CN109300040A - Overseas investment methods of risk assessment and system based on full media big data technology - Google Patents
Overseas investment methods of risk assessment and system based on full media big data technology Download PDFInfo
- Publication number
- CN109300040A CN109300040A CN201810997591.2A CN201810997591A CN109300040A CN 109300040 A CN109300040 A CN 109300040A CN 201810997591 A CN201810997591 A CN 201810997591A CN 109300040 A CN109300040 A CN 109300040A
- Authority
- CN
- China
- Prior art keywords
- risk
- investment
- layer
- forecast model
- training sample
- 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
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000005516 engineering process Methods 0.000 title claims abstract description 28
- 238000012502 risk assessment Methods 0.000 title claims abstract description 28
- 238000012549 training Methods 0.000 claims abstract description 42
- 238000012544 monitoring process Methods 0.000 claims abstract description 11
- 230000006870 function Effects 0.000 claims description 24
- 210000002569 neuron Anatomy 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 11
- 230000004913 activation Effects 0.000 claims description 10
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 238000013139 quantization Methods 0.000 claims description 4
- 230000007787 long-term memory Effects 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 230000002265 prevention Effects 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000003475 lamination Methods 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000004451 qualitative analysis Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/06—Asset management; Financial planning or analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
Abstract
The present invention relates to a kind of overseas investment methods of risk assessment and system based on full media big data technology, the appraisal procedure include: the history investment data of overseas investment country to be obtained, as training sample by cloud computing and web crawlers method；Training sample is normalized, historical risk quantizating index is obtained；According to historical risk quantizating index, risk forecast model is established；Monitor the current risk investment data to investment country；The current risk investment data is normalized, current risk quantizating index is obtained；Based on the risk forecast model and the current risk quantizating index, determine to the investment risk situation to investment country.The present invention can Overall Acquisition overseas investment country history investment data, historical risk quantizating index is obtained, and then establishes risk profile mould, real-time monitoring waits for the current risk investment data of investment country, so that it is determined that improving the accuracy of investment to the investment risk situation to investment country.
Description
Technical field
The present invention relates to overseas investment risk assessment and early warning fields, are based on full media big data more particularly to one kind
The overseas investment methods of risk assessment and system of technology.
Background technique
Statistical data is shown, by year ends 2016, the domestic investor in 2.44 ten thousand, China's Mainland is in global 190 countries
Set up overseas enterprise 3.72 ten thousand, 5 trillion dollars of overseas enterprise's total assets, current year foreign direct investment flow 1961.5 hundred million
Dollar, increase by 34.7% on a year-on-year basis.But the overseas investment decision of Chinese Enterprises has very strong blindness, often only sees overseas
The market business opportunity and preferential policy of country, and the political economy to the state, Social Culture, natural ecology, laws and regulations, finance
The investment environments such as the tax lack comprehensively deep investigation and analysis, lack rational knowledge to the complicated risk factors wherein contained
With effective prevention, war turmoil, terrorist activity, anti-China forces etc. are adversely affected in addition, make China's overseas funded project in recent years
The case to suffer heavy losses happens occasionally.
The derivative approach of overseas investment risk is varied, various information numerous and complicateds, the quantization of overseas investment risk
Assessment difficulty increases by geometric progression." big data " is as a kind of emerging data processing technique and Cognitive Thinking, it is considered to be
The powerful of decision support.Therefore it there is an urgent need to utilize the technological means such as big data, cloud computing, intelligent recognition, integrates all kinds of
Data resource is made overseas investment risk intelligent measurement, identification, prediction, an early warning platform, fast, accurately and comprehensively is felt
Know government investment risk overseas, strengthen the risk prevention system ability of enterprise's overseas investment, reduces because of information asymmetry and decision not section
Economic loss caused by.
Summary of the invention
It is not comprehensive in order to solve risk investment information in order to solve the above problem in the prior art, it is quasi- to improve investment
True property, the present invention provides a kind of overseas investment methods of risk assessment and system based on full media big data technology.
A kind of overseas investment methods of risk assessment based on full media big data technology, the appraisal procedure include:
By cloud computing and web crawlers method, the history investment data of overseas investment country is obtained, as training sample
This；
The training sample is normalized, historical risk quantizating index is obtained；
According to the historical risk quantizating index, risk forecast model is established；
Monitor the current risk investment data to investment country；
The current risk investment data is normalized, current risk quantizating index is obtained；
Based on the risk forecast model and the current risk quantizating index, determine to the throwing to investment country
Provide risk situation.
Optionally, described that the training sample is normalized, historical risk quantizating index is obtained, it is specific to wrap
It includes:
Training sample is divided into 8 dimensions and 54 two-level index；
The data of each two-level index are normalized, the dimensionless number between [0,1] is obtained.
Optionally, it is normalized according to the following formula:
Wherein, xiFor i-th of two-level index numerical value in training sample,Occur for the index in training sample
Minimum value,For the maximum value that the index in training sample occurs, x 'iFor normalization after index value, i=1,2 ...
54。
Optionally, the risk forecast model includes input layer, hidden layer and output layer；Wherein,
The input layer is for receiving historical risk quantizating index or current risk quantizating index；
The hidden layer is used to construct based on deep neural network, and the hidden layer includes A layers, A >=1, and each layer is volume
Any one in lamination, activation primitive layer, pond layer, LSTM layers of shot and long term memory network and full articulamentum；
The output layer is for exporting risk sources and risk assessment grade.
Optionally, the risk sources are divided into 8, respectively society and politics, economy and finance, business environment, law political affairs
Plan, industry, natural environment, natural resources and ecology；
Risk assessment grade is divided into 5 grades, respectively devoid of risk, slight risks, moderate risk, severe risk, extreme wind
Danger.
Optionally, the appraisal procedure further include:
The risk forecast model is analyzed, determines whether the risk forecast model meets the requirements, is wanted if do not met
It asks, then corrects the risk forecast model.
Optionally, the analysis risk forecast model, determines whether the risk forecast model meets the requirements, has
Body includes:
Randomly select k-th of training sample X (k), corresponding desired output D (k) and reality output Y (k)；Wherein,
X (k)=(x1(k), x2(k), x3(k) ..., xn(k))；
D (k)=(d1(k), d2(k), d3(k) ..., dn(k))；
N=1,2 ..., i, xn(k) n-th of index value in k-th of training sample is indicated；dn(k) x is indicatedn(k) corresponding
Desired output；
It is based onCalculate global error E, wherein m indicates training sample
This quantity；
The size for comparing the global error E and preset error threshold, determines whether the risk forecast model meets
It is required that.
Optionally, the amendment risk forecast model, specifically includes:
Utilize the connection weight of hidden layer to output layer, the δ of output layero(k) with the output of hidden layer, error letter is calculated
The partial derivative δ of each neuron of several pairs of hidden layersh(k)；
Utilize the δ of each neuron of output layero(k) with the output of each neuron of hidden layer, connection weight ω is correctedho
(k)；
Utilize the δ of each neuron of output layero(k) come with the output of each neuron of hidden layer, correct output layer threshold value to
(k)；
(k+1) a training sample, corresponding desired output and reality output are extracted, the risk profile mould is reanalysed
Type then terminates to repair until global error is less than preset error threshold or study number is greater than preset study number maximum value
Just.
Optionally, the amendment risk forecast model, further includes:
The connection weight and biasing of each hidden layer are preset, and assigns the random number in a section (- 1,1) respectively, if
Determine error function e, error threshold and study number maximum value M.
In order to solve the above technical problems, the present invention also provides following schemes:
A kind of overseas investment risk evaluating system based on full media big data technology, the assessment system include:
Acquiring unit, for by cloud computing and web crawlers method, the history for obtaining overseas investment country to invest number
According to as training sample；
First processing units obtain historical risk quantizating index for the training sample to be normalized；
Modeling unit, for establishing risk forecast model according to the historical risk quantizating index；
Monitoring unit, for monitoring the current risk investment data to investment country；
The second processing unit obtains current risk amount for the current risk investment data to be normalized
Change index；The modeling unit is also used to determine based on the risk forecast model and the current risk quantizating index to institute
State the investment risk situation to investment country.
According to an embodiment of the invention, the invention discloses following technical effects
The present invention is by cloud computing and web crawlers method, the history investment data of Overall Acquisition overseas investment country, and
It is normalized, obtains historical risk quantizating index, and then establish risk profile mould, real-time monitoring waits for investment country
Current risk investment data is based on the risk forecast model, determines to the investment risk situation to investment country, mentions
Accuracy with high investment.
Detailed description of the invention
Fig. 1 is that the present invention is based on the flow charts of the overseas investment methods of risk assessment of full media big data technology；
Fig. 2 is the model structure schematic diagram of risk forecast model；
Fig. 3 is that the present invention is based on the specific implementation of the overseas investment methods of risk assessment of full media big data technology illustrations
It is intended to；
Fig. 4 is that the present invention is based on the signals of the modular structure of the overseas investment risk evaluating system of full media big data technology
Figure.
Symbol description:
Acquiring unit -1, first processing units -2, modeling unit -3, monitoring unit -4, the second processing unit -5.
Specific embodiment
The preferred embodiment of the present invention described with reference to the accompanying drawings.It will be apparent to a skilled person that this
A little embodiments are used only for explaining technical principle of the invention, it is not intended that limit the scope of the invention.
The object of the present invention is to provide a kind of overseas investment methods of risk assessment based on full media big data technology, lead to
Cloud computing and web crawlers method, the history investment data of Overall Acquisition overseas investment country are crossed, and is normalized,
Historical risk quantizating index is obtained, and then establishes risk profile mould, real-time monitoring waits for the current risk investment number of investment country
According to, it is based on the risk forecast model, it is determining to the investment risk situation to investment country, improve the accurate of investment
Property.
As shown in Figure 1, the present invention is based on the overseas investment methods of risk assessment of full media big data technology to include:
Step 100: by cloud computing and web crawlers method, the history investment data of overseas investment country is obtained, as
Training sample；
Step 200: the training sample being normalized, historical risk quantizating index is obtained；
Step 300: according to the historical risk quantizating index, establishing risk forecast model；
Step 400: current risk investment data of the monitoring to investment country；
Step 500: the current risk investment data being normalized, current risk quantizating index is obtained；
Step 600: being based on the risk forecast model and the current risk quantizating index, determine to described wait invest
The investment risk situation of country.
Wherein, in step 200, described that the training sample is normalized, it obtains historical risk quantization and refers to
Mark, specifically includes:
Step 201: training sample is divided into 8 dimensions and 54 two-level index (as shown in table 1)；
Step 202: the data of each two-level index being normalized, the dimensionless number between [0,1] is obtained.
Table 1
It can be normalized according to formula (1):
Wherein, xiFor i-th of two-level index numerical value in training sample,Occur for the index in training sample
Minimum value,For the maximum value that the index in training sample occurs, x 'iFor normalization after index value, i=1,2 ...
54。
Further, the risk forecast model includes input layer, hidden layer and output layer (as shown in Figure 2).
Wherein, the input layer is for receiving historical risk quantizating index or current risk quantizating index；The hidden layer
For being constructed based on deep neural network, the hidden layer includes A layers, A >=1, and each layer is convolutional layer, activation primitive layer, pond
Change any one in layer, LSTM layers of shot and long term memory network and full articulamentum；The output layer is for exporting risk sources and wind
Dangerous evaluation grade.
(1) convolutional layer:
Convolutional layer is will to input to carry out convolution with specified template convolution device, to obtain for the comprehensive of input matrix
Property characterization.Specific formula is as follows:
Wherein, g (i, j) is the value of the corresponding point exported after convolution, and f (i, j) is corresponding element in input data, h
(i, j) is the element of convolution mask filter corresponding position.
(2) activation primitive layer:
In convolutional layer, each layer of output is all the linear function of upper layer input.Therefore, in order to realize nonlinear transformation,
Basis of the nonlinear function as activation primitive layer is introduced, such deep-neural-network just has better fitting and expression
Ability.Common activation primitive is RELU function, sigmoid function or hyperbolic functions.
RELU function:
RELU function is one kind of activation primitive, and uses a kind of relatively conventional activation primitive at present, passes through RELU
Function can complete data conversion, and obtained data structure size is constant, it is only necessary to and a threshold value can be obtained by activation value,
The operation of a lot of complexity is calculated without spending, formula is as follows:
F (x)=max (0, x) (3).
Sigmoid function:
One real number input is mapped in [0,1] range by Sigmoid function, therefore it has when characterizing probability distribution formula
There is good performance, however its parameter convergence rate is very slow, also will affect trained efficiency, formula is as follows:
Hyperbolic functions:
One real number input is mapped in [- 1,1] range by hyperbolic functions.When input is 0, the output of tanh function is
0, meet the basic demand of our activation primitives.However, tanh function is also likely to be present gradient saturation problem.
(3) pond layer:
The main function of pond layer is compressed to data, and the dimension of data is on the one hand made to become smaller, and simplifies neural network
Computation complexity；On the one hand Feature Compression is carried out, main feature is extracted.
(4) LSTM layers:
LSTM is the upgrading variant for RNN (Recognition with Recurrent Neural Network), this method makes previous information and current times
Business is contacted, and can handle computer to output and input be different length sequence this kind problem.For t moment
Input xt, h can be based on by formula belowt-1And xtContinuously calculate variable below:
Wherein, σ is the other logic sigmoid function of Element-Level, and tanh is antitrigonometric function, and ⊙ is that element respective items multiply
Method.Wxf、Whf、Wxi、Whi、Wxg、WhgRespectively indicate the element that ranks are corresponded in the weight matrix of door, bfIndicate the weight square of door
The f of battle array arranges corresponding slope, biIndicate the corresponding slope of the i-th column of the weight matrix of door, bgIndicate the weight matrix of door
G arranges corresponding slope, ft, itAnd otIt is to forget door, the output of input gate and out gate respectively.ctIt is cell member state value,
htIt is hidden state value, gt is to calculate new ctCandidate value.
(5) full articulamentum:
It acts as all features are connected, output valve is given to classifier.The objective function selection of full articulamentum
Softmax function.Softmax classifier is a kind of effective ways for handling multi-class classification task.F will be exported with support
(xi, W) as (not calibrated and be likely difficult to explain) score of each classification SVM it is different, Softmax classifier can
More intuitive output is provided, i.e. normalization class probability, can be used for describing sample and belong to probability to some classification, it is right
Sample carries out modelling explanation.In Softmax classifier, Function Mapping f (xi；W)=WxiIt remains unchanged, but by these points
Number is construed to the non-normalized log probability to each classification, and defines loss letter using the cross entropy with following form
Number:
fjIndicate j-th of element vector of certain a kind of score f.As previously mentioned, the total loss of data set is LiAll
Mean value and regularization term R (W) in training examples.FunctionReferred to as softmax function: it will
It is 1 that the vector (in z) of one any real value score, which is compressed into a vector value between 0 and 1 and meets summation,.
The risk sources are divided into 8, respectively society and politics, economy and finance, business environment, policy of the law, industry,
Natural environment, natural resources and ecology；
Risk assessment grade is divided into 5 grades, respectively devoid of risk, slight risks, moderate risk, severe risk, extreme wind
Danger.
Wherein, variable y0、y1、y2、y3、y4、y5、y6、y7For showing risk sources, pair of exports coding and assessment result
It should be related to as shown in table 2.Work as yiWhen=0, indicate without corresponding associated risk factors；Work as yiWhen=1, indicate that there are corresponding phases
Close risk factors.
Table 2
Overseas investment country risk is divided into five grades, respectively devoid of risk, slight risks, moderate wind by the present invention
Danger, severe risk, extreme risk.Variable y8、y9、y10、y11For showing Risk-warning grade, exports coding and assessment result
Corresponding relationship it is as shown in table 3:
Table 3
Preferably, the present invention is based on the overseas investment methods of risk assessment of full media big data technology, further includes:
The risk forecast model is analyzed, determines whether the risk forecast model meets the requirements, is wanted if do not met
It asks, then corrects the risk forecast model.
Further, the analysis risk forecast model, determines whether the risk forecast model meets the requirements,
It specifically includes:
Randomly select k-th of training sample X (k), corresponding desired output D (k) and reality output Y (k)；Wherein,
X (k)=(x1(k), x2(k), x3(k) ..., xn(k))；
D (k)=(d1(k), d2(k), d3(k) ..., dn(k))；
N=1,2 ..., i, xn(k) n-th of index value in k-th of training sample is indicated；dn(k) x is indicatedn(k) corresponding
Desired output；
It is based onCalculate global error E, wherein m indicates training sample
This quantity；
The size for comparing the global error E and preset error threshold, determines whether the risk forecast model meets
It is required that.
Optionally, the amendment risk forecast model, specifically includes:
Utilize the connection weight of hidden layer to output layer, the δ of output layero(k) with the output of hidden layer, error letter is calculated
The partial derivative δ of each neuron of several pairs of hidden layersh(k)；
Utilize the δ of each neuron of output layero(k) with the output of each neuron of hidden layer, connection weight ω is correctedho
(k)；
Utilize the δ of each neuron of output layero(k) come with the output of each neuron of hidden layer, correct output layer threshold value to
(k)；
(k+1) a training sample, corresponding desired output and reality output are extracted, the risk profile mould is reanalysed
Type then terminates to repair until global error is less than preset error threshold or study number is greater than preset study number maximum value
Just.
Wherein, the amendment risk forecast model, further includes:
The connection weight and biasing of each hidden layer are preset, and assigns the random number in a section (- 1,1) respectively, if
Determine error function e, error threshold and study number maximum value M.
Compared with prior art, the method that the present invention uses qualitative analysis to combine with quantitative analysis, in original state's family tradition
On the basis of 4 first class index of dangerous assessment models and 18 two-level index, 4 first class index and 37 two-level index have been increased newly,
Keep Assessing country risk and prediction more complete accurate.In addition, this comprehensive study of the present invention has used foreign direct investment reason
By, country risk correlation theory and risk investment early warning correlation theory, the country based on deep neural network is constructed
Risk warning model, and this model is applied to China's overseas direct investment Assessing country risk and prediction (as shown in Figure 3).
As shown in figure 4, the present invention is based on the overseas investment risk evaluating system of full media big data technology, including obtain
Unit 1, first processing units 2, modeling unit 3, monitoring unit 4 and the second processing unit 5.
Wherein, the acquiring unit 1 is used to obtain going through for overseas investment country by cloud computing and web crawlers method
History investment data, as training sample.
The first processing units 2 obtain historical risk quantization and refer to for the training sample to be normalized
Mark.
The modeling unit 3 is used to establish risk forecast model according to the historical risk quantizating index.
The monitoring unit 4 is used to monitor the current risk investment data to investment country.
Described the second processing unit 5 obtains current wind for the current risk investment data to be normalized
Dangerous quantizating index；The modeling unit is also used to determine based on the risk forecast model and the current risk quantizating index
To the investment risk situation to investment country.
Compared with the existing technology, the present invention is based on the overseas investment risk evaluating system of full media big data technology with it is upper
The beneficial effect for stating the overseas investment methods of risk assessment based on full media big data technology is identical, and details are not described herein.
So far, it has been combined preferred embodiment shown in the drawings and describes technical solution of the present invention, still, ability
Field technique personnel are it is easily understood that protection scope of the present invention is expressly not limited to these specific embodiments.Without departing from
Under the premise of the principle of the present invention, those skilled in the art can make equivalent change or replacement to the relevant technologies feature, this
Technical solution after a little changes or replacement will fall within the scope of protection of the present invention.
Claims (10)
1. a kind of overseas investment methods of risk assessment based on full media big data technology, which is characterized in that the appraisal procedure
Include:
By cloud computing and web crawlers method, the history investment data of overseas investment country is obtained, as training sample；
The training sample is normalized, historical risk quantizating index is obtained；
According to the historical risk quantizating index, risk forecast model is established；
Monitor the current risk investment data to investment country；
The current risk investment data is normalized, current risk quantizating index is obtained；
Based on the risk forecast model and the current risk quantizating index, determine to the investment risk to investment country
Situation.
2. the overseas investment methods of risk assessment according to claim 1 based on full media big data technology, feature exist
In, it is described that the training sample is normalized, historical risk quantizating index is obtained, is specifically included:
Training sample is divided into 8 dimensions and 54 two-level index；
The data of each two-level index are normalized, the dimensionless number between [0,1] is obtained.
3. the overseas investment methods of risk assessment according to claim 2 based on full media big data technology, feature exist
In being normalized according to the following formula:
Wherein, xiFor i-th of two-level index numerical value in training sample,The minimum occurred for the index in training sample
Value,For the maximum value that the index in training sample occurs, x 'iFor normalization after index value, i=1,2 ... 54.
4. the overseas investment methods of risk assessment according to claim 1 based on full media big data technology, feature exist
In the risk forecast model includes input layer, hidden layer and output layer；Wherein,
The input layer is for receiving historical risk quantizating index or current risk quantizating index；
The hidden layer be used for based on deep neural network construct, the hidden layer includes A layer, A >=1, each layer be convolutional layer,
Any one in activation primitive layer, pond layer, LSTM layers of shot and long term memory network and full articulamentum；
The output layer is for exporting risk sources and risk assessment grade.
5. the overseas investment methods of risk assessment according to claim 4 based on full media big data technology, feature exist
In the risk sources are divided into 8, respectively society and politics, economy and finance, business environment, policy of the law, industry, natural ring
Border, natural resources and ecology；
Risk assessment grade is divided into 5 grades, respectively devoid of risk, slight risks, moderate risk, severe risk, extreme risk.
6. the overseas investment methods of risk assessment according to claim 4 based on full media big data technology, feature exist
In the appraisal procedure further include:
The risk forecast model is analyzed, determines whether the risk forecast model meets the requirements, if it does not meet the requirements, is then repaired
The just described risk forecast model.
7. the overseas investment methods of risk assessment according to claim 6 based on full media big data technology, feature exist
In the analysis risk forecast model determines whether the risk forecast model meets the requirements, specifically includes:
Randomly select k-th of training sample X (k), corresponding desired output D (k) and reality output Y (k)；Wherein,
X (k)=(x1(k), x2(k), x3(k) ..., xn(k))；
D (k)=(d1(k), d2(k), d3(k) ..., dn(k))；
N=1,2 ..., i, xn(k) n-th of index value in k-th of training sample is indicated；dn(k) x is indicatedn(k) corresponding expectation is defeated
Out；
It is based onCalculate global error E, wherein m indicates number of training
Amount；
The size for comparing the global error E and preset error threshold, determines whether the risk forecast model meets the requirements.
8. the overseas investment methods of risk assessment according to claim 7 based on full media big data technology, feature exist
In the amendment risk forecast model specifically includes:
Utilize the connection weight of hidden layer to output layer, the δ of output layero(k) with the output of hidden layer, error function is calculated to hidden
Hide the partial derivative δ of each neuron of layerh(k)；
Utilize the δ of each neuron of output layero(k) with the output of each neuron of hidden layer, connection weight ω is correctedho(k)；
Utilize the δ of each neuron of output layero(k) come with the output of each neuron of hidden layer, correct output layer threshold value to(k)；
(k+1) a training sample, corresponding desired output and reality output are extracted, reanalyses the risk forecast model, directly
It is less than preset error threshold to global error or study number is greater than preset study number maximum value, then terminates to correct.
9. the overseas investment methods of risk assessment according to claim 8 based on full media big data technology, feature exist
In the amendment risk forecast model, further includes:
The connection weight and biasing of each hidden layer are preset, and assigns the random number in a section (- 1,1) respectively, setting misses
Difference function e, error threshold and study number maximum value M.
10. a kind of overseas investment risk evaluating system based on full media big data technology, which is characterized in that the assessment system
Include:
Acquiring unit, for obtaining the history investment data of overseas investment country by cloud computing and web crawlers method, as
Training sample；
First processing units obtain historical risk quantizating index for the training sample to be normalized；
Modeling unit, for establishing risk forecast model according to the historical risk quantizating index；
Monitoring unit, for monitoring the current risk investment data to investment country；
The second processing unit obtains current risk quantization and refers to for the current risk investment data to be normalized
Mark；The modeling unit is also used to based on the risk forecast model and the current risk quantizating index, determine to it is described to
The investment risk situation of investment country.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810997591.2A CN109300040A (en) | 2018-08-29 | 2018-08-29 | Overseas investment methods of risk assessment and system based on full media big data technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810997591.2A CN109300040A (en) | 2018-08-29 | 2018-08-29 | Overseas investment methods of risk assessment and system based on full media big data technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109300040A true CN109300040A (en) | 2019-02-01 |
Family
ID=65165906
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810997591.2A Pending CN109300040A (en) | 2018-08-29 | 2018-08-29 | Overseas investment methods of risk assessment and system based on full media big data technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109300040A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110009504A (en) * | 2019-04-08 | 2019-07-12 | 湖北风口网络科技有限公司 | A kind of highly-safe inauguration investment system and method |
| CN110414709A (en) * | 2019-06-18 | 2019-11-05 | 重庆金融资产交易所有限责任公司 | Debt risk intelligent Forecasting, device and computer readable storage medium |
| CN111160662A (en) * | 2019-12-31 | 2020-05-15 | 咪咕文化科技有限公司 | Risk prediction method, electronic equipment and storage medium |
| CN111986808A (en) * | 2020-07-30 | 2020-11-24 | 珠海中科先进技术研究院有限公司 | Method, device and medium for health insurance risk assessment and control |
| CN112085384A (en) * | 2020-09-08 | 2020-12-15 | 华侨大学 | Mailing risk evaluation method and system based on combination of fuzzy reasoning and LSTM |
| CN113177700A (en) * | 2021-04-15 | 2021-07-27 | 国任财产保险股份有限公司 | Risk assessment method, system, electronic equipment and storage medium |
| CN115182395A (en) * | 2022-06-29 | 2022-10-14 | 山东大学 | Submarine stratum stability evaluation and prediction method and system |
-
2018
- 2018-08-29 CN CN201810997591.2A patent/CN109300040A/en active Pending
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110009504A (en) * | 2019-04-08 | 2019-07-12 | 湖北风口网络科技有限公司 | A kind of highly-safe inauguration investment system and method |
| CN110414709A (en) * | 2019-06-18 | 2019-11-05 | 重庆金融资产交易所有限责任公司 | Debt risk intelligent Forecasting, device and computer readable storage medium |
| CN111160662A (en) * | 2019-12-31 | 2020-05-15 | 咪咕文化科技有限公司 | Risk prediction method, electronic equipment and storage medium |
| CN111986808A (en) * | 2020-07-30 | 2020-11-24 | 珠海中科先进技术研究院有限公司 | Method, device and medium for health insurance risk assessment and control |
| CN111986808B (en) * | 2020-07-30 | 2023-12-12 | 珠海中科先进技术研究院有限公司 | Health insurance risk assessment and control method, device and medium |
| CN112085384A (en) * | 2020-09-08 | 2020-12-15 | 华侨大学 | Mailing risk evaluation method and system based on combination of fuzzy reasoning and LSTM |
| CN113177700A (en) * | 2021-04-15 | 2021-07-27 | 国任财产保险股份有限公司 | Risk assessment method, system, electronic equipment and storage medium |
| CN113177700B (en) * | 2021-04-15 | 2022-09-13 | 国任财产保险股份有限公司 | Risk assessment method, system, electronic equipment and storage medium |
| CN115182395A (en) * | 2022-06-29 | 2022-10-14 | 山东大学 | Submarine stratum stability evaluation and prediction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109300040A (en) | Overseas investment methods of risk assessment and system based on full media big data technology | |
| CN110044554A (en) | A kind of online test method of the metal pressure container leakage based on acoustic emission signal | |
| CN107909564A (en) | A kind of full convolutional network image crack detection method based on deep learning | |
| CN107067341B (en) | RBFNN power distribution automation system state operation evaluation method based on multi-level entropy weight | |
| CN110689162B (en) | Bus load prediction method, device and system based on user side classification | |
| CN110111113A (en) | A kind of detection method and device of exception transaction node | |
| AU2019100968A4 (en) | A Credit Reporting Evaluation System Based on Mixed Machine Learning | |
| CN109299434B (en) | Cargo customs clearance big data is intelligently graded and sampling observation rate computing system | |
| Matviychuk et al. | Neuro-fuzzy model of country's investment potential assessment | |
| CN111476274B (en) | Big data predictive analysis method, system, device and storage medium | |
| CN112632765B (en) | Combat capability assessment method combining weighting method and SEM method | |
| Liu et al. | Stock price trend prediction model based on deep residual network and stock price graph | |
| Azadeh et al. | The impact of decision-making units features on efficiency by integration of data envelopment analysis, artificial neural network, fuzzy C-means and analysis of variance | |
| CN114764682B (en) | Rice safety risk assessment method based on multi-machine learning algorithm fusion | |
| Gan | Discrete Hopfield neural network approach for crane safety evaluation | |
| Yang | A Study on the Impact of Corporate Financial Accounting Management System on Corporate Innovation Under Sustainable Development Strategy | |
| Luo | Application of BP neural network in economic management of coastal area | |
| CN114580934A (en) | Early warning method for food detection data risk based on unsupervised anomaly detection | |
| Wang et al. | Joint loan risk prediction based on deep learning‐optimized stacking model | |
| Liu et al. | RETRACTED ARTICLE: Company financial path analysis using fuzzy c-means and its application in financial failure prediction | |
| CN106127606A (en) | The treating method and apparatus of oil and gas index data in a kind of exploration and development project | |
| Jeihoonian et al. | Modeling and comparing energy consumption in basic metal industries by neural networks and ARIMA | |
| Hassanzadeh et al. | Developing a new method using Artificial Immune System in order to High Productivity of Inefficient Units in Network DEA approach | |
| CN111062118A (en) | Multilayer soft measurement modeling system and method based on neural network prediction layering | |
| Wang | Financial distress prediction for listed enterprises using fuzzy C-means |
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 |
Application publication date: 20190201 |
|
| RJ01 | Rejection of invention patent application after publication |