A method and device for generating financial product recommendation scripts

Abstract

The present application provides a method and device for generating financial product recommendation scripts, which relate to the field of artificial intelligence and can also be used in the financial field, including: parsing the financial product description text to be parsed according to the rule metadata set corresponding to the financial product description text to be parsed to obtain corresponding financial product information; determining the corresponding financial product recommendation degree based on the financial product information, a pre-built financial product recommendation degree generation model, and the financial product purchase history information of the customer to be recommended; writing the financial product information and the financial product recommendation degree into a pre-built standard script template to obtain the corresponding financial product recommendation script. The present application can obtain financial product information in the financial product description text based on color tags, and generate financial product recommendation scripts based on the financial product information.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence and can be used in the financial field. Specifically, it is a method for generating financial product recommendation scripts. Background Art

[0002] In the financial sector, there are a wide variety of financial product prospectuses. To understand the financial products they recommend to clients, product managers need to read the corresponding prospectuses. However, these prospectuses are often quite long, and crucial information that influences the recommendation process can easily be lost.

[0003] Currently, there are two main methods for extracting this key information: manual extraction, which requires high labor costs; and direct manipulation of product specifications using application programming interfaces (APIs), where underlying programs define data extraction rules to extract relevant information. However, this method requires programming based on the type of product specifications. This requires rewriting the extraction code as new product specifications become available, necessitating a high technical barrier to entry. Without the ability to automatically extract key information from financial product specifications, automatically generating recommendations for these products is out of the question. Summary of the Invention

[0004] In response to the problems in the prior art, the present application provides a method and device for generating financial product recommendation scripts, which can obtain financial product information in the financial product description text based on color labels, and generate financial product recommendation scripts based on the financial product information.

[0005] To solve the above technical problems, this application provides the following technical solutions:

[0006] In a first aspect, the present application provides a method for generating a recommendation script for a financial product, comprising:

[0007] Parsing the financial product description text to be parsed according to the rule metadata set corresponding to the financial product description text to be parsed to obtain corresponding financial product information;

[0008] Determining a corresponding financial product recommendation degree based on the financial product information, a pre-built financial product recommendation degree generation model, and the financial product purchase history information of the customer to be recommended;

[0009] The financial product information and the financial product recommendation level are written into a pre-built standard speech template to obtain corresponding financial product recommendation speech.

[0010] Furthermore, the step of parsing the financial product description text to be parsed according to the rule metadata set corresponding to the financial product description text to be parsed to obtain corresponding financial product information includes:

[0011] Using an application programming interface to read the full amount of data in the financial product description text to be parsed, to obtain a corresponding target metadata set;

[0012] Color-labeling the financial product description text to be parsed according to preset parsing keywords to obtain the financial product description text with color labels;

[0013] The target metadata set is traversed according to the rules in the rule metadata set, and the financial product information specified by the rules is read; wherein the rules are set based on the color tag.

[0014] Furthermore, determining the corresponding financial product recommendation level based on the financial product information, a pre-built financial product recommendation level generation model, and the financial product purchase history information of the customer to be recommended includes:

[0015] Calculate the expected net value and expected net value rate of return of the financial product based on the financial product information;

[0016] The historical purchase information of the financial product, the financial product rate in the financial product information, the net value and the net value yield are input into the financial product recommendation degree generation model to generate the financial product recommendation degree for the customer to be recommended.

[0017] Furthermore, the calculating of the expected net value and expected net value rate of return of the financial product based on the financial product information includes:

[0018] Obtaining a historical net value sequence of the financial product from the financial product information;

[0019] Calculate the corresponding historical net asset value yield sequence based on the historical net asset value sequence;

[0020] The historical net worth sequence and the historical net worth rate of return sequence are fitted using a time series analysis algorithm to obtain the expected net worth and the expected net worth rate of return.

[0021] Furthermore, the step of inputting the historical purchase information of the financial product, the financial product rate in the financial product information, the net value and the net value yield into the financial product recommendation generation model to generate the financial product recommendation level for the customer to be recommended includes:

[0022] Generate a corresponding recommended parameter feature set based on the historical purchase information of the financial product, the financial product rate in the financial product information, the net value and the net value yield;

[0023] The recommendation parameter feature set is input into the financial product recommendation degree generation model to generate the financial product recommendation degree of the financial product for the customer to be recommended.

[0024] Furthermore, the step of constructing the financial product recommendation generation model includes:

[0025] Generate a historical feature set of recommended parameters corresponding to each customer based on the customer's historical financial product purchase information, the financial product rate in the financial product information, the net value and the net value rate of return;

[0026] Mark positive and negative samples of each recommended parameter historical feature set based on each customer's purchase results for the financial product;

[0027] The positive and negative sample labeling results and the recommendation parameter historical feature set are input into a distributed gradient boosting model for training to obtain the financial product recommendation degree generation model.

[0028] Furthermore, the method for generating financial product recommendation scripts further includes:

[0029] Collecting audio data of a financial product account manager using the financial product recommendation script to recommend a corresponding financial product to the customer to be recommended, and generating corresponding audio conversion text;

[0030] The audio-to-text conversion is compared with the financial product recommendation words, and an early warning process is performed based on the comparison result.

[0031] Furthermore, the method for generating financial product recommendation scripts, comparing the audio-to-text conversion with the financial product recommendation scripts and performing early warning processing based on the comparison results, includes:

[0032] Input the audio-to-text conversion into the BERT model for semantic parsing to obtain the corresponding first sentence label;

[0033] Inputting the financial product recommendation phrase into the BERT model for semantic parsing to obtain a corresponding second sentence label;

[0034] Compare the first sentence label with the second sentence label, and issue a warning if they are inconsistent.

[0035] In a second aspect, the present application provides a device for generating a recommendation script for a financial product, comprising:

[0036] a financial product information generating unit, configured to parse the financial product description text to be parsed according to a set of rule metadata corresponding to the financial product description text to be parsed, to obtain corresponding financial product information;

[0037] a recommendation degree calculation unit, configured to determine a corresponding financial product recommendation degree based on the financial product information, a pre-built financial product recommendation degree generation model, and the financial product purchase history information of the customer to be recommended;

[0038] The recommendation speech unit is used to write the financial product information and the recommendation degree of the financial product into a pre-built standard speech template to obtain the corresponding financial product recommendation speech.

[0039] Furthermore, the financial product information generating unit includes:

[0040] A target metadata set generation module is configured to read the full amount of data in the financial product description text to be parsed using an application program interface to obtain a corresponding target metadata set;

[0041] A color label generation module is used to color-label the financial product description text to be parsed according to preset parsing keywords to obtain the financial product description text with color labels;

[0042] A product information extraction module is configured to traverse the target metadata set according to the rules in the rule metadata set and read the financial product information specified by the rules; wherein the rules are set based on the color tag.

[0043] Furthermore, the recommendation degree calculation unit includes:

[0044] A net value parameter calculation module, configured to calculate the expected net value and expected net value rate of return of the financial product based on the financial product information;

[0045] The recommendation degree calculation module is used to input the historical purchase information of the financial product, the financial product rate in the financial product information, the net value and the net value yield into the financial product recommendation degree generation model to generate the financial product recommendation degree of the financial product for the customer to be recommended.

[0046] Furthermore, the net value parameter calculation module includes:

[0047] A historical net value sequence generation module, configured to obtain the historical net value sequence of the financial product from the financial product information;

[0048] A historical net asset value yield sequence generation module is used to calculate the corresponding historical net asset value yield sequence based on the historical net asset value sequence;

[0049] The net value parameter expectation module is used to use a time series analysis algorithm to fit the historical net value sequence and the historical net value yield sequence to obtain the expected net value and the expected net value yield.

[0050] Furthermore, the recommendation degree calculation module includes:

[0051] A parameter feature set generation submodule is configured to generate a corresponding recommended parameter feature set based on the historical purchase information of the financial product, the financial product rate in the financial product information, the net value and the net value yield;

[0052] The recommendation degree calculation submodule is used to input the recommendation parameter feature set into the financial product recommendation degree generation model to generate the financial product recommendation degree of the financial product for the customer to be recommended.

[0053] Furthermore, the financial product recommendation speech generation device further includes:

[0054] A parameter history feature set generating unit, configured to generate a recommended parameter history feature set corresponding to each customer based on the customer's financial product purchase history information, the financial product rate in the financial product information, the net value and the net value rate of return;

[0055] A sample marking unit is used to mark positive and negative samples of each recommended parameter historical feature set based on the purchase results of each customer for the financial product;

[0056] The model training unit is used to input the positive and negative sample labeling results and the recommendation parameter historical feature set into a distributed gradient boosting model for training to obtain the financial product recommendation degree generation model.

[0057] Furthermore, the financial product recommendation speech generation device further includes:

[0058] An audio conversion unit is used to collect audio data when a financial product account manager uses the financial product recommendation words to recommend the corresponding financial product to the recommended customer, and generate corresponding audio conversion text;

[0059] A comparison and warning unit is used to compare the audio-to-text conversion with the financial product recommendation words and perform warning processing based on the comparison result.

[0060] Furthermore, the comparison and warning unit includes:

[0061] A first sentence labeling module is used to input the audio-to-text conversion into a BERT model for semantic analysis to obtain a corresponding first sentence label;

[0062] A second sentence labeling module is used to input the financial product recommendation words into the BERT model for semantic analysis to obtain the corresponding second sentence label;

[0063] The comparison and warning module is used to compare the first sentence label with the second sentence label, and issue a warning if they are inconsistent.

[0064] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for generating financial product recommendation scripts when executing the program.

[0065] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for generating financial product recommendation scripts.

[0066] In a fifth aspect, the present application provides a computer program product, comprising a computer program / instruction, which, when executed by a processor, implements the steps of the method for generating financial product recommendation scripts.

[0067] In response to the problems in the existing technology, the financial product recommendation script generation method and device provided in this application can obtain financial product information in the financial product description text based on color labels, and generate financial product recommendation script based on the financial product information; combining the financial product information extraction method based on color labels with the financial product recommendation degree generation model can parse the financial product description text and provide financial product recommendation scripts to financial product account managers, and even issue warnings for possible verbal errors that may occur when financial product account managers use financial product recommendation scripts to recommend financial products to customers; thereby guiding customers to adapt to and accept the net value transformation of financial products, reasonably assess their own risk tolerance, and enhance customer investment experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0069] Figure 1 This is one of the flow charts of the method for generating financial product recommendation scripts in an embodiment of the present application;

[0070] Figure 2 This is a flowchart for obtaining corresponding financial product information in an embodiment of the present application;

[0071] Figure 3 This is a flowchart for determining the recommendation level of corresponding financial products in an embodiment of the present application;

[0072] Figure 4 This is a flowchart for calculating the expected net value and expected net value rate of return of a financial product in an embodiment of the present application;

[0073] Figure 5 A flowchart for generating a financial product recommendation level for a customer to be recommended according to an embodiment of the present application;

[0074] Figure 6 This is a flowchart of constructing a financial product recommendation generation model in an embodiment of the present application;

[0075] Figure 7 This is the second flowchart of the method for generating financial product recommendation scripts in the embodiment of this application;

[0076] Figure 8 This is a flowchart of the early warning process in the embodiment of the present application;

[0077] Figure 9 This is one of the structural diagrams of the device for generating financial product recommendation scripts in an embodiment of the present application;

[0078] Figure 10 This is a structural diagram of the financial product information generation unit in an embodiment of the present application;

[0079] Figure 11 This is a structural diagram of a recommendation degree calculation unit in an embodiment of the present application;

[0080] Figure 12 This is a structural diagram of the net value parameter calculation module in the embodiment of the present application;

[0081] Figure 13 This is a structural diagram of the recommendation degree calculation module in an embodiment of the present application;

[0082] Figure 14 This is the second structural diagram of the device for generating financial product recommendation scripts in an embodiment of the present application;

[0083] Figure 15 This is the third structural diagram of the device for generating financial product recommendation scripts in an embodiment of the present application;

[0084] Figure 16 This is a structural diagram of the comparison and warning unit in an embodiment of the present application;

[0085] Figure 17 A schematic diagram of the structure of an electronic device in an embodiment of the present application;

[0086] Figure 18 This is a flowchart of a method for generating financial product recommendation scripts in an embodiment of the present application;

[0087] Figure 19 This is one of the model training diagrams in the embodiment of this application;

[0088] Figure 20 This is the second model training diagram in the embodiment of this application. DETAILED DESCRIPTION

[0089] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0090] It should be noted that the method and device for generating financial product recommendation scripts provided in this application can be used in the financial field, and can also be used in any field other than the financial field. The application field of the method and device for generating financial product recommendation scripts provided in this application is not limited.

[0091] The acquisition, storage, use and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0092] In one embodiment, see Figure 1 In order to obtain financial product information from financial product descriptions based on color tags and generate recommendation scripts for the financial products based on this financial product information, this application provides a method for generating financial product recommendation scripts. This method first extracts financial product information from financial product descriptions based on color tags, and then generates recommendation scripts for the financial products based on this financial product information, so that financial product account managers can recommend the financial products to potential customers (potential purchasers of the financial products) according to the recommendation scripts.

[0093] It should be noted that when generating recommendation scripts for financial products, the financial product recommendation script generation method provided in this application combines the color labeling method with the financial product recommendation degree generation model to parse the financial product description text and provide financial product recommendation scripts for financial product account managers. Subsequently, text comparison technology will be used to issue warnings for possible verbal errors (such as reading errors) that may occur when financial product account managers use the financial product recommendation scripts to recommend financial products to potential customers, in order to guide customers to adapt to and accept the net value transformation of financial products, reasonably assess their own risk tolerance, and enhance customers' investment experience.

[0094] The method for generating financial product recommendation scripts provided in this application includes:

[0095] S101: Parsing the financial product description text to be parsed according to a rule metadata set corresponding to the financial product description text to be parsed to obtain corresponding financial product information;

[0096] Specifically, in one embodiment, see Figure 2Step S101 includes: using an application program interface to read the full amount of data in the financial product description text to be parsed to obtain a corresponding target metadata set (S201); color-labeling the financial product description text to be parsed according to preset parsing keywords to obtain a financial product description text with color labels (S202); traversing the target metadata set according to the rules in the rule metadata set, and reading the financial product information specified by the rules (S203); wherein the rules are set based on the color labels.

[0097] Specifically, the process from step S201 to step S203 is as follows:

[0098] 1. Use an application programming interface (e.g., Java POI) to read the full data of the financial product description text (also known as the product manual) to be parsed and obtain the corresponding target metadata set. The target metadata set includes the full data of the financial product description text.

[0099] 2. For the product description text, set the color and font color of each cell in each table to obtain the color-tagged financial product description text. Then, set a rule set for the product description template based on the color and font color of each cell, thereby obtaining a rule metadata set. This rule metadata set contains at least the information represented by each font color in each cell.

[0100] 3. Select the instruction manual template corresponding to the financial product description text according to business needs, and extract the rule metadata set based on the instruction manual template.

[0101] 3. Use the rule metadata dataset to model the rule set.

[0102] 4. Use the rule set model to collect information of the target metadata set.

[0103] 5. The extracted financial product information is stored in storage media such as files and databases.

[0104] Specific examples are as follows:

[0105] In this example, the rules in the rule metadata set are as follows:

[0106] 1) The orange background color of the table is used to determine the range of information to be extracted, which is the cell to the right of the cell. The font color "black" is fixed information, the font color "green" is the entry item, and the format is "[table name / column name]"

[0107] 2) The table background color "dark blue" is used to determine the range of information to be extracted, which is the vertical range on the right side of the unit. The font color "white" in the table is fixed information, and the font color "yellow" is the entry item. The format is "[table name]"

[0108] 3) The table background color "light blue" is used to determine the range of information to be extracted, which is the entire column below the cell. The font color "black" in the table is fixed information, and the font color "orange" is the input item. The format is "[column name]"

[0109] A portion of a product brochure template provided by a financial product issuing company is shown in Table 1 below:

[0110] Table 1

[0111]

[0112] If the business scenario requires extracting the fund name and Class A subscription fee information from the product manual, the problem can be transformed into:

[0113] 1) Extract the cell information on the right side of the "Fund Name" table and store the relevant information in the JJZC field in the BASE_INFO table of the database.

[0114] 2) Extract all values ​​from the "Subscription Amount (M)" and "Rate" columns in the right range of the "Class A Subscription Fee" table and store the relevant information in the MONEY and RATE fields in the SGJE_INFO table. Business personnel can then modify the rule template as follows by adjusting the table background color (see Table 2):

[0115] Table 2

[0116]

[0117] When a financial product issuing company provides a new product brochure, the steps to extract key information are:

[0118] For an example of a product manual, see Table 3:

[0119] Table 3

[0120]

[0121] The first step is to use Java POI to parse the uploaded product manual and obtain the target metadata set.

[0122] The second step is to find the corresponding template based on the product manual template type in the target metadata set and extract the rule metadata set. An example of the rule metadata set is:

[0123] {

[0124] …………

[0125] Product brochure template:……;

[0126] Rule 1:

[0127] {

[0128] Fixed information: Fund manager;

[0129] Extract range: right;

[0130] Format: BASE_INFO / JJZC;

[0131] Range append: []

[0132] };

[0133] Rule 2:

[0134] {

[0135] Fixed information: Class A subscription fee;

[0136] Extraction range: vertical range on the right;

[0137] Format: SGJE_INFO;

[0138] other:[{

[0139] Fixed information: subscription amount (M);

[0140] Extract range: entire column;

[0141] Format: SGJE_INFO / MONEY;

[0142] Range appended: []

[0143] },{

[0144] Fixed information: rate;

[0145] Extract range: entire column;

[0146] Format: SGJE_INFO / RATE;

[0147] Range append: []

[0148] };

[0149] …………

[0150] }

[0151] The third step is to traverse the rule metadata data set, find rule 1, and extract information such as "fund manager", "right side", "BASE_INFO / JJZC", etc.

[0152] The fourth step is to traverse the target metadata set to be extracted according to the rules, find the data in the "right" cell of the "Fund Manager" table, that is, "XX Hybrid Securities Investment Fund", and insert it into the JJZC field in the BASE_INFO table.

[0153] Step 5: Continue traversing Rule 2, Rule 3, and so on until all rules are traversed.

[0154] From the above description, it can be seen that the financial product recommendation script generation method provided in this application can parse the financial product description text to be parsed according to the rule metadata set corresponding to the financial product description text to be parsed to obtain the corresponding financial product information.

[0155] S102: Determine the corresponding financial product recommendation level based on the financial product information, a pre-built financial product recommendation level generation model, and the financial product purchase history information of the customer to be recommended;

[0156] Specifically, in one embodiment, see Figure 3 Step S102 includes: calculating the expected net value and expected net value rate of return of the financial product based on the financial product information (S301); inputting the historical purchase information of the financial product, the financial product rate in the financial product information, the net value and the net value rate of return into the financial product recommendation degree generation model to generate the financial product recommendation degree of the financial product for the customer to be recommended (S302).

[0157] Further, see Figure 4 Step S301 includes: obtaining the historical net value sequence of the financial product from the financial product information (S401); calculating the corresponding historical net value yield sequence based on the historical net value sequence (S402); and fitting the historical net value sequence and the historical net value yield sequence using a time series analysis algorithm to obtain the expected net value and the expected net value yield (S403).

[0158] In specific implementation, statistical and time series analysis techniques are used to analyze product net value and returns:

[0159] Step 1: Integrate the acquired net worth data {x t ,x t-1 ,x t-2 ,…x t-n For example, the net value of a certain net value financial product was 1 on August 1, 2019, and changed three times within a month: 1.002 on August 5, 2019, 1.006 on August 12, 2019, and 1.005 on August 30, 2019. The net value data obtained through integration is {1, 1.002, 1.006, 1.005}.

[0160] Step 2: Calculate the net asset value rate of return R t , where x t is the current net value, x t-n is the initial net value, and n is the number of days the product has been in operation. Based on the example in step 1, if an investor subscribes to the net value financial product for the initial offering, the net value yield on August 30, 2018 is (1.005-1)×365 / 30=6.08%.

[0161] Step 3: Draw the net value and net value yield curves based on the net value sequence and the calculated net value yield sequence to reflect the changes and discover trend patterns.

[0162] Step 4: For sequences with trend patterns, use time series analysis and prediction technology, such as LSTM and other model technologies, to predict the future net value of the product x t+m =f1(x t ,x t-1 ,…x t-n ) and net asset value rate of return R t+m =f2(R t ,R t-1 ,R t-2 ,…,R t-n ).

[0163] From the above description, it can be seen that the financial product recommendation script generation method provided in this application can calculate the expected net value and expected net value yield of the financial product based on the financial product information.

[0164] Further, see Figure 5 Step S302 includes: generating a corresponding recommendation parameter feature set based on the historical purchase information of the financial product, the financial product rate in the financial product information, the net value and the net value yield (S501); inputting the recommendation parameter feature set into the financial product recommendation degree generation model to generate the financial product recommendation degree of the financial product for the customer to be recommended (S502).

[0165] It should be noted that, in one embodiment, see Figure 6 The steps of constructing the financial product recommendation generation model include:

[0166] S601: Generate a historical feature set of recommended parameters for each customer based on each customer's historical financial product purchase information, the financial product rates in the financial product information, the net value and net value yield rate; for example, integrate the product's net value data (denoted as N), net value yield data (denoted as R), and purchase cost data (denoted as C), and also integrate data such as whether the customer holds other investment products while holding the product (denoted as Q), and product portfolio adjustments (denoted as T) to form a modeling data feature set (denoted as X), where X = {N, R, C, Q, T};

[0167] S602: Label each recommended parameter historical feature set as positive or negative samples based on each customer's purchase results for the financial product. For example, positive and negative samples are identified based on whether the customer holds the product. That is, if the customer has historically held the product, the customer is labeled as 1, otherwise, the customer is labeled as 0. These serve as positive and negative sample labels, denoted as Y.

[0168] S603: Input the positive and negative sample labeling results and the recommendation parameter historical feature set into the distributed gradient boosting model for training to obtain the financial product recommendation generation model. For example, model training is performed according to the modeling process of classic machine learning models (such as XGBoost, logistic regression, etc.); for example, model data (X, Y) is preprocessed first, and then the model is established and model training is performed, such as Figure 19 and Figure 20 shown.

[0169] The trained model is then used to predict customers who do not hold the product to obtain the product's recommendation potential (i.e., the recommendation probability predicted by the model), and then combined with the marketer's experience to determine whether the product needs to be recommended to the customer.

[0170] From the above description, it can be seen that the financial product recommendation script generation method provided in this application can input the financial product historical purchase information, the financial product rate in the financial product information, the net value and the net value yield into the financial product recommendation degree generation model to generate the financial product recommendation degree for the customer to be recommended.

[0171] S103: Writing the financial product information and the financial product recommendation level into a pre-built standard speech template to obtain corresponding financial product recommendation speech.

[0172] Specifically, in one embodiment, see Figure 7 The method for generating financial product recommendation scripts further includes:

[0173] S701: Collecting audio data when a financial product account manager uses the financial product recommendation script to recommend a corresponding financial product to the customer to be recommended, and generating corresponding audio conversion text;

[0174] S702: Compare the audio-to-text conversion with the financial product recommendation words, and perform early warning processing based on the comparison result.

[0175] Further, see Figure 8 Step S702 includes: inputting the audio-to-text conversion into the BERT model for semantic parsing to obtain a corresponding first sentence label (S801); inputting the financial product recommendation words into the BERT model for semantic parsing to obtain a corresponding second sentence label (S802); comparing the first sentence label with the second sentence label, and issuing an early warning if they are inconsistent (S803).

[0176] Understandably, see Figure 18 , the above steps S701 to S702 can be specifically executed as follows:

[0177] Step 1: Set up a standard speech template.

[0178] Step 2: The financial product account manager enters the financial product code on the front end of the webpage;

[0179] Step 3: The backend generates a parameter list based on the financial product code, for example, {"pdid":xxxxx,"jjlx":xxxxx,"jjzc":xxxxx,………………}. The corresponding values ​​in the list are read and combined with the standard script template to form a financial product recommendation script, which is then transmitted to the front-end page and provided to the financial product account manager.

[0180] Step 4: The financial product account manager reads the financial product recommendation script aloud. The front-end records the sound through a high-definition camera and uses ASR technology to convert the acquired voice into structured data. At the same time, the financial product recommendation script is also structured.

[0181] Step 5: Build BERT text models for the structured data corresponding to the words read by the financial product account manager in step 4 and the structured data corresponding to the words recommended by the financial product, respectively. Use a short sentence sliding window to segment them according to different semantics and label the sentences separately.

[0182] Therefore, at the tag level, the audio-converted text can be compared with the financial product recommendation language to see if they are consistent. If they are inconsistent, an early warning message can be issued.

[0183] Finally, for example, the recommendation script for financial products could be:

[0184] "You are about to purchase a ×××× fund, managed by ×××× Fund Company, with a product code of ××××, and issued through public offering. The product's net asset value is ××××, and its net asset value yield is ××××. Analysis shows that this product will maintain a stable growth trend and is ×××× compatible with your investment preferences, so we recommend purchasing this product. Do you understand and agree with the above?"

[0185] From the above description, it can be seen that the financial product recommendation script generation method provided by this application can obtain financial product information in the financial product description text based on color labels, and generate financial product recommendation script based on the financial product information; combining the financial product information extraction method based on color labels with the financial product recommendation degree generation model can parse the financial product description text and provide financial product recommendation scripts to financial product account managers, and even issue warnings for possible verbal errors that may occur when financial product account managers use financial product recommendation scripts to recommend financial products to customers; thereby guiding customers to adapt to and accept the net value transformation of financial products, reasonably assess their own risk tolerance, and enhance customer investment experience.

[0186] Based on the same inventive concept, the embodiments of the present application also provide a financial product recommendation speech generation device, which can be used to implement the method described in the above embodiments, as described in the following embodiments. Since the principle of solving the problem by the financial product recommendation speech generation device is similar to that of the financial product recommendation speech generation method, the implementation of the financial product recommendation speech generation device can refer to the implementation of the method based on software performance benchmark determination, and the repetitions will not be repeated. As used below, the term "unit" or "module" can be a combination of software and / or hardware that implements the predetermined function. Although the system described in the following embodiments is preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and conceived.

[0187] In one embodiment, see Figure 9 In order to obtain financial product information in the financial product description text based on color labels and generate recommendation words for the financial products based on the financial product information, the present application provides a financial product recommendation wording generation device, including: a financial product information generation unit 901, a recommendation degree calculation unit 902 and a recommendation wording unit 903.

[0188] The financial product information generating unit 901 is configured to parse the financial product description text to be parsed according to the rule metadata set corresponding to the financial product description text to be parsed, and obtain the corresponding financial product information;

[0189] A recommendation degree calculation unit 902 is configured to determine a corresponding financial product recommendation degree based on the financial product information, a pre-built financial product recommendation degree generation model, and the financial product purchase history information of the customer to be recommended;

[0190] The recommendation speech unit 903 is used to write the financial product information and the financial product recommendation level into a pre-built standard speech template to obtain the corresponding financial product recommendation speech.

[0191] In one embodiment, see Figure 10 The financial product information generating unit 901 includes: a target metadata set generating module 1001, a color label generating module 1002 and a product information extracting module 1003.

[0192] The target metadata set generation module 1001 is configured to read the full amount of data in the financial product description text to be parsed using an application programming interface to obtain a corresponding target metadata set;

[0193] A color label generating module 1002 is configured to color-label the financial product description text to be parsed according to preset parsing keywords to obtain the financial product description text with color labels;

[0194] The product information extraction module 1003 is configured to traverse the target metadata set according to the rules in the rule metadata set and read the financial product information specified by the rules; wherein the rules are set based on the color tags.

[0195] In one embodiment, see Figure 11 The recommendation degree calculation unit 902 includes: a net value parameter calculation module 1101 and a recommendation degree calculation module 1102.

[0196] The net value parameter calculation module 1101 is used to calculate the expected net value and expected net value rate of return of the financial product based on the financial product information;

[0197] The recommendation degree calculation module 1102 is used to input the financial product historical purchase information, the financial product rate in the financial product information, the net value and the net value yield into the financial product recommendation degree generation model to generate the financial product recommendation degree for the customer to be recommended.

[0198] In one embodiment, see Figure 12 The net value parameter calculation module 1101 includes: a historical net value sequence generation module 1201, a historical net value yield sequence generation module 1202 and a net value parameter expectation module 1203.

[0199] A historical net value sequence generation module 1201 is configured to obtain the historical net value sequence of the financial product from the financial product information;

[0200] A historical net asset value yield sequence generation module 1202 is configured to calculate a corresponding historical net asset value yield sequence based on the historical net asset value sequence;

[0201] The net value parameter expectation module 1203 is used to fit the historical net value sequence and the historical net value rate of return sequence using a time series analysis algorithm to obtain the expected net value and the expected net value rate of return.

[0202] In one embodiment, see Figure 13 The recommendation degree calculation module 1102 includes: a parameter feature set generation submodule 1301 and a recommendation degree calculation submodule 1302.

[0203] The parameter feature set generation submodule 1301 is configured to generate a corresponding recommended parameter feature set based on the historical purchase information of the financial product, the financial product rate in the financial product information, the net value and the net value rate of return;

[0204] The recommendation degree calculation submodule 1302 is configured to input the recommendation parameter feature set into the financial product recommendation degree generation model to generate the financial product recommendation degree of the financial product for the customer to be recommended.

[0205] In one embodiment, see Figure 14 The financial product recommendation speech generation device also includes: a parameter history feature set generation unit 1401, a sample marking unit 1402 and a model training unit 1403.

[0206] The parameter history feature set generating unit 1401 is configured to generate a recommended parameter history feature set corresponding to each customer based on the customer's financial product purchase history information, the financial product rate in the financial product information, the net value and the net value rate of return;

[0207] The sample marking unit 1402 is used to mark positive and negative samples of each recommendation parameter historical feature set according to the purchase results of each customer for the financial product;

[0208] The model training unit 1403 is used to input the positive and negative sample labeling results and the recommendation parameter historical feature set into a distributed gradient boosting model for training to obtain the financial product recommendation degree generation model.

[0209] In one embodiment, see Figure 15 The financial product recommendation speech generation device also includes: an audio conversion unit 1501 and a comparison and warning unit 1502.

[0210] The audio conversion unit 1501 is used to collect audio data when the financial product account manager recommends the corresponding financial product to the recommended customer using the financial product recommendation script, and generate corresponding audio conversion text;

[0211] The comparison and warning unit 1502 is used to compare the audio-to-text conversion with the financial product recommendation words and perform warning processing based on the comparison result.

[0212] In one embodiment, see Figure 16 The comparison and warning unit 1502 includes: a first sentence marking module 1601, a second sentence marking module 1602 and a comparison and warning module 1603.

[0213] The first sentence labeling module 1601 is used to input the audio-to-text conversion into the BERT model for semantic analysis to obtain the corresponding first sentence label;

[0214] The second sentence labeling module 1602 is configured to input the financial product recommendation phrase into the BERT model for semantic analysis to obtain a corresponding second sentence label;

[0215] The comparison and warning module 1603 is used to compare the first sentence label with the second sentence label, and issue a warning if they are inconsistent.

[0216] From a hardware perspective, in order to obtain financial product information from a financial product description based on color tags and generate a recommendation for the financial product based on the financial product information, the present application provides an embodiment of an electronic device for implementing all or part of the method for generating a recommendation for a financial product. The electronic device specifically includes the following:

[0217] Processor, memory, communication interface and bus; wherein the processor, memory and communication interface communicate with each other via the bus; the communication interface is used to realize information transmission between the financial product recommendation speech generation device and related equipment such as the core business system, user terminal and related database; the logic controller can be a desktop computer, tablet computer and mobile terminal, etc., but this embodiment is not limited to this. In this embodiment, the logic controller can be implemented with reference to the embodiment of the financial product recommendation speech generation method and the embodiment of the financial product recommendation speech generation device in the embodiment, and their contents are merged here, and the repeated parts are not repeated.

[0218] It is understandable that the user terminal may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, a smart watch, a smart bracelet, etc.

[0219] In practical applications, portions of the method for generating financial product recommendation scripts can be executed on the electronic device as described above, or all operations can be performed on the client device. The specific selection can be based on the processing capabilities of the client device and the limitations of the user's usage scenario. This application does not impose any restrictions on this. If all operations are performed on the client device, the client device may also include a processor.

[0220] The client device may include a communication module (i.e., a communication unit) that can establish a communication connection with a remote server to implement data transmission with the server. The server may include a server on the task scheduling center side, and in other implementation scenarios, may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, a server cluster consisting of multiple servers, or a server structure of a distributed device.

[0221] Figure 17 Schematic block diagram of the system structure of the electronic device 9600 according to an embodiment of the present application. Figure 17 As shown, the electronic device 9600 may include a central processing unit 9100 and a memory 9140; the memory 9140 is coupled to the central processing unit 9100. It is worth noting that the Figure 17 is exemplary; other types of structures may also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0222] In one embodiment, the function of generating a method for generating a recommendation script for a financial product may be integrated into the central processing unit 9100. The central processing unit 9100 may be configured to perform the following control:

[0223] S101: Parsing the financial product description text to be parsed according to a rule metadata set corresponding to the financial product description text to be parsed to obtain corresponding financial product information;

[0224] S102: Determine the corresponding financial product recommendation level based on the financial product information, a pre-built financial product recommendation level generation model, and the financial product purchase history information of the customer to be recommended;

[0225] S103: Writing the financial product information and the financial product recommendation level into a pre-built standard speech template to obtain corresponding financial product recommendation speech.

[0226] From the above description, it can be seen that the financial product recommendation script generation method provided by this application can obtain financial product information in the financial product description text based on color labels, and generate financial product recommendation script based on the financial product information; combining the financial product information extraction method based on color labels with the financial product recommendation degree generation model can parse the financial product description text and provide financial product recommendation scripts to financial product account managers, and even issue warnings for possible verbal errors that may occur when financial product account managers use financial product recommendation scripts to recommend financial products to customers; thereby guiding customers to adapt to and accept the net value transformation of financial products, reasonably assess their own risk tolerance, and enhance customer investment experience.

[0227] In another embodiment, the financial product recommendation speech generation device can be configured separately from the central processing unit 9100. For example, the data composite transmission device financial product recommendation speech generation device can be configured as a chip connected to the central processing unit 9100, and the function of the financial product recommendation speech generation method can be realized through the control of the central processing unit.

[0228] like Figure 17 As shown, the electronic device 9600 may further include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It is worth noting that the electronic device 9600 does not necessarily have to include Figure 17 In addition, the electronic device 9600 may also include all components shown in Figure 17 For components not shown, reference may be made to the prior art.

[0229] like Figure 17 As shown, the central processing unit 9100 is sometimes also referred to as a controller or operation control, and may include a microprocessor or other processor device and / or logic device. The central processing unit 9100 receives input and controls the operation of various components of the electronic device 9600.

[0230] Memory 9140 can be, for example, one or more of a cache, flash memory, hard drive, removable media, volatile memory, non-volatile memory, or other suitable devices. It can store the aforementioned failure-related information and also store programs that execute the relevant information. The CPU 9100 can execute the programs stored in memory 9140 to implement information storage or processing.

[0231] The input unit 9120 provides input to the central processing unit 9100. The input unit 9120 may be, for example, a keypad or touch input device. The power supply 9170 is used to provide power to the electronic device 9600. The display 9160 is used to display objects such as images and text. The display may be, for example, an LCD display, but is not limited thereto.

[0232] The memory 9140 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), or a SIM card. Alternatively, it may be a memory that retains information even when power is off, can be selectively erased, and is provided with more data. Examples of such memory are sometimes referred to as EPROMs. The memory 9140 may also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142 for storing application programs and function programs or processes for executing the operation of the electronic device 9600 by the central processing unit 9100.

[0233] The memory 9140 may also include a data storage unit 9143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various driver programs for the electronic device's communication functions and / or for executing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0234] The communication module 9110 is a transmitter / receiver 9110 that sends and receives signals via an antenna 9111. The communication module (transmitter / receiver) 9110 is coupled to the central processor 9100 to provide input signals and receive output signals, which may be the same as in a conventional mobile communication terminal.

[0235] Based on different communication technologies, multiple communication modules 9110 can be provided in the same electronic device, such as a cellular network module, a Bluetooth module, and / or a wireless local area network module. The communication module (transmitter / receiver) 9110 is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide audio output via the speaker 9131 and receive audio input from the microphone 9132, thereby implementing common telecommunication functions. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. Furthermore, the audio processor 9130 is coupled to the central processing unit 9100, enabling local recording via the microphone 9132 and playback of stored audio via the speaker 9131.

[0236] Embodiments of the present application also provide a computer-readable storage medium capable of implementing all steps of the method for generating financial product recommendation scripts in the above-mentioned embodiment, where the execution subject is a server or a client. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the computer program implements all steps of the method for generating financial product recommendation scripts in the above-mentioned embodiment, where the execution subject is a server or a client. For example, when the processor executes the computer program, the following steps are implemented:

[0237] S101: Parsing the financial product description text to be parsed according to a rule metadata set corresponding to the financial product description text to be parsed to obtain corresponding financial product information;

[0238] S102: Determine the corresponding financial product recommendation level based on the financial product information, a pre-built financial product recommendation level generation model, and the financial product purchase history information of the customer to be recommended;

[0239] S103: Writing the financial product information and the financial product recommendation level into a pre-built standard speech template to obtain corresponding financial product recommendation speech.

[0240] From the above description, it can be seen that the financial product recommendation script generation method provided by this application can obtain financial product information in the financial product description text based on color labels, and generate financial product recommendation script based on the financial product information; combining the financial product information extraction method based on color labels with the financial product recommendation degree generation model can parse the financial product description text and provide financial product recommendation scripts to financial product account managers, and even issue warnings for possible verbal errors that may occur when financial product account managers use financial product recommendation scripts to recommend financial products to customers; thereby guiding customers to adapt to and accept the net value transformation of financial products, reasonably assess their own risk tolerance, and enhance customer investment experience.

[0241] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, apparatus, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0242] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (apparatus), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as a combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0243] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0244] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0245] Specific embodiments are used in the present invention to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

Claims

1. A method for generating a recommendation script for a financial product, characterized in that: include: Parsing the financial product description text to be parsed according to the rule metadata set corresponding to the financial product description text to be parsed to obtain corresponding financial product information; Determining a corresponding financial product recommendation degree based on the financial product information, a pre-built financial product recommendation degree generation model, and the financial product purchase history information of the customer to be recommended; Writing the financial product information and the financial product recommendation level into a pre-built standard script template to obtain corresponding financial product recommendation scripts; Collecting audio data of a financial product account manager using the financial product recommendation script to recommend a corresponding financial product to the customer to be recommended, and generating corresponding audio conversion text; The audio-to-text conversion is compared with the financial product recommendation words, and an early warning process is performed based on the comparison result.

2. The method for generating financial product recommendation scripts according to claim 1, characterized in that: The step of parsing the financial product description text to be parsed according to the rule metadata set corresponding to the financial product description text to be parsed to obtain corresponding financial product information includes: Using an application programming interface to read the full amount of data in the financial product description text to be parsed, to obtain a corresponding target metadata set; Color-labeling the financial product description text to be parsed according to preset parsing keywords to obtain the financial product description text with color labels; The target metadata set is traversed according to the rules in the rule metadata set, and the financial product information specified by the rules is read; wherein the rules are set based on the color tag.

3. The method for generating financial product recommendation scripts according to claim 1, characterized in that: The determining of the corresponding financial product recommendation degree based on the financial product information, a pre-built financial product recommendation degree generation model, and the financial product purchase history information of the customer to be recommended includes: Calculate the expected net value and expected net value rate of return of the financial product based on the financial product information; The financial product historical purchase information, the financial product rate in the financial product information, the expected net value and the expected net value yield are input into the financial product recommendation generation model to generate the financial product recommendation level for the customer to be recommended.

4. The method for generating financial product recommendation scripts according to claim 3, characterized in that: Calculating the expected net value and expected net value rate of return of the financial product based on the financial product information includes: Obtaining a historical net value sequence of the financial product from the financial product information; Calculate the corresponding historical net asset value yield sequence based on the historical net asset value sequence; The historical net worth sequence and the historical net worth rate of return sequence are fitted using a time series analysis algorithm to obtain the expected net worth and the expected net worth rate of return.

5. The method for generating financial product recommendation scripts according to claim 3, characterized in that: The step of inputting the historical purchase information of the financial product, the financial product rate in the financial product information, the expected net value, and the expected net value rate of return into the financial product recommendation degree generation model to generate the financial product recommendation degree for the customer to be recommended includes: Generate a corresponding recommendation parameter feature set based on the historical purchase information of the financial product, the financial product rate in the financial product information, the expected net value, and the expected net value rate of return; The recommendation parameter feature set is input into the financial product recommendation degree generation model to generate the financial product recommendation degree of the financial product for the customer to be recommended.

6. The method for generating financial product recommendation scripts according to claim 3, characterized in that: The steps of constructing the financial product recommendation generation model include: Generate a historical feature set of recommended parameters for each customer based on the customer's historical financial product purchase information, the financial product rate, net value, and net value yield in the customer's financial product information; Mark positive and negative samples of each recommended parameter historical feature set based on each customer's purchase results for the financial product; The positive and negative sample labeling results and the recommendation parameter historical feature set are input into a distributed gradient boosting model for training to obtain the financial product recommendation degree generation model.

7. The method for generating financial product recommendation scripts according to claim 1, characterized in that: The comparing the audio-to-text conversion with the financial product recommendation words and performing early warning processing based on the comparison result includes: Input the audio-to-text conversion into the BERT model for semantic parsing to obtain the corresponding first sentence label; Inputting the financial product recommendation phrase into the BERT model for semantic parsing to obtain a corresponding second sentence label; Compare the first sentence label with the second sentence label, and issue a warning if they are inconsistent.

8. A device for generating a recommendation script for a financial product, characterized in that: include: a financial product information generating unit, configured to parse the financial product description text to be parsed according to a set of rule metadata corresponding to the financial product description text to be parsed, to obtain corresponding financial product information; a recommendation degree calculation unit, configured to determine a corresponding financial product recommendation degree based on the financial product information, a pre-built financial product recommendation degree generation model, and the financial product purchase history information of the customer to be recommended; A recommendation script unit, configured to write the financial product information and the financial product recommendation level into a pre-built standard script template to obtain a corresponding financial product recommendation script; An audio conversion unit is used to collect audio data when a financial product account manager uses the financial product recommendation words to recommend the corresponding financial product to the recommended customer, and generate corresponding audio conversion text; A comparison and warning unit is used to compare the audio-to-text conversion with the financial product recommendation words and perform warning processing based on the comparison result.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the method for generating financial product recommendation scripts according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for generating financial product recommendation scripts as described in any one of claims 1 to 7 are implemented.

11. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method for generating financial product recommendation scripts as described in any one of claims 1 to 7 are implemented.

Images