TWM577148U - Electronic device for evaluating financial risk - Google Patents

Abstract

An electronic device for evaluating financial risk is provided. A data mining module stores hyperlink list and economic indicator list, access a hyperlink in the hyperlink list, and determines correlation of the hyperlink according to the economic indicator list. An information acquisition module obtains webpage information corresponding to the hyperlink. A regression analysis module performs a regression analysis to generate a first factor list based on the economic indicator and the webpage information. A machine learning module performs a machine learning algorithm to generate a second factor list based on the economic indicator and the webpage information. A filtering module generates a sufficient factor list based on the first factor list and the second factor list. A risk identification module generates a risk identification model for evaluating financial risk status according to the webpage information and the sufficient factor list.

Description

Electronic device for assessing financial risks

This new creation is about an electronic device for assessing financial risks.

Due to the frequent changes in the financial market, in the past, when financial market events occurred, the information collection and analysis of financial risks were conducted by means of manual Internet search. For example, countries with high inflation rates, high real interest rates, or negative real interest rates are among countries with higher financial risks. Therefore, banks will pay special attention to financial products related to the country. However, manual information collection and analysis requires a lot of human resources and time. In addition, people with different backgrounds may have an unobjective view of financial information, resulting in financial risk assessment results that are inconsistent due to the different people performing the assessment.

The novel creation provides an electronic device for evaluating financial risks, including: a storage unit and a processor. The storage unit stores a plurality of modules. The processor is coupled to the storage unit and accesses and executes the plurality of modules, wherein the modules include a data exploration module, an information acquisition module, a regression analysis module, a machine learning module, a screening module, and a risk identification Module. The data exploration module stores a hyperlink list and a list of economic indicators. The data exploration module accesses the hyperlink in the hyperlink list and determines the relevance of the hyperlink according to the economic indicator list. The information acquisition module obtains webpage information corresponding to the hyperlink. The regression analysis module performs a regression analysis based on the economic indicator list and the web page information to generate a first factor list. The machine learning module executes a machine learning algorithm based on a list of economic indicators and web page information to generate a second factor list. The screening module generates a list of significant factors based on the first factor list and the second factor list. The risk identification module generates a risk identification model based on the webpage information and the significant factor list, and evaluates the financial risk status according to the risk identification model.

Based on the above, the novel creation can automatically collect webpage information associated with economic indicators using techniques similar to web crawlers. On the other hand, this new creation can use regression analysis and machine learning algorithms to select the main economic indicators that affect financial risks, and then based on these economic indicators to establish a risk identification model with high accuracy to assess financial risk status.

The above described features and advantages of the present invention will become more apparent and understood from the following description.

1 is a schematic diagram of an electronic device 10 for assessing financial risk in accordance with an embodiment of the present invention. Referring to FIG. 1 , the electronic device 10 may include a processor 100 and a storage unit 300 . The processor 100 is coupled to the storage unit 300 and accesses and executes a plurality of modules stored in the storage unit 300. The modules include a data mining module 310, an information acquisition module 320, a regression analysis module 330, and a machine. The learning module 340, the screening module 350, and the risk identification module 360. The functions of each of the above modules will be described later in this document.

The processor 100 can be, for example, a central processing unit (CPU), or other programmable general purpose or special purpose microprocessor, digital signal processor (DSP), A novel controller, an application specific integrated circuit (ASIC) or the like or a combination of the above elements is not limited to this.

The storage unit 300 is configured to store various software, materials, and various types of codes required for the operation of the electronic device 10. The storage unit 300 can be, for example, any type of fixed or removable random access memory (RAM), read-only memory (ROM), flash memory (flash memory). ), hard disk drive (HDD), solid state drive (SSD) or the like or a combination of the above elements, the novel creation is not limited thereto.

The data mining module 310 is configured to store a hyperlink list and a list of economic indicators, wherein the hyperlink list stores a hyperlink of one or more credible financial market information websites. For example, the hyperlink list may include, for example, a website hyperlink, a website of a well-known financial information company, a well-known financial news website, and the like. The list of economic indicators stores one or more keywords that can be used to determine the economic indicators of the state of the financial market. Economic indicators can include unemployment rate, consumer confidence index, low-income population, internal conflict events, external conflict events, domestic Gross domestic product, real GDP growth rate, annual inflation rate, exchange rate, exchange rate stability, loan interest rate, real interest rate, reserve currency stock, internal debt, debt indicator, fiscal deficit, liability structure, investment amount, double exchange rate, income level, This new type of creation is not limited to keywords related to factors such as current account balance and trade dependency.

The data mining module 310 can access the hyperlinks in the hyperlink list and determine the relevance of the hyperlink based on the list of economic indicators. Specifically, the electronic device 10 can obtain the webpage information corresponding to the hyperlink in the hyperlink list through the information acquiring module 320. The data mining module 310 can determine the relevance of the hyperlink to the keywords in the economic indicator list according to the webpage information and the economic indicator list. If the relevance of a hyperlink to the keyword in the economic indicator list is higher than a predetermined threshold, the data mining module 310 adds the hyperlink in the webpage information to the hyperlink list. In this manner, the data mining module 310 can associate the webpage information corresponding to the hyperlink with the economic indicator, and store the sub-hyperlink in the webpage corresponding to the hyperlink into the hyperlink list. Repeatedly, the data mining module 310 will add all the hyperlinks associated with the economic indicators and the sub-hyperlinks under the hyperlink to the hyperlink list to generate an updated hyperlink list. . Then, the information obtaining module 320 can obtain the webpage information based on the updated hyperlink list, so as to achieve a deep search for the webpage.

The regression analysis module 330 can perform a regression analysis based on the economic indicator list and the web page information to generate a first factor list. Specifically, FIG. 2 illustrates a flow chart of the regression analysis module 330 performing regression analysis according to an embodiment of the present invention. In this embodiment, it is assumed that the economic indicator list includes (N-1) economic indicators, where N is an arbitrary positive integer.

In step S21, the regression analysis module 330 performs a variance analysis (ANOVA) and an F-check on the economic indicators of the i-th (i is one of (1 to (N-1)) in the economic indicator list according to the webpage information. Produce test results corresponding to economic indicators. In step S22, the regression analysis module 330 determines whether the saliency of the i-th economic indicator is higher than a threshold according to the test result. If yes, the process proceeds to step S23. Otherwise, the process proceeds to step S24. In step S23, the regression analysis module 330 adds the i-th economic indicator to a first factor list. The first factor list may record one or more economic indicators, and the economic indicators are more significant to the webpage information than the thresholds described in step S22. In other words, if web page information is related to financial risk, the economic indicators recorded in the first factor list represent economic indicators that significantly affect financial risk. In step S24, the regression analysis module 330 adds i to 1. In step S25, the regression analysis module 330 determines whether i is equal to N. If yes, step S21 is re-executed. Otherwise, the process of regression analysis ends.

Please return to Figure 1. The machine learning module 240 can execute a machine learning algorithm based on the economic indicator list and the web page information to generate a second factor list, wherein the machine learning algorithm can be, for example, a neural network algorithm or any kind of machine learning algorithm. Creation is not limited to this. Taking the machine learning algorithm as a neural network algorithm as an example, FIG. 3 illustrates a flow chart of the machine learning module 340 executing a machine learning algorithm according to an embodiment of the present invention.

In step S31, the machine learning module 240 generates tag data based on economic indicators in the historical web page information and economic indicator list. The historical web page information may be, for example, a web page selected from historical web pages that is highly correlated with financial risk. In step S32, the machine learning module 240 trains the neural network based on the tag data, wherein the manner of training the neural network can be performed by the user of the electronic device 10 according to the needs of the conventional neural network training method. Training, this new creation is not limited to this. After training the neural network, in step S33, the machine learning module 240 generates a second factor list based on the neural network and the web page information. Specifically, the machine learning module 240 can classify a web page information into a corresponding economic indicator using a neural network. In other words, the second factor list records the economic indicators corresponding to the web page information.

Please return to Figure 1. After the regression analysis module 330 establishes the first factor list and the machine learning module 340 establishes the second factor list, the screening module 350 can generate a significant factor list based on the first factor list and the second factor list. Specifically, the screening module 350 may select an economic indicator corresponding to the union of the first factor list and the second factor list, and generate a significant factor list by using the economic indicators.

One or more economic indicators recorded in the list of significant factors represent economic indicators that significantly affect financial risk. Based on this, the risk identification module 360 can generate a more accurate risk identification model according to the webpage information and the significant factor list, and evaluate the financial risk status according to the risk identification model. Since the input parameters of the risk identification model (ie, economic indicators that have significant influence on financial risk) have been screened, the risk identification model is less susceptible to noise in the process of assessing financial risk status (eg : Interference with economic indicators that have no influence on financial risks. On the other hand, screening economic indicators allows for input parameters for a small number of risk identification models. In other words, when the risk identification module 360 or other electronic device uses the risk identification model of the novel creation to evaluate the financial risk situation, less computational complexity will be used.

In summary, the novel creation can automatically collect webpage information associated with economic indicators using a web crawler-like technique, and perform a deeper search on hyperlinks in related webpage information. Through regression analysis, we select economic indicators that significantly affect the degree of risk from among many economic indicators. On the other hand, the novel creation can use machine learning algorithms to classify webpage information to link webpage information to corresponding economic indicators. Finally, the regression analysis is compared with the results of the machine learning algorithm to select the main economic indicators that significantly affect financial risk. In this way, financial industry personnel can establish a high-accuracy risk identification model to assess financial risk status based on these selected economic indicators.

Although the present invention has been disclosed in the above embodiments, it is not intended to limit the novel creation, and any person skilled in the art can make some changes without departing from the spirit and scope of the novel creation. Retouching, the scope of protection of this new creation is subject to the definition of the scope of the patent application attached.

10‧‧‧Electronic devices for assessing financial risks

100‧‧‧ processor

300‧‧‧ storage unit

310‧‧‧Data Exploration Module

320‧‧‧Information acquisition module

330‧‧‧Regression Analysis Module

340‧‧‧ machine learning module

350‧‧‧Screening module

360‧‧‧ Risk Identification Module

S21, S22, S23, S24, S25, S31, S32, S33‧‧

1 is a schematic diagram of an electronic device for assessing financial risk in accordance with an embodiment of the present invention. 2 is a flow chart showing the regression analysis module performing regression analysis according to an embodiment of the present invention. 3 is a flow chart showing a machine learning module executing a machine learning algorithm according to an embodiment of the present invention.

Claims (7)

An electronic device for evaluating a financial risk, comprising: a storage unit storing a plurality of modules; and a processor coupled to the storage unit and accessing and executing the plurality of modules, wherein the modules include: data exploration a module, storing a hyperlink list and a list of economic indicators, the data exploration module accessing the hyperlink in the hyperlink list and determining the relevance of the hyperlink according to the economic indicator list; the information acquisition module acquires the corresponding Hyperlink webpage information; regression analysis module, performing regression analysis based on the economic indicator list and the webpage information to generate a first factor list; a machine learning module, executing a machine learning algorithm based on the economic indicator list and the webpage information Generating a second factor list; a screening module, generating a list of significant factors based on the first factor list and the second factor list; and a risk identification module, generating a risk identification model according to the webpage information and the significant factor list, and according to The risk identification model assesses the state of financial risk. The electronic device for assessing financial risk as described in claim 1, wherein the data mining module further determines the relevance of the hyperlink according to the webpage information and the economic indicator list. An electronic device for assessing financial risk as described in claim 2, wherein the data mining module adds a second hyperlink corresponding to the webpage information to the hyperlink list based on the correlation being higher than a threshold . The electronic device for assessing financial risk according to claim 1, wherein the regression analysis module is configured to perform: performing a variance analysis and an F-check on the economic indicator in the economic indicator list according to the webpage information, thereby Generating a test result corresponding to the economic indicator; and adding the economic indicator to the first factor list based on the test result. An electronic device for assessing financial risk as recited in claim 1, wherein the machine learning algorithm comprises a neural network algorithm, and the machine learning module is configured to perform: based on historical web page information and the list of economic indicators The economic indicator generates label data; trains the neural network based on the label data; and generates the second factor list based on the neural network and the webpage information. The electronic device for assessing financial risk according to claim 1, wherein the screening module selects an economic indicator corresponding to the union of the first factor list and the second factor list to generate the significant factor list. For example, the electronic device for assessing financial risks mentioned in the scope of patent application, wherein the economic indicators in the list of economic indicators include unemployment rate, consumer confidence index, low-income population, internal conflict events, external conflict events, domestic production Gross value, real GDP growth rate, annual inflation rate, exchange rate, exchange rate stability, loan interest rate, real interest rate, reserve currency stock, internal debt, debt indicator, fiscal deficit, liability structure, investment amount, double exchange rate, income level, often One of account balance and trade dependency.