f (xi, 1) is the SHAP value of xi, j, that is, the contribution value of the 1 st feature in the ith analyte flow line to the final predicted probability value yi, when f (xi, 1)>0, the feature promotes the prediction probability value, namely the forward effect; conversely, the characteristic causes the prediction probability value to be reduced, namely, the characteristic has adverse effect; f (xi, 2) is the contribution value of the 2 nd feature in the ith analyte flow line to the final prediction probability value yi; f (xi, k) is the contribution value of the kth feature in the ith analyte flow line to the final prediction probability value yi, wherein k is greater than 2; y is base The average value of the target variables of all the analyte flow lines corresponding to the whole bargained predictive mixed classification model is the baseline of the whole bargained predictive mixed classification model.
The greatest advantage of the SHAP value is that the SHAP can reflect the influence of the features in each sample, and also shows the positive and negative of the influence. Therefore, what is finally given to the user side is that, in addition to the minimum and maximum bargaining reduction (i.e., the target bargaining interval), the features for increasing the bargaining amplitude are listed in order from large to small according to the shape value, that is, the feature data types corresponding to the bargaining prediction result are subjected to influence sorting to output a corresponding feature data type influence sorting result for additional reference during bargaining of the carrier.
It should be noted that, in the implementation process, the candidate logistics line that has been recently recruited/commissioned and failed in bargaining needs to be filtered out from the logistics line to be analyzed in advance, that is, for the logistics line that has been recently adjusted in price, the filtering is performed temporarily, and the bargaining analysis is not performed. In a specific implementation process, the bargaining list corresponding to the bargaining prediction result can be sent to the user side for business personnel to bargain, and the bargaining list issued to the business personnel in each area where the logistics route is located can contain attribute information corresponding to the abnormal freight transportation logistics route, such as second bargaining guidance information of business type, transportation type, bank level, area where the logistics route is located, corresponding cargo quantity and the like. Besides, the bargaining list may also include the current first bargaining instruction information of the logistics route, such as the current at-rate, minimum bargaining amplitude, maximum bargaining amplitude, etc. Further, the bargaining price reducing frame and the carrier will feedback can be obtained, for example, after the regional service personnel are collected in the link and the regional service personnel take the regional service personnel to the issued list, the bargaining price reducing frame is compared with the actual bargaining price reducing frame of the carrier, namely the actual bargaining price reducing frame of each carrier of each logistics route, and the collection result is used for updating the carrier side data.
As shown in fig. 3, in a specific application process, the following steps are included: steps 301 to 303: firstly, acquiring contract data of a three-party and self-operated line, carrying data of the three-party and self-operated line, future line cargo quantity prediction data of a logistics line, price per kilometer of a friend, oil price, high-speed normal government strategy and other data; step 304: carrying out bargaining price reduction prediction on the logistics route based on the current price by utilizing an AI classification model (namely a bargaining price prediction mixed classification model); step 305: screening out abnormal freight rate logistics routes with large prediction reduction amplitude; step 306: sending to service personnel; steps 307-309: the service personnel contacts the current line carrier to carry out line bargaining based on the bargaining and amplitude reduction prediction result, if the carrier agrees to bargaining, the service personnel can carry out a countersign contract to carry out logistics line cost reduction; if the carrier does not agree with bargaining, the logistics line can be brought into a re-bidding pool, and the carrier is replaced when the batch is to be recruited.
As shown in fig. 4, in a complete implementation flow, the following steps are included: steps 401 to 403: acquiring line side data, carrier side data and external associated data; step 404: performing data integration and updating; steps 405-406: executing feature engineering of a hybrid classification model (namely a bargaining prediction hybrid classification model), and constructing a training sample and a prediction sample of the hybrid classification model; step 407: constructing a mixed classification model based on the training samples and the prediction samples; step 408: carrying out mixed classification model tuning; step 409: performing mixed classification model evaluation; step 410: predicting the price reduction amplitude of the analyte flow line; step 411: filtering out the analyte flow line which is adopted recently or unsuccessful in bargaining at present; step 412: issuing a list to service personnel for bargaining; step 413: and feeding back the price reduction amplitude and the carrier will so as to enable the user to perform data integration and updating subsequently. In addition, the price reduction range uploaded by business personnel and the carrier will feedback result can be obtained and continuously used for data integration and updating.
The method for predicting the price interval of the logistics route comprises the steps of obtaining characteristic data corresponding to a logistics route to be analyzed, inputting the characteristic data into a preset bargaining prediction mixed classification model for bargaining amplitude reduction prediction analysis, and outputting prediction probability values corresponding to different bargaining intervals; the bargaining forecasting mixed classification model comprises a bargaining gradient classification model for extracting high-dimensional combination characteristics and a bargaining amplitude reduction forecasting model for performing characteristic data correlation analysis; and finally outputting a corresponding bargaining prediction result based on the size of the prediction probability value. The bargaining forecasting mixed classification model can be used for carrying out accurate bargaining analysis on the logistics route, and outputting a bargaining analysis result in time to guide the bargaining process of the logistics route, so that the accuracy and timeliness of the bargaining amplitude reduction forecasting of the logistics route are effectively improved, and the transportation cost of the logistics route is reduced.
Corresponding to the method for predicting the price interval of the logistics route, the invention also provides a device for predicting the price interval of the logistics route. Since the embodiment of the device is similar to the embodiment of the method, the description is simple, and the related points can be referred to the description of the embodiment of the method, and the embodiment of the price interval prediction device of the logistics route described below is only schematic. Fig. 5 is a schematic structural diagram of a price interval prediction apparatus for a logistics route according to an embodiment of the present disclosure.
The price interval prediction device of logistics circuits specifically comprises the following parts:
a feature data acquiring unit 501, configured to acquire feature data corresponding to a flow line to be analyzed;
a bargaining price reducing amplitude prediction analysis unit 502, configured to input the feature data into a preset bargaining price prediction mixed classification model for bargaining price reducing amplitude prediction analysis, and output prediction probability values corresponding to different bargaining intervals respectively; the bargaining price forecasting mixed classification model comprises a bargaining price gradient classification model for extracting high-dimensional combination characteristics and a bargaining price reducing estimation model for performing characteristic data correlation analysis;
a bargaining prediction result output unit 503, configured to output a corresponding bargaining prediction result based on the magnitude of the prediction probability value.
Further, the bargained price reduction prediction analysis unit is specifically configured to:
inputting the feature data into the bargaining gradient classification model for high-dimensional combined feature extraction to obtain high-dimensional combined feature data output by the bargaining gradient classification model; the high-dimensional combined feature data is a high-dimensional feature vector represented based on index values of leaf nodes of a decision tree structure in the bargained gradient classification model;
inputting the high-dimensional combined feature data and the feature data into the bargaining price reducing estimation model for feature data correlation analysis to obtain a prediction probability value corresponding to each bargaining interval output by the bargaining price reducing estimation model; wherein the feature data is low-dimensional discrete feature data associated with the analyte flow line;
the bargaining gradient classification model and the bargaining amplitude reduction estimation model are obtained by training based on sample characteristic data corresponding to a sample logistics line and a bargaining prediction result corresponding to the sample characteristic data.
Further, the step of inputting the high-dimensional combined feature data and the feature data into the negotiated price amplitude reduction estimation model to perform feature data association analysis to obtain a predicted probability value corresponding to each negotiated price interval output by the negotiated price amplitude reduction estimation model specifically includes:
based on the bargaining price reduction pre-estimation model, performing feature combination on the high-dimensional combined feature data and the feature data according to different feature data types to realize that the feature data of the same type are classified into the same domain, and determining corresponding hidden vectors aiming at the domain; and determining a prediction probability value corresponding to each bargaining interval based on the hidden vector.
Further, the bargaining prediction result output unit is specifically configured to: and normalizing the predicted probability values, determining the target probability value corresponding to each bargaining interval, determining the bargaining interval corresponding to the maximum target probability value as a target bargaining interval, and taking the target bargaining interval as first bargaining guide information in the bargaining prediction result.
Further, the bargaining prediction result output unit is specifically further configured to: and determining an abnormal freight rate logistics route in the logistics route to be analyzed based on the prediction probability value, and taking attribute information corresponding to the abnormal freight rate logistics route as second bargaining guide information in the bargaining prediction result.
Further, the price interval prediction device for the logistics route further includes:
the attribution analysis unit is used for carrying out attribution analysis on a plurality of characteristic data types corresponding to the bargaining prediction result and determining a sharp value corresponding to each characteristic data type;
and the characteristic influence sequencing unit is used for sequencing the influence of a plurality of characteristic data types corresponding to the bargaining prediction result based on the value of the Charprily value and outputting a corresponding characteristic data type influence sequencing result.
Further, the feature data includes: the system comprises identity identification characteristic data of a carrier corresponding to the analyte flow line, cross characteristic data between the carrier and the analyte flow line, current bargaining time characteristic data and cost reference factor characteristic data.
The price interval prediction device of the logistics route, which is disclosed by the embodiment of the disclosure, acquires the characteristic data corresponding to the logistics route to be analyzed, inputs the characteristic data into a preset bargaining price prediction mixed classification model for bargaining price reduction prediction analysis, and outputs prediction probability values corresponding to different bargaining intervals respectively; the bargaining forecasting mixed classification model comprises a bargaining gradient classification model for extracting high-dimensional combination characteristics and a bargaining amplitude reduction forecasting model for performing characteristic data correlation analysis; and finally outputting a corresponding bargaining prediction result based on the size of the prediction probability value. The bargaining forecasting mixed classification model can be used for carrying out accurate bargaining analysis on the logistics route, and outputting a bargaining analysis result in time to guide the bargaining process of the logistics route, so that the accuracy and timeliness of the bargaining amplitude reduction forecasting of the logistics route are effectively improved, and the transportation cost of the logistics route is reduced.
Corresponding to the provided method for predicting the price interval of the logistics line, the disclosure also provides electronic equipment. Since the embodiment of the electronic device is similar to the embodiment of the method described above, the description is relatively simple, and please refer to the description of the embodiment of the method described above for relevant points, and the electronic device described below is only exemplary. Fig. 6 is a schematic physical structure diagram of an electronic device according to an embodiment of the disclosure. The electronic device may include: a processor (processor) 601, a memory (memory) 602 and a communication bus 603, wherein the processor 601 and the memory 602 communicate with each other through the communication bus 603 and communicate with the outside through the communication interface 604. Processor 601 may invoke logic instructions in memory 602 to perform a method of price interval prediction for a logistics route, the method comprising: acquiring characteristic data corresponding to a flow line of an analyte; inputting the characteristic data into a preset bargaining forecasting mixed classification model for bargaining price reduction forecasting analysis, and outputting forecasting probability values corresponding to different bargaining intervals respectively; the bargaining price forecasting mixed classification model comprises a bargaining price gradient classification model for extracting high-dimensional combination characteristics and a bargaining price reducing estimation model for performing characteristic data correlation analysis; and outputting a corresponding bargaining prediction result based on the size of the prediction probability value.
Furthermore, the logic instructions in the memory 602 may be implemented in software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solutions of the present disclosure, which are essential or part of the technical solutions contributing to the prior art, may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a computer, or a network device) to execute all or part of the steps of the methods described in the embodiments of the present disclosure. And the aforementioned storage medium includes: various media capable of storing program codes, such as a Memory chip, a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
In another aspect, the present disclosure also provides a computer program product, where the computer program product includes a computer program stored on a processor-readable storage medium, where the computer program includes program instructions, and when the program instructions are executed by a computer, the computer is capable of executing the method for predicting the price interval of the logistics route provided by the above-mentioned method embodiments. The method comprises the following steps: acquiring characteristic data corresponding to a flow line of an analyte; inputting the characteristic data into a preset bargaining forecast mixed classification model for bargaining amplitude reduction forecast analysis, and outputting forecast probability values corresponding to different bargaining intervals respectively; the bargaining forecasting mixed classification model comprises a bargaining gradient classification model for extracting high-dimensional combination features and a bargaining amplitude reduction estimation model for performing feature data association analysis; and outputting a corresponding bargaining prediction result based on the size of the prediction probability value.
In another aspect, the disclosed embodiment further provides a processor-readable storage medium, where a computer program is stored on the processor-readable storage medium, and when the computer program is executed by a processor, the computer program is implemented to perform the method for predicting the price interval of the logistics route provided in the foregoing embodiments. The method comprises the following steps: acquiring characteristic data corresponding to a flow line of an analyte; inputting the characteristic data into a preset bargaining forecast mixed classification model for bargaining amplitude reduction forecast analysis, and outputting forecast probability values corresponding to different bargaining intervals respectively; the bargaining forecasting mixed classification model comprises a bargaining gradient classification model for extracting high-dimensional combination features and a bargaining amplitude reduction estimation model for performing feature data association analysis; and outputting a corresponding bargaining prediction result based on the size of the prediction probability value.
The processor-readable storage medium may be any available medium or data storage device that can be accessed by a processor, including, but not limited to, magnetic memory (e.g., floppy disks, hard disks, magnetic tape, magneto-optical disks (MOs), etc.), optical memory (e.g., CDs, DVDs, BDs, HVDs, etc.), and semiconductor memory (e.g., ROMs, EPROMs, EEPROMs, non-volatile memories (NAND FLASH), solid State Disks (SSDs)), etc.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one position, or may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment may be implemented by software plus a necessary general hardware platform, and may also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a computer, or a network device) to execute the method according to the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present disclosure, not to limit it; although the present disclosure has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present disclosure.

Claims (10)

1. A method for predicting a price interval of a logistics route is characterized by comprising the following steps:
acquiring characteristic data corresponding to a flow line of an analyte;
inputting the characteristic data into a preset bargaining forecasting mixed classification model for bargaining price reduction forecasting analysis, and outputting forecasting probability values corresponding to different bargaining intervals respectively; the bargaining forecasting mixed classification model comprises a bargaining gradient classification model for extracting high-dimensional combination features and a bargaining amplitude reduction estimation model for performing feature data association analysis;
and outputting a corresponding bargaining prediction result based on the size of the prediction probability value.
2. The method for predicting the price intervals of the logistics lines according to claim 1, wherein the characteristic data is input into a preset bargaining forecast mixed classification model for bargaining amplitude reduction forecast analysis, and forecast probability values corresponding to different bargaining intervals are output, and the method specifically comprises the following steps:
inputting the feature data into the bargaining gradient classification model for high-dimensional combined feature extraction to obtain high-dimensional combined feature data output by the bargaining gradient classification model; the high-dimensional combined feature data is a high-dimensional feature vector represented based on index values of leaf nodes of a decision tree structure in the bargained gradient classification model;
inputting the high-dimensional combined feature data and the feature data into the bargaining price reducing estimation model for feature data correlation analysis to obtain a prediction probability value corresponding to each bargaining interval output by the bargaining price reducing estimation model; wherein the feature data is low-dimensional discrete feature data associated with the analyte flow line;
the bargaining gradient classification model and the bargaining amplitude reduction estimation model are obtained by training based on sample characteristic data corresponding to a sample logistics line and a bargaining prediction result corresponding to the sample characteristic data.
3. The method for predicting the price interval of the logistics route according to claim 2, wherein the step of inputting the high-dimensional combined feature data and the feature data into the negotiated price amplitude reduction estimation model for feature data association analysis to obtain the predicted probability value corresponding to each negotiated price interval output by the negotiated price amplitude reduction estimation model specifically comprises the steps of:
based on the bargaining price reduction pre-estimation model, performing feature combination on the high-dimensional combined feature data and the feature data according to different feature data types to realize that the feature data of the same type are classified into the same domain, and determining corresponding hidden vectors aiming at the domain; and determining a prediction probability value corresponding to each bargaining interval based on the hidden vector.
4. The method for predicting the price interval of the logistics route according to claim 1, wherein outputting a corresponding bargained prediction result based on the prediction probability value specifically comprises: and normalizing the prediction probability values, determining the target probability values corresponding to the bargaining intervals, determining the bargaining interval corresponding to the maximum target probability value as a target bargaining interval, and taking the target bargaining interval as first bargaining guide information in the bargaining prediction result.
5. The method for predicting the price interval of the logistics route according to claim 1, wherein a corresponding bargained prediction result is output based on the prediction probability value, and specifically comprising: and determining an abnormal freight rate logistics route in the logistics routes to be analyzed based on the prediction probability value, and taking attribute information corresponding to the abnormal freight rate logistics route as second bargaining guide information in the bargaining prediction result.
6. The method for predicting the price interval of the logistics route according to claim 1, further comprising, after outputting the corresponding bargained prediction result based on the magnitude of the prediction probability value:
attributing and analyzing a plurality of characteristic data types corresponding to the bargaining prediction result, and determining a sharp value corresponding to each characteristic data type;
and based on the value of the Charprili value, carrying out influence sequencing on a plurality of characteristic data types corresponding to the bargaining prediction result, and outputting a corresponding characteristic data type influence sequencing result.
7. A price interval prediction device for a logistics route, comprising:
the characteristic data acquisition unit is used for acquiring characteristic data corresponding to a flow line of the analyte;
the bargaining price reducing amplitude prediction analysis unit is used for inputting the characteristic data into a preset bargaining price prediction mixed classification model for carrying out bargaining price reducing amplitude prediction analysis and outputting prediction probability values respectively corresponding to different bargaining intervals; the bargaining price forecasting mixed classification model comprises a bargaining price gradient classification model for extracting high-dimensional combination characteristics and a bargaining price reducing estimation model for performing characteristic data correlation analysis;
and the bargaining prediction result output unit is used for outputting a corresponding bargaining prediction result based on the prediction probability value.
8. The price interval prediction device of a logistics route of claim 7, wherein the bargained price reduction prediction analysis unit is specifically configured to:
inputting the feature data into the bargaining gradient classification model for high-dimensional combined feature extraction to obtain high-dimensional combined feature data output by the bargaining gradient classification model; the high-dimensional combined feature data is a high-dimensional feature vector represented based on index values of leaf nodes of a decision tree structure in the bargained gradient classification model;
inputting the high-dimensional combined feature data and the feature data into the bargaining price reducing estimation model for feature data correlation analysis to obtain a prediction probability value corresponding to each bargaining interval output by the bargaining price reducing estimation model; wherein the feature data is low-dimensional discrete feature data associated with the analyte flow line;
the bargaining gradient classification model and the bargaining amplitude reduction estimation model are obtained by training based on sample characteristic data corresponding to a sample logistics line and a bargaining prediction result corresponding to the sample characteristic data.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the computer program performs the steps of the method for price interval prediction of a logistics route according to any of the claims 1 to 6.
10. A processor-readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the steps of the method for predicting a price interval of a logistics route of any one of claims 1 to 6.
CN202210976114.4A 2022-08-15 2022-08-15 Logistics line price interval prediction method and device Pending CN115271826A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210976114.4A CN115271826A (en) 2022-08-15 2022-08-15 Logistics line price interval prediction method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210976114.4A CN115271826A (en) 2022-08-15 2022-08-15 Logistics line price interval prediction method and device

Publications (1)

Publication Number Publication Date
CN115271826A true CN115271826A (en) 2022-11-01

Family

ID=83751352

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210976114.4A Pending CN115271826A (en) 2022-08-15 2022-08-15 Logistics line price interval prediction method and device

Country Status (1)

Country Link
CN (1) CN115271826A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115619507A (en) * 2022-12-05 2023-01-17 阿里健康科技(杭州)有限公司 Method, device and equipment for determining target resource exchange amount of data object

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115619507A (en) * 2022-12-05 2023-01-17 阿里健康科技(杭州)有限公司 Method, device and equipment for determining target resource exchange amount of data object

Similar Documents

Publication Publication Date Title
Vargas et al. Deep leaming for stock market prediction using technical indicators and financial news articles
CN108509492B (en) Big data processing and system based on real estate industry
CN109636482B (en) Data processing method and system based on similarity model
CN112270546A (en) Risk prediction method and device based on stacking algorithm and electronic equipment
CN110930038A (en) Loan demand identification method, loan demand identification device, loan demand identification terminal and loan demand identification storage medium
CN111951050A (en) Financial product recommendation method and device
CN114581249B (en) Financial product recommendation method and system based on investment risk bearing capacity assessment
CN113537960A (en) Method, device and equipment for determining abnormal resource transfer link
CN115545886A (en) Overdue risk identification method, overdue risk identification device, overdue risk identification equipment and storage medium
CN115271826A (en) Logistics line price interval prediction method and device
Alizadeh et al. Development of a customer churn model for banking industry based on hard and soft data fusion
CN111210347A (en) Transaction risk early warning method, device, equipment and storage medium
CN111667307B (en) Method and device for predicting financial product sales volume
CN115034762A (en) Post recommendation method and device, storage medium, electronic equipment and product
Gurav et al. Predict stock market’s fluctuating behaviour: Role of investor’s sentiments on stock market performance
CN112396507A (en) Shadow division-based integrated SVM personal credit evaluation method
Li et al. Research on listed companies’ credit ratings, considering classification performance and interpretability
Singh et al. Utilising Sentiment Analysis To Predict Market Movements In Currency Exchange Rates
Alwadei et al. On the Sensitivity of Residual Networks for Time Series Classification
Suguna CHURN ANALYSIS FOR MAXIMIZING PROFIT USING STATISTICAL MODEL IN TELECOM INDUSTRY
CN118037429A (en) Multitasking prediction method for enterprise credit rating and credit expectation
CN118333739A (en) Method for constructing retail credit risk prediction model and retail credit business Scoremult model
CN118211064A (en) Training method and device of visit rate prediction model
CN118195756A (en) Data analysis method for resource allocation and electronic equipment
CN118333737A (en) Method for constructing retail credit risk prediction model and consumer credit business Scorebetai model

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