Ethereum address entity information mining method and system

Abstract

The invention provides an Ethereum address entity information mining method and system, and the method comprises the steps: sorting Ethereum addresses in an Ethereum address library based on the number of comments of the Ethereum addresses, and reading the comments in the Ethereum addresses one by one based on the positions of the sorted Ethereum addresses; entity names in the comments are extracted through a preset first neural network model; extracting an attribution name in the comment through a preset second neural network model; each entity label is correspondingly provided with a label set, each label in the label set corresponds to a mapping value, the words in the comments are matched with each label in each label set, and mapping values of the words in the comments to each label set are obtained; determining whether the word is an entity label based on the mapping value corresponding to the word in the comment; and performing word segmentation on the comments, constructing a word set based on part-of-speech, and determining entity behavior terms in the word set based on word frequencies of words in the word set.

Description

Technical Field

[0001] The present invention relates to the field of data mining technology, and in particular to a method and system for mining Ethereum address entity information. Background Art

[0002] Ethereum Browser, as a window to blockchain technology, opens a convenient door for users to explore the Ethereum ecosystem in depth. This tool integrates a wealth of functions, allowing even blockchain novices to easily get started and gain a deep understanding of the Ethereum blockchain. Through the Ethereum Browser, users can query their accounts instantly. Users only need to enter an Ethereum address to track all transaction records of the address, including key information such as transaction time, amount, counterparty, and transaction status, which greatly facilitates asset management and transaction verification.

[0003] In addition, the Ethereum browser also provides diversified services such as block information query, network status monitoring and activity log recording, which fully meets the diverse needs of users for blockchain information. Its interface is friendly and easy to operate, so both technical experts and ordinary users can quickly get started.

[0004] However, for each Ethereum address, its entity information is often difficult to obtain directly through the Ethereum browser. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides an Ethereum address entity information mining method to eliminate or improve one or more defects existing in the prior art.

[0006] One aspect of the present invention provides a method for mining Ethereum address entity information, the method comprising the steps of:

[0007] Determine the number of comments of an Ethereum address in a pre-built Ethereum address library through an Ethereum browser, sort the Ethereum addresses in the Ethereum address library based on the number of comments of the Ethereum addresses, and read the comments in the Ethereum addresses one by one based on the position of the Ethereum addresses in the sort;

[0008] The entity names in the comments are extracted by using a preset first neural network model to complete the extraction of the entity names;

[0009] Extracting the place of origin name in the comments by using a preset second neural network model to complete the extraction of the place of origin name;

[0010] Each entity tag is correspondingly provided with a tag set, each tag in the tag set corresponds to a mapping value, a word in the comment is matched with each tag in each tag set, a mapping value of the word in the comment for each tag set is obtained, and based on the mapping value corresponding to the word in the comment, whether the word is an entity tag is determined, thereby completing the extraction of the entity tag;

[0011] The comments are segmented, and a word set is constructed based on the part of speech. The entity behavior terms in the word set are determined based on the word frequency of the words in the word set, and the entity behavior terms are extracted.

[0012] Adopting the above scheme, this scheme first determines the priority extraction of the Ethereum address of the comments based on the number of comments of the Ethereum address, and completes the extraction of comments for the corresponding Ethereum address, and further extracts the content in the comments, and extracts the entity name and the place of origin name through the neural network model; matches the words of the comments with the preset tag set, determines the mapping value corresponding to the word, determines whether the word is an entity tag based on the mapping value, and completes the extraction of the entity tag; after word segmentation, the word frequency determines the entity behavior terminology in the word set, and completes the extraction of the entity behavior terminology; this scheme determines the entity information by deeply mining the comments of the Ethereum address.

[0013] In some embodiments of the present invention, in the step of sorting the Ethereum addresses in the Ethereum address library based on the number of comments of the Ethereum addresses, and reading the comments in the Ethereum addresses one by one based on the position of the Ethereum addresses in the sort, the Ethereum addresses are placed on the X-axis of the coordinate system based on the sort of the Ethereum addresses in the Ethereum address library, the number of comments of the Ethereum addresses is used as the X-axis of the coordinate system, an image of the coordinate system is constructed, the first Ethereum address on the X-axis whose number of comments is less than a first threshold is determined, and the reading of comments is stopped at this Ethereum address.

[0014] When the above scheme is used, if the number of comments is too small, it means that the Ethereum address has no practical significance. This scheme determines the Ethereum address that does not read comments based on the number of comments of the Ethereum address, thereby improving the reading efficiency.

[0015] In some embodiments of the present invention, after the step of reading the comments in the Ethereum address one by one based on the position of the Ethereum address in the sorting, it also includes: unifying the language of the comments into English.

[0016] In some embodiments of the present invention, in the step of extracting entity names from comments by using a preset first neural network model to complete the extraction of entity names, multiple entity names extracted from the same comment are obtained, the position interval of each entity name is determined in the comment, and the entity names with overlapping position intervals are merged.

[0017] In some embodiments of the present invention, in the steps of extracting entity names in comments by using a preset first neural network model to complete the extraction of entity names and extracting place names in comments by using a preset second neural network model to complete the extraction of place names, denoising is performed by using the LLAMA3 large model to delete the noisy entity names and place names.

[0018] In some embodiments of the present invention, in the step of matching a word in a comment with each tag in each tag set to obtain a mapping value of the word in the comment for each tag set, the largest mapping value of the word in the comment among all tags in a tag set is used as the mapping value of the word.

[0019] In some embodiments of the present invention, in the step of determining whether a word in a comment is an entity tag based on a mapping value corresponding to the word, and completing the extraction of the entity tag, the mapping value of the word is compared with a preset mapping threshold to determine whether the word is an entity tag.

[0020] In some embodiments of the present invention, in the step of segmenting comments, constructing a word set based on parts of speech, determining entity behavior terms in the word set based on the word frequency of the words in the word set, and completing the extraction of entity behavior terms, the words in the word set that have a frequency greater than a preset threshold are used as entity behavior terms.

[0021] In the steps of segmenting the comments, constructing a word set based on parts of speech, determining entity behavior terms in the word set based on the word frequency of the words in the word set, and completing the extraction of entity behavior terms, verbs and nouns in the comments are extracted to construct a word set.

[0022] The second aspect of the present invention also provides an Ethereum address entity information mining system, which includes a computer device, the computer device includes a processor and a memory, the memory stores computer instructions, the processor is used to execute the computer instructions stored in the memory, and when the computer instructions are executed by the processor, the system implements the steps implemented by the method described above.

[0023] The third aspect of the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps implemented by the aforementioned Ethereum address entity information mining method.

[0024] Additional advantages, purposes, and features of the present invention will be described in part in the following description, and will become apparent to those skilled in the art after studying the following, or may be learned from the practice of the present invention. The purposes and other advantages of the present invention can be specifically pointed out and obtained in the specification and the accompanying drawings.

[0025] Those skilled in the art will appreciate that the objectives and advantages that can be achieved with the present invention are not limited to the above specific description, and the above and other objectives that can be achieved by the present invention will be more clearly understood from the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The drawings described herein are used to provide a further understanding of the present invention, constitute a part of this application, and do not constitute a limitation of the present invention.

[0027] Figure 1 A schematic diagram of a first embodiment of the method for mining entity information of Ethereum addresses of the present invention;

[0028] Figure 2 This is a schematic diagram of a second implementation of the Ethereum address entity information mining method of the present invention. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0030] It should also be noted that, in order to avoid obscuring the present invention due to unnecessary details, only structures and / or processing steps closely related to the solutions according to the present invention are shown in the accompanying drawings, while other details that are not closely related to the present invention are omitted.

[0031] like Figure 1 As shown, the present invention proposes a method for mining Ethereum address entity information, the steps of which include:

[0032] Step S100, determining the number of comments of the Ethereum addresses in the pre-built Ethereum address library through the Ethereum browser, sorting the Ethereum addresses in the Ethereum address library based on the number of comments of the Ethereum addresses, and reading the comments in the Ethereum addresses one by one based on the positions of the Ethereum addresses in the sorting;

[0033] In the specific implementation process, in many Ethereum browsers, users can comment on an Ethereum address. This solution uses some information in the comments to infer the possible entity information of the Ethereum address.

[0034] In the specific implementation process, in the step of sorting the Ethereum addresses in the Ethereum address library based on the number of comments on the Ethereum addresses, the Ethereum addresses in the Ethereum address library are sorted in order from large to small.

[0035] In the specific implementation process, the Ethereum address library is a pre-stored collection of a large number of Ethereum addresses.

[0036] In the specific implementation process, this scheme first builds the Ethereum client locally and synchronizes all blocks. Then, this scheme extracts all Ethereum addresses in the block to form an Ethereum address library. However, this database is very large. If the Ethereum address in the Ethereum address library is used to query the Ethereum browser for comments on the address, it will cause the Ethereum browser to crash and trigger the anti-crawler mechanism. Therefore, this scheme is based on the principle that the possibility of comments on an address is positively correlated with its turnover, that is, the higher the turnover, the more important the address is, and the higher the possibility of comments, and vice versa. Therefore, when querying comments from each Ethereum browser, this scheme queries from high turnover addresses to low turnover addresses. At the same time, this scheme designs an indicator called the number of comments per thousand addresses, which means the number of comments found for every 1,000 comment queries. According to the assumptions of this scheme, if this scheme uses the number of queried addresses as the x-axis and this indicator as the y-axis, its function graph should be similar to an inverse proportional function. This scheme generates the function graph in real time when running the crawler program to query each Ethereum browser. When this solution notices that the image tends to a number infinitely close to 0 and remains stable for a period of time, we believe that the query has ended and the crawler should be exited to avoid any further waste of resources.

[0037] Step S210, extracting entity names in the comments through a preset first neural network model to complete the extraction of entity names;

[0038] Step S220, extracting the place name in the comment by using a preset second neural network model, thereby completing the extraction of the place name;

[0039] In the specific implementation process, the first neural network model and can adopt a name database, a spacy library model or a LLAMA3 large model.

[0040] In the specific implementation process, the place names are very different from the entity names, because the database of entity names is often difficult to obtain in full, while the database of place names is easier to obtain than the database of entity names. Therefore, place names are generally much less than personal names. Therefore, this solution can cover most places by relying on the place name database without much need for AI algorithms to infer place names.

[0041] Step S230: Each entity tag is correspondingly provided with a tag set, each tag in the tag set corresponds to a mapping value, a word in the comment is matched with each tag in each tag set, a mapping value of the word in the comment for each tag set is obtained, and based on the mapping value corresponding to the word in the comment, whether the word is an entity tag is determined, thereby completing the extraction of the entity tag;

[0042] In the specific implementation process, since each entity tag often has multiple names, and multiple names correspond to one entity tag, this solution constructs multiple names of an entity tag as tags as a tag set, and sets different mapping values ​​based on the correlation between each tag and the entity tag to reflect the degree of relationship between each tag and the entity tag, and determines whether the word in the comment is an entity tag based on the mapping value corresponding to the word.

[0043] In the specific implementation process, when a word in a comment corresponds to multiple tags in a tag set, the word has multiple mapping values ​​corresponding to multiple tags. The maximum value is taken as the mapping value corresponding to the word, and the mapping value of the word is compared with the preset mapping judgment threshold. If the mapping value of the word is greater than the preset mapping judgment threshold, the word is judged as an entity tag.

[0044] Step S240, segment the comments, construct a word set based on the part of speech, determine the entity behavior terms in the word set based on the word frequency of the words in the word set, and complete the extraction of the entity behavior terms.

[0045] In the specific implementation process, in the step of segmenting the comments, the spacy library model segments the comments.

[0046] In the specific implementation process, dictionary matching, morphological rules or LLAMA3 model are used to identify parts of speech;

[0047] Dictionary matching: Match the words in the text with the entries in the dictionary through a predefined dictionary, and determine the part of speech of the word based on the part-of-speech tagging information in the dictionary. This method relies on the completeness and accuracy of the dictionary, and may not work well for words that are not included in the dictionary or for words with multiple meanings.

[0048] Morphological rules: Use the morphological features of words (such as prefixes, suffixes, and roots) to infer the part of speech; for example, in English, the "-ly" suffix usually indicates an adverb, while the "-ness" suffix often indicates a noun.

[0049] Adopting the above scheme, this scheme first determines the priority extraction of the Ethereum address of the comments based on the number of comments of the Ethereum address, and completes the extraction of comments for the corresponding Ethereum address, and further extracts the content in the comments, and extracts the entity name and the place of origin name through the neural network model; matches the words of the comments with the preset tag set, determines the mapping value corresponding to the word, determines whether the word is an entity tag based on the mapping value, and completes the extraction of the entity tag; after word segmentation, the word frequency determines the entity behavior terminology in the word set, and completes the extraction of the entity behavior terminology; this scheme determines the entity information by deeply mining the comments of the Ethereum address.

[0050] In some embodiments of the present invention, in the step of sorting the Ethereum addresses in the Ethereum address library based on the number of comments of the Ethereum addresses, and reading the comments in the Ethereum addresses one by one based on the position of the Ethereum addresses in the sort, the Ethereum addresses are placed on the X-axis of the coordinate system based on the sort of the Ethereum addresses in the Ethereum address library, the number of comments of the Ethereum addresses is used as the X-axis of the coordinate system, an image of the coordinate system is constructed, the first Ethereum address on the X-axis whose number of comments is less than a first threshold is determined, and the reading of comments is stopped at this Ethereum address.

[0051] In the specific implementation process, the Ethereum addresses in the Ethereum address database are sorted in order from large to small, and the number of comments on the Ethereum addresses gradually decreases in the extension direction of the X-axis of the coordinate system.

[0052] When the above scheme is used, if the number of comments is too small, it means that the Ethereum address has no practical significance. This scheme determines the Ethereum address that does not read comments based on the number of comments of the Ethereum address, thereby improving the reading efficiency.

[0053] In some embodiments of the present invention, after the step of reading the comments in the Ethereum address one by one based on the position of the Ethereum address in the sorting, it also includes: unifying the language of the comments into English.

[0054] In the specific implementation process, in the step of unifying the language of comments into English, after this solution extracts comments from each Ethereum address, in order to facilitate subsequent data processing, what this solution will do is to unify the language of all comments. This solution unifies all languages ​​into English. This solution will use online translation to unify the language. This solution uses the method of merging and translating similar languages ​​to reduce the number of requests for this solution. Specifically, it first traverses all comments and uses the langid library to classify the languages ​​of all comments. For comments in the same language, this solution uses a delimited string to concatenate them into a large string and then translate them. Assume that there are currently three sentences in this solution. This solution uses a delimited string to concatenate the three sentences and then translate them. Since the segmented string is composed of special characters, the translator will not translate it, but will only translate it. This solution only needs to segment the translated string to obtain the corresponding translation content.

[0055] Generally speaking, online translation can support translation of up to 5,000 words, so this method can be used to concatenate strings to a length of no more than 5,000, which greatly improves translation efficiency. Better yet, a pipeline approach can be used to process data, that is, the query and translation of comment data can be performed simultaneously, without having to wait until all data queries are completed before starting translation.

[0056] Specifically, we currently have three sentences A, B, and C. This solution uses the separator string [SEP] to concatenate the three sentences into A[SEP]B[SEP]C and then translate them. Since the separator string [SEP] consists of special characters, the translator will not translate it, but only translate A, B, and C. This solution only needs to split the translated string according to [SEP] to get the corresponding translation content of A, B, and C.

[0057] In some embodiments of the present invention, in the step of extracting entity names from comments by using a preset first neural network model to complete the extraction of entity names, multiple entity names extracted from the same comment are obtained, the position interval of each entity name is determined in the comment, and the entity names with overlapping position intervals are merged.

[0058] In practice, entity tags refer to words that describe which organization the address belongs to. For example, the Ethereum address belongs to a wallet manufacturer or a user of an exchange.

[0059] In the specific implementation process, if "zhang han sheng" exists in the comment, the extracted entity names include "zhang han" and "han sheng". If there is no "zhang han sheng" in the entity name database, but there are "hansheng" and "zhang han", the entity names need to be merged;

[0060] This solution is handled by superimposing name intervals. This solution takes the index intervals of all names in the comments. If the comment is "This address belongs to zhang han sheng", then the position interval of "zhang han" is [24,32], and the position interval of "han sheng" is [30,38]; if there is an overlap between the two position intervals, they are merged to [24,38], and the merged entity name "zhang han sheng" is obtained.

[0061] By adopting the above scheme, the present scheme can merge entity names according to their positions in the comments, which can reduce the number of entity names to be extracted on the one hand, and improve the accuracy of the entity names finally extracted on the other hand.

[0062] like Figure 2As shown, in some embodiments of the present invention, after determining whether the word is an entity tag based on the mapping value corresponding to the word in the comment and completing the step of extracting the entity tag, it also includes step S231, determining the position in the comment where the extracted entity name is located, and whether there is a noise word in a preset noise word set within a threshold distance range before and after the comment position. If so, the entity tag is deleted.

[0063] The threshold distance may be 4 words.

[0064] In the specific implementation process, after completing the extraction of entity names, this solution no longer uses LLAMA3 to remove noise, but removes noise based on some known rules of this solution, because the entity tag features are relatively typical and only appear in special semantics. This method can effectively improve the speed of noise filtering. This solution constructs a noise exclusion database based on the characteristics of the data. These words basically appear within 4 words before and after the label matching position. Therefore, for each matching position, as long as there is a word hit in the noise exclusion database, we will exclude the match and remove some entity tags that are not reference-oriented.

[0065] In some embodiments of the present invention, in the steps of extracting entity names in comments by using a preset first neural network model to complete the extraction of entity names and extracting place names in comments by using a preset second neural network model to complete the extraction of place names, denoising is performed by using the LLAMA3 large model to delete the noisy entity names and place names.

[0066] In some embodiments of the present invention, in the step of matching a word in a comment with each tag in each tag set to obtain a mapping value of the word in the comment for each tag set, the largest mapping value of the word in the comment among all tags in a tag set is used as the mapping value of the word.

[0067] In some embodiments of the present invention, in the step of determining whether a word in a comment is an entity tag based on a mapping value corresponding to the word, and completing the extraction of the entity tag, the mapping value of the word is compared with a preset mapping threshold to determine whether the word is an entity tag.

[0068] In the specific implementation process, if an entity tag is coin base Exchange, the tag set corresponding to the entity tag includes fishcoin, coinbase and Exchange, the mapping value corresponding to fishcoin is 100, the mapping value corresponding to coinbase is 90, and the mapping value corresponding to Exchange is 10. If coinbase and Exchange are included in a comment, the corresponding mapping values ​​are 90 and 10. The maximum mapping value is 90, and 90 is compared with the preset mapping threshold. If 90 is greater than the preset mapping threshold, coinbase is determined to be the entity tag; if only Exchange is included in a comment, the corresponding mapping value is 10, the maximum mapping value is 10, and 10 is compared with the preset mapping threshold. If 10 is not greater than the preset mapping threshold, there is no entity tag in the comment.

[0069] By adopting the above scheme, this scheme can determine the entity label of a commenting employee through the mapping value through the label set corresponding to the preset entity label, thereby improving the extraction accuracy.

[0070] In some embodiments of the present invention, in the step of segmenting comments, constructing a word set based on parts of speech, determining entity behavior terms in the word set based on the word frequency of the words in the word set, and completing the extraction of entity behavior terms, the words in the word set that have a frequency greater than a preset threshold are used as entity behavior terms.

[0071] In the steps of segmenting the comments, constructing a word set based on parts of speech, determining entity behavior terms in the word set based on the word frequency of the words in the word set, and completing the extraction of entity behavior terms, verbs and nouns in the comments are extracted to construct a word set.

[0072] An embodiment of the present invention also provides an Ethereum address entity information mining system, which includes a computer device, wherein the computer device includes a processor and a memory, wherein the memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps implemented by the method described above.

[0073] The embodiment of the present invention also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps implemented by the aforementioned Ethereum address entity information mining method are implemented. The computer-readable storage medium can be a tangible storage medium, such as a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a floppy disk, a hard disk, a removable storage disk, a CD-ROM, or any other form of storage medium known in the technical field.

[0074] It should be understood by those skilled in the art that the exemplary components, systems and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software or a combination of the two. Whether it is performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present invention are programs or code segments used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link via a data signal carried in a carrier.

[0075] It should be clear that the present invention is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present invention.

[0076] In the present invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with features of other embodiments or replace features of other embodiments.

[0077] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the embodiments of the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for mining entity information of Ethereum addresses, characterized in that: The steps of the method include: Determine the number of comments of an Ethereum address in a pre-built Ethereum address library through an Ethereum browser, sort the Ethereum addresses in the Ethereum address library based on the number of comments of the Ethereum addresses, and read the comments in the Ethereum addresses one by one based on the position of the Ethereum addresses in the sort; The entity names in the comments are extracted by using a preset first neural network model to complete the extraction of the entity names; Extracting the place of origin name in the comments by using a preset second neural network model to complete the extraction of the place of origin name; Each entity tag is correspondingly provided with a tag set, each tag in the tag set corresponds to a mapping value, a word in the comment is matched with each tag in each tag set, a mapping value of the word in the comment for each tag set is obtained, and based on the mapping value corresponding to the word in the comment, whether the word is an entity tag is determined, thereby completing the extraction of the entity tag; The comments are segmented, and a word set is constructed based on the part of speech. The entity behavior terms in the word set are determined based on the word frequency of the words in the word set, and the entity behavior terms are extracted.

2. The Ethereum address entity information mining method according to claim 1, characterized in that: In the steps of sorting the Ethereum addresses in the Ethereum address library based on the number of comments of the Ethereum addresses, and reading the comments in the Ethereum addresses one by one based on the positions of the Ethereum addresses in the sort, the Ethereum addresses are placed on the X-axis of the coordinate system based on the sort of the Ethereum addresses in the Ethereum address library, the number of comments of the Ethereum addresses is used as the X-axis of the coordinate system, an image of the coordinate system is constructed, the first Ethereum address on the X-axis whose number of comments is less than a first threshold is determined, and the reading of comments is stopped at the Ethereum address.

3. The Ethereum address entity information mining method according to claim 1, characterized in that: After the step of reading the comments in the Ethereum address one by one based on the position of the Ethereum address in the sorting, it also includes: unifying the language of the comments into English.

4. The method for mining Ethereum address entity information according to any one of claims 1 to 3, characterized in that: In the step of extracting entity names from comments by using a preset first neural network model to complete the extraction of entity names, multiple entity names extracted from the same comment are obtained, the position interval of each entity name is determined in the comment, and entity names with overlapping position intervals are merged.

5. The Ethereum address entity information mining method according to claim 1, characterized in that: In the steps of extracting entity names from comments by using the preset first neural network model to complete the extraction of entity names and extracting place names from comments by using the preset second neural network model to complete the extraction of place names, denoising is performed by using the LLAMA3 large model to delete the noisy entity names and place names.

6. The Ethereum address entity information mining method according to claim 1, characterized in that: In the step of matching the words in the comment with each tag in each tag set to obtain the mapping value of the words in the comment for each tag set, the maximum mapping value of the word in the comment among all tags in a tag set is taken as the mapping value of the word.

7. The Ethereum address entity information mining method according to claim 1 or 6, characterized in that: In the step of determining whether a word is an entity tag based on a mapping value corresponding to a word in a comment and completing the extraction of the entity tag, the mapping value of the word is compared with a preset mapping threshold to determine whether the word is an entity tag.

8. The Ethereum address entity information mining method according to claim 1, characterized in that: After determining whether the word is an entity tag based on the mapping value corresponding to the word in the comment, the step of completing the entity tag extraction also includes the step of determining the position of the extracted entity name in the comment, and whether there is a noise word in a preset noise word set within a threshold distance range before and after the comment position. If so, deleting the entity tag.

9. The Ethereum address entity information mining method according to claim 1, characterized in that: In the step of segmenting the comments, constructing a word set based on parts of speech, determining the entity behavior terms in the word set based on the word frequency of the words in the word set, and completing the extraction of the entity behavior terms, the words in the word set that are greater than a preset word frequency threshold are used as entity behavior terms.

10. An Ethereum address entity information mining system, characterized in that: The system includes a computer device, which includes a processor and a memory. The memory stores computer instructions. The processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps implemented by the method as described in any one of claims 1 to 9.

Images