CN110210959A - Analysis method, device and the storage medium of financial data - Google Patents
Analysis method, device and the storage medium of financial data Download PDFInfo
- Publication number
- CN110210959A CN110210959A CN201910500343.7A CN201910500343A CN110210959A CN 110210959 A CN110210959 A CN 110210959A CN 201910500343 A CN201910500343 A CN 201910500343A CN 110210959 A CN110210959 A CN 110210959A
- Authority
- CN
- China
- Prior art keywords
- data
- financial data
- model
- score
- financial
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- 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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/12—Accounting
- G06Q40/125—Finance or payroll
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Technology Law (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The embodiment of the present application provides analysis method, device and the storage medium of a kind of financial data, is related to Data Analysis Services field.Method includes: the financial data for obtaining enterprise；By the first preset data handle model treatment described in financial data, obtain for indicate the financial data whether Yi Chang the first result.Since data target is easy to occur being arranged not reasonable, therefore its upper limit in the accuracy of judgement is very low.Compared to this, since the first preset data processing model can have the higher upper limit in the accuracy of judgement, therefore model treatment financial data is handled by the first preset data, obtain the first result can accurate response data it is whether abnormal.This improves judge financial data whether Yi Chang accuracy.
Description
Technical field
This application involves Data Analysis Services field, in particular to a kind of analysis method of financial data, device and
Storage medium.
Background technique
The financial data of enterprise can directly or indirectly reflect the financial situation of enterprise, wherein directly anti-in financial data
The data type for reflecting the financial situation of enterprise can be for example: net assets income ratio, the turnover of total assets, invested assets return rate
With the turnover of total assets etc.；And the data type of the reversed financial situation for reflecting enterprise can be for example among financial data: enterprise
Negative press quantity, negative press quantity growth rate, position vacant quantity and position vacant quantity growth rate etc..
Currently, can be by the way that the financial data of enterprise be matched with the data target being artificially correspondingly arranged, to judge financial number
According to whether there is exception.Although this mode can be realized judge whether financial data is abnormal, if but data target is artificially arranged
It is not reasonable, it is easy to appear the erroneous judgement to financial data, influences the accuracy of judgement.
Summary of the invention
The application is to provide analysis method, device and the storage medium of a kind of financial data, to improve the financial number of judgement
According to whether Yi Chang accuracy.
In a first aspect, the embodiment of the present application provides a kind of analysis method of financial data, which comprises
Obtain the financial data of enterprise；
Financial data described in model treatment is handled by the first preset data, is obtained for whether indicating the financial data
The first abnormal result.
In the embodiment of the present application, since data target is easy to occur being arranged not reasonable, therefore it is in the standard of judgement
The upper limit in true property is very low, is easy to appear judgement inaccuracy.Compared to this, since the first preset data processing model is in judgement
It can have the higher upper limit in accuracy, therefore the first preset data used to handle model treatment financial data, the first knot of acquisition
Fruit can more acurrate response data it is whether abnormal.This improves judge financial data whether Yi Chang accuracy.
With reference to first aspect, in the first possible implementation, the method also includes:
The financial data is matched with preset data target, is obtained for indicating whether the financial data is abnormal
Second result；
According to first result and described second as a result, determining whether the financial data is abnormal.
In the embodiment of the present application, since it is by the way of by data target matching and data processing model processing cooperation
To determine whether financial data abnormal, further increase judge financial data whether Yi Chang accuracy.
With reference to first aspect or the first possible implementation of first aspect, in second of possible implementation
In, financial data described in model treatment is handled by the first preset data, is obtained for indicating whether the financial data is abnormal
The first result, comprising:
A kind of financial data is separately input at least two first preset data processing models, is obtained each
First abnormality score of the first preset data processing model output；Wherein, at least two first preset data processing
The type of model is different；
According to first abnormality score, abnormal score is determined, wherein the exception score is for indicating described first
As a result.
In the embodiment of the present application, since model can be handled to this by different types of at least two first preset data
Financial data is handled, and is realized and is cooperated determining abnormal score using different types of first preset data processing model, makes
Obtaining abnormal score can more accurately reflect whether financial data is abnormal.
The possible implementation of second with reference to first aspect is looked forward to obtaining in the third possible implementation
After the financial data of industry, the method also includes:
Determine the degree of correlation between a kind of financial data and corresponding regression data, wherein the regression data
For other financial datas relevant to the financial data are returned to the financial data and are generated；
The degree of correlation is separately input at least one second preset data processing model, it is pre- to obtain each described second
If the second abnormality score of data processing model output；Wherein, the type of second preset data processing model and described the
The type that one preset data handles model is identical, and the parameter of one species model is different；
It is corresponding, according to first abnormality score, determine abnormal score, comprising:
According to first abnormality score and second abnormality score, the abnormal score is determined.
In the embodiment of the present application, since the degree of correlation variation between two data is also able to reflect out one of data
It is whether abnormal, therefore by participating in calculating financial data to the degree of correlation of relevant other financial datas, so that determining
Abnormal score more accurately reflects whether financial data is abnormal.
The third possible implementation with reference to first aspect, in the fourth possible implementation, according to described
First abnormality score and second abnormality score determine the abnormal score, comprising:
The highest abnormality score of score value is determined from first abnormality score and second abnormality score；Alternatively,
First abnormality score and second abnormality score are averaging, determine average mark, wherein the score value is highest
Abnormality score or the average mark are the abnormal score.
In the embodiment of the present application, due to determining that the mode of abnormality score includes selecting a variety of sides such as best result or averaging
Formula, so that determining that the mode of abnormality score can select according to actual needs, so that the abnormality score determined more pastes
Close actual conditions.
The third possible implementation with reference to first aspect, in a fifth possible implementation, described in foundation
The step of degree of correlation corresponding each second preset data processing model includes:
Obtain the history financial data collection of the enterprise；
Determine that the history financial data concentrates the degree of correlation of each financial data with corresponding regression data；
According to all degrees of correlation determined, the model parameter of each second preset data processing model is determined.
In the embodiment of the present application, since the processing of the second preset data can be established by the history financial data collection of enterprise
Model, so that the second preset data processing model established just can more meet the actual conditions of enterprise, so that second is pre-
If data processing model can export more accurate result for the financial data of the enterprise.
Second aspect, the embodiment of the present application provide a kind of analytical equipment of financial data, and described device includes:
Data acquisition module, for obtaining the financial data of enterprise；
Data processing module obtains for handling financial data described in model treatment by the first preset data and is used for table
Show the financial data whether Yi Chang the first result.
In conjunction with second aspect, in the first possible implementation,
The data processing module is also used to match the financial data with preset data target, obtains and is used for table
Show the financial data whether Yi Chang the second result；According to first result and described second as a result, determining the finance
Whether data are abnormal.
In conjunction with the possible implementation of the first of second aspect or second aspect, in second of possible implementation
In,
The data processing module is preset for a kind of financial data to be separately input at least two described first
Data processing model obtains the first abnormality score of each first preset data processing model output；Wherein, at least two
The type of the first preset data processing model is different；According to first abnormality score, abnormal score is determined,
In, the exception score is for indicating first result.
In conjunction with second of possible implementation of second aspect, in the third possible implementation, in the number
After the financial data for obtaining module acquisition enterprise,
The data processing module is also used to determine the phase between a kind of financial data and corresponding regression data
Guan Du, wherein the regression data is to return other financial datas relevant to the financial data to the financial data
And it generates；The degree of correlation is separately input at least one second preset data processing model, it is pre- to obtain each described second
If the second abnormality score of data processing model output；Wherein, the type of second preset data processing model and described the
The type that one preset data handles model is identical, and the parameter of one species model is different；
Corresponding, the data processing module is used for according to first abnormality score and second abnormality score, really
Make the abnormal score.
In conjunction with the third possible implementation of second aspect, in the fourth possible implementation,
The data processing module, due to determining score value from first abnormality score and second abnormality score
Highest abnormality score；Alternatively, being averaging to first abnormality score and second abnormality score, average mark is determined
Number, wherein the highest abnormality score of score value or the average mark are the abnormal score.
In conjunction with the third possible implementation of second aspect, in a fifth possible implementation,
The data processing module is also used to obtain the history financial data collection of the enterprise；Determine the history finance
The degree of correlation of each financial data and corresponding regression data in data set；According to all degrees of correlation determined, institute is determined
State the model parameter of the corresponding each second preset data processing model of all degrees of correlation.
The third aspect, the embodiment of the present application provide a kind of electronic equipment, comprising: communication interface, memory lead to described
The processor that letter interface is connected with the memory；
The memory, for storing program；
The processor, for calling and running described program, with execute as first aspect or first aspect is any can
The analysis method of financial data described in the implementation of energy.
Fourth aspect, the embodiment of the present application provide a kind of computer-readable storage media, store on the storage medium
Have program code, when said program code is run by the computer, execute as first aspect or first aspect is any can
The analysis method of financial data described in the implementation of energy.
To enable the above objects, features, and advantages of the application to be clearer and more comprehensible, preferred embodiment is cited below particularly, and cooperate
Appended attached drawing, is described in detail below.
Detailed description of the invention
Technical solution in ord to more clearly illustrate embodiments of the present application, below will be to needed in the embodiment attached
Figure is briefly described, it should be understood that the following drawings illustrates only some embodiments of the application, therefore is not construed as pair
The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 shows a kind of first pass figure of the analysis method of financial data provided by the embodiments of the present application；
Fig. 2 shows a kind of second flow charts of the analysis method of financial data provided by the embodiments of the present application；
Fig. 3 shows the structural block diagram of a kind of electronic equipment provided by the embodiments of the present application；
Fig. 4 shows a kind of structural block diagram of the analytical equipment of financial data provided by the embodiments of the present application.
Specific embodiment
Below in conjunction with attached drawing in the embodiment of the present application, technical solutions in the embodiments of the present application is described.
It should also be noted that similar label and letter indicate similar terms in following attached drawing, therefore, once a certain Xiang Yi
It is defined in a attached drawing, does not then need that it is further defined and explained in subsequent attached drawing.Meanwhile the application's
In description, term " first ", " second " etc. are only used for distinguishing description, are not understood to indicate or imply relative importance.
Referring to Fig. 1, the embodiment of the present application provides a kind of analysis method of financial data, the analysis side of the financial data
Method can be executed by electronic equipment, which can be terminal or server, wherein terminal can be PC
(personal computer, PC), tablet computer, smart phone, personal digital assistant (personal digital
Assistant, PDA) etc.；Server can be network server, database server, Cloud Server or by multiple child servers
The server set of composition at etc..
Specifically, the analysis method of the financial data may include: step S100 and step S200.
Step S100: the financial data of enterprise is obtained.
Step S200: financial data described in model treatment is handled by the first preset data, is obtained for indicating the wealth
Be engaged in data whether Yi Chang the first result.
Successively step S100 and step S200 will be described in detail below.
For the financial data convenient for accurately handling enterprise, electronic equipment can establish that analysis handles the financial data
One preset data handles model.
Due to enterprise financial data type can there are many, such as a plurality of types of financial datas can be respectively net
Return on Assets, the turnover of total assets, invested assets return rate, the turnover of total assets, enterprise's negative press quantity, negative press
Quantity growth rate, position vacant quantity and position vacant quantity growth rate.It is different for whether accurate every kind of financial data of determination has
Often, electronic equipment can establish the corresponding first preset data processing model of every kind of financial data respectively, so that every kind established
First preset data processing model can targetedly handle a kind of corresponding financial data.Wherein, due to establishing every kind first
The process that preset data handles model is roughly the same, for ease of understanding this programme, and the present embodiment will be to establish a certain financial number
It is illustrated for handling model according to corresponding first preset data.
Specifically, when electronic equipment can be preset by obtaining from the modes such as the extraction of the database of enterprise or user's input
The history financial data collection of Jian Duanneigai enterprise, wherein history financial data concentration may include the enterprise in preset time period
Multiple financial datas of interior generation, for example, preset time period can be away from the trimestral duration of current date, history financial data
Concentration may be embodied in the daily enterprise's negative press quantity of the three Ge Yuenei enterprise.
Certainly, the duration of preset time period is not limited to enumerating for the present embodiment, can also carry out according to the actual situation
Selection, if such as the characteristics of financial data to need to analyze its variation in longer period can determine if it is different
Often, then settable long of preset time period, such as away from current date half a year even 1 year.
In the present embodiment, due to the difference of actual demand, model treatment one can be handled using first preset data
Kind financial data, or a kind of financial data of model treatment can also be handled using at least two first preset datas, and at least
The type of two the first preset data processing models is different.For example, if actual demand is on the basis of guaranteeing accuracy
The computational load for reducing electronic equipment, can be according to the data type of the financial data, using most suitable processing data class
The first preset data of one kind of type handles the model treatment financial data, for example uses mixed Gauss model.In another example if practical
Demand is to need very high accuracy, then can handle model using mutually different at least two first preset data of type
The financial data is handled, for example uses Gauss model, mixed Gauss model, one-class SVM (one-class Support
Vector Machine, one class-support vector machine) in model and i-forest (Isolation forest isolates forest) model
At least two.
Therefore, electronic equipment can establish the first of quantity meet demand using history financial data collection according to actual needs
Preset data handles model.It wherein, is the technical solution convenient for being fully understood by the application, the present embodiment will be to electronic equipment such as
What is established for each first preset data processing model using history financial data collection is illustrated.
If it is Gauss model that the first preset data, which handles model, Gauss model can be as shown in following formula (1):
In formula (1), p (x) is the output of Gauss model, and it is respectively mean value and mark that μ, δ, which are the model parameter of Gauss model,
It is quasi- poor.
Electronic equipment can calculate whole wealth by calculating whole financial datas that history financial data is concentrated
The mean value and standard deviation for data of being engaged in.So, the mould of the Gauss model is arranged according to calculated mean value and standard deviation for electronic equipment
Shape parameter is just realized and establishes first preset data processing model.
If it is mixed Gauss model that the first preset data, which handles model, Gauss model can be as shown in following formula (2):
In formula (1), p (x) is the output of mixed Gauss model, and K is the number for constituting the Gauss model of mixed Gauss model,
And K is the natural number greater than 1, ωiFor the weight of i Gauss model, μi, σiFor the model parameter of i-th of Gauss model,
Respectively mean value and standard deviation.
Electronic equipment is calculated also by whole financial datas that history financial data is concentrated, and also calculates whole wealth
The mean value and standard deviation for data of being engaged in.So, electronic equipment is according in calculated mean value and standard deviation setting mixed Gauss model
The model parameter of each Gauss model is also realized and establishes first preset data processing model.
If the first preset data handle model be one-class SVM model or i-forest model, can be first by history
Financial data is concentrated normally to be extracted with abnormal financial data respectively, using normal financial data as forward direction sample to the
One preset data processing model is trained, and handles mould to the first preset data using abnormal financial data as negative sense sample
Type is trained.Determine that (weight is the processing of the first preset data to the weight in the first preset data processing model by training
The model parameter of model), first preset data processing model is established to realize.
In the present embodiment, since regression relation is also able to reflect between the financial data of the other types of financial data and enterprise
Whether the financial data is abnormal out, for example, under normal circumstances, the turnover of total assets can use how regression parameter (determines
Regression parameter is in subsequent detailed description) calculate how the regression data returned to invested assets return rate (specifically determines for 0.7
Regression data also will be in subsequent detailed description), it is that difference 0.5 is invested assets return between 0.65 with invested assets return rate
The degree of correlation of rate and regression data, and the degree of correlation is that 0.5 expression invested assets return rate is normal.And if the degree of correlation changes to 2
When, illustrate that invested assets return rate is possible to exception occurred.Therefore, electronic equipment can also be according to the correlation of financial data
Degree establishes the second preset data for handling the degree of correlation and handles model, to react corresponding experience by the processing degree of correlation
Whether data are abnormal.
It should be noted that it is corresponding to can establish every kind of financial data when establishing the first preset data processing model
First preset data handles model.It is handled unlike model from the first preset data is established, due to being not every kind of financial number
According to there is relevant other financial datas, the financial data of not associated other financial datas can not calculate correlation
Degree.It therefore, can be to every kind of financial number with relevant other financial datas when establishing the second preset data processing model
Model is handled according to corresponding second preset data is established.On this basis, electronic equipment can be first to a variety of financial numbers of enterprise
According to being screened, to determine that there is every kind of financial data of relevant other financial datas.
Specifically, electronic equipment it is complete can to obtain enterprise by from the modes such as the extraction of the database of enterprise or user's input
The financial data of portion's type.Electronic equipment can be averaging such as arithmetic mean or weighting to the financial data of all categories
It is average, to determine degree of correlation reference value.On this basis, electronic equipment can utilize covariance matrix, all categories
Every kind of financial data and degree of correlation reference value in financial data are calculated.For example, covariance matrix can be such as following formula (3) institute
Show:
Σ=E [(x-u) ' (x-u)] (3)
In formula (3), x is any financial data in the financial data of all categories, and u is degree of correlation reference value.
Electronic equipment can determine the reference degree of correlation of the every kind of financial data compared with degree of correlation reference value by calculating.Electricity
Sub- equipment can be screened to calculated with reference to the degree of correlation by presetting relevance threshold, be greater than correlation to filter out
Spend reference the degree of correlation of threshold value, and be greater than each of relevance threshold with reference to the corresponding financial data of the degree of correlation can be with
Relevant other financial datas a kind of financial data.
In every kind of financial data of the other financial datas for determining to have associated, electronic equipment can be according to every
The degree of correlation between kind financial data and corresponding regression data establishes the corresponding second preset data processing mould of every kind of degree of correlation
Type.
For example, the stronger financial data of correlation includes: that net assets income ratio, the turnover of total assets and enterprise are negatively new
Hear quantity.Electronic equipment can generate regression data and returning to the turnover of total assets to net assets income ratio, determine
The degree of correlation between net assets income ratio and corresponding regression data, and according to the degree of correlation of net assets income ratio corresponding second
Preset data handles model.And electronic equipment can also be given birth to and returning to net assets income ratio to the turnover of total assets
At regression data, the degree of correlation between the turnover of total assets and corresponding regression data is determined, and according to the turnover of total assets
Corresponding second preset data of the degree of correlation handle model.It is understood that being received for enterprise's negative press quantity and net assets
Beneficial rate, and for the turnover of total assets and enterprise's negative press quantity, each correlation can also be established based on above-mentioned process
Corresponding second preset data handles model.Therefore 3 kinds of financial datas are based on, it can establish 6 kinds of degrees of correlation corresponding second
Preset data handles model.
Wherein, substantially due to the process of the corresponding second preset data processing model of the degree of correlation for establishing every kind of financial data
Identical, this programme for ease of understanding, the present embodiment is still to establish corresponding second according to the degree of correlation of a certain financial data
It is illustrated for preset data processing model.
Specifically, electronic equipment is also with logical based on determining that the reference degree of correlation of the financial data is greater than relevance threshold
It crosses from the modes such as the extraction of the database of enterprise or user's input, obtains the history financial data of the financial data in preset time period
Collection.
Further, electronic equipment can first determine that history financial data concentrates each financial data and corresponding time
Return the degree of correlation between data, to establish corresponding second preset data processing mould using all degrees of correlation determined
Type.
As the optional way for determining the degree of correlation, electronic equipment can use history financial data and concentrate at least partly
Financial data, determine relevant other financial datas revert to this at least partly financial data match needed for recurrence join
Number, for example, net profit growth rate be A, and net assets growth rate be B, using least square method determine each B revert to often
Regression parameter needed for a A matching is a ' and b ', then is integrated to the multiple a ' and multiple b ' that determine, can be determined most
Whole regression parameter a and b.Electronic equipment calculates the financial data using the regression parameter of each financial data,
That determines each financial data returns to data, for example, net profit growth rate is A's in the case where regression parameter is a and b
Returning to data A ' is aB+b.To obtain each financial data return between data and the financial data go out difference be just each
The degree of correlation between financial data and corresponding regression data, for example, A '-A is then in the case where returning to data A ' is aB+b
The degree of correlation between net profit growth rate and corresponding regression data.
In the present embodiment, also due to the difference of actual demand, handles a kind of the second present count of the degree of correlation of financial data
It can be selected according to actual needs according to the quantity of processing model.For example, it is also possible to handle model using second preset data
The corresponding degree of correlation is handled, or can also be using the corresponding degree of correlation of at least two second preset datas processing model treatment.
Wherein, for convenient for actual implementation, the second preset data handles model can be identical as the first preset data processing type of model,
And the model parameter of the first preset data of one species processing model and the second preset data processing model is different.For example, first
Preset data processing model is that Gauss model is obtained through financial data training, and the second preset data handles model as Gauss model warp
Degree of correlation training obtains.Since the type that the first preset data handles model and the second preset data processing model is all Gaussian mode
Type, therefore its type is identical.And since trained financial data and the degree of correlation are two kinds of entirely different data, after training
First preset data handles model and the model parameter of the second preset data processing model is not identical.
It should be noted that being also the technical solution convenient for being fully understood by the application, the present embodiment will be to electronic equipment
How to be established for every kind of the second preset data processing model and is illustrated using all degrees of correlation determined.
If it is Gauss model that the second preset data, which handles model, electronic equipment can be calculated all degrees of correlation, with
Calculate the mean value and standard deviation of all degrees of correlation.So, the height is arranged according to calculated mean value and standard deviation in electronic equipment
The model parameter of this model just realizes and establishes second preset data processing model.
If it is mixed Gauss model that the second preset data, which handles model, electronic equipment can also be counted all degrees of correlation
It calculates, also calculates the mean value and standard deviation of all degrees of correlation.So, electronic equipment is arranged according to calculated mean value and standard deviation
The model parameter of each Gauss model in mixed Gauss model also achieves and establishes second preset data processing model.
If the second preset data handle model be one-class SVM model or i-forest model, can also be first by institute
There is the normal and abnormal degree of correlation in the degree of correlation to extract respectively, is preset the normal degree of correlation as positive sample to second
Data processing model is trained, and is instructed using the abnormal degree of correlation as negative sense sample to the second preset data processing model
Practice.Changing the weight in the second preset data processing model by training, (weight is the mould that the second preset data handles model
Shape parameter), second preset data processing model is established to realize.
It should be noted that the first preset data processing model and/or the can be established according to actual needs in practice
Two preset datas handle model.For example, if needing to determine whether financial data is different in practice by analyzing financial data itself
Often, the first preset data processing model is established using history financial data collection, without establishing the degree of correlation corresponding second
Preset data handles model.In another example if needing to determine by the degree of correlation of analyzing financial data in practice, financial data is
No exception establishes the second preset data processing model using all degrees of correlation that history financial data collection is determined, thus nothing
The corresponding first preset data processing model of financial data need to be established.In another example if need in practice by analyzing financial data and
The degree of correlation of analyzing financial data can be established using history financial data collection and be corresponded in conjunction with determining whether financial data is abnormal
The first preset data handle model while, also established using all degrees of correlation that history financial data collection is determined corresponding
Second preset data handles model.
After establishing the first preset data processing model and/or the second preset data processing model, electronic equipment can
Determine whether financial data abnormal using the first preset data processing model and/or the second preset data processing model, i.e., it is electric
The step S100 and step S200 of the analysis method of financial data can be executed sequentially in sub- equipment.
Step S100: the financial data of enterprise is obtained.
Electronic equipment can be by extracting or the modes such as user's input obtain the original wealth of the enterprise from the database of enterprise
Business data.If inherently a kind of financial data for needing to carry out anomaly analysis of the raw financial data, then after electronic equipment
It is continuous the raw financial data directly to be inputted into corresponding first preset data processing model.Such as raw financial data is net profit
Profit, electronic equipment can be handled in model so that the net profit is directly input to corresponding first preset data.If the needs into
The financial data of row anomaly analysis, which needs to calculate by a variety of raw financial datas, to be obtained, then electronic equipment can be to original wealth
Business data are calculated, to determine corresponding financial data.Such as needing to carry out the financial data of anomaly analysis is net assets
Earning rate, electronic equipment can by by be initial data net profit compared with the net assets for initial data, to calculate
The net assets income ratio.
Further, financial data is either directly acquired, or obtains finance by calculating original financial data
Data, after obtaining financial data, electronic equipment can continue to execute step S200.
Step S200: financial data described in model treatment is handled by the first preset data, is obtained for indicating the wealth
Be engaged in data whether Yi Chang the first result.
In the present embodiment, financial data can be input to the corresponding first preset data processing model of financial data, with
Obtain the first abnormality score of the first preset data processing model output.
In the case where the quantity of the corresponding first preset data processing model of financial data is at least two, due to various
The scoring criterion that first preset data handles model is different, so that the first of various first preset datas processing model output is abnormal
The difference of score may be very big.It is a kind of for the first default of mixed Gauss model for example, calculated for same financial data
The scoring criterion of data processing model can be between 0-1, and the first abnormality score exported illustrates the finance number closer to 1
A possibility that according to exception, is bigger, on the contrary then smaller；And another the first preset data for i-forest model handles model
Scoring criterion then can be between 0-100, and the first abnormality score exported illustrates financial data exception closer to 100
Possibility is bigger, on the contrary then smaller.
To guarantee to handle each first abnormality score of model output based on various first preset datas, can accurately determine
Abnormal score out, wherein the exception score be used to indicate financial data whether Yi Chang the first result.Electronic equipment can will be each
First abnormality score of kind the first preset data processing model output is transformed under same standard on data.Optionally, electronic equipment
Minimum or maximum scoring criterion can be determined from the scoring criterion that various first preset datas handle model, and is successively determined
Multiple proportion between the every kind of scoring criterion and the minimum or maximum scoring criterion of other first preset data processing models out.
In this way, every kind of multiple proportion the first abnormality score corresponding with the multiple proportion is multiplied by electronic equipment, it is just abnormal by various first
Score is transformed under same standard on data.
For example, based on type be i-forest model the first preset data handle model scoring criterion 0-100 it
Between, and the first preset data that type is mixed Gauss model handles the scoring criterion of model between 0-1, electronic equipment can be with
The scoring criterion and type for determining the first preset data processing model that type is mixed Gauss model are i-forest model
The first preset data processing model scoring criterion between multiple proportion be 100 times.If type is mixed Gauss model
The first abnormality score that first preset data handles model output is 0.9, and type is the first present count of i-forest model
The first abnormality score according to processing model output is 88, which is 90 by electronic equipment, so that
Score after converting is in same standard on data with score into 90 into 88.
Certainly, if the scoring criterion of various first preset datas processing model is identical or the first preset data handles model
Quantity be one, then without to the first preset data processing model export the first abnormality score carry out score conversion.
Further, after obtaining the first abnormality score, electronic equipment determines abnormal score using first abnormality score.
As the first optional way for determining abnormal score using the first abnormality score:
If it is one that the first preset data, which handles model, electronic equipment can be directly defeated by the first preset data processing model
Whether first abnormality score out is abnormal for assessing financial data, i.e., first abnormality score can be used as abnormal score.
As second of optional way for determining abnormal score using the first abnormality score:
If it is at least two that the first preset data, which handles model, and the marking of at least two first preset datas processing model
When standard difference, electronic equipment needs to determine abnormal score using the first abnormality score being transformed under same standard on data.
And if the first preset data processing model is at least two, but the scoring criterion phase of at least two first preset datas processing model
Meanwhile the first abnormality score that electronic equipment can directly be exported using each first preset data processing model determines exception
Score.Wherein, it is stated to avoid tiring out, the present embodiment will directly be exported with electronic equipment using each first preset data processing model
The first abnormality score determine to be illustrated for abnormal score.
Illustratively, electronic equipment can select an abnormality score from multiple first abnormality scores.For example, electronics
Equipment can select the highest abnormality score of score value from multiple first abnormality scores.Wherein, the highest abnormality score of the score value
It is whether abnormal for assessing financial data as abnormal score.Alternatively, electronic equipment can also take the mode of averaging to more
A first abnormality score is averaging, and determines average mark.Wherein, which is then used as abnormal score, for assessing wealth
Whether data of being engaged in are abnormal.For example, the degree of reliability that electronic equipment can handle model according to each first preset data is pre-
The weight (its weight of the better model of reliability can be bigger) of each first preset data processing model is first set, is determined
(the corresponding weight of each first abnormality score is that export this first different to the product of each first abnormality score and corresponding weight out
The weight of the first preset data processing model of ordinary index), and all products are added up and are averaging, so that it is determined that average mark out
Number.Certainly, the mode of averaging is not limited to the average weighted mode that the present embodiment enumerates, can also be according to actual needs
The mode such as arithmetic average, square mean number, exponential average of selection.
In the present embodiment, for convenient for whether extremely to determine the financial data using abnormal score, electronic equipment can be preparatory
Outlier threshold score is set.Wherein, the mode that outlier threshold score is arranged can be using static direct set-up mode, Huo Zheye
Dynamic set-up mode can be used, such as be dynamically generated the outlier threshold score according to the abnormal score that history is determined.
Electronic equipment can be by the exception score compared with the outlier threshold score, by judging whether the exception score is greater than exception
Threshold score and determine whether the financial data abnormal.If it is determined that abnormal score is greater than outlier threshold score, electronic equipment is determined
The financial data is abnormal, and the warning information of financial data exception occurs to the manager of enterprise；It is on the contrary, it is determined that it is normal,
Without alarm.
It, can be with benefit in addition to determining whether financial data is abnormal using the first preset data processing model in the present embodiment
The model treatment degree of correlation is handled with the second preset data to determine whether financial data is abnormal.
Specifically, electronic equipment can use the financial data regression parameter and other wealth relevant to the financial data
Business data determine the regression data of the financial data, and seek difference to the financial data to the regression data and determine related
Degree.Model is handled to which the degree of correlation is input to corresponding second preset data by electronic equipment, it is default that second can be obtained
Second abnormality score of data processing model.
In the present embodiment, the case where the quantity of the corresponding second preset data processing model of the degree of correlation is at least two
Under, also due to the scoring criterion of at least two second preset datas processing model is different, so that at least two second preset datas
The difference for handling the second abnormality score of model output may be very big.Also it is handled to guarantee to be based at least two second preset datas
Second abnormality score of model output, can accurately determine abnormal score, and electronic equipment can also be pre- at least two second
If the second abnormality score of data processing model output is transformed under same standard on data.Wherein, before conversion regime can refer to
Understanding is stated, is not repeated herein.
It will also be appreciated that if the scoring criterion of at least two second preset datas processing model is identical or second is default
The quantity of data processing model is one, then without carrying out score conversion to the second abnormality score.
Further, electronic equipment can also determine abnormal score using the second abnormality score.
As the first optional way for determining abnormal score using the second abnormality score:
If the quantity that the second preset data handles model is one, electronic equipment can be directly by the processing of the second preset data
Whether the second abnormality score of model output is abnormal for assessing financial data, wherein second abnormality score can also be used as
Abnormal score.
As second of optional way for determining abnormal score using the second abnormality score:
If the quantity that the second preset data handles model is at least two, and at least two second preset datas handle model
Scoring criterion difference when, electronic equipment needs to determine exception using the second abnormality score for being transformed under same standard on data
Score.And the quantity of the second preset data processing model is at least two, but at least two second preset datas processing model
When scoring criterion is identical, the second abnormality score that each second preset data processing model directly exports is can also be used in electronic equipment
Determine abnormal score.It is stated to avoid tiring out, the present embodiment will be directly defeated using the second preset data processing model with electronic equipment
The second abnormality score out is determined to be illustrated for abnormal score.
Illustratively, electronic equipment can select two abnormality scores from multiple second abnormality scores.For example, electronics
Equipment can also select the highest abnormality score of score value from multiple second abnormality scores.Wherein, the highest exception of the score value point
Whether number is abnormal for assessing financial data as abnormal score.Alternatively, electronic equipment can also take the mode pair of averaging
Multiple second abnormality scores are averaging, and determine average mark.Wherein, which is then used as abnormal score, for assessing
Whether financial data is abnormal.Also for example, electronic equipment can be pre- according to the degree of reliability of each second preset data processing model
The weight (its weight of the better model of reliability can be bigger) of each second preset data processing model is first set, is determined every
(the corresponding weight of each second abnormality score is to export this second abnormal point to the product of a second abnormality score and corresponding weight
The weight of several the second preset data processing models), and all products are added up and are averaging, so that it is determined that average mark out.
Certainly, the mode of averaging is also not limited to the average weighted mode that the present embodiment enumerates, can also be according to actual needs
The mode such as arithmetic average, square mean number, exponential average of selection.
In the present embodiment, electronic equipment can also be by the exception score compared with the outlier threshold score of setting, to pass through
Judge whether the exception score is greater than outlier threshold score and determines whether the financial data is abnormal.Wherein, outlier threshold score
Set-up mode can refer to aforementioned understanding, be not repeated herein.If it is determined that abnormal score is greater than outlier threshold score, electronics is set
It is standby to determine that the financial data is abnormal, and the warning information of financial data exception occurs to the manager of enterprise；It is on the contrary, it is determined that its
Normally, without alarm.
In addition, the present embodiment can also be common using the first preset data processing model and the second preset data processing model
Determine abnormal score.
Specifically, when the first preset data handles the scoring criterion difference of model and the second preset data processing model,
First abnormality score and the second abnormality score can also be transformed under same score criteria by electronic equipment, after using conversion
Abnormal score is determined in first abnormality score and the conversion of the second abnormality score.Wherein, the mode of conversion can refer to aforementioned understanding,
It is not repeated herein.Certainly, identical as the second preset data processing scoring criterion of model in the first preset data processing model
When, electronic equipment can also be handled directly at the first abnormality score and the second preset data of model output using the first preset data
The second abnormality score for managing model output, determines abnormal score.
Illustratively, the present embodiment is also to handle model using the first preset data processing model and the second preset data
For scoring criterion is identical, determining abnormal score is illustrated.
Electronic equipment can determine the highest abnormality score of score value from the first abnormality score and the second abnormality score；Or
Person, electronic equipment can also be averaging the first abnormality score and the second abnormality score, determine average mark.Wherein it is determined that
The highest abnormality score of score value and the mode of averaging can be not repeated herein refering to aforementioned understanding out.Likewise, the score value
Highest abnormality score or average mark can be used to determine whether the financial data to be abnormal as abnormal score.Certainly, really
Whether Yi Chang concrete mode can also be also not repeated the fixed financial data herein refering to aforementioned understanding.
It is true using the first preset data processing model and/or the second preset data processing model to realize in the present embodiment
It is higher and higher to determine the whether abnormal accuracy of financial data, electronic equipment can constantly to the first preset data handle model and/or
Second preset data processing model is updated iteration.It below will be made of how to update the first preset data processing model, Yi Jiru
What updates the second preset data processing model is illustrated respectively.
It is updated exemplary approach as to the first preset data processing model, update week has been preset in electronic equipment
The duration of phase, period can be set according to actual needs.Electronic equipment is determining that current point in time is rising for current period
When time point beginning or termination time point, electronic equipment can handle model to the first preset data and be updated, i.e. electronic equipment
The history financial data collection for being also unused for updating the first preset data processing model can be extracted from the database of enterprise, with
The first preset data processing model is updated using the history financial data collection.
If it is Gauss model or mixed Gauss model that the first preset data, which handles model, electronic equipment passes through to history finance
Whole financial datas in data set are calculated, and the mean value and standard deviation of whole financial datas can be calculated.So, electronics
Equipment can be arranged in the model parameter or mixed Gauss model of the Gauss model often using the mean value and standard deviation currently determined
The model parameter of a Gauss model realizes the update to the first preset data processing model.Alternatively, electronic equipment will also be current
The mean value determined and history mean value are averaging and the average mean determined, and by the standard deviation currently determined and history
The average that standard deviation is averaging and determines is poor, to be arranged using the average mean currently determined and average standard deviation
The model parameter of each Gauss model in the model parameter or mixed Gauss model of the Gauss model also achieves default to first
The update of data processing model.It is understood that the mode being averaging can use such as arithmetic average, weighted average, put down
The modes such as square average, exponential average.
If the second preset data handle model be one-class SVM model or i-forest model, can also be first by history
Financial data is concentrated normally to be extracted with abnormal financial data respectively, using normal financial data as forward direction sample to the
One preset data processing model is trained, and handles mould to the first preset data using abnormal financial data as negative sense sample
Type is trained.Updating the weight in the first preset data processing model by training, (weight is the processing of the first preset data
The model parameter of model), to realize the update to the first preset data processing model.
And as exemplary approach is updated to the second preset data processing model, electronic equipment can also be in current time
When point is the start time point of current period or terminates time point, the is also unused for updating to extract from the database of enterprise
Two preset datas handle the history financial data collection of model, to be updated at the second preset data using the history financial data collection
Manage model.
Specifically, electronic equipment also can use history financial data concentration at least partly financial data determine it is new
Regression parameter, and the regression parameter that history is determined is updated using new regression parameter, to obtain updated time
Return parameter.Electronic equipment utilizes updated regression parameter, can determine that history financial data concentrates all financial datas
The degree of correlation.In this way, electronic equipment using all degrees of correlation for determining can the second preset data processing model be updated.
If it is Gauss model or mixed Gauss model that the second preset data, which handles model, electronic equipment passes through to all correlations
Degree is calculated, and the mean value and standard deviation of all degrees of correlation can be calculated.So, electronic equipment can be using currently determining
The model parameter of each Gauss model in the model parameter or mixed Gauss model of the Gauss model is arranged in mean value and standard deviation, real
The update to the second preset data processing model is showed.Alternatively, electronic equipment is also by the mean value currently determined and history mean value
The average mean for being averaging and determining, and the standard deviation currently determined and historical standard deviation are averaging and determined
Average is poor, thus using the average mean currently determined and average standard deviation be arranged the Gauss model model parameter or
The model parameter of each Gauss model in mixed Gauss model also achieves the update to the second preset data processing model.It can
Such as arithmetic average, weighted average, square mean number, exponential average can also be used in a manner of it is understood that being averaging
Etc. modes.
If it is one-class SVM model or i-forest model that the second preset data, which handles model, will can also first own
The normal and abnormal degree of correlation extracts respectively in the degree of correlation, using the normal degree of correlation as positive sample to the second present count
It is trained according to processing model, and the second preset data processing model is instructed using the abnormal degree of correlation as negative sense sample
Practice.Updating the weight in the second preset data processing model by training, (weight is the mould that the second preset data handles model
Shape parameter), to realize the update to the second preset data processing model.
Referring to Fig. 2, based on the same inventive concept, it, can be with benefit on the basis of determining abnormal score using model
It is matched with preset index with financial data and determines matching score, and more accurately determined by abnormal score and matching score
Whether financial data is abnormal.Corresponding, the analysis method of the financial data can also include: step S101 and step S201.
Step S101: the financial data is matched with preset data target, is obtained for indicating the financial data
Whether Yi Chang matching score.
Step S201: according to the abnormal score and the matching score, determine whether the financial data is abnormal.
Successively step S101 and step S201 will be described in detail below.
Electronic equipment can preset data target corresponding with financial data, and the data target is arranged in electronic equipment
Mode can be artificial setting or be automatically generated according to history exception score.Wherein, data target can be value range, finance
Data indicate that the financial data may be normal in the range of being located at the data target, otherwise, it means that the financial data may be different
Often.Data target based on setting, electronic equipment can execute step S101.
Step S101: the financial data is matched with preset data target, is obtained for indicating the financial data
Whether Yi Chang the second result.
Electronic equipment handles model treatment financial data using the first preset data processing model and/or the second preset data
While, electronic equipment also matches the financial data with preset data target, to judge whether financial data is located at data
In the range of index.
If it is determined that financial data is located in the range of data target, electronic equipment can be according to preset score create-rule
It generates for indicating that the financial data may matching score be, for example, normally 100 points.It should be noted that being generated based on score
The matching score that rule generates can be located under same score criteria with abnormal score.
If it is determined that financial data is not located in the range of data target, electronic equipment can be generated according to preset score and be advised
The matching score then generated for indicating that the financial data may be abnormal is, for example, 0 point.
After determining matching score, electronic equipment can continue to execute step S201.
Step S201: according to first result and described second as a result, determining whether the financial data is abnormal.
Electronic equipment directly can determine final score using abnormal score and matching score.Optionally, electronic equipment
The highest score of score value can be determined from abnormal score and matching score；Alternatively, electronic equipment can also be to abnormal score
It is averaging with matching score, determines average.Wherein it is determined that the highest score of score value and the mode of averaging can join out
Aforementioned understanding is read, is not repeated herein.Likewise, the highest score of the score value or average can be used as final score
In determining whether the financial data is abnormal.Certainly, using final score determine the financial data whether Yi Chang concrete mode
It can be also not repeated herein refering to aforementioned understanding.
Referring to Fig. 3, based on the same inventive concept, the embodiment of the present application provides a kind of electronic equipment 10, which is set
Standby 10 may include the communication interface 11 for being connected to the database of enterprise, the one or more processors for executing program instructions
12, bus 13 and various forms of memories 14, for example, disk, ROM or RAM, or any combination thereof.Illustratively, it calculates
Machine platform can also include be stored in ROM, RAM or other kinds of non-transitory storage medium, or any combination thereof in journey
Sequence instruction.
For memory 14 for storing program, the program that processor 12 is used to call and in run memory 14 is aforementioned to execute
Financial data analysis method.
Referring to Fig. 4, the embodiment of the present application provides a kind of analytical equipment 100 of financial data, the analysis of financial data
Device 100 is applied to electronic equipment, and the analytical equipment 100 of the financial data includes:
Data acquisition module 110, for obtaining the financial data of enterprise.
Data processing module 120 is used for for handling financial data described in model treatment by the first preset data
Indicate the financial data whether Yi Chang the first result.
Optionally, the data processing module 120 is also used to match the financial data with preset data target,
Obtain for indicate the financial data whether Yi Chang the second result；According to first result and described second as a result, really
Whether the fixed financial data is abnormal.
Optionally, the data processing module 120, for a kind of financial data to be separately input at least two institutes
The first preset data processing model is stated, the first abnormality score of each first preset data processing model output is obtained；Its
In, the type of at least two first preset data processing models is different；According to first abnormality score, determine
Abnormal score, wherein the exception score is for indicating first result.
It should be noted that due to it is apparent to those skilled in the art that, for the convenience and letter of description
Clean, system, the specific work process of device and unit of foregoing description can be with reference to corresponding in preceding method embodiment
Journey, details are not described herein.
The computer that some embodiments of the application additionally provide a kind of non-volatile program code that computer is executable can
Storage medium is read, is stored with program code on the computer readable storage medium, execution when which is run by computer
The step of analysis method of the financial data of any of the above-described embodiment.
In detail, which can be general storage medium, such as mobile disk, hard disk, on the storage medium
Program code when being run, the step of being able to carry out the analysis method of the above-mentioned financial data for applying example.
The program code product of the analysis method of financial data provided by the embodiment of the present application, including store program generation
The computer readable storage medium of code, the instruction that program code includes can be used for executing the method in previous methods embodiment, have
Body, which is realized, can be found in embodiment of the method, and details are not described herein.
In conclusion therefore it is in the accuracy of judgement since data target is easy to occur being arranged not reasonable
The upper limit is very low.Compared to this, since the first preset data processing model can have the higher upper limit in the accuracy of judgement, therefore
Handle model treatment financial data by the first preset data, obtain abnormal score can accurate response data it is whether abnormal.Cause
This, improve judge financial data whether Yi Chang accuracy.
More than, the only specific embodiment of the application, but the protection scope of the application is not limited thereto, and it is any to be familiar with
Those skilled in the art within the technical scope of the present application, can easily think of the change or the replacement, and should all cover
Within the protection scope of the application.Therefore, the protection scope of the application should be subject to the protection scope in claims.
Claims (10)
1. a kind of analysis method of financial data, which is characterized in that the described method includes:
Obtain the financial data of enterprise；
Financial data described in model treatment is handled by the first preset data, is obtained for indicating whether the financial data is abnormal
The first result.
2. the analysis method of financial data according to claim 1, which is characterized in that the method also includes:
The financial data is matched with preset data target, obtain for indicate the financial data whether Yi Chang second
As a result；
According to first result and described second as a result, determining whether the financial data is abnormal.
3. the analysis method of financial data according to claim 1 or 2, which is characterized in that at the first preset data
Manage model treatment described in financial data, obtain for indicate the financial data whether Yi Chang the first result, comprising:
A kind of financial data is separately input at least two first preset data processing models, is obtained each described
First preset data handles the first abnormality score of model output；Wherein, at least two first preset datas handle model
Type it is different；
According to first abnormality score, abnormal score is determined, wherein the exception score is for indicating first knot
Fruit.
4. the analysis method of financial data according to claim 3, which is characterized in that obtain enterprise financial data it
Afterwards, the method also includes:
Determine the degree of correlation between a kind of financial data and corresponding regression data, wherein the regression data is will
Other financial datas relevant to the financial data are returned to the financial data and are generated；
The degree of correlation is separately input at least one second preset data processing model, obtains each second present count
According to the second abnormality score of processing model output；Wherein, second preset data handles the type and described first of model in advance
If the type of data processing model is identical, and the parameter of one species model is different；
It is corresponding, according to first abnormality score, determine abnormal score, comprising:
According to first abnormality score and second abnormality score, the abnormal score is determined.
5. the analysis method of financial data according to claim 4, which is characterized in that according to first abnormality score and
Second abnormality score determines the abnormal score, comprising:
The highest abnormality score of score value is determined from first abnormality score and second abnormality score；Alternatively, to institute
It states the first abnormality score and second abnormality score is averaging, determine average mark, wherein the highest exception of score value
Score or the average mark are the abnormal score.
6. the analysis method of financial data according to claim 4, which is characterized in that establish each second present count
Include: according to the step of processing model
Obtain the history financial data collection of the enterprise；
Determine that the history financial data concentrates the degree of correlation of each financial data with corresponding regression data；
According to all degrees of correlation determined, the model parameter of each second preset data processing model is determined.
7. a kind of analytical equipment of financial data, which is characterized in that described device includes:
Data acquisition module, for obtaining the financial data of enterprise；
Data processing module is obtained for handling financial data described in model treatment by the first preset data for indicating
State financial data whether Yi Chang the first result.
8. the analytical equipment of financial data according to claim 7, which is characterized in that
The data processing module is also used to match the financial data with preset data target, obtains for indicating
State financial data whether Yi Chang the second result；According to first result and described second as a result, determining the financial data
It is whether abnormal.
9. the analytical equipment of financial data according to claim 7 or 8, which is characterized in that
The data processing module, for a kind of financial data to be separately input at least two first preset datas
Model is handled, the first abnormality score of each first preset data processing model output is obtained；Wherein, described at least two
The type that first preset data handles model is different；According to first abnormality score, abnormal score is determined, wherein
The exception score is for indicating first result.
10. a kind of computer-readable storage media, which is characterized in that program code is stored on the storage medium, when described
When program code is run by the computer, the analysis method of the financial data as described in any claim of claim 1-6 is executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910500343.7A CN110210959A (en) | 2019-06-10 | 2019-06-10 | Analysis method, device and the storage medium of financial data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910500343.7A CN110210959A (en) | 2019-06-10 | 2019-06-10 | Analysis method, device and the storage medium of financial data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110210959A true CN110210959A (en) | 2019-09-06 |
Family
ID=67791855
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910500343.7A Pending CN110210959A (en) | 2019-06-10 | 2019-06-10 | Analysis method, device and the storage medium of financial data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110210959A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111612040A (en) * | 2020-04-24 | 2020-09-01 | 平安直通咨询有限公司上海分公司 | Financial data anomaly detection method based on isolated forest algorithm and related device |
| CN111882415A (en) * | 2020-07-24 | 2020-11-03 | 未鲲(上海)科技服务有限公司 | Training method and related device of quality detection model |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101071477A (en) * | 2006-05-10 | 2007-11-14 | 何千军 | Financial analysis system and method based on expert system and nonlinear technology |
| CN101246584A (en) * | 2008-03-14 | 2008-08-20 | 山东浪潮齐鲁软件产业股份有限公司 | Tax estimation system model |
| CN102496073A (en) * | 2011-12-20 | 2012-06-13 | 南京审计学院 | Financial fraud audit assisting system and audit assisting method |
| CN106803204A (en) * | 2017-01-17 | 2017-06-06 | 上海聚赟大数据有限公司 | A kind of floater's promise breaking analysis method based on machine learning |
| CN107818416A (en) * | 2017-11-01 | 2018-03-20 | 众安信息技术服务有限公司 | The method and apparatus for generating Corporate Finance index forecast model |
| CN108090138A (en) * | 2017-11-29 | 2018-05-29 | 链家网（北京）科技有限公司 | The monitoring method and system of a kind of data warehouse |
| CN108460521A (en) * | 2018-02-02 | 2018-08-28 | 广州供电局有限公司 | The recommendation method and system of the audit target |
| CN108629675A (en) * | 2018-04-27 | 2018-10-09 | 信阳师范学院 | A kind of Multiple trees financial alert method |
| CN108876600A (en) * | 2018-08-20 | 2018-11-23 | 平安科技（深圳）有限公司 | Warning information method for pushing, device, computer equipment and medium |
| CN108876166A (en) * | 2018-06-27 | 2018-11-23 | 平安科技（深圳）有限公司 | Financial risk authentication processing method, device, computer equipment and storage medium |
| CN109800928A (en) * | 2019-03-25 | 2019-05-24 | 广州大学 | A kind of investment combination prediction technique and system based on Accounting Information Relevance |
-
2019
- 2019-06-10 CN CN201910500343.7A patent/CN110210959A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101071477A (en) * | 2006-05-10 | 2007-11-14 | 何千军 | Financial analysis system and method based on expert system and nonlinear technology |
| CN101246584A (en) * | 2008-03-14 | 2008-08-20 | 山东浪潮齐鲁软件产业股份有限公司 | Tax estimation system model |
| CN102496073A (en) * | 2011-12-20 | 2012-06-13 | 南京审计学院 | Financial fraud audit assisting system and audit assisting method |
| CN106803204A (en) * | 2017-01-17 | 2017-06-06 | 上海聚赟大数据有限公司 | A kind of floater's promise breaking analysis method based on machine learning |
| CN107818416A (en) * | 2017-11-01 | 2018-03-20 | 众安信息技术服务有限公司 | The method and apparatus for generating Corporate Finance index forecast model |
| CN108090138A (en) * | 2017-11-29 | 2018-05-29 | 链家网（北京）科技有限公司 | The monitoring method and system of a kind of data warehouse |
| CN108460521A (en) * | 2018-02-02 | 2018-08-28 | 广州供电局有限公司 | The recommendation method and system of the audit target |
| CN108629675A (en) * | 2018-04-27 | 2018-10-09 | 信阳师范学院 | A kind of Multiple trees financial alert method |
| CN108876166A (en) * | 2018-06-27 | 2018-11-23 | 平安科技（深圳）有限公司 | Financial risk authentication processing method, device, computer equipment and storage medium |
| CN108876600A (en) * | 2018-08-20 | 2018-11-23 | 平安科技（深圳）有限公司 | Warning information method for pushing, device, computer equipment and medium |
| CN109800928A (en) * | 2019-03-25 | 2019-05-24 | 广州大学 | A kind of investment combination prediction technique and system based on Accounting Information Relevance |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111612040A (en) * | 2020-04-24 | 2020-09-01 | 平安直通咨询有限公司上海分公司 | Financial data anomaly detection method based on isolated forest algorithm and related device |
| CN111612040B (en) * | 2020-04-24 | 2024-04-30 | 平安直通咨询有限公司上海分公司 | Financial data anomaly detection method and related device based on isolated forest algorithm |
| CN111882415A (en) * | 2020-07-24 | 2020-11-03 | 未鲲(上海)科技服务有限公司 | Training method and related device of quality detection model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108833458B (en) | Application recommendation method, device, medium and equipment | |
| CN109784779B (en) | Financial risk prediction method, device and storage medium | |
| CN109885452A (en) | Method for monitoring performance, device and terminal device | |
| CN108898476A (en) | A kind of loan customer credit-graded approach and device | |
| EP3961507A1 (en) | Optimal policy learning and recommendation for distribution task using deep reinforcement learning model | |
| CN111178537B (en) | Feature extraction model training method and device | |
| CN107832291A (en) | Client service method, electronic installation and the storage medium of man-machine collaboration | |
| CN110503564A (en) | Save case processing method, system, equipment and storage medium from damage based on big data | |
| CN112597745A (en) | Intelligent chart generation method and device, computer system and readable storage medium | |
| CN111651989B (en) | Named entity recognition method and device, storage medium and electronic device | |
| CN110210959A (en) | Analysis method, device and the storage medium of financial data | |
| CN110263136B (en) | Method and device for pushing object to user based on reinforcement learning model | |
| CN107665383A (en) | A kind of construction method and computing device of resume processing delay model | |
| CN117193980A (en) | Task remaining duration calculation method and device | |
| CN109785155A (en) | Method and Related product based on medical insurance reimbursement model adjustment medical insurance strategy | |
| CN116151991A (en) | Enterprise estimation method, enterprise estimation device, electronic equipment, storage medium and product | |
| CN115269085A (en) | Mobile bank page display method and device | |
| CN114298825A (en) | Method and device for extremely evaluating repayment volume | |
| CN114092057A (en) | Project model construction method and device, terminal equipment and storage medium | |
| WO2020037922A1 (en) | Stock index forecasting method, device, and storage medium | |
| CN117815674B (en) | Game information recommendation method and device, computer readable medium and electronic equipment | |
| CN113205217B (en) | Data processing method, device, equipment and storage medium | |
| CN116862561B (en) | Product heat analysis method and system based on convolutional neural network | |
| CN118278750B (en) | Power grid planning risk identification method, device, equipment and storage medium based on big data technology | |
| CN110109976A (en) | Data processing method, device, system and storage medium |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190906 |