Component modularized financial product design information classification method and system

Abstract

The invention provides a component modularized financial product design information classification method and system, and belongs to the field of information classification. The problem of complicated information classification is solved; the method specifically comprises the following steps: S1, obtaining classification information and financial product design information; s2, acquiring financial product information, a product set and a classification set; performing word frequency analysis on the financial product information in the product set to obtain a product keyword list, and performing screening in combination with the classification set to obtain a product feature word list; s3, analyzing the financial product design information through the product feature word list; obtaining conformance values of the financial product design information and the product feature words to obtain a conformance value list; s4, judging the category of the financial product design information according to the conformity value list; the financial product design information is classified by analyzing the financial product information, so that the classification difficulty of the financial product design information is reduced.

Description

Technical Field

[0001] The present invention provides a component modular financial product design information classification method and system, relating to the field of information classification. Background Art

[0002] The existing classification methods and systems for financial product design information have the following deficiencies: Ambiguous or inconsistent classification standards: Different financial institutions or departments may adopt different classification standards, resulting in the same type of information being classified differently in different systems, increasing the difficulty of information integration and sharing. For example, some institutions may classify products based on their investment period, while others may classify them based on their risk level. Single classification dimension: Some classification methods only consider a single dimension, such as classification by product type or customer group, ignoring other important factors that may affect information classification. This may lead to incomplete information classification and fail to meet diverse information query and analysis needs; Unclear classification hierarchy: The classification hierarchy is not set reasonably, resulting in information classification being too complex or too simple, and unable to effectively reflect the hierarchical relationship of information. Summary of the Invention

[0003] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a component modular financial product design information classification method and system, aiming to solve the problem of complex information classification.

[0004] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for classifying component modular financial product design information, the classification method comprising: Step S1: Obtain classification information and financial product design information; Step S2: Based on the classification information, obtain financial product information, product sets, and classification sets; perform word frequency analysis on the financial product information in the product set to obtain a product keyword list; combine the classification set to filter the product keyword list, extract product feature words, and obtain a product feature word list; Step S3: Analyze the financial product design information using the product feature word list; obtain the matching values ​​between the financial product design information and the product feature words based on the frequency of occurrence of the product feature words in the financial product design information, and calculate the matching values ​​to obtain a matching value list; Step S4: Determine the category of the financial product design information based on the matching value list, and obtain the category of the financial product design information by combining the category set in the classification information.

[0005] Furthermore, the specific steps of step S1 are as follows: Step S11: Obtain financial product information jr, treat the financial product information of the same category as a product set, denoted as LB, obtain the number of elements t in the product set, and obtain LB={jr(1), jr(2), ..., jr(t)}; Step S12: Integrate the product sets of the same classification dimension to form a classification set FL, and record the number of product sets n; obtain FL={LB(1), LB(2), ...LB(n)}; Step S13: Obtain the classification dimension m; perform statistics on the classification set according to the classification dimension to form classification information, which is recorded as FX, FX={FL(1), FL(2), ..., FL(m)}.

[0006] Furthermore, the specific steps of step S2 are as follows: Step S21: Obtaining financial product information within the product set through the product set, performing word frequency analysis on the financial product information to obtain a word frequency list of the financial product information, and processing the word frequency list of the financial product information within the product set to obtain a product keyword list; Step S22: Filter the product keyword list according to the classification set, obtain the number n of product sets in the classification set, obtain the product keyword list corresponding to the product set, and record it as cgj (1) to cgj (n); compare the product keyword list and remove repeated product keywords, as follows: Get the number of product keywords cgs of cgj (1); extract the product keywords of cgj (1), and record them as ccj (1) to ccj (cgs); Search cgj(2), ..., cgj(n) with ccj(1) to ccj(cgs) respectively; if ccj(1) to ccj(cgs) do not exist in cgj(2), ..., cgj(n), then determine that they are product feature words, count the product feature words, and obtain a product feature word list; Obtain the product feature word list from cgj(2) to cgj(n).

[0007] Furthermore, the specific steps of step S21 are as follows: Step S211: performing text preprocessing on the financial product information by removing emoticons, uniform resource locators, and punctuation marks to obtain an initial text; Step S212: Use Jieba to perform word segmentation on the initial text, split the text into words or phrases, obtain a word segmentation list, obtain a stop word list, filter the word segmentation list according to the stop word list, count the word frequencies in the filtered word segmentation list, and obtain a word frequency list; Step S213: Obtain a word frequency list of financial product information in the product set, analyze the word frequency list, and obtain a product keyword list.

[0008] Furthermore, the specific steps of step S211 are as follows: Remove emoticons: Get the computer code bq1 for the first emoticon, and bq2 for the last emoticon. Use regular expressions to delete characters in the range [bq1, bq2] from the financial product information. Remove Uniform Resource Locators: Traverse the financial product information and compare it with "http". If any character "http" is found, record the character position zf1. At the same time, find the position where the first space appears after the character, record it as zf2, and delete the statements in the financial product information where the characters are between zf1 and zf2. Remove punctuation marks: Save the punctuation marks in a list, use regular expressions to match the characters and punctuation marks in the financial product information, and delete them if they match.

[0009] Furthermore, the specific steps of step S212 are as follows: The specific steps of filtering the word list according to the stop word list are as follows: looping through the word list, obtaining the word, matching the word with the stop word list, and deleting the word if it can be retrieved in the stop word list; After the screening is completed, the number of occurrences of the segmentation words in the segmentation list is counted, as follows: the segmentation list is converted into a set to obtain a segmentation set, the segmentation elements in the segmentation set are traversed, the segmentation list is cyclically counted according to the segmentation elements to obtain the number of occurrences of the segmentation elements, and the segmentation elements and the number of occurrences of the segmentation elements form a word frequency list.

[0010] Furthermore, the specific steps of step S213 are as follows: Get the number t of financial product information in the product set and obtain the corresponding word frequency list, recorded as cp(c); Take the first word frequency list cp(1) as the standard list, obtain the number of elements ysg in the standard list, and record the elements in the standard list as bzl(1), bzl(2), ... bzl(ysg); perform cyclic judgment on cp(1), cp(2), ..., cp(t) according to bzl(1) to bzl(ysg); If bzl(1) always exists in cp(1) to cp(t), then bzl(1) is determined to be a product keyword; Make judgments on bzl (2) to bzl (ysg); The product keywords are counted according to the judgment result, and the product keywords are sorted in descending order based on the number of occurrences of elements in the word frequency list; thus, a product keyword list is obtained.

[0011] Furthermore, the specific steps of step S3 are as follows: Step S31: obtaining the number of feature words tgs in the product feature word list; recording the feature words in the product feature word list as tn, obtaining the ranking of the feature words in the product feature word list, and obtaining the ranking value px(tn); Step S32: cyclically judging the feature words in the financial product design information, recording the number of times the feature words appear in the financial product design information, and obtaining the frequency of occurrence pl(tn); combining the number of feature words tgs in the product feature word list and the ranking value px(tn) of the feature words to calculate the matching value xfz between the financial product design information and the product feature words; ; Step S33: Calculate the matching values ​​of the n product feature word lists and the financial product design information to obtain xfz(1) to xfz(n), analyze the matching values, and obtain a matching value list cfzl.

[0012] Furthermore, the specific steps of step S33 are as follows: Step S331: Obtain matching values ​​xfz(1) to xfz(n), accumulate the matching values ​​to obtain the matching total value xzz; calculate the matching judgment value glz(i) based on the proportional relationship between the matching values ​​and the matching total value and the variance of the matching values; ; xfz(i) represents the consistency value between the i-th product feature word list and the financial product design information; Step S332: Obtain the maximum value among the matching judgment values ​​to obtain the maximum judgment value max, and analyze and judge the maximum judgment value: If max>50%, the matching value calculation is valid, the matching values ​​are counted, and a matching value list cfzl is obtained; then step S4 is entered; If max≤50%, the matching value calculation is invalid and needs to be recalculated; go to step S333; Step S333: Update the product feature word list, mask the last digit of the product feature word list, obtain the frequency of occurrence of the feature word in the financial product design information pl (tn), the number of feature words in the product feature word list tgs-1, and the ranking value px (tn) of the feature word, calculate, obtain the updated matching value, and repeat steps S331-S332.

[0013] A modular financial product design information classification system, the classification system comprising: Information acquisition module: obtain classification information and financial product design information; Information processing module: Based on the classification information, obtain financial product information, product sets, and classification sets; perform word frequency analysis on the financial product information in the product set to obtain a product keyword list; combine the classification set to filter the product keyword list, extract product feature words, and obtain a product feature word list; Information analysis module: Analyzes financial product design information using a product feature word list; calculates the matching values ​​between the financial product design information and the product feature words based on the frequency of occurrence of the product feature words in the financial product design information, and calculates the matching values ​​to obtain a matching value list; Information classification module: judges the category of financial product design information based on the matching value list, and obtains the category of financial product design information by combining the classification set in the classification information.

[0014] Compared with the prior art, the present invention has the following beneficial effects: Clear classification standards: This invention conducts vertical analysis on similar products to obtain the common features of products in the same category and obtain product keywords. At the same time, vertical analysis is conducted on products in different categories, retaining the unique keywords of similar products and filtering out the common keywords to obtain classification standards. Wide classification dimensions: This invention analyzes classified products in different dimensions and obtains unique classification standards according to different classification dimensions, ensuring classification accuracy in different classification dimensions; Clear classification hierarchy: The present invention classifies financial product design information through a three-layer structure, namely, extracting classification standards according to products of the same category, screening classification standards according to products of different categories under the same classification dimension, and classifying financial product design information of different dimensions according to the classification dimension. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings: Figure 1 Schematic diagram of the method of the present invention; Figure 2 This is a schematic diagram of keyword processing in the present invention; Figure 3 Schematic diagram of feature word processing in the present invention Figure 4 Schematic diagram of information classification of the present invention Figure 5 Schematic diagram of the system of the present invention. DETAILED DESCRIPTION

[0016] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments. Example 1

[0017] See also Figure 1 , a component modular financial product design information classification method includes: Step S1: Obtain classification information and financial product design information; Step S11: Obtain financial product information jr, treat the financial product information of the same category as a product set, denoted as LB, obtain the number of elements t in the product set, and obtain LB={jr(1), jr(2), ..., jr(t)}; It should be noted that bonds, bank loans, financial leasing IPOs, fixed increases, and equity crowdfunding are all financing products. Here, LB represents the product set of financing products, jr (1), jr (2), ..., jr (t) represent bonds, bank loans, financial leasing IPOs, fixed increases, and equity crowdfunding; Step S12: Integrate the product sets of the same classification dimension to form a classification set FL, and record the number of product sets n; obtain FL={LB(1), LB(2), ...LB(n)}; It should be noted that: according to the functional dimension, they can be divided into financing products, investment products, and risk management products. Here, FL represents the classification set of functional dimensions, and LB (1), LB (2), ... LB (n) represent the collection of financing products, investment products, and risk management products. Step S13: Obtain the classification dimension m; perform statistics on the classification set according to the classification dimension to form classification information, which is recorded as FX, FX={FL(1), FL(2), ..., FL(m)}.

[0018] It should be noted that financial products can be classified according to multiple dimensions such as function, risk-return characteristics and trading venue. Among them, they can be further divided into financing products, investment products and risk management products according to function.

[0019] Step S2: Based on the classification information, obtain financial product information, product sets, and classification sets; perform word frequency analysis on the financial product information in the product set to obtain a product keyword list; combine the classification set to filter the product keyword list, extract product feature words, and obtain a product feature word list; Step S21: Obtaining financial product information within the product set through the product set, performing word frequency analysis on the financial product information to obtain a word frequency list of the financial product information, and processing the word frequency list of the financial product information within the product set to obtain a product keyword list; Step S211: performing text preprocessing on the financial product information by removing emoticons, uniform resource locators, and punctuation marks to obtain an initial text; Step S2111: Remove emoticons: Obtain the computer code bq1 for the first emoticon, obtain the computer code bq2 for the last emoticon, and delete the characters in the range [bq1, bq2] from the financial product information using a regular expression. Step S2112: Remove Uniform Resource Locators: The financial product information is traversed and compared with "http". If any character "http" is found, the character position zf1 is recorded. The first occurrence of a space after the character is found and recorded as zf2. The statements with character positions zf1 to zf2 in the financial product information are deleted. It should be noted that the Uniform Resource Locator is an identification method used to completely describe the address of web pages and other resources on the Internet, and has nothing to do with text content.

[0020] Step S2113: Remove punctuation marks: save the punctuation marks in a list, match the characters and punctuation marks in the financial product information using regular expressions, and delete them if they match; Step S212: Use Jieba to perform word segmentation on the initial text, split the text into words or phrases, obtain a word segmentation list, obtain a stop word list, filter the word segmentation list according to the stop word list, count the word frequencies in the filtered word segmentation list, and obtain a word frequency list; Step S2121: Filter the word list according to the stop word list as follows: loop through the word list to obtain the word, match the word with the stop word list, and delete the word if the word can be retrieved in the stop word list; Step S2122: After the screening is completed, the number of occurrences of the segmented words in the segmented word list is counted, specifically as follows: the segmented word list is converted into a set to obtain a segmented word set, the segmented word elements in the segmented word set are traversed, and the segmented word list is cyclically counted according to the segmented word elements to obtain the number of occurrences of the segmented word elements. The segmented word elements and the number of occurrences of the segmented word elements are used to form a word frequency list; It should be noted that the word frequency list format is as follows: [{word segmentation element 1: number of times 1}, {word segmentation element 2: number of times 2}, {word segmentation element 3: number of times 3}].

[0021] Note: The stopword list is a tool used in natural language processing (NLP) and information retrieval to filter out common words that contribute less to the semantics of the text. A common stopword list can be used, or a custom stopword list can be defined according to requirements.

[0022] import re import jieba # Original text text = "你好!欢迎访问 https: / / example.com。这是一个测试,包含数字123和标点!!!" # 1. Remove emoji text = re.sub(r'[\U00010000-\U0010ffff]', '', text) # 2. Remove URL text = re.sub(r'http\S+', '', text) # 3. Remove punctuation text = re.sub(r'[,。!?;:“”‘’()《》]', '', text) # 4. Remove numbers text = re.sub(r'\d+', '', text) # 5. Word segmentation and stopword processing stopwords = {"的", "了", "在"} words = jieba.lcut(text) filtered_words = [word for word in words if word not in stopwords] # Output result print(filtered_words) # Output: ['你好', '欢迎访问', '这是一个测试', '包含', '和标点'] Please refer to Figure 2 ; Step S213: Obtain the word frequency list of the financial product information in the product set, and analyze the word frequency list to obtain the product keyword list; Step S2131: Obtain the number t of the financial product information in the product set, and obtain the corresponding word frequency list, denoted as cp(c), where 1 ≤ c ≤ t; Step S2132: Take the first word frequency list cp(1) as the standard list, obtain the number of elements ysg in the standard list, and record the elements in the standard list as bzl(1), bzl(2), ... bzl(ysg); perform cyclic judgment on cp(1), cp(2), ..., cp(t) according to bzl(1), bzl(2), ..., bzl(ysg); If bzl(1) always exists in cp(1), cp(2), ..., cp(t), then bzl(1) is determined to be a product keyword; Similarly, make judgments on bzl(2), ..., bzl(ysg); The product keywords are counted according to the judgment result, and the product keywords are sorted in descending order based on the number of occurrences of elements in the word frequency list; thus, a product keyword list is obtained.

[0023] It should be noted that the word frequency list is: [{segmentation element 1: number of times 1}, {segmentation element 2: number of times 2}, {segmentation element 3: number of times 3}]. By judging the segmentation elements, the product keywords are obtained and the keywords are sorted according to the number.

[0024] See also Figure 3 ; Step S22: Filter the product keyword list according to the classification set to obtain a product feature word list; Step S221: Obtain the number n of product sets in the classification set, and obtain the product keyword list corresponding to the product set, which is recorded as cgj(1), cgj(2), ..., cgj(n); Step S222: Traverse cgj(1), cgj(2), ..., cgj(n), compare the product keyword list, and remove repeated product keywords, as follows: Get the number cgs of product keywords of cgj(1); extract the product keywords of cgj(1), record them as ccj(1), ccj(2), ..., ccj(cgs); Search cgj(2), ..., cgj(n) with ccj(1), ccj(2), ..., ccj(cgs) respectively; if ccj(1), ccj(2), ..., ccj(cgs) does not exist in cgj(2), ..., cgj(n), it is judged to be a product feature word, and the product feature words are counted to obtain a product feature word list; Similarly, obtain the product feature word list of cgj(2), ..., cgj(n); In the process of searching cgj(2), cgj(2) is searched one by one through ccj(1), ccj(2), ..., ccj(cgs).

[0025] It should be noted that the product feature word list is obtained by traversing and judging on the basis of the product keyword list, which does not affect the original order of the product keyword list. The arrangement order of the product feature word list is consistent with that of the product keyword list, and is sorted in descending order of word frequency.

[0026] Step S3: Analyze the financial product design information using the product feature word list; obtain the matching values ​​between the financial product design information and the product feature word list based on the frequency of occurrence of the product feature words in the financial product design information; and calculate the matching values ​​to obtain a matching value list; Step S31: obtaining the number of feature words tgs in the product feature word list; recording the feature words in the product feature word list as tn, obtaining the ranking of the feature words in the product feature word list, and obtaining the ranking value px(tn); Step S32: cyclically judge the feature words in the financial product design information, record the number of times the feature words appear in the financial product design information, and obtain the frequency of occurrence pl(tn); combine the number of feature words tgs in the product feature word list and the ranking value px(tn) of the feature words to calculate and obtain the matching value xfz between the financial product design information and the product feature words; ; For example, if there is a product feature word list [A, B, C], its ranking values ​​are 1, 2, 3; the frequency of occurrence of A, B, and C in the financial product design information is 10, 20, and 30 respectively; Then calculate according to the formula: , we get xfz=30.

[0027] Step S33: According to steps S31-S32, the matching values ​​of the n product feature word lists and the financial product design information are calculated to obtain xfz(1), xfz(2), ..., xfz(n), and the matching values ​​are analyzed to obtain a matching value list cfzl; Step S331: Obtain matching values ​​xfz(1), xfz(2), ..., xfz(n), accumulate the matching values ​​to obtain the matching total value xzz; calculate the matching judgment value glz(i) based on the proportional relationship between the matching values ​​and the matching total value and the variance of the matching values; ; It should be noted that the formula multiplies the proportional relationship between the matching value and the total matching value by the variance of the matching value to obtain the matching judgment value. The proportional relationship between the matching value and the total matching value determines the proportion of matching values, and the variance is used to weight the proportion. (Variance is a statistic that measures the degree of dispersion of a set of data. By introducing variance, errors caused by small differences in matching values ​​are avoided, and matching values ​​with large differences are given higher weights to improve calculation accuracy.) Based on the judgment value, a more intuitive judgment of the matching value between the product feature word list and the financial product design information is made. It should be noted that variance is a statistic that measures the degree of dispersion of a set of data. The coincidence judgment value is more accurate when calculated by variance. Step S332: Obtain the maximum value among the matching judgment values ​​to obtain the maximum judgment value max, and analyze and judge the maximum judgment value: If max>50%, the matching value calculation is valid, the matching values ​​are counted, and a matching value list cfzl is obtained; then step S4 is entered; If max≤50%, the matching value calculation is invalid and needs to be recalculated; go to step S333; Step S333: Update the product feature word list, mask the last digit in the product feature word list, obtain the frequency of occurrence of the feature word in the financial product design information (pl(tn), the number of feature words in the product feature word list (tgs-1), and the ranking value (px(tn)) of the feature word, calculate the updated matching value, and repeat steps S331-S332. See also Figure 4 ; Step S4: Determine the category of the financial product design information according to the matching value list, and obtain the category of the financial product design information by combining the classification set in the classification information; Step S41: Obtain the product set corresponding to the largest matching value according to the matching value list, and obtain the category of the financial product design information; If the matching value list is [10, 20, 30], the product set corresponding to the matching value list is financing products, investment products, and risk management products; then the category of the financial product design information is determined to be risk management products; Step S42: According to the classification information, obtain the classification dimension m, obtain the corresponding matching value list under different classification dimensions, obtain the category lb of the financial product design information according to the product set corresponding to the largest matching value, and obtain lb(1), lb(2), ..., lb(m); which constitute the category to which the financial product design information belongs.

[0028] For example, the classification dimensions are based on function, risk-return characteristics and trading venues. Obtain a list of matching values ​​under the functional dimension, and determine that the category of the financial product design information is a risk management product based on the product set corresponding to the largest matching value; Obtain a list of matching values ​​under the risk-return characteristic dimension, and determine that the category of the financial product design information is a high-risk product based on the product set corresponding to the largest matching value; Obtain a list of matching values ​​under the trading venue, and determine whether the category of the financial product design information is an on-exchange trading product based on the product set corresponding to the largest matching value; Risk management products, high-risk products, and exchange-traded products constitute the categories to which financial product design information belongs.

[0029] It should be noted that the matching value list is the matching values ​​of different categories and financial product design information under the same classification dimension. For example, according to the function dimension, they are divided into financing products, investment products, and risk management products. According to the matching value, the financial product design information belongs to financing products; according to the risk-return characteristic dimension, the financial product design information belongs to low-risk products; the classification results of all classification dimensions are combined to obtain the classification of financial product design information.

[0030] Step S43: Encrypt and store the financial product design information according to the category to which it belongs, as follows: Determine the number of encryption times based on the classification dimension m; that is, for classification based on the two dimensions of function and risk-return characteristics, perform two encryptions; Obtain the categories lb(1), lb(2), ..., lb(m) of the financial product design information under different classification dimensions; encode the text of the category to obtain the encoding value; For example, the code corresponding to the functional dimension is: 11100101100010101001111111101000100000111011110111100111101110110100111001011011101010100110; Extend the length of the coded value in proportion to the financial product design information to obtain the encrypted value; It should be noted that geometric extension refers to extending the length of the code value to be consistent with the length of the text of the financial product design information. For example, if the text length of the financial product design information is 10, and the code length is 3, the specific value is 101, then the extended value is 101 101 101 1; Map the financial product design information according to the numerical value at the current position of the password value, map the information corresponding to 0 to the horizontal axis, and the information corresponding to 1 to the vertical axis; alternately store the information on the horizontal and vertical axes to complete the first encryption and obtain the encrypted text. Combined with the classification dimension m, obtain m encrypted values. For each encrypted value (a total of m), repeat the mapping and alternate storage process of the first encryption.

[0031] In each repetition, determine how to map the encrypted text to the horizontal and vertical axes according to the encrypted value. Each repetition will generate a new encrypted text, use the encrypted text as the input for the next encryption and map it with the encrypted value; after completing m encryptions, obtain the encrypted information. For example: the information is "A B C D 1 2 3 4", and the encrypted value is 10101100; then A B C D 1 2 3 4 respectively correspond to 10101100; the information on the horizontal axis is "B D 3 4", and the information on the vertical axis is "A C 1 2"; after alternate storage, it is "B A D C 3 1 4 2". Embodiment 2

[0032] Please refer to Figure 5 , a component modular financial product design information classification system includes: Information acquisition module: acquire classification information and financial product design information. Information processing module: according to the classification information, acquire financial product information, product set, and classification set; perform word frequency analysis on the financial product information in the product set to obtain a product keyword list, and combine with the classification set to screen the product keyword list and extract product feature words to obtain a product feature word list. Information analysis module: analyze the financial product design information through the product feature word list; calculate the matching value between the financial product design information and the product feature word according to the appearance frequency of the product feature word in the financial product design information, and count the matching values to obtain a matching value list. Information classification module: judge the category of the financial product design information according to the matching value list, and combine with the classification set in the classification information to obtain the category of the financial product design information. The above formulas are all dimensionless and take their numerical values for calculation. The formula is obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation. If there are weight coefficients and proportionality coefficients, the set values are for quantifying each parameter to obtain a specific numerical value for subsequent comparison. Regarding the magnitudes of the weight coefficients and proportionality coefficients, as long as they do not affect the proportional relationship between the parameters and the quantified numerical values.

[0033] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the above-described embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-described embodiments within the technical scope disclosed by the present invention, or replace some of the technical features therein with equivalents. Such modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for classifying component modular financial product design information, characterized in that: Classification methods include: Step S1: Obtain classification information and financial product design information; Step S2: Based on the classification information, obtain financial product information, product sets, and classification sets; perform word frequency analysis on the financial product information in the product set to obtain a product keyword list; combine the classification set to filter the product keyword list, extract product feature words, and obtain a product feature word list; Step S3: Analyze the financial product design information using the product feature word list; obtain the matching values ​​between the financial product design information and the product feature words based on the frequency of occurrence of the product feature words in the financial product design information, and calculate the matching values ​​to obtain a matching value list; Step S4: Determine the category of the financial product design information based on the matching value list, and obtain the category of the financial product design information by combining the category set in the classification information.

2. A method for classifying modular financial product design information according to claim 1, characterized in that: The specific steps of step S1 are as follows: Step S11: Obtain financial product information jr, treat the financial product information of the same category as a product set, denoted as LB, obtain the number of elements t in the product set, and obtain LB={jr(1), jr(2), ..., jr(t)}; Step S12: Integrate the product sets of the same classification dimension to form a classification set FL, and record the number of product sets n; obtain FL={LB(1), LB(2), ...LB(n)}; Step S13: Obtain the classification dimension m; perform statistics on the classification set according to the classification dimension to form classification information, which is recorded as FX, FX={FL(1), FL(2), ..., FL(m)}.

3. The method for classifying modular financial product design information according to claim 1, characterized in that: The specific steps of step S2 are as follows: Step S21: Obtaining financial product information within the product set through the product set, performing word frequency analysis on the financial product information to obtain a word frequency list of the financial product information, and processing the word frequency list of the financial product information within the product set to obtain a product keyword list; Step S22: Filter the product keyword list according to the classification set, obtain the number n of product sets in the classification set, obtain the product keyword list corresponding to the product set, and record it as cgj (1) to cgj (n); compare the product keyword list and remove repeated product keywords, as follows: Get the number of product keywords cgs of cgj (1); extract the product keywords of cgj (1), and record them as ccj (1) to ccj (cgs); Search cgj(2), ..., cgj(n) with ccj(1) to ccj(cgs) respectively; if ccj(1) to ccj(cgs) do not exist in cgj(2), ..., cgj(n), then determine that they are product feature words, count the product feature words, and obtain a product feature word list; Obtain the product feature word list from cgj(2) to cgj(n).

4. A method for classifying modular financial product design information according to claim 3, characterized in that: The specific steps of step S21 are as follows: Step S211: performing text preprocessing on the financial product information by removing emoticons, uniform resource locators, and punctuation marks to obtain an initial text; Step S212: Use Jieba to perform word segmentation on the initial text, split the text into words or phrases, obtain a word segmentation list, obtain a stop word list, filter the word segmentation list according to the stop word list, count the word frequencies in the filtered word segmentation list, and obtain a word frequency list; Step S213: Obtain a word frequency list of financial product information in the product set, analyze the word frequency list, and obtain a product keyword list.

5. A modular financial product design information classification method and system according to claim 4, characterized in that: The specific steps of step S211 are as follows: Remove emoticons: Get the computer code bq1 for the first emoticon, and bq2 for the last emoticon. Use regular expressions to delete characters in the range [bq1, bq2] from the financial product information. Remove Uniform Resource Locators: Traverse the financial product information and compare it with "http". If any character "http" is found, record the character position zf1. At the same time, find the position where the first space appears after the character, record it as zf2, and delete the statements in the financial product information where the characters are between zf1 and zf2. Remove punctuation marks: Save the punctuation marks in a list, use regular expressions to match the characters and punctuation marks in the financial product information, and delete them if they match.

6. A method for classifying modular financial product design information according to claim 4, characterized in that: The specific steps of step S212 are as follows: The specific steps of filtering the word list according to the stop word list are as follows: looping through the word list, obtaining the word, matching the word with the stop word list, and deleting the word if it can be retrieved in the stop word list; After the screening is completed, the number of occurrences of the segmentation words in the segmentation list is counted, as follows: the segmentation list is converted into a set to obtain a segmentation set, the segmentation elements in the segmentation set are traversed, the segmentation list is cyclically counted according to the segmentation elements to obtain the number of occurrences of the segmentation elements, and the segmentation elements and the number of occurrences of the segmentation elements form a word frequency list.

7. A method for classifying modular financial product design information according to claim 4, characterized in that: The specific steps of step S213 are as follows: Get the number t of financial product information in the product set and obtain the corresponding word frequency list, recorded as cp(c); Take the first word frequency list cp(1) as the standard list, obtain the number of elements ysg in the standard list, and record the elements in the standard list as bzl(1), bzl(2), ... bzl(ysg); perform cyclic judgment on cp(1), cp(2), ..., cp(t) according to bzl(1) to bzl(ysg); If bzl(1) always exists in cp(1) to cp(t), then bzl(1) is determined to be a product keyword; Make judgments on bzl (2) to bzl (ysg); The product keywords are counted according to the judgment result, and the product keywords are sorted in descending order based on the number of occurrences of elements in the word frequency list; thus, a product keyword list is obtained.

8. The method for classifying modular financial product design information according to claim 1, characterized in that: The specific steps of step S3 are as follows: Step S31: obtaining the number of feature words tgs in the product feature word list; recording the feature words in the product feature word list as tn, obtaining the ranking of the feature words in the product feature word list, and obtaining the ranking value px(tn); Step S32: cyclically judging the feature words in the financial product design information, recording the number of times the feature words appear in the financial product design information, and obtaining the frequency of occurrence pl(tn); combining the number of feature words tgs in the product feature word list and the ranking value px(tn) of the feature words to calculate the matching value xfz between the financial product design information and the product feature words; ； Step S33: Calculate the matching values ​​of the n product feature word lists and the financial product design information to obtain xfz(1) to xfz(n), analyze the matching values, and obtain a matching value list cfzl.

9. A method for classifying modular financial product design information according to claim 8, characterized in that: The specific steps of step S33 are as follows: Step S331: Obtain matching values ​​xfz(1) to xfz(n), accumulate the matching values ​​to obtain the matching total value xzz; calculate the matching judgment value glz(i) based on the proportional relationship between the matching values ​​and the matching total value and the variance of the matching values; ； xfz(i) represents the consistency value between the i-th product feature word list and the financial product design information; Step S332: Obtain the maximum value among the matching judgment values ​​to obtain the maximum judgment value max, and analyze and judge the maximum judgment value: If max>50%, the matching value calculation is valid, and the matching values ​​are counted to obtain the matching value list cfzl; Go to step S4; If max≤50%, the matching value calculation is invalid and needs to be recalculated; go to step S333; Step S333: Update the product feature word list, mask the last digit of the product feature word list, obtain the frequency of occurrence of the feature word in the financial product design information pl (tn), the number of feature words in the product feature word list tgs-1, and the ranking value px (tn) of the feature word, calculate, obtain the updated matching value, and repeat steps S331-S332.

10. A modular financial product design information classification system, applicable to the modular financial product design information classification method according to any one of claims 1 to 9, characterized in that: Classification systems include: Information acquisition module: obtain classification information and financial product design information; Information processing module: Based on the classification information, obtain financial product information, product sets, and classification sets; perform word frequency analysis on the financial product information in the product set to obtain a product keyword list; combine the classification set to filter the product keyword list, extract product feature words, and obtain a product feature word list; Information analysis module: Analyzes financial product design information using a product feature word list; calculates the matching values ​​between the financial product design information and the product feature words based on the frequency of occurrence of the product feature words in the financial product design information, and calculates the matching values ​​to obtain a matching value list; Information classification module: judges the category of financial product design information based on the matching value list, and obtains the category of financial product design information by combining the classification set in the classification information.

Images