TW202230262A - Recommended financial product and risk control system and implementation method thereof - Google Patents

Abstract

Recommended financial product and risk control system and implementation method thereof are disclosed. The present invention is configured to capture a voiceprint information under test by a personality analysis device, and analyze and compare the voiceprint information under test to obtain a voiceprint classification feature of the voiceprint information under test by a voiceprint analysis model and then search a personality analysis information according to the voiceprint classification feature. Further, the personality analysis information is analyzed to obtain a user's financial credit information to determine a user's potential credit risk as an evaluating reference for loan, and is compared to obtain a financial product information. In this way, the user can obtain a selection reference for financial products through the personality analysis information.

Description

Financial product recommendation and risk control system and method

The present invention relates to a voiceprint recognition technology, in particular to a method of capturing a user's voiceprint information through a character analysis device to analyze the user's character, determine the user's credit risk and recommend suitable financial products according to the user's character. Financial product recommendation and risk control system and financial product recommendation method.

According to press, there are currently many financial product projects, and consumers can invest in different financial product projects. In order to increase the transaction volume of financial products, financial product providers usually contact consumers through direct contact or telephone interviews through salespeople. The interview content confirms consumers' needs for specific financial products, and then recommends suitable financial products to consumers. Alternatively, some financial product providers will collect consumers' needs through indirect contact (questionnaires or chatbots) with consumers. However, , whether it is direct contact or indirect contact, consumers will be vigilant or reserved, which makes the interview results inaccurate. Therefore, financial product providers cannot provide financial products suitable for consumers, resulting in financial products that are not consumed. However, some of the reasons why financial product providers cannot provide consumers with suitable financial products may be due to the influence of consumers’ personalities. As mentioned above, financial product providers in the interview During the process, the consumer's personality will be actively analyzed. However, consumers may still deceive or conceal during the interview process. Based on this, how to avoid consumers from concealing their true personality, accurately analyze the consumer's personality, and provide the most suitable consumer advice. Financial commodities, this is a problem to be solved.

In view of the above-mentioned problems, the inventors of the present invention have conducted research and improvements on voiceprint processing methods, financial commodity matching methods and character analysis methods based on years of experience in related industries; for this reason, the main purpose of the present invention is to provide a By collecting the user's voiceprint information, the financial product recommendation and risk control system and method suitable for the user's personality are analyzed.

In order to achieve the above-mentioned purpose, the financial product recommendation and risk control system and financial product recommendation method of the present invention mainly use a character analysis device to capture a voiceprint information to be measured, and then analyze and compare it based on a voiceprint analysis model. The voiceprint information to be tested is analyzed to obtain the voiceprint classification features of the voiceprint information to be tested, and a corresponding character analysis information is queried according to the voiceprint classification features, and then a corresponding financial information is matched with the character analysis information. Commodity information, the present invention mainly establishes a voiceprint analysis model through supervised machine learning. The voiceprint analysis model uses at least voiceprint sample information and at least one personality test information as training parameters. In any emotional state of the user, the personality analysis device will The user's character can be accurately analyzed, and the user's credit risk can be judged with the character analysis information, and the most suitable financial product information can be matched. Accordingly, the present invention can achieve the following effects: (1) The voiceprint analysis model uses voiceprint sample information and personality test information as training parameters, which can predict the potential intelligence of the test object, such as consumer spending preferences, consumer investment preferences, or consumer credit risk, and risk tolerance status and other information; (2) Predicting the user's personality through supervised machine learning can eliminate emotional factors; (3) Personality test information can be electronic test questionnaires or physical test questionnaires. The content of the questionnaires involves psychological tests. Therefore, the voiceprint classification features analyzed by the voiceprint analysis model and the corresponding personality analysis information, the analysis results are very suitable for the user's personality; (4) The present invention estimates the true character of the consumer by capturing the voiceprint information, which can be provided to the financial service end, so that the financial service end can judge the credit risk of the consumer, as a reference for whether or not to borrow; (5) The financial service terminal can also provide the most suitable financial products according to the true personality of consumers, which can increase the sales volume of financial products.

In order to enable your examiners to clearly understand the purpose, technical features and effects of the present invention, the following descriptions are combined with the diagrams for illustration, please refer to.

Please refer to "Figure 1", the figure shows the composition diagram (1) of the present invention. As shown in the figure, the financial product recommendation and risk control system 1 of the present invention is mainly composed of a character analysis device 10, at least one The client information device 20 and at least one financial server device 30 are composed. Under normal operation, at least one client information device 20 completes information connection with the personality analysis device 10; the personality analysis device 10 can collect at least voiceprint information and a personality test information corresponding to each voiceprint information, and use the voiceprint information and the personality test information as training parameters to establish a voiceprint analysis model through a supervised machine learning algorithm. The personality analysis device 10 can be a computer device, In addition, after the training of the voiceprint analysis model is completed, the personality analysis device 10 can classify the newly input voiceprint information through the voiceprint analysis model, so that different voiceprint information can be classified into appropriate voiceprint classification features. And according to the voiceprint classification features, a character analysis information is compared; the client information device 20 can be operated by the user, and the client information device 20 can encode the user's voice signal into a voiceprint that can be processed and analyzed by the machine Information (digital or analog), the client information device 20 can be a computer equipment, a mobile communication device, a wearable electronic device or an embedded system, any one that can collect sound signals and execute application programs can be implemented. It is limited, and the user of the client information device 20 can be a consumer with financial investment needs or borrowing needs; the financial service end device 30 is installed on the financial service end, so that the financial service end can be used for credit reporting. During the operation, you can obtain the character analysis information compared with the user terminal, and use the character analysis information to compare the financial credit evaluation information or financial product information. The financial service terminal can use the financial credit evaluation information as an evaluation reference for credit reporting operations. , and can also recommend the most suitable financial product information for the user-end character.

Please refer to "Fig. 2" again, which shows a schematic diagram (2) of the composition of the present invention, and please refer to "Fig. 1" in conjunction. As shown in the figure, the character analysis device 10 of the present invention includes an arithmetic processing module In the group 101, there is another voiceprint processing module 102, a data analysis module 103, a data storage module 104 and an information transmission module 105 to complete the information connection with the computing processing module 101; (1) The operation processing module 101 is used to run the character analysis device 10 and drive the above modules, and has functions such as logical operation, temporary storage of operation results, and storage of execution instruction positions, and can be a central processing unit. (Central Processing Unit, CPU) or a microcontroller (Microcontroller Unit, MCU); (2) The voiceprint processing module 102 can perform a voiceprint processing operation after receiving the voiceprint information. The voiceprint processing operation at least includes a noise reduction process and a captured voiceprint sample, and the voiceprint information The voiceprint sample information is generated after being processed by the voiceprint processing operation for the data analysis module 103 to capture the voiceprint sample information for subsequent processing and analysis, wherein the voiceprint information can be digital or analog uncompressed, lossy Information such as compression, lossless compression, or a combination thereof; (3) The data analysis module 103 can execute a supervised machine learning algorithm, such as: artificial neural network (Artificial Neural Network, ANN), support vector machine (support vector machine, SVM), nearest neighbor method (k nearest neighbor, KNN), Gaussian Mixture Model, Naive Bayes Classifier, or Decision tree, but not limited thereto, and the data analysis module 103 can Capture voiceprint sample information and personality test information, and users can label the personality test information in advance, and analyze the voiceprint sample information and personality through a supervised machine learning algorithm based on the labels set by the user. Test the correlation of information to establish a voiceprint analysis model, in which several voiceprint classification features can be generated according to the annotation results. The voiceprint analysis model can be based on training parameters (voiceprint sample information and personality test information) and annotation results. The annotation results analyze several voiceprint classification features, and the voiceprint analysis model can be based on the physical characteristics of the voiceprint information (training data) as a basis for judgment, such as: voiceprint frequency, volume, waveform, wave number, sound speed or speech rate, etc., the above examples are only examples, not limited to this, and each voiceprint classification feature can correspond to specific character analysis information, wherein, the voiceprint classification feature forms at least two or more categories, preferably There are seven types of voiceprint classification features, and each classification feature represents a different character. Furthermore, the data analysis module 103 can further combine the corresponding financial product information or a financial credit evaluation information according to the character analysis information; (4) The data storage module 104 includes a voiceprint database 1041 that stores at least voiceprint sample information; a personality test database 1042 that stores at least one personality test information; a model database 1043 that stores voiceprints Analysis model; a character database 1044 stores at least one character analysis information, each character analysis information includes at least one consumption habit information, at least one investment habit information and at least one credit risk information, the consumption habit information is for influencing consumption behavior Personality characteristics, such as being prone to impulse purchase of products, hesitant about products, and inquiring about product evaluations before consumption. The investment habit information described is the personality characteristics that affect investment behavior. Personality impulses are easy to follow other people's blind investment, poor money concept, and blind investment; the credit risk information mentioned is the personality characteristics that affect the financial credit evaluation level, for example, a highly responsible personality, and the loan will be repaid on time after the loan is approved, so the financial credit On the contrary, a character with a poor sense of responsibility has a poor financial credit rating; a financial product database 1045 stores at least one piece of financial product information, and the character analysis information corresponds to relevant financial products based on parameters such as consumption habit information and investment habit information. Information, wherein each financial product information is set with a matching label, and each matching label corresponds to the associated character analysis information, so that the character analysis device 10 can search for the matching label through the matching label. Financial commodity information, and the financial commodity information includes investment insurance policies, savings insurance, futures commodities, foreign exchange commodities, option commodities, fund commodities, stock commodities, and bond commodities, and each financial commodity information has corresponding consumption habits and investment habits; a financial credit database 1046 stores at least one financial credit evaluation information, the financial credit evaluation information has a plurality of financial credit evaluation levels, and different financial credit evaluations can be compared according to the credit risk information of the character analysis information grade for reference by the financial service end in reviewing the loan; (5) The information transmission module 105 can be connected to each client information device 20 through Internet information, so as to receive and transmit information to the client information device 20 .

Please refer to "Fig. 3", which shows the flow chart of the implementation of the present invention, and please refer to "Fig. 1" ~ "Fig. 2" in conjunction with the financial product recommendation and risk control system 1 of the present invention, mainly through The character analysis device 10 establishes a voiceprint analysis model to analyze a character analysis information of each voiceprint information, and compares at least one financial product information matched with the character analysis information, wherein the financial product recommendation method is as follows: (1) 1. Retrieve the voiceprint information to be tested and the personality test information S1: please refer to "Fig. 4" in combination, which shows the schematic diagram (1) of the implementation of the present invention. The sum and plural character test information D2 are input into a character analysis device 10; (2) A voiceprint analysis model S2 is established: After the personality analysis device 10 captures the voiceprint information D1 and the personality test information D2, the voiceprint processing module 102 can perform a voiceprint processing operation on the voiceprint information D1. In the voiceprint processing operation, the voiceprint information D1 can be processed to reduce noise and capture voiceprint samples. When the voiceprint information D1 completes the voiceprint processing operation, the voiceprint sample information D1' can be generated. The character analysis device 10 can The voiceprint sample information D1' and the personality test information D2 are stored in the data storage module 104, and the data analysis module 103 can retrieve the voiceprint sample information D1' and the personality test information D2 in the data storage module 104, wherein, The voiceprint sample information D1' is used as input data, the personality test information D2 is used as target data, and the personality test information D2 is marked, so that the data analysis module 103 executes a supervised machine learning algorithm, by calculating the voiceprint sample information D1' and The weight of the corresponding personality test information D2 is used to establish a voiceprint analysis model, and the voiceprint analysis model is stored in the data storage module 104 for the data analysis module 103 to compare and classify; (3) 1. Capture the voiceprint information to be measured S3: Please refer to "Figure 5". The figure shows the schematic diagram (2) of the implementation of the present invention. As shown in the figure, after the voiceprint analysis model is established, the client After the information device 20 is connected to the personality analysis device 10, the client information device 20 can further transmit a voiceprint information D3 to be tested to the personality analysis device 10; (4) A voiceprint processing operation S4 is performed on the voiceprint information to be measured: after the personality analysis device 10 receives the voiceprint information D3 to be measured, the voiceprint processing module 102 can perform a voiceprint processing operation on the voiceprint information D3 to be measured, so as to Generate a voiceprint sample information D3' to be tested; (5) The voiceprint analysis model compares the information of the voiceprint sample to be tested S5: After the data analysis module 103 obtains the information of the voiceprint sample to be tested D3', it can compare the information of the voiceprint sample to be tested D3' through the voiceprint analysis model The voiceprint classification feature, and the data analysis module 103 can search for a corresponding character analysis information in the data storage module 104 based on the voiceprint classification feature; (6) 1. Comparing financial credit information with the character analysis information S6: The data analysis module 103 then uses the character analysis information as a matching condition to search the financial credit database 1046, and uses matching to obtain a financial information corresponding to the character analysis information For credit evaluation information D4, the computing processing module 101 then transmits the matched financial credit evaluation information D4 to the financial server device 30, so that the financial server can use the financial credit evaluation information D4 as a reference for evaluating whether to borrow or not; (7) 1. Comparing financial product information with the character analysis information S7: The data analysis module 103 then uses the character analysis information as a matching condition to search the financial product database 1045, and uses matching to obtain a financial product corresponding to the character analysis information Commodity information D5, the arithmetic processing module 101 then transmits the matching financial commodity information D5 to the client information device 20.

Continuing from the above, the financial product recommendation and risk control system 1 of the present invention can provide the financial product demander or financial product provider to input the voiceprint information D3 to be measured, and then the personality analysis device 10 can analyze the voiceprint information D3 to be measured. The character analysis information D4 is generated, and then the character analysis information D4 is used to determine the user's financial credit evaluation information and financial commodity information. The financial service terminal can use the financial credit evaluation information as a reference for the user's credit evaluation, effectively improving the accuracy of the credit evaluation. In addition, the financial service terminal can analyze the information according to the personality, and match the most appropriate financial product combination. If the user is a financial product demander, the financial product demander can choose the most suitable financial product from their own personality characteristics. If the user is a financial product provider, the user can find a combination of financial products suitable for the user according to the user's voiceprint data, so as to achieve the matching between the demander and the product provider.

Please refer to "Fig. 6", which is another embodiment (1) of the present invention. Please refer to "Fig. 3" ~ "Fig. 5" in conjunction with the step "Comparing financial product information with personality analysis information S6", After the data analysis module 103 matches the financial product information D5, the arithmetic processing module 101 can simultaneously transmit the character analysis information and the financial product information D5 to the client information device 20, and the user can check the character through the client information device 20 Analyze whether the information conforms to one's own personality. If it does not conform to one's own personality, the user can create a correction message D6 through the client information device 20, and transmit the correction message D6 to the personality analysis device 10, so that the data analysis module 103 can be based on the correction. The message D6 corrects the label given by the under-tested voiceprint sample information D3' to correct the voiceprint analysis model, wherein the correction message D6 can set matching character analysis information, so that the data analysis module 103 can correct it according to The character analysis information of the message D6 finds out the corresponding label, and further corrects the voiceprint analysis model.

Please refer to FIG. 7, another embodiment (2) of the present invention is shown in the figure. As shown in the figure, the financial service end device 30 has a communication capture module 301, and the financial service end device 30 is information connected to the In the personality analysis device 10 , the communication capture module 301 can capture a communication signal (ie, a call sound signal) of the telephone system, and convert the format of the communication signal into voiceprint information to be tested for the personality analysis device 10 to perform For analysis and comparison, the financial industry can capture the voice of the interviewee (that is, the voiceprint information to be tested) through telephone interviews, and after the analysis by the personality analysis device 10, the interviewee can be quickly matched suitable financial instruments.

From the above, it can be seen that the financial product recommendation and risk control system and method of the present invention includes a personality analysis device and at least one client information device. During implementation, the financial product recommendation and risk control system is mainly captured by the personality analysis device. Take a voiceprint information to be tested, analyze and compare the voiceprint information to be tested based on a voiceprint analysis model, to analyze the voiceprint classification features of the voiceprint information to be tested, and query a corresponding voiceprint according to the voiceprint classification features. Character analysis information, the character analysis device then matches a corresponding financial credit evaluation information or a financial commodity information according to the character analysis information, so that the financial service end can use the financial credit evaluation information as an auxiliary judgment reference for credit investigation, the financial service end Corresponding financial product information is also provided to the user according to the user's character; accordingly, after the present invention is implemented, it can indeed achieve a financial product that analyzes the financial product suitable for the user's character by collecting the user's voiceprint information. The purpose of recommendation and risk control system and financial product recommendation method.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the scope of the present invention; any person who is familiar with this technique, without departing from the spirit and scope of the present invention, makes equal changes and modifications, all It should be covered within the patent scope of the present invention.

To sum up, the invention has the patent requirements of "industrial applicability", "novelty" and "progressiveness"; the applicant should file an application for an invention patent with the Jun Bureau in accordance with the provisions of the Patent Law.

1: Financial product recommendation and risk control system 10: Personality Analysis Device 101: Operation processing module 102: Voiceprint processing module 103: Data Analysis Module 104:Data storage module 1041: Voiceprint Database 1042: Personality Test Database 1043: Model Database 1044: Personality Database 1045: Financial Commodities Database 1046: Financial Credit Database 105: Information transmission module 106: Quiz Module 20: Client information device 30: Financial server device 301: Communication capture module D1: Voiceprint information D2: Personality Test Information D1’: Voiceprint sample information D3: Voiceprint information to be tested D4: Financial credit assessment information D3’: Voiceprint sample information to be tested D5: Financial Commodities Information D6: Correction message S1: Retrieve voiceprint information and personality test information to be tested S2: Establish a voiceprint analysis model S3: Retrieve voiceprint information to be tested S4: Perform voiceprint processing on the voiceprint information to be measured S5: The voiceprint analysis model compares the sample information of the voiceprint to be tested S6: Compare financial credit information with personality analysis information S7: Compare financial product information with personality analysis information

Figure 1 is a schematic diagram (1) of the composition of the present invention. Figure 2 is a schematic diagram (2) of the composition of the present invention. Fig. 3 is a flow chart of the implementation of the present invention. FIG. 4 is a schematic diagram (1) of the implementation of the present invention. FIG. 5 is a schematic diagram (2) of the implementation of the present invention. FIG. 6 is another embodiment (1) of the present invention. FIG. 7 is another embodiment (2) of the present invention.

S1: Retrieve voiceprint information and personality test information to be tested

S2: Establish a voiceprint analysis model

S3: Retrieve voiceprint information to be tested

S4: Perform voiceprint processing on the voiceprint information to be measured

S5: The voiceprint analysis model compares the sample information of the voiceprint to be tested

S6: Compare credit risk and financial product information with personality analysis information

Claims (10)

A financial product recommendation and risk control method. A character analysis device is used to capture voiceprint information to be tested, and a voiceprint analysis model is used to analyze corresponding financial credit assessment information or financial product information, which includes: The personality analysis device inputs the plural voiceprint information and the plural personality test information into the personality analysis device; The personality analysis device performs a voiceprint processing operation on the voiceprint information to generate voiceprint sample information, and after marking each of the personality test information, based on the voiceprint sample information and the marked personality test information, establishes the voiceprint analysis model; a client information device, transmitting the voiceprint information to be tested to the character analysis device, and the character analysis device performs the voiceprint processing operation on the voiceprint information to be tested to generate a voiceprint sample information to be tested; After acquiring the information of the voiceprint sample to be tested, the character analysis device compares the voiceprint classification features of the information of the voiceprint sample to be tested through the voiceprint analysis model; and The character analysis device compares a corresponding character analysis information based on the voiceprint classification feature, and matches the corresponding financial credit evaluation information or the financial product information based on the character analysis information. The financial product recommendation and risk control method according to claim 1, wherein after the personality analysis device receives a correction message, the personality analysis device is based on modifying the annotation of the information of the voiceprint sample to be tested. The financial product recommendation and risk control method according to claim 1, wherein the voiceprint processing operation includes a noise reduction process and a captured voiceprint sample. The financial product recommendation and risk control method according to claim 1, wherein the personality analysis information includes one or a combination of consumption habit information, investment habit information and credit risk information. A financial product recommendation and risk control system, comprising: a client information device, storing a voiceprint information to be tested; a personality analysis device, which completes the information connection with the client information device, the personality analysis device includes an arithmetic processing module, and a voiceprint processing module and a data analysis module to complete the information connection with the arithmetic processing module; the arithmetic processing module for running the character analysis device; The voiceprint processing module processes the captured voiceprint information to be measured into voiceprint sample information based on the voiceprint processing operation; The data analysis module compares the voiceprint information to be tested based on the voiceprint analysis model to compare the voiceprint classification feature of the voiceprint sample information, and the voiceprint classification feature corresponds to character analysis information, and The data analysis module matches the character analysis information to generate corresponding financial credit evaluation information or financial product information; and A financial service end device completes information connection with the character analysis device, and the financial service end device is used for receiving the financial credit evaluation information. The financial product recommendation and risk control system according to claim 5, wherein a data storage module completes information connection with the computing processing module, and the data storage module includes a voiceprint database and a personality test database , a model database, a character database and a financial credit database. The financial product recommendation and risk control system according to claim 5, wherein the financial server device includes a communication capture module, and the communication capture module captures a communication signal format and converts it into the voiceprint to be tested News. The financial product recommendation and risk control system according to claim 5, wherein the data analysis module uses the character analysis information to match the corresponding financial product information, and then transmits the financial product information to the client Information device. The financial product recommendation and risk control system according to claim 5, wherein the voiceprint processing operation includes a noise reduction process and a captured voiceprint sample. The financial product recommendation and risk control system according to claim 5, wherein the personality analysis information includes a consumption habit information and an investment habit information.

Images