A credit assessment and credit application system and method based on information data
Through the credit assessment and credit application system based on information data, a credit assessment model in multiple scenarios was established, and the problem of imperfect credit assessment of farmers' micro-loans was solved, the satisfaction of farmers' financing needs and the effective supply of financial resources was achieved, and the improvement of the Management efficiency of financial institutions.
Patent Information
- Application Number
- CN202011030817.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-09-27
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2040-09-27
AI Technical Summary
The credit assessment system for farmers' micro-loans is incomplete, which makes the supply of financial resources unable to meet the financial needs of farmers' business development, and it is difficult for financial institutions to obtain farmers' transaction and management data in a timely manner for effective assessment.
Credit assessment and credit application system based on information data, obtain real-name authentication and data authorization information by receiving farmers' financing application requests, collect and analyze external information data, establish a credit assessment model in multiple scenarios, and match the matching conditions based on the model results. The credit product realizes real-time data interaction with financial institutions and credit application.
It has improved the financing availability and convenience of farmers and new agricultural operators, promoted the development of inclusive finance and rural economy, and improved the management efficiency and financing risk control of financial institutions.
Smart Images

Figure CN112328868B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of credit assessment technology and financial technology output, and in particular to a credit assessment and credit application system and method based on transaction and management information data. Background Art
[0002] The agricultural population is huge, accounting for more than 40% of China's total population. At the same time, due to the difficulties and characteristics of small loans for farmers, such as incomplete financial information and scattered loan objects, the credit evaluation system for small loans for farmers is very imperfect. At the same time, the credit evaluation system for farmers in many financial institutions cannot be improved and upgraded in a timely manner, resulting in the inability of the existing rural financial resource supply to meet the capital needs of farmers' business development. In recent years, the business model of farmers has been constantly developing, and new agricultural business entities such as family farms, farmers' professional cooperatives, and leading enterprises have gradually emerged. At the same time, farmers are also encouraged to use information technology, not only using e-commerce trading platforms to realize online buying and selling transactions, but also starting to use production management platforms to realize online management of production data and improve operating efficiency. However, it is difficult for actual financial institutions to connect to the platforms used by farmers to obtain system data in a timely manner, thus lacking the collection and evaluation of effective transaction and management data. Summary of the invention
[0003] In order to solve the technical problems in the prior art, such as the incomplete financial information and scattered loan objects of small loans for farmers, which lead to the imperfect credit evaluation system for small loans for farmers, and the inability of many financial institutions to improve and upgrade the credit evaluation system for farmers in a timely manner, resulting in the inability of the existing rural financial resource supply to meet the capital needs of farmers' business development, the present invention provides a credit evaluation and credit application system and method based on information data, attempts to build a credit evaluation system based on farmers' transaction and management data, and connects with financial institutions to complete data information interaction and real-time credit application through the Internet. It can not only improve the availability and convenience of financing for farmers and new agricultural operators, obtain the required funds to engage in their various production and operation activities, promote inclusive finance and rural economic development, but also improve the management efficiency and financing risk control of financial institutions.
[0004] In order to achieve this technical effect, the present invention adopts the following technical solution.
[0005] A credit assessment and credit application method based on information data, the method comprising the steps of:
[0006] A. Receive the financing application request initiated by the farmer and obtain the corresponding real-name authentication and data authorization information;
[0007] B. Collect and store external information data returned by the big data platform according to the data authorization information, and parse the external information data;
[0008] C. Select the required indicators from the indicator library, combine the indicators into model relationships through operators, and establish a credit assessment model for multiple scenarios;
[0009] D. Based on the results of the credit assessment model, eligible credit products are matched to farmers according to the credit product routing matching rules.
[0010] In addition, collecting and storing external information data returned by the big data platform according to the data authorization information, and parsing the external information data include:
[0011] B1. Request the external data platform interface to obtain external information data;
[0012] B2. According to the description of the external information data, query the corresponding indicator and establish a relationship between the external information data and the indicator;
[0013] B3. Store the external information data that establishes a relationship with the indicators into the database together with the indicators.
[0014] In addition, the querying of corresponding indicators according to the description of the external information data and establishing a relationship between the external information data and the indicators includes:
[0015] B21. Searching for a corresponding indicator according to the description of the external information data. If the corresponding indicator can be found, establishing a relationship between the external information data and the indicator;
[0016] B22. If no indicator can be searched, the external information data description is segmented, and the obtained segmented words are used to search the indicator system according to the keywords, and indicators that hit multiple keywords are selected, and the indicators and keywords and metadata corresponding to the keywords are returned to establish a corresponding relationship between the external information data and the indicators;
[0017] B23. If the indicator is not obtained in the end, the external information data will be placed in the exception pool for exception handling.
[0018] In addition, the steps of storing the external information data related to the indicator together with the indicator into the database include:
[0019] B301. Perform keyword comparison on multiple external information data descriptions that establish relationships with the same indicator to obtain external information data descriptions with high relevance;
[0020] B302. For highly relevant external information data descriptions, perform logic detection according to metadata corresponding to keywords, and perform information data fusion on the highly relevant external information data descriptions.
[0021] In addition, the information data fusion of the highly correlated external information data descriptions includes clustering the relationships between the highly correlated external information data into absorption-type relationships, offsetting relationships, corrective relationships and illogical relationships, and using the external information data descriptions generated later to absorb, offset, correct or parallelly process the external information data descriptions generated earlier according to the time relationship, so as to eliminate the redundancy of the external information data descriptions and reduce errors.
[0022] In addition, the method further includes: searching for corresponding indicators according to the description of the external information data, and establishing a relationship between the external information data and the indicators.
[0023] B2-3, a series of preprocessing is performed on the external information data and the indicators to obtain the preprocessed indicator data, and the preprocessing indicator data is standardized to obtain the standardized indicator data, wherein the preprocessing includes a missing value processing flow and an abnormal value processing flow, wherein:
[0024] The missing value processing flow includes: for the case where the corresponding indicator is queried from the external information data description but the metadata is missing, one or more of the following methods are adopted: directly deleting the indicator with the missing value; filling the missing value according to the similarity between the indicators; filling the missing value according to the correlation between the indicators;
[0025] The outlier processing flow includes: outlier detection, outlier screening and outlier post-processing, wherein the outlier post-processing includes one or more of deleting outliers, taking default values, and correcting outliers using average values.
[0026] In addition, the credit assessment model under multiple scenarios established by selecting the required indicators from the indicator library and combining the indicators into a model relational expression through operators includes:
[0027] C1. According to the type of farmers, determine the model relationship corresponding to the type of farmers;
[0028] C2, extracting the set of indicators in the model relation through regular expressions;
[0029] C3. Divide the indicator set into multiple groups, obtain the metadata value corresponding to each indicator in each group of indicators through the thread pool at the same time, and assemble all the metadata values into a table;
[0030] C4. Input the model relationship and metadata value group table into the general calculation framework, and finally output the credit limit as the credit assessment model.
[0031] In addition, based on the results of the credit assessment model, the credit products that meet the conditions are matched for farmers according to the credit product routing matching rules, including:
[0032] According to the credit product matching routing rules, the financing application request and the predetermined credit product information are matched with the credit product, the real-name authentication information of the farmer's financing application request and the farmer's credit assessment model are pushed to the automatic acceptance interface of the corresponding financial institution channel and the credit application is initiated, and the credit application processing result returned by the automatic acceptance interface of the financial institution channel is received.
[0033] In addition, the credit product matching routing rules include general routing and special routing, where:
[0034] The general routing includes: determining financial institution channels that can provide credit products, financial institution channels that meet the current farmer type, and financial institution channels whose assessment amounts of the credit assessment model are within the limit;
[0035] The special routing includes: firstly determining whether there is a special routing rule; if there is a special routing rule, firstly determining whether there is a financial institution channel that meets the conditions according to the special routing rule, and if so, selecting the financial institution channel according to the priority; if there is no financial institution channel that meets the conditions in the special routing rule, then adopting the general routing to select the financial institution channel;
[0036] If there are no qualified financial institution channels under both general routing and special routing rules, it is determined that there are no available financial institution channels. When there are no available financial institution channels, the status of the financing application request is adjusted to application processing, and the push message is transferred to manual processing.
[0037] The present invention also includes a credit evaluation and credit application system based on information data, the system includes an authentication and authorization module, a data acquisition module, a data storage module, a data indicator selection module, a credit evaluation module and a credit application module, wherein:
[0038] The authentication and authorization module is the system's interface to farmers, used to receive financing application requests initiated by farmers and obtain corresponding real-name authentication and data authorization information;
[0039] The data acquisition module and the data storage module are used to collect and store external information data returned by the big data platform;
[0040] The data indicator selection module is used to analyze the external information data, specifically including: according to the description of the external information data, querying the corresponding indicator, and establishing a relationship between the external information data and the indicator;
[0041] The credit assessment module is used to select required indicators from the indicator library, combine the indicators into model relations through operators, and establish credit assessment models under multiple scenarios;
[0042] The credit application module is the system's acceptance interface for farmers and financial institution channels. It is used to match eligible credit products for farmers based on the results of the credit assessment model and in accordance with the credit product routing matching rules, initiate credit applications to the acceptance interface of the corresponding financial institution channel, receive the credit application processing results returned by the automatic acceptance interface of the financial institution channel, and feedback the credit application processing results to the farmers.
[0043] The present invention combines the actual use scenarios or types of farmers, supports the extraction of indicators and model calculations of information data such as multi-party transactions and management through technical means, and assists financial institutions in completing pre-loan credit assessments and credit line recommendations. It can objectively and reasonably obtain farmers' credit scores and credit ratings and conduct credit line assessments, and ultimately achieve intelligent matching of credit products by the system and initiate credit applications, which can better meet farmers' loan needs and improve user experience, effectively solving the problem of financing difficulties in the agricultural industry. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] Figure 1 The figure is a simplified structural diagram of a credit evaluation and credit application system based on information data according to a specific embodiment of the present invention.
[0045] Figure 2 The figure is a flow chart of a credit assessment and credit application method based on information data according to a specific embodiment of the present invention.
[0046] Figure 3 The present invention is a partial structural diagram of a credit assessment and credit application system based on information data according to a specific embodiment of the present invention.
[0047] Figure 4 The present invention is a partial flowchart of a credit assessment and credit application method based on information data according to a specific embodiment of the present invention. DETAILED DESCRIPTION
[0048] The present invention will be described in detail below in conjunction with the accompanying drawings.
[0049] Detailed exemplary embodiments are disclosed below. However, the specific structural and functional details disclosed herein are only for the purpose of describing exemplary embodiments.
[0050] However, it should be understood that the present invention is not limited to the specific exemplary embodiments disclosed, but covers all modifications, equivalents and alternatives falling within the scope of the present disclosure. In the description of all the figures, the same reference numerals represent the same elements.
[0051] With reference to the drawings, the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the present invention, so they have no substantial technical significance. Any structural modification, change in proportional relationship, or adjustment of size, without affecting the effects and purposes that can be achieved by the present invention, should still fall within the scope of the technical contents disclosed by the present invention. At the same time, the position limiting terms quoted in this specification are only for the convenience of description, and are not used to limit the scope of the implementation of the present invention. The change or adjustment of the relative relationship, without substantially changing the technical contents, should also be regarded as the scope of the implementation of the present invention.
[0052] It should also be understood that the term "and / or" as used herein includes any and all combinations of one or more of the associated listed items. It should also be understood that when a component or unit is referred to as being "connected" or "coupled" to another component or unit, it can be directly connected or coupled to the other component or unit, or there can be intermediate components or units. In addition, other words used to describe the relationship between components or units should be understood in the same manner (e.g., "between" versus "directly between," "adjacent" versus "directly adjacent," etc.).
[0053] Figure 2 The figure is a flow chart of a credit assessment and credit application method based on information data included in a specific embodiment of the present invention. As shown in the figure, the method of the specific embodiment of the present invention includes the steps of:
[0054] A. Receive the financing application request initiated by the farmer and obtain the corresponding real-name authentication and data authorization information;
[0055] B. Collect and store external information data returned by the big data platform according to the data authorization information, and parse the external information data;
[0056] C. Select the required indicators from the indicator library, combine the indicators into model relationships through operators, and establish a credit assessment model for multiple scenarios;
[0057] D. Based on the results of the credit assessment model, eligible credit products are matched to farmers according to the credit product routing matching rules.
[0058] The data authorization information includes authorization for use of transaction and management data, such as password information, login password, verification information, etc. In a specific embodiment of the present invention, the credit assessment and credit application system based on information data first receives the financing application request of the farmer, processes the real-name authentication according to the application process to obtain authentication information, and then obtains data authorization information according to the real-name authentication. The real-name authentication information includes verification of the identity information of the individual or the enterprise, and the data authorization information also includes the return value for use of transaction and management data.
[0059] Specifically, the above-mentioned farmers can be individuals or agriculture-related enterprises applying for financing. When an applicant applies for financing from a financial institution, the applicant needs to provide identity proof and other certificate data, the applicant's mobile phone number, bank card account information and sign a credit investigation authorization letter, and the above data information provided by the applicant will be authenticated by real name and authorized to use the data information.
[0060] Among them, the credit assessment model establishes a farmer credit limit assessment model in the multi-dimensional and multi-scenario system, including credit score and credit limit. The multi-dimensional system includes basic information, stability, performance credit, asset evaluation, and behavioral preference. The multi-scenario system includes e-commerce trading platform, breeding management platform, purchase and sales platform, etc. The credit assessment model constructs a scoring card model based on the farmer data under the trading and management platform. The score of the scoring interval or type in which the indicator variable value is located is obtained by performing logical operations on them to obtain the final total score of the farmer in the constructed scoring card, which is the user credit assessment model. The credit level is divided according to the score, and then the final credit limit is obtained according to the credit rating and credit limit calculation model.
[0061] Specifically, the credit assessment model built by the above-mentioned multi-dimensional system combined with the multi-scenario system gives an example of the dimensions in the e-commerce transaction platform scenario, including basic information, stability (including transaction duration, etc.), performance credit (including transaction payment usage, etc.), asset evaluation (including account balance usage, etc.), and behavioral preferences (including transaction amount, etc.). The dimensions in the breeding management platform scenario include basic information, stability (including platform usage, etc.), performance credit (including financial data such as accounts receivable), asset evaluation (including financial data such as bank accounts), and behavioral preferences (including breeding production data, etc.). The data indicators selected from the actual use scenarios of farmers are substituted into the model relationship and calculated to obtain a credit assessment model including credit score, credit rating and credit limit.
[0062] The collection and storage collects and stores the external information data returned by the big data platform, including initiating the acquisition of basic data, transaction and management data authorized by farmers from the big data platform, including the internal data platform and the third-party platform system, and receiving the returned data; the internal data platform mainly converts the data into the external information data description required for the indicator through the sql engine, and the data format from the third-party platform system is often not uniform, for example, data is obtained by real-time requesting the third-party platform system interface, and the returned data needs to be parsed. In the present invention, whether it is a more standardized internal data platform (existing data) or from a third-party platform system, it needs to be taken into consideration, but the difficulty of the processing flow is different. Therefore, for the credit evaluation and credit application system based on information data, it can be regarded as external information data.
[0063] In addition, the external information data includes basic data, transaction records, management data, execution information, debt status, dispute information, etc. The data acquisition module obtains a large amount of external information data such as authorized basic data, transaction and management data, and then sends it to the data storage module to store the returned data, and at the same time sends it to the data indicator selection module. Among them, the existing data from the internal data platform is converted into standardized data required for the indicator through the SQL engine. The specific steps are as follows:
[0064] 1. Write SQL statements based on indicators;
[0065] 2. Use the SQL engine to parse SQL statements;
[0066] 3. Generate information data description corresponding to the indicator.
[0067] In addition, for external information data from a third-party platform system, in a specific embodiment of the present invention, the external information data returned by the big data platform is collected and stored according to the data authorization information, and the external information data is parsed, including:
[0068] B1. Request the external data platform interface to obtain external information data;
[0069] B2. According to the description of the external information data, query the corresponding indicator and establish a relationship between the external information data and the indicator;
[0070] B3. Store the external information data that establishes a relationship with the indicators into the database together with the indicators.
[0071] Specifically, parsing external information data of the third-party platform system may include the following steps.
[0072] 1. Request to connect to the external interface of the third-party platform system and obtain external data (data in JSON or XML format); the third-party data platform can be a distributed data storage system, such as a cloud storage data system, etc. The access and connection to the external interface of the third-party platform system is performed by computer automatic capture or push acceptance.
[0073] 2. Extract external information data through FastJson and dom4j; in the big data system, external information data of the same layer are processed together. Here, the data of the same layer are processed together to meet the convenience of unified parsing of external information data.
[0074] 3. Circularly read the specific external information data in each layer. This step may further include the following sub-steps.
[0075] 3.1 Obtaining external information data description; the external information data description can be a brief record of the external information data and a summary of the core content by the third-party platform system. However, in some loose third-party platform systems, the external information data description may be lacking. Therefore, the present invention can skip this step and directly search for the full text of the external information data in subsequent steps. Therefore, when the external information data description is not available, the entire external information data can also be regarded as the external data information description. This is only a difference in workload, but not much difference for the overall solution.
[0076] 3.2 According to the description of external information data, query the corresponding indicator (full word match: external data source name or external data description). If the indicator can be queried, establish a relationship between the external data and the current indicator.
[0077] 3.3 If the indicator is not found, the external data description is segmented; then the obtained segmented words are used to search the indicator system according to the keywords, and the indicators that hit multiple keywords are selected (while saving the relationship between the indicator and the external data description), and the indicator is returned; a description of the external information data or a corresponding relationship between the external information data and the current indicator is established; then step 3.1 is repeated.
[0078] 3.4 If the indicator is not obtained in the end, the external data description is put into the exception pool for subsequent processing. The processing of the exception pool can be, for example, discarding, or performing a missing value processing process, or pushing it to manual processing.
[0079] 3.5 The external information data corresponding to the indicators, the description of the external information data and the keywords in the external data are stored in the database together with the indicators.
[0080] Figure 4This is a partial flow chart of a credit assessment and credit application method based on information data according to a specific embodiment of the present invention. As shown in the figure, in a specific embodiment of the present invention, the description of external information data, querying corresponding indicators, and establishing a relationship between the external information data and the indicators include:
[0081] B21. Searching for a corresponding indicator according to the description of the external information data. If the corresponding indicator can be found, establishing a relationship between the external information data and the indicator;
[0082] B22. If no indicator can be searched, the description of the external information data or the full text of the external information data is segmented, and the obtained segmented words are used to search the indicator system according to the keywords, and indicators that hit multiple keywords are selected, and the indicators and keywords and metadata corresponding to the keywords are returned to establish a corresponding relationship between the external information data and the indicators;
[0083] B23. If the indicator is not obtained in the end, the external information data will be placed in the exception pool for exception handling.
[0084] The word segmentation method, for example, uses the HMM (hidden Markov chain) method to segment Chinese words, divides the external information data description into sentence-initial words, sentence-midwords, sentence-final words, and independent words, and then decomposes the external information data description or the entire external information data into individual words based on the logical relationship among the sentence-initial words, sentence-midwords, sentence-final words, and independent words.
[0085] Then, the search is performed based on the correspondence between the indicators and the keywords. For example, judicial risk indicators generally include special keywords such as "court", "ruling", "execution", "request", and "compensation". Based on these special keywords, the relationship between the external information data description and the judicial risk indicator can be established. After that, some common keywords can be obtained to obtain the metadata corresponding to the keywords, such as the specific compensation amount, execution date, etc.
[0086] In addition, in a specific embodiment of the present invention, the step of storing the external information data that establishes a relationship with the indicator together with the indicator in the database includes:
[0087] B301. Perform keyword comparison on multiple external information data descriptions that establish relationships with the same indicator to obtain external information data descriptions with high relevance;
[0088] B302. For highly relevant external information data descriptions, perform logic detection according to metadata corresponding to keywords, and perform information data fusion on the highly relevant external information data descriptions.
[0089] An important improvement of the present invention is to improve the accuracy of the credit assessment model, so it is very important to fuse the information data of the highly relevant external information data description; and this is also one of the main advantages of the present invention over the prior art.
[0090] The so-called information data fusion mainly comes from external information data from different sources, which may have a large number of redundant parts. For example, an execution information, compensation information or debt repayment information may be obtained from different sources as external data sources, or an execution information, compensation information or debt repayment information may have different intermediate stages, and there may be certain logical relationships between the intermediate stages. Taking judicial compensation information as an example, the first instance and the second instance of the case, or even the retrial and subsequent judicial execution, may result in a relationship of mutual absorption, correction, or offset. If the redundancy between these external information data from different information sources and the logical relationship between these external information data are not considered, the credit assessment results of a farmer may seriously deviate from the actual situation. Therefore, the present invention strives to improve this point.
[0091] In a specific implementation of the present invention, keyword comparison is first performed on multiple external information data descriptions that establish a relationship with the same indicator to obtain highly relevant external information data descriptions. For example, data from different third-party platform systems, if they focus on the same information, may have a large number of identical or similar keywords; and data from different stages of the same external information data may also be closely related due to some numbers, party information, and case information. Therefore, keyword comparison is first performed in the present invention. These keywords can be personnel keywords, date keywords, coding keywords, etc. Through keyword comparison, multiple highly relevant external information data descriptions can be obtained.
[0092] Next, the present invention performs logic detection on highly relevant external information data descriptions according to metadata corresponding to keywords, and performs information data fusion on the highly relevant external information data descriptions. The information data fusion refers to performing logic detection and calculation based on the read keywords and corresponding metadata to eliminate redundant external information data.
[0093] Therefore, in a specific embodiment of the present invention, information data fusion of the highly correlated external information data descriptions includes clustering the relationships between the highly correlated external information data into absorption-type relationships, offsetting relationships, corrective relationships and illogical relationships, and according to the time-related relationships, using the external information data descriptions generated later to absorb, offset, correct or parallelly process the external information data descriptions generated earlier, so as to eliminate the redundancy of the external information data descriptions and reduce errors.
[0094] For example, after searching, two highly related external information data descriptions are found. In the first external information data description, the information obtained by searching keywords and related metadata is that a debt dispute arises between A and B, and after a mediation agreement, A pays RMB 100,000 to B. In the second external information data description, the information obtained by searching keywords and related metadata is that B applied to the court to enforce the aforementioned mediation agreement of RMB 100,000, and the execution was completed. Therefore, the logical relationship between these two external information data descriptions is an absorption relationship. The second external information data description absorbs the first external information data description, and the two external information descriptions cannot be accumulated.
[0095] Let's take another example to illustrate. After searching, two highly related external information data descriptions are found. The first external information data description is that the court ruled in the first instance that A should compensate B RMB 100,000; the second external information data description is that the court ruled in the second instance to revoke the original first instance judgment. Therefore, the logical relationship between the two external information data descriptions is an offsetting relationship. The second external information data description offsets the first external information data description. Both external information data descriptions can be excluded from the calculation scope of the credit assessment model.
[0096] Let's take another example to illustrate. After searching, two highly related external information data descriptions are found. The first external information data description is that A pays off a debt of 100,000 yuan to B; in the second external information data description, A applies for execution to reverse 50,000 yuan. Therefore, the logical relationship between the two external information data descriptions is a correction relationship, so both external information data descriptions need to be included in the calculation scope of the credit assessment model.
[0097] From the above examples, it can be seen that a prominent improvement of the present invention is that by eliminating the redundant parts of external information data from different sources, for example, an execution information, compensation information or debt repayment information may be obtained from different sources as external data sources, or there may be different intermediate stages for an execution information, compensation information or debt repayment information, and there may be certain logical relationship calculations between the intermediate stages to obtain an accurate credit assessment model. This correction is particularly important for the current information explosion stage. For example, the same external information data may be forwarded or generated tens of thousands of times, or different stages between the same facts may also lead to very different results. When using a computer to directly capture external information data from a big data network, this large amount of information redundancy and the logical relationship between information will lead to seriously distorted evaluation results if not processed, and may even be completely meaningless to financial institutions.
[0098] In addition, in a specific implementation of the present invention, the querying of the corresponding indicator according to the description of the external information data and establishing a relationship between the external information data and the indicator further includes:
[0099] B2-3, a series of preprocessing is performed on the external information data and the indicators to obtain the preprocessed indicator data, and the preprocessing indicator data is standardized to obtain the standardized indicator data, wherein the preprocessing includes a missing value processing flow and an abnormal value processing flow, wherein:
[0100] The missing value processing flow includes: for the case where the corresponding indicator is queried from the external information data description but the metadata is missing, one or more of the following methods are adopted: directly deleting the indicator with the missing value; filling the missing value according to the similarity between the indicators; filling the missing value according to the correlation between the indicators;
[0101] The outlier processing flow includes: outlier detection, outlier screening and outlier post-processing, wherein the outlier post-processing includes one or more of deleting outliers, taking default values, and correcting outliers using average values.
[0102] Specifically, the data preprocessing process includes a missing value processing process and an outlier processing process for the index data to be extracted, which is mainly to convert the obtained original information data into formatted information data that can be used for model relationship calculation. Missing value processing is first performed before the credit assessment model is calculated. The methods for missing value processing include the following: 1. Directly delete the index containing missing values and its corresponding external information data, which is the simplest way. 2. Fill the missing values according to the similarity between the index items. 3. Fill the missing values according to the correlation between the index items. The second step is to process outliers, and the outlier processing includes the steps of outlier detection, outlier screening, and outlier processing. The processing of outliers includes one or more of deleting outliers, taking default values, correcting outliers with average values, or not processing them. For example, if the number of transactions of a user in a certain month is 0, but the transaction amount of that month is not 0, the transaction value is judged to be an outlier through association analysis. Then the outlier can be corrected by selecting the mean value based on the data of the user's index before and after the month.
[0103] Therefore, compared with the prior art, the present invention has the robustness to accommodate missing data or obviously erroneous data and is highly intelligent.
[0104] In addition, the standardization of the data is often used in the processing of certain comparison and evaluation indicators. It removes the unit restrictions of external information data and converts it into dimensionless pure numerical values to facilitate comparison and weighting of indicator data of different units or magnitudes, which is required for model relationship calculations.
[0105] In addition, in a specific embodiment of the present invention, the steps of selecting the required indicators from the indicator library, combining the indicators into a model relation through operators, and establishing a credit assessment model under multiple scenarios include:
[0106] C1. According to the type of farmers, determine the model relationship corresponding to the type of farmers;
[0107] C2, extracting the set of indicators in the model relation through regular expressions;
[0108] C3. Divide the indicator set into multiple groups, obtain the metadata value corresponding to each indicator in each group of indicators through the thread pool at the same time, and assemble all the metadata values into a table;
[0109] C4. Input the model relationship and metadata value group table into the general calculation framework, and finally output the credit limit as the credit assessment model.
[0110] For example, the indicator library is:
[0111]
[0112] The credit assessment and credit application system based on information data in the specific implementation of the present invention will automatically select indicators and combine formulas according to different customer types (the customer type is taken from the corresponding real-name authentication and data authorization information obtained in the financing application request); for example: 30 types such as growers, pig feed dealers, agricultural supplies dealers, and fat pig farming; if the customer type is a grower, the system automatically selects plantArea, AssetTotal, DebtTotal and other indicators for combination calculation as the model relationship.
[0113] For example, a model relational expression related to a basic quota is MAX(MIN(indicator 1*weight A, (indicator 2*weight B-indicator 3*weight C)), MIN(indicator 4*weight D, indicator 5*weight E)). First, the set of indicators in the model relational expression is extracted through a regular expression, and then the data is read from the database. For example, the indicator set is divided into multiple groups, and the metadata value corresponding to each indicator in each group of indicators is obtained simultaneously through the thread pool, and all the metadata values are assembled into a table. The metadata value is stored synchronously with the indicator and external information data in the aforementioned step B.
[0114] In addition, in a specific implementation of the present invention, the matching of qualified credit products for farmers based on the results of the credit assessment model and the credit product routing matching rules includes:
[0115] According to the credit product matching routing rules, the financing application request and the predetermined credit product information are matched with the credit product, the real-name authentication information of the farmer's financing application request and the farmer's credit assessment model are pushed to the automatic acceptance interface of the corresponding financial institution channel and the credit application is initiated, and the credit application processing result returned by the automatic acceptance interface of the financial institution channel is received.
[0116] In addition, in a specific implementation of the present invention, the credit product matching routing rules include general routing and special routing, wherein:
[0117] The general routing includes: determining financial institution channels that can provide credit products, financial institution channels that meet the current farmer type, and financial institution channels whose assessment amounts of the credit assessment model are within the limit;
[0118] The special routing includes: firstly determining whether there is a special routing rule; if there is a special routing rule, firstly determining whether there is a financial institution channel that meets the conditions according to the special routing rule, and if so, selecting the financial institution channel according to the priority; if there is no financial institution channel that meets the conditions in the special routing rule, then adopting the general routing to select the financial institution channel;
[0119] If there are no qualified financial institution channels under both general routing and special routing rules, it is determined that there are no available financial institution channels. When there are no available financial institution channels, the status of the financing application request is adjusted to application processing, and the push message is transferred to manual processing.
[0120] Specifically, in a specific embodiment of the present invention, after receiving the financing application request initiated by the farmer, the authentication and authorization information is extracted to verify the authentication and authorization information. After receiving the credit assessment model result sent by the credit assessment module, according to the credit situation included in the credit assessment model result, and according to the credit product matching routing rules, the financing application request and the credit product information matching credit product, the farmer information and the farmer credit assessment result are sent to the interface of the corresponding financial institution channel and the credit application is initiated, and the credit application result returned by the financial institution is received.
[0121] The credit product matching routing rules are mainly divided into: 1. General routing, that is, the rules for most routing; 2. Special routing, individual special routing rules. The two ends of the routing are the platform side and the channel side. The platform side belongs to the input end of the routing, which mainly transmits valid information; and the channel side is the output end, as long as the financial institution channels that can be screened for credit products are determined. The channels that initially meet the current conditions are judged according to priority, and can provide credit products; financial institution channels that support the current customer type; financial institution channels whose credit assessment amount is within the limit, etc. After filtering through the above conditions, the selected financial institution channels all meet the requirements, so it is necessary to set the priority weight in the existing data: only one channel will be screened out through the above two conditions, and this channel is the best channel.
[0122] When encountering unexpected situations, not all financing applications are applicable to general routing rules. In the actual operation process, special routing rules can be configured separately for certain scenarios or situations. Applicable scenarios: If the eligible financial institution channels are A, C and D, under normal circumstances, the system will automatically give priority to financial institution channel A according to the general routing rules, because this financial institution channel is more satisfied with the quota and credit assessment results; however, for the use of farmers, since the business outlets of financial institution channel A cannot cover the area where the farmer is located, in order to meet the user experience as the highest rule, although financial institution channel A has been met, financial institution channel C is still selected instead of A for the convenience of farmers. For this situation, the system configures special routing rules based on the farmer's area and the channel service area. The overall judgment logic is as follows: 1. Prioritize whether there is a special route; 2. If there is a special rule, first determine whether there is a channel that meets the conditions in the rule, and if so, take the channel with the highest priority; 3. If there is no available channel for the special route, determine whether general routing is required; if necessary, use the general route to select the channel, otherwise it is determined that there is no available channel. 4. If there is no available channel, the application will become an application in process, and the system will push a message to inform the platform to switch to manual processing.
[0123] Figure 1 A structural diagram of a credit evaluation and credit application system based on information data according to a specific embodiment of the present invention; Figure 3 This is a partial structural diagram of a credit evaluation and credit application system based on information data according to a specific embodiment of the present invention. As shown in the figure, the specific embodiment of the present invention also includes a credit evaluation and credit application system based on information data, the system includes an authentication and authorization module, a data acquisition module, a data storage module, a data indicator selection module, a credit evaluation module and a credit application module, wherein:
[0124] The authentication and authorization module is the system's interface to farmers, used to receive financing application requests initiated by farmers and obtain corresponding real-name authentication and data authorization information;
[0125] The data acquisition module and the data storage module are used to collect and store external information data returned by the big data platform;
[0126] The data indicator selection module is used to analyze the external information data, specifically including: according to the description of the external information data, querying the corresponding indicator, and establishing a relationship between the external information data and the indicator;
[0127] The credit assessment module is used to select required indicators from the indicator library, combine the indicators into model relations through operators, and establish credit assessment models under multiple scenarios;
[0128] The credit application module is the system's interface between farmers and financial institutions. It is used to match eligible credit products for farmers based on the results of the credit assessment model and in accordance with the credit product routing matching rules, initiate credit applications to the automatic acceptance interface of the corresponding financial institution channel, receive the credit application processing results returned by the automatic acceptance interface of the financial institution channel, and feedback the credit application processing results to the farmers.
[0129] The credit assessment and credit application system based on information data can be implemented in the form of a network server, such as a centralized data processing system or a distributed data processing system.
[0130] Specifically, the credit application matching credit products can be based on the credit assessment model and the information in the loan request sent by the farmer. The matching system has preset credit product routing rules, including the application requirements or conditions of the credit products of the financial institution and the requirements for docking transaction and management data, and the strategies adopted in combination with the actual use scenarios and risk points of the farmers. For example: Financial institutions such as Rural Credit Microfinance follow the requirements of the breeding scenario based on the farmers' breeding transactions or management data and corresponding risk points. When the credit assessment results of the farmers meet the credit requirements of the financial institutions, the system will send an application for a breeding loan credit product to the Rural Credit Microfinance Financial Institution, and after receiving the credit application results returned by Rural Credit Microfinance, it will be fed back to the farmers.
[0131] The present invention combines the actual use scenarios or types of farmers, supports the extraction of indicators and model calculations of information data such as multi-party transactions and management through technical means, and assists financial institutions in completing pre-loan credit assessments and credit line recommendations. It can objectively and reasonably obtain farmers' credit scores and credit ratings and conduct credit line assessments, and ultimately achieve intelligent matching of credit products by the system and initiate credit applications, which can better meet farmers' loan needs and improve user experience, effectively solving the problem of financing difficulties in the agricultural industry.
[0132] The above description shows and describes several preferred embodiments of the present invention. It should be understood that the present invention is not limited to the form disclosed in this specification, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the invention concept described in this specification through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art do not deviate from the spirit and scope of the present invention, and should be within the scope of protection of the claims attached to the present invention.
Claims
1. A credit assessment and credit application method based on information data, characterized in that: The method comprises the steps of: A. Receive the financing application request initiated by the farmer and obtain the corresponding real-name authentication and data authorization information; B. Collect and store external information data returned by the big data platform according to the data authorization information, and parse the external information data; C. Select the required indicators from the indicator library, combine the indicators into model relationships through operators, and establish a credit assessment model for multiple scenarios; D. Based on the credit assessment model, matching qualified credit products for farmers according to the credit product routing matching rules; Wherein, collecting and storing external information data returned by the big data platform according to the data authorization information, and parsing the external information data includes: B1. Request the external data platform interface to obtain external information data; B2. According to the description of the external information data, query the corresponding indicator and establish a relationship between the external information data and the indicator; B3. Store the external information data related to the indicators into the database together with the indicators; The step of storing the external information data that establishes a relationship with the indicator together with the indicator into the database includes: B301. Perform keyword comparison on multiple external information data descriptions that establish relationships with the same indicator to obtain external information data descriptions with high relevance; B302. For highly relevant external information data descriptions, perform logic detection according to metadata corresponding to keywords, and perform information data fusion on the highly relevant external information data descriptions; The information data fusion of the highly correlated external information data descriptions includes clustering the relationships between the highly correlated external information data into absorption relationships, offset relationships, correction relationships and illogical relationships, and absorbing, offsetting, correcting or parallelizing the external information data descriptions generated earlier according to the temporal relationship, so as to eliminate the redundancy of the external information data descriptions and reduce errors; The method of selecting the required indicators from the indicator library, combining the indicators into a model relational expression through operators and establishing a credit assessment model under multiple scenarios includes: C1. According to the type of farmers, determine the model relationship corresponding to the type of farmers; C2, extracting the set of indicators in the model relation through regular expressions; C3. Divide the indicator set into multiple groups, obtain the metadata value corresponding to each indicator in each group of indicators through the thread pool at the same time, and assemble all the metadata values into a table; C4, input the model relational expression and metadata value group table into the general calculation framework, and finally output the credit limit as the credit assessment model; Based on the results of the credit assessment model, the credit products that meet the conditions are matched to farmers according to the credit product routing matching rules, including: According to the credit product matching routing rules, the financing application request and the predetermined credit product information are matched with the credit product, the real-name authentication information of the farmer's financing application request and the farmer's credit assessment model are pushed to the automatic acceptance interface of the corresponding financial institution channel and the credit application is initiated, and the credit application processing result returned by the automatic acceptance interface of the financial institution channel is received.
2. The credit assessment and credit application method based on information data according to claim 1, characterized in that: The querying of the corresponding indicator according to the description of the external information data and establishing a relationship between the external information data and the indicator comprises: B21. Searching for a corresponding indicator according to the description of the external information data. If the corresponding indicator can be found, establishing a relationship between the external information data and the indicator; B22. If no indicator can be searched, the external information data description is segmented, and the obtained segmented words are used to search the indicator system according to the keywords, and indicators that hit multiple keywords are selected, and the indicators and keywords and metadata corresponding to the keywords are returned to establish a corresponding relationship between the external information data and the indicators; B23. If the indicator is not obtained in the end, the external information data will be placed in the exception pool for exception handling.
3. The credit assessment and credit application method based on information data according to claim 1, characterized in that: After querying the corresponding indicator according to the description of the external information data and establishing a relationship between the external information data and the indicator, the method further includes: B2-3. Preprocess the external information data and indicators to obtain preprocessed indicator data, and standardize the preprocessed indicator data to obtain standardized indicator data, wherein the preprocessing includes a missing value processing flow and an outlier processing flow, wherein: The missing value processing flow includes: for the case where the corresponding indicator is queried from the external information data description but the metadata is missing, one or more of the following methods are adopted: directly deleting the indicator with the missing value; filling the missing value according to the similarity between the indicators; filling the missing value according to the correlation between the indicators; The outlier processing flow includes: outlier detection, outlier screening and outlier post-processing, wherein the outlier post-processing includes one or more of deleting outliers, taking default values, and correcting outliers using average values.
4. The credit assessment and credit application method based on information data according to claim 1, characterized in that: The credit product matching routing rules include general routing and special routing, where: The general routing includes: determining financial institution channels that can provide credit products, financial institution channels that meet the current farmer type, and financial institution channels whose assessment amounts of the credit assessment model are within the limit; The special routing includes: firstly determining whether there is a special routing rule; if there is a special routing rule, firstly determining whether there is a financial institution channel that meets the conditions according to the special routing rule, and if so, selecting the financial institution channel according to the priority; if there is no financial institution channel that meets the conditions in the special routing rule, then adopting the general routing to select the financial institution channel; If there are no qualified financial institution channels under both general routing and special routing rules, it is determined that there are no available financial institution channels. When there are no available financial institution channels, the status of the financing application request is adjusted to application processing, and the push message is transferred to manual processing.
5. A credit evaluation and credit application system based on information data using the credit evaluation and credit application method based on information data as described in claim 1, characterized in that: The system includes an authentication and authorization module, a data collection module, a data storage module, a data indicator selection module, a credit assessment module and a credit application module, wherein: The authentication and authorization module is the system's interface to farmers, used to receive financing application requests initiated by farmers and obtain corresponding real-name authentication and data authorization information; The data acquisition module and the data storage module are used to collect and store external information data returned by the big data platform; The data indicator selection module is used to analyze the external information data, specifically including: according to the description of the external information data, querying the corresponding indicator, and establishing a relationship between the external information data and the indicator; The credit assessment module is used to select required indicators from the indicator library, combine the indicators into model relations through operators, and establish credit assessment models under multiple scenarios; The credit application module is the system's automatic acceptance interface for farmers and financial institution channels. It is used to match eligible credit products for farmers based on the results of the credit assessment model and in accordance with the credit product routing matching rules, initiate a credit application to the automatic acceptance interface of the corresponding financial institution channel, receive the credit application processing results returned from the automatic acceptance interface of the financial institution channel, and feedback the credit application processing results to the farmers.
