CN113095604B - Fusion method, device and equipment of product data and storage medium - Google Patents

Fusion method, device and equipment of product data and storage medium Download PDF

Info

Publication number
CN113095604B
CN113095604B CN202110639687.3A CN202110639687A CN113095604B CN 113095604 B CN113095604 B CN 113095604B CN 202110639687 A CN202110639687 A CN 202110639687A CN 113095604 B CN113095604 B CN 113095604B
Authority
CN
China
Prior art keywords
data
product data
classified product
fusion
classified
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.)
Active
Application number
CN202110639687.3A
Other languages
Chinese (zh)
Other versions
CN113095604A (en
Inventor
沈嘉良
王遥
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.)
Ping An Technology Shenzhen Co Ltd
Original Assignee
Ping An Technology Shenzhen 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 Ping An Technology Shenzhen Co Ltd filed Critical Ping An Technology Shenzhen Co Ltd
Priority to CN202110639687.3A priority Critical patent/CN113095604B/en
Publication of CN113095604A publication Critical patent/CN113095604A/en
Application granted granted Critical
Publication of CN113095604B publication Critical patent/CN113095604B/en
Priority to PCT/CN2022/071476 priority patent/WO2022257457A1/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • 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
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • 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

Landscapes

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

Abstract

The invention relates to the field of big data, and provides a method, a device, equipment and a storage medium for fusing product data, which are used for improving the value prediction accuracy of the product data. The fusion method of the product data comprises the following steps: classifying the product data to be processed based on asset classes to obtain a plurality of classified product data sets; respectively carrying out normalization fusion processing on the plurality of classified product data sets to obtain each normalization fusion data; respectively performing value autoregressive prediction on a plurality of classified product data sets through a structural vector autoregressive model, target index data and normalized fusion data to obtain initial value prediction data; performing inverse normalization processing on the initial value prediction data to obtain candidate value prediction data; and performing correlation fusion processing on the candidate value prediction data to obtain target value prediction data. In addition, the invention also relates to a block chain technology, and the product data to be processed can be stored in the block chain.

Description

Fusion method, device and equipment of product data and storage medium
Technical Field
The present invention relates to the field of big data prediction estimation, and in particular, to a method, an apparatus, a device, and a storage medium for fusing product data.
Background
With the development of computer technology, the demand for asset valuation is increasing. At present, value prediction is generally carried out on product data to be processed through an asset management system by adopting a cash flow conversion method or a market price comparison method to obtain the product value data.
However, since the cash flow conversion method or the market price comparison method performs value prediction from a single influence factor, value prediction cannot be accurately performed on different types and varying product data, and the prediction flexibility is low, resulting in low accuracy of value prediction of product data.
Disclosure of Invention
The invention provides a method, a device, equipment and a storage medium for fusing product data, which are used for improving the value prediction accuracy of the product data.
The invention provides a product data fusion method in a first aspect, which comprises the following steps:
the method comprises the steps of obtaining product data to be processed, and classifying the product data to be processed based on asset classes to obtain a plurality of classified product data sets;
respectively carrying out normalization fusion processing on the plurality of classified product data sets through a preset normalization model to obtain normalization fusion data corresponding to each classified product data;
obtaining target index data corresponding to each classified product data, and performing value autoregressive prediction on the classified product data sets through a preset structure vector autoregressive model, the target index data corresponding to each classified product data and normalized fusion data corresponding to each classified product data to obtain initial value prediction data corresponding to each classified product data set, wherein the target index data comprise economic index data and market index data;
performing inverse normalization processing on the initial value prediction data corresponding to each classified product data set through the normalization model to obtain candidate value prediction data corresponding to each classified product data set;
and performing correlation fusion processing on the candidate value prediction data corresponding to each classified product data set to obtain target value prediction data.
Optionally, in a first implementation manner of the first aspect of the present invention, the performing, by using a preset normalization model, normalization fusion processing on the multiple classified product data sets respectively to obtain normalization fusion data corresponding to each classified product data includes:
acquiring asset price data, normalized reference index data and normalized reference index coefficients of the normalized reference index data, which correspond to the classified product data sets respectively, wherein the normalized reference index data is used for indicating index factors influencing the normalized fusion processing of the asset price data on the basis of a preset reference price index;
and respectively carrying out value data operation based on normalization on the plurality of classified product data sets based on the asset price data, the normalized reference index data and the normalized reference index coefficient at the moment through a preset normalization model to obtain normalized fusion data corresponding to each classified product data set.
Optionally, in a second implementation manner of the first aspect of the present invention, the obtaining target index data corresponding to each classified product data, and performing value autoregressive prediction on the multiple classified product data sets through a preset structure vector autoregressive model, the target index data corresponding to each classified product data, and normalized fusion data corresponding to each classified product data, respectively, to obtain initial value prediction data corresponding to each classified product data set, where the target index data includes economic index data and market index data, includes:
acquiring target index data, disturbance item data and independent variable coefficients which respectively correspond to the plurality of classified product data sets, wherein the target index data comprises economic index data and market index data which respectively correspond to the plurality of classified product data sets;
and respectively carrying out structural vector autoregression operation processing on the plurality of classified product data sets through a preset structural vector autoregression model, target index data, disturbance item data and an independent variable coefficient corresponding to each classified product data and normalized fusion data corresponding to each classified product data to obtain initial value prediction data corresponding to each classified product data set.
Optionally, in a third implementation manner of the first aspect of the present invention, the performing, by using the normalization model, inverse normalization processing on the initial value prediction data corresponding to each classified product data set to obtain candidate value prediction data corresponding to each classified product data set includes:
acquiring forecast reference index data of initial value forecast data corresponding to each classified product data set and a forecast reference index coefficient of the forecast reference index data, wherein the forecast reference index data is used for indicating an index factor which influences the inverse normalization processing of asset price data on the basis of a preset reference price index;
and performing operation based on inverse normalization on the basis of the prediction reference index data, the prediction reference index coefficient, a preset inverse normalization formula and the initial value prediction data corresponding to each classified product data set through the normalization model to obtain candidate value prediction data corresponding to each classified product data set.
Optionally, in a fourth implementation manner of the first aspect of the present invention, the performing correlation fusion processing on candidate value prediction data corresponding to each classified product data set to obtain target value prediction data includes:
performing correlation analysis on the candidate value prediction data corresponding to each classified product data set to obtain a correlation coefficient corresponding to each classified product data set;
and carrying out weighted summation on the candidate value prediction data corresponding to each classified product data set through the correlation coefficient corresponding to each classified product data set to obtain target value prediction data.
Optionally, in a fifth implementation manner of the first aspect of the present invention, the obtaining product data to be processed and classifying the product data to be processed based on asset classes to obtain multiple classified product data sets includes:
receiving an asset valuation request sent by a target terminal, reading product data to be processed from a preset database based on the asset valuation request, and performing data preprocessing on the product data to be processed to obtain preprocessed product data;
and sequentially carrying out multi-level convolution feature extraction, attention mechanism-based feature fusion, asset class probability calculation and asset class classification on the preprocessed product data through a preset classification model to obtain a plurality of classified product data sets.
Optionally, in a sixth implementation manner of the first aspect of the present invention, after performing correlation fusion processing on candidate value prediction data corresponding to each classified product data set to obtain target value prediction data, the method further includes:
obtaining an error value based on the target value prediction data, matching a target optimization strategy corresponding to the error value, and executing an optimization process based on the target optimization strategy.
The second aspect of the present invention provides a product data fusion apparatus, including:
the classification module is used for acquiring product data to be processed and classifying the product data to be processed based on asset classes to obtain a plurality of classified product data sets;
the normalized fusion module is used for respectively carrying out normalized fusion processing on the plurality of classified product data sets through a preset normalized model to obtain normalized fusion data corresponding to each classified product data;
the prediction module is used for acquiring target index data corresponding to each classified product data, and performing value autoregressive prediction on the classified product data sets through a preset structure vector autoregressive model, the target index data corresponding to each classified product data and normalized fusion data corresponding to each classified product data to obtain initial value prediction data corresponding to each classified product data set, wherein the target index data comprise economic index data and market index data;
the inverse normalization module is used for performing inverse normalization processing on the initial value prediction data corresponding to each classified product data set through the normalization model to obtain candidate value prediction data corresponding to each classified product data set;
and the correlation fusion module is used for performing correlation fusion processing on the candidate value prediction data corresponding to each classified product data set to obtain target value prediction data.
Optionally, in a first implementation manner of the second aspect of the present invention, the normalization fusion module is specifically configured to:
acquiring asset price data, normalized reference index data and normalized reference index coefficients of the normalized reference index data, which correspond to the classified product data sets respectively, wherein the normalized reference index data is used for indicating index factors influencing the normalized fusion processing of the asset price data on the basis of a preset reference price index;
and respectively carrying out value data operation based on normalization on the plurality of classified product data sets based on the asset price data, the normalized reference index data and the normalized reference index coefficient at the moment through a preset normalization model to obtain normalized fusion data corresponding to each classified product data set.
Optionally, in a second implementation manner of the second aspect of the present invention, the prediction module is specifically configured to:
acquiring target index data, disturbance item data and independent variable coefficients which respectively correspond to the plurality of classified product data sets, wherein the target index data comprises economic index data and market index data which respectively correspond to the plurality of classified product data sets;
and respectively carrying out structural vector autoregression operation processing on the plurality of classified product data sets through a preset structural vector autoregression model, target index data, disturbance item data and an independent variable coefficient corresponding to each classified product data and normalized fusion data corresponding to each classified product data to obtain initial value prediction data corresponding to each classified product data set.
Optionally, in a third implementation manner of the second aspect of the present invention, the inverse normalization module is specifically configured to:
acquiring forecast reference index data of initial value forecast data corresponding to each classified product data set and a forecast reference index coefficient of the forecast reference index data, wherein the forecast reference index data is used for indicating an index factor which influences the inverse normalization processing of asset price data on the basis of a preset reference price index;
and performing operation based on inverse normalization on the basis of the prediction reference index data, the prediction reference index coefficient, a preset inverse normalization formula and the initial value prediction data corresponding to each classified product data set through the normalization model to obtain candidate value prediction data corresponding to each classified product data set.
Optionally, in a fourth implementation manner of the second aspect of the present invention, the relevance fusion module is specifically configured to:
performing correlation analysis on the candidate value prediction data corresponding to each classified product data set to obtain a correlation coefficient corresponding to each classified product data set;
and carrying out weighted summation on the candidate value prediction data corresponding to each classified product data set through the correlation coefficient corresponding to each classified product data set to obtain target value prediction data.
Optionally, in a fifth implementation manner of the second aspect of the present invention, the classification module is specifically configured to:
receiving an asset valuation request sent by a target terminal, reading product data to be processed from a preset database based on the asset valuation request, and performing data preprocessing on the product data to be processed to obtain preprocessed product data;
and sequentially carrying out multi-level convolution feature extraction, attention mechanism-based feature fusion, asset class probability calculation and asset class classification on the preprocessed product data through a preset classification model to obtain a plurality of classified product data sets.
Optionally, in a sixth implementation manner of the second aspect of the present invention, the product data fusion apparatus further includes:
and the optimization execution module is used for acquiring an error value based on the target value prediction data, matching a target optimization strategy corresponding to the error value, and executing an optimization process based on the target optimization strategy.
A third aspect of the present invention provides a product data fusion apparatus, including: a memory and at least one processor, the memory having instructions stored therein; the at least one processor invokes the instructions in the memory to cause the fusion device of the product data to perform the fusion method of the product data described above.
A fourth aspect of the present invention provides a computer-readable storage medium having stored therein instructions, which, when run on a computer, cause the computer to execute the above-described fusion method of product data.
In the embodiment of the invention, the classification based on the asset type is carried out on the product data to be processed, so that the product data to be processed with different asset types can be conveniently and effectively processed in a subsequent targeted manner; by respectively carrying out normalization fusion processing on a plurality of classified product data sets, a plurality of influence factors of different asset types are integrated, prediction factors of values of different types and changed product data are enriched, and the value prediction accuracy of the different types and changed product data is improved; respectively performing value autoregressive prediction on a plurality of classified product data sets through a preset structure vector autoregressive model, target index data corresponding to each classified product data and normalized fusion data corresponding to each classified product data, and performing inverse normalization processing on initial value prediction data corresponding to each classified product data set, so that structural relation calculation among standard asset price indexes (namely target index data) is realized, value change of assets under future user names is estimated according to the predicted initial value prediction data, and the value change meets risk measurement standards such as sub-additivity, interchangeability, consistency and the like; and performing correlation fusion processing on candidate value prediction data corresponding to each classified product data set to enable the candidate value prediction data to meet risk measurement standards such as additivity, interchangeability and consistency, and accordingly improving the value prediction accuracy of the product data.
Drawings
FIG. 1 is a schematic diagram of an embodiment of a fusion method of product data according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of another embodiment of a fusion method of product data according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of an embodiment of a product data fusion device according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of another embodiment of a product data fusion device according to an embodiment of the present invention;
fig. 5 is a schematic diagram of an embodiment of a product data fusion device in the embodiment of the present invention.
Detailed Description
The embodiment of the invention provides a method, a device, equipment and a storage medium for fusing product data, and improves the value prediction accuracy of the product data.
The terms "first," "second," "third," "fourth," and the like in the description and in the claims, as well as in the drawings, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It will be appreciated that the data so used may be interchanged under appropriate circumstances such that the embodiments described herein may be practiced otherwise than as specifically illustrated or described herein. Furthermore, the terms "comprises," "comprising," or "having," and any variations thereof, are intended to cover non-exclusive inclusions, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
For convenience of understanding, a specific flow of an embodiment of the present invention is described below, and referring to fig. 1, an embodiment of a method for fusing product data in an embodiment of the present invention includes:
101. and acquiring product data to be processed, and classifying the product data to be processed based on asset classes to obtain a plurality of classified product data sets.
It is to be understood that the execution subject of the present invention may be a fusion device of product data, and may also be a terminal or a server, which is not limited herein. The embodiment of the present invention is described by taking a server as an execution subject.
Wherein the products in the product data to be processed are specific items and financial products that can be used for asset valuation, such as: mortgage (house, car) and insurance policy, product data to be processed including asset base information and historical price data for specific items, asset base information and historical price data for financial products, and other data that can be used for asset valuation. The asset categories include a main category and a plurality of sub-categories, such as: the main category is a (insurance policy), and the corresponding subcategories are a1 (social insurance policy) and a2 (business insurance policy).
The server can extract the initial product data to be evaluated from a preset database after obtaining the user authorization, or the server can receive the initial product data to be evaluated sent by a user side (target terminal); after the server obtains initial product data to be evaluated, performing data preprocessing of security detection and duplicate removal on the initial product data to be evaluated to obtain preprocessed product data; the server can obtain a plurality of classified product data by performing asset class classification based on keywords on the preprocessed product data; the server can also classify the assets of the preprocessed product data by a preset classification model, so that a plurality of classified product data are obtained.
The server may perform the keyword-based asset class classification on the preprocessed product data to obtain a plurality of classified product data, and the performing process may specifically include: the server sequentially carries out word segmentation and keyword extraction on the preprocessed product data to obtain target keywords; the method comprises the steps of matching a target keyword with preset asset class vocabularies to obtain target asset class vocabularies, obtaining asset classes of the target asset class vocabularies, using resource classes as tags of preprocessed product data corresponding to the target keyword, classifying the preprocessed product data according to the tags to obtain a plurality of classified product data sets, wherein each classified product data set comprises classified product data of a plurality of subcategories, and one classified product data set corresponds to one main class.
102. And respectively carrying out normalization fusion processing on the plurality of classified product data sets through a preset normalization model to obtain normalization fusion data corresponding to each classified product data.
The server respectively carries out normalization fusion processing on a plurality of classified product data sets through a preset normalization model, and the specific implementation process comprises the following steps: the server acquires a prediction factor corresponding to each classified product data, wherein the prediction factor comprises index data, price data and coefficients which affect the overall fluctuation trend and amplitude of the assets of each classified product data in different periods, and the prediction factor comprises normalized reference index data and normalized reference index coefficients of the normalized reference index data; the server calls a pre-created normalization model to perform linear regression processing and operation processing on the plurality of classified product data respectively based on the prediction factors, so that normalized fusion data corresponding to the classified product data are obtained, wherein the normalization model is used for performing normalized fusion on the overall fluctuation trend and amplitude of the assets of the classified product data in different periods.
The specific implementation process of calling the pre-created normalization model by the server and respectively performing linear regression processing and operation processing on the plurality of classified product data based on the prediction factors so as to obtain the normalized fusion data corresponding to each classified product data can be as follows: the server calls a pre-created normalization model to perform linear regression processing and operation processing based on sub-categories on the classified product data based on the prediction factors to obtain main category normalized estimation data corresponding to the sub-categories, and sums the main category normalized estimation data corresponding to the sub-categories to obtain normalized fusion data corresponding to the classified product data, wherein the normalized fusion data are as follows: taking a classified product data set B taking the classified product data set as a main category B as an example, the classified product data set B comprises classified product data B11 corresponding to a subcategory B1 and classified product data B21 corresponding to a subcategory B2, a main category corresponding to a subcategory B1 and a subcategory B2 is B, the server calls a pre-created normalization model and performs linear regression processing and operation processing based on the subcategory on the classified product data B11 based on a prediction factor to obtain main category normalized estimation data B1 corresponding to the main category B of the subcategory B11, similarly, main category normalized estimation data B2 can be obtained, and the main category normalized estimation data B1 and the main category normalized estimation data B2 are added to obtain normalized fusion data corresponding to the classified product data set B.
By respectively carrying out normalization fusion processing on a plurality of classified product data sets, integrating a plurality of influence factors of different asset types, enriching prediction factors of values of different types and changed product data, and improving the value prediction accuracy of different types and changed product data.
103. And obtaining target index data corresponding to each classified product data, and performing value autoregressive prediction on a plurality of classified product data sets through a preset structure vector autoregressive model, the target index data corresponding to each classified product data and normalized fusion data corresponding to each classified product data to obtain initial value prediction data corresponding to each classified product data set, wherein the target index data comprise economic index data and market index data.
The server matches index data stored in a preset database according to the asset classes corresponding to the classified product data sets to obtain economic index data and market index data, namely target index data, corresponding to the classified product data sets respectively, and obtains factor factors of value autoregressive prediction, wherein the factor factors comprise disturbance terms, independent variable coefficients, model parameters and the like; the server performs fusion processing based on autoregression on a plurality of classified product data sets respectively through a preset structure vector autoregression model based on target index data, factor factors and normalized fusion data corresponding to each classified product data set to obtain initial value prediction data corresponding to each classified product data set, the initial value prediction data is asset value data corresponding to a prediction moment (namely the normalized fusion processing moment t + 1), and the initial value prediction data corresponding to each classified product data set comprises a structural relation among standard asset price indexes. Among other things, economic indicator data may include, but is not limited to, Consumer Price Index (CPI), Gross Domestic Product (GDP), and industrial production (industrial production), and market indicator data may include, but is not limited to, liquidity, industrial status, production cost, resource cost, and market demand.
The value autoregressive prediction is carried out on a plurality of classified product data sets through a preset structure vector autoregressive model, the structural relation calculation among standard asset price indexes is realized, the value change of the assets under the future user name is estimated according to the predicted initial value prediction data, the value change meets the risk measurement standards of sub-additivity, interchangeability, consistency and the like, and the value prediction accuracy of the product data is improved.
104. And performing inverse normalization processing on the initial value prediction data corresponding to each classified product data set through a normalization model to obtain candidate value prediction data corresponding to each classified product data set.
The server obtains prediction reference index data corresponding to initial value prediction data corresponding to each classified product data set and a prediction reference index coefficient of the prediction reference index data, determines the initial value prediction data, the prediction reference index data and the prediction reference index coefficient corresponding to each classified product data set as anti-normalization operation factors, calls a normalization model, and operates the operation factors based on a preset anti-normalization formula to obtain candidate value prediction data corresponding to each classified product data set, wherein the prediction reference index data are used for indicating index factors influencing anti-normalization processing of asset price data based on a preset reference price index.
105. And performing correlation fusion processing on the candidate value prediction data corresponding to each classified product data set to obtain target value prediction data.
The concrete implementation process of the server performing correlation fusion processing on the candidate value prediction data corresponding to each classified product data set may be as follows: the server can carry out numerical measurement on the structural relation (correlation) between the candidate value prediction data corresponding to each classified product data set to obtain a correlation coefficient corresponding to each classified product data set; the server can also extract data related to the structural relation from the calculation data generated in the calculation process of the candidate value prediction data corresponding to each classified product data set, so as to obtain the correlation coefficient corresponding to each classified product data set; and the server determines the relevant coefficients corresponding to the classified product data sets as the weighting coefficients corresponding to the classified product data sets, and calculates the sum of the classified product data sets according to the weighting coefficients so as to obtain target value prediction data.
By carrying out correlation fusion processing on candidate value prediction data corresponding to each classified product data set, the candidate value prediction data meet risk measurement standards such as sub-additivity, interchangeability and consistency, and the value prediction accuracy of the product data is improved.
In the embodiment of the invention, the classification based on the asset type is carried out on the product data to be processed, so that the product data to be processed with different asset types can be conveniently and effectively processed in a subsequent targeted manner; by respectively carrying out normalization fusion processing on a plurality of classified product data sets, a plurality of influence factors of different asset types are integrated, prediction factors of values of different types and changed product data are enriched, and the value prediction accuracy of the different types and changed product data is improved; respectively performing value autoregressive prediction on a plurality of classified product data sets through a preset structure vector autoregressive model, target index data corresponding to each classified product data and normalized fusion data corresponding to each classified product data, and performing inverse normalization processing on initial value prediction data corresponding to each classified product data set, so that structural relation calculation among standard asset price indexes (namely target index data) is realized, value change of assets under future user names is estimated according to the predicted initial value prediction data, and the value change meets risk measurement standards such as sub-additivity, interchangeability, consistency and the like; and performing correlation fusion processing on candidate value prediction data corresponding to each classified product data set to enable the candidate value prediction data to meet risk measurement standards such as additivity, interchangeability and consistency, and accordingly improving the value prediction accuracy of the product data.
Referring to fig. 2, another embodiment of the method for fusing product data according to the embodiment of the present invention includes:
201. and acquiring product data to be processed, and classifying the product data to be processed based on asset classes to obtain a plurality of classified product data sets.
Specifically, the server receives an asset valuation request sent by a target terminal, reads product data to be processed from a preset database based on the asset valuation request, and performs data preprocessing on the product data to be processed to obtain preprocessed product data; and sequentially carrying out multi-level convolution feature extraction, attention mechanism-based feature fusion, asset class probability calculation and asset class classification on the preprocessed product data through a preset classification model to obtain a plurality of classified product data sets.
The method comprises the steps that a server receives an asset valuation request sent by a target terminal, analyzes the asset valuation request to obtain asset valuation key information, the asset valuation key information comprises product data requirements and requirement information of asset value prediction, and product data to be processed corresponding to the asset valuation key information is read from a preset database; and sequentially carrying out data cleaning, data conversion, duplicate removal fusion and safety detection on the product data to be processed (namely the data preprocessing comprises the data cleaning, the data conversion, the duplicate removal fusion and the safety detection), thereby obtaining the preprocessed product data.
The method comprises the steps that a server calls a preset classification model, based on a multi-hierarchy convolutional neural network algorithm which is constructed in advance, hierarchy convolutional feature extraction is carried out on preprocessed product data to obtain each hierarchy feature, based on a preset attention mechanism, attention-based fusion is carried out on each hierarchy feature to obtain a target feature, asset class probability calculation is carried out on the target feature through a classification network in the preset classification model, a target asset class is determined according to the asset class probability obtained through calculation, the preprocessed product data are classified according to the target asset class to obtain a plurality of classified product data sets, and the convolutional neural network algorithms of each hierarchy can be the same or different.
The feature extraction is carried out on the preprocessed product data through a multi-level convolutional neural network algorithm, and the special benefit and the total energy accuracy of the extracted preprocessed product data are improved. The product data to be processed is classified based on the asset types, so that the product data to be processed of different asset types can be processed in a subsequent targeted and effective manner, and the value prediction accuracy of the product data is improved.
202. And respectively carrying out normalization fusion processing on the plurality of classified product data sets through a preset normalization model to obtain normalization fusion data corresponding to each classified product data.
Specifically, the server acquires asset price data, normalized reference index data and normalized reference index coefficients of the normalized reference index data at moments corresponding to a plurality of classified product data sets respectively, wherein the normalized reference index data is used for indicating index factors influencing the normalized fusion processing of the asset price data on the basis of a preset reference price index; and respectively carrying out value data operation based on normalization on a plurality of classified product data sets through a preset normalization model based on the asset price data at the moment, the normalized reference index data and the normalized reference index coefficient to obtain normalized fusion data corresponding to each classified product data set.
Wherein, the normalized reference index data is used for indicating index factors which influence the normalized fusion processing of the asset price data on the basis of a preset reference price index, such as: the index factors corresponding to the room price include the section and the age of the house.
For example, a classified product dataset a with a classified product dataset as a main category a is taken as an example, the classified product dataset a comprises classified product data a11 corresponding to a subcategory a1 and classified product data a21 corresponding to a subcategory a2, and the server acquires asset price data a1 of the classified product data a11 at time ttNormalized reference index data x1t,kAnd normalized reference index coefficient c1t,kAnd the time-of-day asset price data a2 of the classified product data a21 at time ttNormalized reference index data x2t,kAnd normalized reference index coefficient c2t,kBased on the preset calculation formula 1 and the time asset price data a1 through a preset normalization modeltNormalized reference index data x1t,kAnd normalized reference index coefficient c1t,kCalculating normalized fusion data A corresponding to the classified product data a111tWherein, the preset calculation formula 1 is:
Figure 394222DEST_PATH_IMAGE001
k and K represent the index number of the normalized reference index data and pass through a preset normalization modelBased on the preset calculation formula 2, the time asset price data a2tNormalized reference index data x2t,kAnd normalized reference index coefficient c2t,kCalculating normalized fusion data A corresponding to the classified product data a212tWherein, the preset calculation formula 2 is: a. the2t=a2t-
Figure 146278DEST_PATH_IMAGE002
c2t,kx2t,kK and K represent the index number of the normalized reference index data, and the fusion data A is normalized1tAnd normalizing the fused data A2tAnd adding to realize the normalized fusion processing of the classified product data set A, and obtaining the normalized fusion data corresponding to the classified product data set A.
203. And obtaining target index data corresponding to each classified product data, and performing value autoregressive prediction on a plurality of classified product data sets through a preset structure vector autoregressive model, the target index data corresponding to each classified product data and normalized fusion data corresponding to each classified product data to obtain initial value prediction data corresponding to each classified product data set, wherein the target index data comprise economic index data and market index data.
Specifically, the server acquires target index data, disturbance item data and independent variable coefficients which correspond to a plurality of classified product data sets respectively, wherein the target index data comprises economic index data and market index data which correspond to the plurality of classified product data sets respectively; and respectively carrying out structural vector autoregression operation processing on a plurality of classified product data sets through a preset structural vector autoregression model, target index data, disturbance item data and an independent variable coefficient corresponding to each classified product data and normalized fusion data corresponding to each classified product data to obtain initial value prediction data corresponding to each classified product data set.
For example, a classified product data set a with a classified product data set as a main category a and a classified product data set B with a classified product data set as a main category B are taken as examples, and normalized fusion data corresponding to the classified product data set a is anAnd the normalized fusion data corresponding to the classified product data set B is BnThe server obtains target index data x corresponding to the classified product data set AApData of disturbance term εANormalized fusion data is AnCorresponding independent variable coefficient cnAnd target index data xApCorresponding independent variable coefficient epAnd acquiring target index data x corresponding to the classified product data set BBpData of disturbance term εBNormalized fusion data is BnCoefficient of independent variation gnAnd target index data xBpCoefficient of independent variable hpThe server passes a calculation formula in a preset structure vector autoregressive model
Figure 968740DEST_PATH_IMAGE003
Wherein n represents the period number changing along with time, P and P represent the index number of the target index data, and the calculation processing of structural vector autoregression is carried out to obtain the initial value prediction data A corresponding to the classified product data set At+1The server passes a calculation formula in a preset structure vector autoregressive model
Figure 173457DEST_PATH_IMAGE004
Wherein n represents the number of periods over time, fnRepresents normalized fusion data as AnPerforming structural vector autoregression operation processing based on independent variable coefficients corresponding to initial value prediction data operation, wherein P and P represent the index number of target index data, and obtaining initial value prediction data B corresponding to classified product data set Bt+1。
204. And performing inverse normalization processing on the initial value prediction data corresponding to each classified product data set through a normalization model to obtain candidate value prediction data corresponding to each classified product data set.
Specifically, the server acquires prediction reference index data of initial value prediction data corresponding to each classified product data set and a prediction reference index coefficient of the prediction reference index data, wherein the prediction reference index data is used for indicating an index factor which influences the inverse normalization processing of the asset price data on the basis of a preset reference price index; and performing operation based on inverse normalization on the basis of the prediction reference index data, the prediction reference index coefficient, a preset inverse normalization formula and initial value prediction data corresponding to each classified product data set through a normalization model to obtain candidate value prediction data corresponding to each classified product data set.
For example, data A is predicted by classifying the initial value corresponding to product data set At+1(corresponding to Main Category A, which includes subcategory a1 and subcategory a 2) As an illustration, the server obtains initial value prediction data At+1Prediction reference index data y1 of the corresponding subcategory a1 at the prediction time t +1t+1,lAnd a prediction reference index coefficient m1t+1,lAnd initial value prediction data At+1Prediction reference index data y2 of the corresponding subcategory a2 at the prediction time t +1t+1,lAnd a prediction reference index coefficient m2t+1,lAnd through a normalization model, based on a preset inverse normalization formula 1:
Figure 372357DEST_PATH_IMAGE005
wherein L (letter L) and L denote the index numbers of the prediction reference index data, and the initial value prediction data At+1Prediction reference index data y1t+1,lAnd a prediction reference index coefficient m1t+1,lCalculating candidate value prediction data a1 corresponding to the A subclass 1 of the classified product data sett+1And based on a preset inverse normalization formula 2 through a normalization model:
Figure 41236DEST_PATH_IMAGE006
wherein L and L represent the number of indices for predicting the reference index data, and the initial value prediction data At+1Prediction reference index data y2t+1,lAnd a prediction reference index coefficient m2t+1,lCalculating candidate value prediction data a2 corresponding to the A subclass 2 of the classified product data sett+1Candidate value prediction data a1t+1And candidate value prediction data a2t+1Determining candidates corresponding to classified product data set AValue prediction data.
205. And performing correlation fusion processing on the candidate value prediction data corresponding to each classified product data set to obtain target value prediction data.
Specifically, the server performs relevance analysis on candidate value prediction data corresponding to each classified product data set to obtain a relevance coefficient corresponding to each classified product data set; and carrying out weighted summation on the candidate value prediction data corresponding to each classified product data set through the correlation coefficient corresponding to each classified product data set to obtain target value prediction data.
The server calls a preset correlation analysis algorithm to measure the structural relationship (correlation) between candidate value prediction data corresponding to each classified product data set to obtain a correlation coefficient corresponding to each classified product data set, determines the correlation coefficient corresponding to each classified product data set as a weighting coefficient corresponding to each classified product data set, and calculates the sum of each classified product data set according to the weighting coefficient to realize the correlation fusion processing of the candidate value prediction data corresponding to each classified product data set, thereby obtaining target value prediction data.
By carrying out correlation fusion processing on candidate value prediction data corresponding to each classified product data set, the candidate value prediction data meet risk measurement standards such as sub-additivity, interchangeability and consistency, and the value prediction accuracy of the product data is improved.
206. Obtaining an error value based on the target value prediction data, matching a target optimization strategy corresponding to the error value, and executing an optimization process based on the target optimization strategy.
The optimization strategy comprises an optimization scheme of the normalized model, a creating-frowns optimization scheme of the structure vector autoregressive model and an optimization scheme of an execution process corresponding to the target value prediction data. And the server acquires real value data corresponding to the target value prediction data and calculates an error value between the target value prediction data and the real value data. Comparing and analyzing a preset range value and an error value to obtain a target range value corresponding to the error value, generating a key value or a structured query statement of the target range value, retrieving a preset optimization strategy hash table through the key value or the structured query statement to obtain a target optimization strategy, acquiring a normalization adjustment model parameter, a normalization adjustment factor, a structure vector autoregressive adjustment model parameter and a structure vector autoregressive adjustment factor which are required by optimization and sent by a target terminal based on the target optimization strategy, recreating a normalization model through the normalization adjustment model parameter and the normalization adjustment factor, and recreating a structure vector autoregressive model through the structure vector autoregressive adjustment model parameter and the structure vector autoregressive adjustment factor; and receiving an execution script based on the target optimization strategy and sent by the target terminal, and adjusting and executing the flow nodes, the data processing mode, the execution program and the like of the execution process corresponding to the target value prediction data through the execution script.
By optimizing the normalization model, the construction of the structural vector autoregressive model and the execution process corresponding to the target value prediction data, the accuracy of the normalization model and the structural vector autoregressive model is improved, the accuracy of the target value prediction data is improved, and the value prediction accuracy of the product data is improved.
In the embodiment of the invention, the product data to be processed of different asset types can be conveniently, subsequently, pertinently and effectively processed, multiple influence factors of different asset types are integrated, the value prediction factors of different types and changed product data are enriched, the value prediction accuracy of different types and changed product data is improved, the structural relation calculation among standard asset price indexes (namely target index data) is realized, the value change of assets under the name of a future user is estimated according to the predicted initial value prediction data, and the risk measurement standards such as sub-additivity, interchangeability and consistency are met; the risk measurement standards such as additivity, interchangeability and consistency are met, so that the value prediction accuracy of the product data is improved, the accuracy of the normalization model and the structure vector autoregressive model can be improved by optimizing the establishment of the normalization model and the structure vector autoregressive model and the execution process corresponding to the target value prediction data, the accuracy of the target value prediction data is improved, and the value prediction accuracy of the product data is improved.
With reference to fig. 3, the method for fusing product data in the embodiment of the present invention is described above, and a device for fusing product data in the embodiment of the present invention is described below, where an embodiment of the device for fusing product data in the embodiment of the present invention includes:
the classification module 301 is configured to obtain product data to be processed, and perform asset class-based classification on the product data to be processed to obtain a plurality of classified product data sets;
the normalized fusion module 302 is configured to perform normalized fusion processing on the multiple classified product data sets through a preset normalized model, so as to obtain normalized fusion data corresponding to each classified product data;
the prediction module 303 is configured to obtain target index data corresponding to each classified product data, perform value autoregressive prediction on the multiple classified product data sets through a preset structure vector autoregressive model, the target index data corresponding to each classified product data, and normalized fusion data corresponding to each classified product data, and obtain initial value prediction data corresponding to each classified product data set, where the target index data includes economic index data and market index data;
the inverse normalization module 304 is configured to perform inverse normalization processing on the initial value prediction data corresponding to each classified product data set through the normalization model to obtain candidate value prediction data corresponding to each classified product data set;
and the correlation fusion module 305 is configured to perform correlation fusion processing on the candidate value prediction data corresponding to each classified product data set to obtain target value prediction data.
The function implementation of each module in the product data fusion device corresponds to each step in the product data fusion method embodiment, and the function and implementation process thereof are not described in detail herein.
In the embodiment of the invention, the classification based on the asset type is carried out on the product data to be processed, so that the product data to be processed with different asset types can be conveniently and effectively processed in a subsequent targeted manner; by respectively carrying out normalization fusion processing on a plurality of classified product data sets, a plurality of influence factors of different asset types are integrated, prediction factors of values of different types and changed product data are enriched, and the value prediction accuracy of the different types and changed product data is improved; respectively performing value autoregressive prediction on a plurality of classified product data sets through a preset structure vector autoregressive model, target index data corresponding to each classified product data and normalized fusion data corresponding to each classified product data, and performing inverse normalization processing on initial value prediction data corresponding to each classified product data set, so that structural relation calculation among standard asset price indexes (namely target index data) is realized, value change of assets under future user names is estimated according to the predicted initial value prediction data, and risk measurement standards such as additivity, interchangeability, consistency and the like are met; and performing correlation fusion processing on candidate value prediction data corresponding to each classified product data set to enable the candidate value prediction data to meet risk measurement standards such as sub-additivity, interchangeability and consistency, and the like, so that the value prediction accuracy of the product data is improved.
Referring to fig. 4, another embodiment of the product data fusion apparatus according to the embodiment of the present invention includes:
the classification module 301 is configured to obtain product data to be processed, and perform asset class-based classification on the product data to be processed to obtain a plurality of classified product data sets;
the normalized fusion module 302 is configured to perform normalized fusion processing on the multiple classified product data sets through a preset normalized model, so as to obtain normalized fusion data corresponding to each classified product data;
the prediction module 303 is configured to obtain target index data corresponding to each classified product data, perform value autoregressive prediction on the multiple classified product data sets through a preset structure vector autoregressive model, the target index data corresponding to each classified product data, and normalized fusion data corresponding to each classified product data, and obtain initial value prediction data corresponding to each classified product data set, where the target index data includes economic index data and market index data;
the inverse normalization module 304 is configured to perform inverse normalization processing on the initial value prediction data corresponding to each classified product data set through the normalization model to obtain candidate value prediction data corresponding to each classified product data set;
a correlation fusion module 305, configured to perform correlation fusion processing on candidate value prediction data corresponding to each classified product data set to obtain target value prediction data;
and the optimization execution module 306 is configured to obtain an error value based on the target value prediction data, match a target optimization strategy corresponding to the error value, and execute an optimization process based on the target optimization strategy.
Optionally, the normalization fusion module 302 may be further specifically configured to:
acquiring asset price data, normalized reference index data and normalized reference index coefficients of the normalized reference index data which correspond to the classified product data sets respectively, wherein the normalized reference index data is used for indicating index factors which influence the normalized fusion processing of the asset price data on the basis of a preset reference price index;
and respectively carrying out value data operation based on normalization on a plurality of classified product data sets through a preset normalization model based on the asset price data at the moment, the normalized reference index data and the normalized reference index coefficient to obtain normalized fusion data corresponding to each classified product data set.
Optionally, the prediction module 303 may be further specifically configured to:
acquiring target index data, disturbance item data and independent variable coefficients which respectively correspond to a plurality of classified product data sets, wherein the target index data comprises economic index data and market index data which respectively correspond to the plurality of classified product data sets;
and respectively carrying out structural vector autoregression operation processing on a plurality of classified product data sets through a preset structural vector autoregression model, target index data, disturbance item data and an independent variable coefficient corresponding to each classified product data and normalized fusion data corresponding to each classified product data to obtain initial value prediction data corresponding to each classified product data set.
Optionally, the denormalization module 304 may be further specifically configured to:
acquiring prediction reference index data of initial value prediction data corresponding to each classified product data set and a prediction reference index coefficient of the prediction reference index data, wherein the prediction reference index data is used for indicating an index factor which influences the inverse normalization processing of the asset price data on the basis of a preset reference price index;
and performing operation based on inverse normalization on the basis of the prediction reference index data, the prediction reference index coefficient, a preset inverse normalization formula and initial value prediction data corresponding to each classified product data set through a normalization model to obtain candidate value prediction data corresponding to each classified product data set.
Optionally, the relevance fusion module 305 may further be specifically configured to:
performing correlation analysis on candidate value prediction data corresponding to each classified product data set to obtain correlation coefficients corresponding to each classified product data set;
and carrying out weighted summation on the candidate value prediction data corresponding to each classified product data set through the correlation coefficient corresponding to each classified product data set to obtain target value prediction data.
Optionally, the classification module 301 may be further specifically configured to:
receiving an asset valuation request sent by a target terminal, reading product data to be processed from a preset database based on the asset valuation request, and performing data preprocessing on the product data to be processed to obtain preprocessed product data;
and sequentially carrying out multi-level convolution feature extraction, attention mechanism-based feature fusion, asset class probability calculation and asset class classification on the preprocessed product data through a preset classification model to obtain a plurality of classified product data sets.
The function implementation of each module and each unit in the product data fusion device corresponds to each step in the product data fusion method embodiment, and the function and implementation process thereof are not described in detail herein.
In the embodiment of the invention, the product data to be processed of different asset types can be conveniently, subsequently, pertinently and effectively processed, multiple influence factors of different asset types are integrated, the value prediction factors of different types and changed product data are enriched, the value prediction accuracy of different types and changed product data is improved, the structural relation calculation among standard asset price indexes (namely target index data) is realized, the value change of assets under the name of a future user is estimated according to the predicted initial value prediction data, and the risk measurement standards such as sub-additivity, interchangeability and consistency are met; the risk measurement standards such as additivity, interchangeability and consistency are met, so that the value prediction accuracy of the product data is improved, the accuracy of the normalization model and the structure vector autoregressive model can be improved by optimizing the establishment of the normalization model and the structure vector autoregressive model and the execution process corresponding to the target value prediction data, the accuracy of the target value prediction data is improved, and the value prediction accuracy of the product data is improved.
Fig. 3 and fig. 4 describe the fusion device of the product data in the embodiment of the present invention in detail from the perspective of the modular functional entity, and the fusion device of the product data in the embodiment of the present invention is described in detail from the perspective of hardware processing.
Fig. 5 is a schematic structural diagram of a product data fusion device according to an embodiment of the present invention, where the product data fusion device 500 may generate relatively large differences due to different configurations or performances, and may include one or more processors (CPUs) 510 (e.g., one or more processors) and a memory 520, and one or more storage media 530 (e.g., one or more mass storage devices) for storing applications 533 or data 532. Memory 520 and storage media 530 may be, among other things, transient or persistent storage. The program stored on the storage medium 530 may include one or more modules (not shown), each of which may include a series of instruction operations in the fusion device 500 for product data. Still further, the processor 510 may be configured to communicate with the storage medium 530 to execute a series of instruction operations in the storage medium 530 on the fusion device 500 of product data.
The product data fusion appliance 500 may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input-output interfaces 560, and/or one or more operating systems 531, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will appreciate that the configuration of the fusion device for product data shown in fig. 5 does not constitute a limitation of the fusion device for product data, and may include more or fewer components than those shown, or some components may be combined, or a different arrangement of components.
The present application further provides a fusion device of product data, including: a memory and at least one processor, the memory having instructions stored therein, the memory and the at least one processor interconnected by a line; at least one processor calls instructions in the memory to cause the fusion device of the product data to perform the steps in the fusion method of the product data.
The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium, and which may also be a volatile computer-readable storage medium, having stored therein instructions, which, when executed on a computer, cause the computer to perform the steps of the fusion method of product data.
Further, the computer-readable storage medium 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, and the like; the storage data area may store data created according to the use of the blockchain node, and the like.
The block chain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism, an encryption algorithm and the like. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a read-only memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
The above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will 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 invention.

Claims (10)

1. A method for fusing product data, the method comprising:
the method comprises the steps of obtaining product data to be processed, and classifying the product data to be processed based on asset classes to obtain a plurality of classified product data sets;
respectively carrying out normalization fusion processing on the plurality of classified product data sets through a preset normalization model to obtain normalization fusion data corresponding to each classified product data;
obtaining target index data corresponding to each classified product data, and performing value autoregressive prediction on the classified product data sets through a preset structure vector autoregressive model, target index data corresponding to each classified product data and normalized fusion data corresponding to each classified product data to obtain initial value prediction data corresponding to each classified product data set, wherein the target index data comprise economic index data and market index data, and when performing value autoregressive prediction on each classified product data set, normalized fusion data corresponding to the classified product data set and normalized fusion data corresponding to other classified product data sets need to be combined at the same time;
performing inverse normalization processing on the initial value prediction data corresponding to each classified product data set through the normalization model to obtain candidate value prediction data corresponding to each classified product data set;
and performing correlation fusion processing on the candidate value prediction data corresponding to each classified product data set to obtain target value prediction data.
2. The product data fusion method of claim 1, wherein the obtaining normalized fusion data corresponding to each classified product data by performing normalized fusion processing on the plurality of classified product data sets through a preset normalization model comprises:
acquiring asset price data, normalized reference index data and normalized reference index coefficients of the normalized reference index data, which correspond to the classified product data sets respectively, wherein the normalized reference index data is used for indicating index factors influencing the normalized fusion processing of the asset price data on the basis of a preset reference price index;
and respectively carrying out value data operation based on normalization on the plurality of classified product data sets based on the asset price data, the normalized reference index data and the normalized reference index coefficient at the moment through a preset normalization model to obtain normalized fusion data corresponding to each classified product data set.
3. The method for fusing product data according to claim 1, wherein the obtaining of the target index data corresponding to each classified product data performs value autoregressive prediction on the plurality of classified product data sets through a preset structure vector autoregressive model, the target index data corresponding to each classified product data and the normalized fusion data corresponding to each classified product data, so as to obtain initial value prediction data corresponding to each classified product data set, and the target index data includes economic index data and market index data, and includes:
acquiring target index data, disturbance item data and independent variable coefficients which respectively correspond to the plurality of classified product data sets, wherein the target index data comprises economic index data and market index data which respectively correspond to the plurality of classified product data sets;
and respectively carrying out structural vector autoregression operation processing on the plurality of classified product data sets through a preset structural vector autoregression model, target index data, disturbance item data and an independent variable coefficient corresponding to each classified product data and normalized fusion data corresponding to each classified product data to obtain initial value prediction data corresponding to each classified product data set, wherein when carrying out value autoregression prediction on each classified product data set, normalized fusion data corresponding to the classified product data set and normalized fusion data corresponding to other classified product data sets need to be simultaneously combined.
4. The method for fusing product data according to claim 1, wherein the performing, by the normalization model, inverse normalization processing on the initial value prediction data corresponding to each classified product data set to obtain candidate value prediction data corresponding to each classified product data set includes:
acquiring forecast reference index data of initial value forecast data corresponding to each classified product data set and a forecast reference index coefficient of the forecast reference index data, wherein the forecast reference index data is used for indicating an index factor which influences the inverse normalization processing of asset price data on the basis of a preset reference price index;
and performing operation based on inverse normalization on the basis of the prediction reference index data, the prediction reference index coefficient, a preset inverse normalization formula and the initial value prediction data corresponding to each classified product data set through the normalization model to obtain candidate value prediction data corresponding to each classified product data set.
5. The method according to claim 1, wherein the performing correlation fusion processing on the candidate value prediction data corresponding to each classified product data set to obtain target value prediction data includes:
performing correlation analysis on the candidate value prediction data corresponding to each classified product data set to obtain a correlation coefficient corresponding to each classified product data set;
and carrying out weighted summation on the candidate value prediction data corresponding to each classified product data set through the correlation coefficient corresponding to each classified product data set to obtain target value prediction data.
6. The method for fusing product data according to claim 1, wherein the obtaining product data to be processed and classifying the product data to be processed based on asset classes to obtain a plurality of classified product data sets comprises:
receiving an asset valuation request sent by a target terminal, reading product data to be processed from a preset database based on the asset valuation request, and performing data preprocessing on the product data to be processed to obtain preprocessed product data;
and sequentially carrying out multi-level convolution feature extraction, attention mechanism-based feature fusion, asset class probability calculation and asset class classification on the preprocessed product data through a preset classification model to obtain a plurality of classified product data sets.
7. The method for fusing product data according to any one of claims 1 to 6, wherein after performing correlation fusion processing on the candidate value prediction data corresponding to each classified product data set to obtain target value prediction data, the method further comprises:
obtaining an error value based on the target value prediction data, matching a target optimization strategy corresponding to the error value, and executing an optimization process based on the target optimization strategy.
8. A product data fusion apparatus, characterized in that the product data fusion apparatus comprises:
the classification module is used for acquiring product data to be processed and classifying the product data to be processed based on asset classes to obtain a plurality of classified product data sets;
the normalized fusion module is used for respectively carrying out normalized fusion processing on the plurality of classified product data sets through a preset normalized model to obtain normalized fusion data corresponding to each classified product data;
the prediction module is used for acquiring target index data corresponding to each classified product data, and performing value autoregressive prediction on the classified product data sets through a preset structure vector autoregressive model, the target index data corresponding to each classified product data and the normalized fusion data corresponding to each classified product data to obtain initial value prediction data corresponding to each classified product data set, wherein the target index data comprise economic index data and market index data, and when the value autoregressive prediction is performed on each classified product data set, the normalized fusion data corresponding to the classified product data set and the normalized fusion data corresponding to other classified product data sets need to be combined at the same time;
the inverse normalization module is used for performing inverse normalization processing on the initial value prediction data corresponding to each classified product data set through the normalization model to obtain candidate value prediction data corresponding to each classified product data set;
and the correlation fusion module is used for performing correlation fusion processing on the candidate value prediction data corresponding to each classified product data set to obtain target value prediction data.
9. A fusion apparatus of product data, characterized by comprising: a memory and at least one processor, the memory having instructions stored therein;
the at least one processor invokes the instructions in the memory to cause the fusion device of product data to perform the fusion method of product data of any of claims 1-7.
10. A computer-readable storage medium having instructions stored thereon, wherein the instructions, when executed by a processor, implement a fusion method of product data according to any one of claims 1-7.
CN202110639687.3A 2021-06-09 2021-06-09 Fusion method, device and equipment of product data and storage medium Active CN113095604B (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202110639687.3A CN113095604B (en) 2021-06-09 2021-06-09 Fusion method, device and equipment of product data and storage medium
PCT/CN2022/071476 WO2022257457A1 (en) 2021-06-09 2022-01-12 Product data fusion method, apparatus and device, and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110639687.3A CN113095604B (en) 2021-06-09 2021-06-09 Fusion method, device and equipment of product data and storage medium

Publications (2)

Publication Number Publication Date
CN113095604A CN113095604A (en) 2021-07-09
CN113095604B true CN113095604B (en) 2021-09-10

Family

ID=76664513

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110639687.3A Active CN113095604B (en) 2021-06-09 2021-06-09 Fusion method, device and equipment of product data and storage medium

Country Status (2)

Country Link
CN (1) CN113095604B (en)
WO (1) WO2022257457A1 (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113095604B (en) * 2021-06-09 2021-09-10 平安科技（深圳）有限公司 Fusion method, device and equipment of product data and storage medium

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110110886A (en) * 2019-03-21 2019-08-09 平安直通咨询有限公司上海分公司 Information forecasting method, device, computer equipment and storage medium
CN111695052A (en) * 2020-06-12 2020-09-22 上海智臻智能网络科技股份有限公司 Label classification method, data processing device and readable storage medium
CN112818215A (en) * 2021-01-12 2021-05-18 平安科技（深圳）有限公司 Product data processing method, device, equipment and storage medium

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050209959A1 (en) * 2004-03-22 2005-09-22 Tenney Mark S Financial regime-switching vector auto-regression
US20110166883A1 (en) * 2009-09-01 2011-07-07 Palmer Robert D Systems and Methods for Modeling Healthcare Costs, Predicting Same, and Targeting Improved Healthcare Quality and Profitability
US11669914B2 (en) * 2018-05-06 2023-06-06 Strong Force TX Portfolio 2018, LLC Adaptive intelligence and shared infrastructure lending transaction enablement platform responsive to crowd sourced information
CN109376899A (en) * 2018-08-31 2019-02-22 阿里巴巴集团控股有限公司 A kind of economic fluctuation data determination method and device
CN110826803A (en) * 2019-11-06 2020-02-21 广东电力交易中心有限责任公司 Electricity price prediction method and device for electric power spot market
CN110827091A (en) * 2019-11-12 2020-02-21 成都航天科工大数据研究院有限公司 Industrial raw material price prediction method
CN111563774A (en) * 2020-05-08 2020-08-21 上海腾暨物联网科技有限公司 Method and system for constructing coal price index prediction and supply-demand relation index
CN113095604B (en) * 2021-06-09 2021-09-10 平安科技（深圳）有限公司 Fusion method, device and equipment of product data and storage medium

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110110886A (en) * 2019-03-21 2019-08-09 平安直通咨询有限公司上海分公司 Information forecasting method, device, computer equipment and storage medium
CN111695052A (en) * 2020-06-12 2020-09-22 上海智臻智能网络科技股份有限公司 Label classification method, data processing device and readable storage medium
CN112818215A (en) * 2021-01-12 2021-05-18 平安科技（深圳）有限公司 Product data processing method, device, equipment and storage medium

Also Published As

Publication number Publication date
CN113095604A (en) 2021-07-09
WO2022257457A1 (en) 2022-12-15

Similar Documents

Publication Publication Date Title
Roy et al. Statistical methods in QSAR/QSPR
CN104025130B (en) Method, system and the equipment of importance between computational entity
US8489502B2 (en) Methods and systems for multi-credit reporting agency data modeling
US9489627B2 (en) Hybrid clustering for data analytics
CN111324642A (en) Model algorithm type selection and evaluation method for power grid big data analysis
CN111724238B (en) Method, device and equipment for evaluating product recommendation accuracy and storage medium
CN111507470A (en) Abnormal account identification method and device
CN112990386B (en) User value clustering method and device, computer equipment and storage medium
Yan et al. Data envelopment analysis classification machine
CN113095604B (en) Fusion method, device and equipment of product data and storage medium
Cheng et al. Data mining for fast and accurate makespan estimation in machining workshops
Darena et al. Machine learning-based analysis of the association between online texts and stock price movements
Dong et al. Forecasting Chinese stock market prices using Baidu search index with a learning-based data collection method
Steyn et al. Sentiment, emotions and stock market predictability in developed and emerging markets
Tian et al. Credit risk models using rule-based methods and machine-learning algorithms
Lee et al. Application of machine learning in credit risk scorecard
Beninel et al. Transfer learning using logistic regression in credit scoring
Li et al. Research on listed companies’ credit ratings, considering classification performance and interpretability
Li Exploring the Potential of Machine Learning Techniques for Predicting Travel Insurance Claims: A Comparative Analysis of Four Models
Chen et al. MDS: a novel method for class imbalance learning
Ashofteh Big data for credit risk analysis: Efficient machine learning models using pyspark
Raghunath et al. Applications of Data Mining in Predicting Stock Values
US20160042091A1 (en) System And Method Of Forming An Index
Łapczyński et al. The number of clusters in hybrid predictive models: does it really matter?
Devaguptam et al. Automated Health Insurance Processing Framework with Intelligent Fraud Detection, Risk Classification and Premium Prediction

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
REG Reference to a national code

Ref country code: HK

Ref legal event code: DE

Ref document number: 40056720

Country of ref document: HK