Intelligent semantic analysis system and method based on large data model

Abstract

The invention relates to the technical field of semantic analysis, and discloses an intelligent semantic analysis system and method based on a data large model, and the method comprises the steps: converting an audio file into a text, dividing the text into a character group and a word group, numbering the character group and the word group, and extracting repeated content; training a deep learning model to process the character group and the word group to form a standard text; analyzing the audio to determine a language region, obtaining a dialect database and extracting dialect content; combining the dialects, generating a to-be-analyzed text, and performing semantic analysis and classification; and finally, obtaining the expression meaning of the standard text according to the analysis result. The existing marked dialects are combined and substituted into the standard text, so that potential multiple explanations caused by dialect differences in the text can be covered and revealed, all expression meanings existing in the standard text can be obtained, and the process can ensure that any possible expression meaning is not missed, so that the accuracy of the expression meanings is improved. And the final semantic analysis is more comprehensive and accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of semantic analysis, and in particular to an intelligent semantic analysis system and method based on a large data model. Background Art

[0002] With the continuous development of information technology, speech recognition and natural language processing technologies have been widely used in many fields. At present, there are many speech recognition systems on the market, which can convert audio into text and perform simple processing on the text; most of the existing speech recognition systems rely on pre-trained language models, which have good results when processing standard Mandarin or specific languages.

[0003] However, the existing traditional speech recognition systems often have reduced recognition accuracy when dealing with dialects, accents, mixed languages, and the speech of patients with stuttering. Especially in multi-dialect areas, there are obvious differences in users' language habits and expressions, and traditional speech recognition systems find it difficult to accurately capture and process these differences. In addition, when dealing with homophones and repeated words, existing technologies often cannot correctly understand and process them in the context, resulting in a decrease in the accuracy of the final semantic analysis results.

[0004] Therefore, it is necessary to develop a new intelligent semantic analysis method that can accurately convert audio into text and process the text multiple times, focusing on dialects and repeated words, and can process and analyze specific semantics through big data and deep learning technology to perform accurate and comprehensive semantic analysis. Summary of the invention

[0005] The present invention provides an intelligent semantic analysis system and method based on a large data model, which promotes solving the problems mentioned in the above background technology.

[0006] The present invention provides the following technical solution: an intelligent semantic analysis method based on a large data model, comprising:

[0007] Obtain the audio file to be analyzed and record it as the audio to be processed;

[0008] According to the content of the audio to be processed, the audio to be processed is converted into a text form and recorded as the text to be processed;

[0009] Analyze the text to be processed and divide the text content in the text to be processed into character groups and word groups;

[0010] Number the word groups and phrase groups separately;

[0011] According to the number, the text to be processed is processed to obtain repeated text segments and repeated word segments;

[0012] A repeated word deep learning model is obtained through data training;

[0013] Through the repeated word deep learning model, a first processing operation and a second processing operation are respectively performed on the repeated text segment and the repeated word segment;

[0014] Recording the to-be-processed text that has completed the first processing operation and the second processing operation as a standard text;

[0015] Analyze the audio to be processed, obtain the region to which the language in the audio to be processed belongs, and record it as the marked region;

[0016] Obtain all dialect databases existing in the marked area, recorded as marked database;

[0017] According to the marked database, the dialect text content existing in the standard text is obtained and recorded as the marked dialect;

[0018] Get the number of marked dialects;

[0019] The marked dialects are combined according to the quantity, and the combined marked dialects are substituted into the marked text to obtain the text to be analyzed;

[0020] Perform semantic analysis on the text to be analyzed, and classify the text to be analyzed according to the results of the semantic analysis;

[0021] According to the classification results of the text to be analyzed, the meaning of the standard text expression is obtained.

[0022] Optionally, analyzing the text to be processed and dividing the text contents in the text to be processed into text groups includes:

[0023] Traverse the characters in the text to be processed;

[0024] Obtain different characters in the text to be processed and classify the same characters into the same character group;

[0025] Get the words in the text to be processed and group the same words into the same word group.

[0026] Optionally, numbering the character groups and word groups separately includes:

[0027] Get the divided text groups, and number each text group in sequence;

[0028] The characters in the same character group are numbered uniformly, and the number is consistent with the number of the character group;

[0029] Obtain the divided word groups, and number each word group in alphabetical order;

[0030] The words in the same word group are numbered uniformly, and the number is consistent with the number of the word group to which the word belongs.

[0031] Optionally, the processing of the text to be processed according to the number to obtain repeated text segments and repeated word segments includes:

[0032] Traverse the text to be processed according to the text number;

[0033] Obtaining continuous text content with the same number in the text to be processed, dividing the text content into a text segment, and recording it as a repeated text segment;

[0034] Get all repeated text segments in the text to be processed;

[0035] Traverse the text to be processed according to the word number;

[0036] Obtaining continuous word contents with the same number in the text to be processed, dividing the word contents into a word segment, and recording it as a repeated word segment;

[0037] Get all repeated word segments in the text to be processed.

[0038] Optionally, the performing of the first processing operation and the second processing operation on the repeated text segment and the repeated word segment respectively through the repeated word deep learning model includes:

[0039] Input the text to be processed into the repeated word deep learning model to determine the meaning of the repeated word segments in the text to be processed;

[0040] If the repeated text segment does not affect the meaning of the text to be processed, the repeated text segment is retained;

[0041] If the repeated text segment does not have any expression meaning in the text to be processed, then mark the repeated text segment;

[0042] Traverse all repeated text segments and obtain all marked repeated text segments;

[0043] For each marked repeated text segment, one character is retained, and the remaining characters with the same number in the repeated text segment are deleted, thereby completing the first processing operation;

[0044] Input the text to be processed into the repeated word deep learning model to determine the meaning of the repeated word segments in the text to be processed;

[0045] If the repeated word segment does not affect the meaning of the text to be processed, the repeated word segment is retained;

[0046] If the repeated word segment does not have any expression meaning in the text to be processed, then mark the repeated word segment;

[0047] Traverse all repeated word segments and obtain all marked repeated word segments;

[0048] For each marked repeated word segment, one word is retained, and the remaining words with the same number in the repeated word segment are deleted, completing the second processing operation.

[0049] Optionally, the marking dialects are combined according to the quantity, and the combined marking dialects are substituted into the marked text to obtain the text to be analyzed, including:

[0050] If there is no marked dialect in the standard text, the standard text is used as the text to be analyzed;

[0051] If there are marked dialects in the standard text, the number of marked dialects is obtained, recorded as N;

[0052] Combine the marked dialects N times:

[0053] In the first combination, N marked dialects are selected for combination;

[0054] Among the marked dialects, select N marked dialects at random until all combinations of N marked dialects are traversed;

[0055] Record the selected marked dialect each time, obtain the expression meaning of each marked dialect, and substitute it into the standard text as the text to be analyzed;

[0056] In the second combination, N-1 marked dialects are selected for combination;

[0057] Among the marked dialects, select N-1 marked dialects at random until all combinations of N-1 marked dialects are traversed;

[0058] Record the selected marked dialect each time, obtain the expression meaning of each marked dialect, and substitute it into the standard text as the text to be analyzed;

[0059] In the third combination, N-2 marked dialects are selected for combination;

[0060] Among the marked dialects, select N-2 marked dialects at random until all combinations of N-2 marked dialects are traversed;

[0061] Record the selected marked dialect each time, obtain the expression meaning of each marked dialect, and substitute it into the standard text as the text to be analyzed;

[0062] For the Nth combination, one marked dialect is selected for combination;

[0063] Among the marked dialects, select any one of them, until all marked dialects are traversed;

[0064] The selected marked dialect each time is recorded, the expression meaning of each marked dialect is obtained, and it is substituted into the standard text as the text to be analyzed.

[0065] Optionally, performing semantic analysis on the text to be analyzed and classifying the text to be analyzed according to the semantic analysis result includes:

[0066] Get all the text to be analyzed;

[0067] Analyze the expression meaning of each text to be analyzed in turn;

[0068] If the text to be analyzed has an expression meaning, the text to be analyzed is marked and the expression meaning of the text to be analyzed is recorded;

[0069] If the text to be analyzed does not have any expression meaning, the text to be analyzed will be deleted.

[0070] Optionally, obtaining the standard text expression meaning according to the classification result of the text to be analyzed includes:

[0071] Obtain the expression meanings of all recorded texts to be analyzed, and form an expression meaning set of standard texts;

[0072] Selecting one expression meaning from the expression meaning set of the standard text as the first expression meaning, and sending it to the user through the user selection module;

[0073] Obtaining the user's judgment on the meaning of the first expression;

[0074] If the user makes a correct judgment on the first expression meaning, the first expression meaning is used as the expression meaning of the standard text, and other expression meanings in the expression meaning set are deleted;

[0075] If the user makes an incorrect judgment on the first expression meaning, the first expression meaning is deleted from the expression meaning set;

[0076] Selecting another expression meaning from the expression meaning set, recording it as a second expression meaning, and sending it to the user through the user selection module again;

[0077] Obtaining the user's judgment on the meaning of the second expression;

[0078] If the user makes a correct judgment on the second expression meaning, the second expression meaning is used as the expression meaning of the standard text, and other expression meanings in the expression meaning set are deleted;

[0079] If the user makes an incorrect judgment on the second expression meaning, the second expression meaning is deleted from the expression meaning set;

[0080] Selecting another expression meaning from the expression meaning set, recording it as a third expression meaning, and sending it to the user through the user selection module again;

[0081] Obtaining the user's judgment on the meaning of the third expression;

[0082] If the user makes a correct judgment on the third expression meaning, the third expression meaning is used as the expression meaning of the standard text, and other expression meanings in the expression meaning set are deleted;

[0083] If the user makes an incorrect judgment on the third expression meaning, the third expression meaning will be deleted from the expression meaning set, and the fourth expression meaning will be selected again and sent to the user. Based on the user's judgment, it will be determined whether to continue selecting expression meanings from the expression meaning set and sending them to the user, until all expression meanings in the expression meaning set are traversed, and the expression meaning of the standard text is obtained based on the user's judgment.

[0084] A system for implementing the intelligent semantic analysis based on the data big model, comprising:

[0085] Audio processing module: processes the audio to be processed and converts the audio to be processed into text to be processed;

[0086] Text segmentation module: divides the text into character groups and word groups according to the text content;

[0087] Text processing module: number the texts and process them according to the numbers to generate standard texts;

[0088] Audio analysis module: analyzes the region to which the language in the audio to be processed belongs based on the audio to be processed;

[0089] Data acquisition module: obtains the dialect database of the marked area through the big data model;

[0090] Dialect marking module: obtains the dialect text content in the standard text and marks it;

[0091] Combination module: record any combination of marked dialects;

[0092] Meaning analysis module: determines whether the text to be analyzed has expression meaning;

[0093] Classification module: classifies the text to be analyzed according to the semantic analysis results of the text to be analyzed;

[0094] User selection module: provides the user with the expression meaning of the recorded standard text for the user to choose;

[0095] Data recording module: stores standard texts that express the meaning of records.

[0096] The present invention has the following beneficial effects:

[0097] 1. Converting the audio to be processed into text can avoid the noise or unclear problems that may appear in the audio affecting the accuracy of subsequent analysis. The text content is easier to retrieve and analyze, and it is also convenient to provide the processing results to the user for user judgment.

[0098] 2. Through numbering, the standardization and normalization of text content is achieved. Through numbering, the same elements in the text can be represented uniformly, reducing repeated descriptions, simplifying data structure, and helping to improve data processing efficiency.

[0099] 3. For repeated text segments and repeated word segments, deletion operations are performed based on whether they have normal expression meanings in the standard text. This can effectively remove redundant information in the text and retain the main expression meaning, making the text more concise and easy to understand; and by deleting repeated content that does not affect the expression meaning, the possibility of misunderstanding and ambiguity can be reduced; in addition, this deletion operation also optimizes the length of the text, improves the processing efficiency and storage efficiency of the text, and saves computing resources and time.

[0100] 4. By combining the existing marked dialects and substituting them into the standard text, we can cover and reveal the potential multiple interpretations in the text caused by dialect differences, and obtain all the expression meanings in the standard text. This process can ensure that no possible expression meaning is missed, making the final semantic analysis more comprehensive and accurate.

[0101] 5. By introducing a user feedback mechanism, it helps to adjust the analysis results based on the user's judgment and opinions, so that the expression meaning of the finalized standard text can better meet the user's personalized needs and make the analysis results more in line with the actual human expression. BRIEF DESCRIPTION OF THE DRAWINGS

[0102] Figure 1 This is a system block diagram of the present invention.

[0103] Figure 2 It is a flowchart of the present invention. DETAILED DESCRIPTION

[0104] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0105] Embodiment 1, an intelligent semantic analysis method based on a large data model, comprising:

[0106] Obtain the audio file to be analyzed and record it as the audio to be processed;

[0107] According to the content of the audio to be processed, the audio to be processed is converted into text form and recorded as the text to be processed; converting the audio to be processed into text form can avoid the noise or unclear problems that may appear in the audio affecting the accuracy of subsequent analysis, and the content in text form is easier to retrieve and analyze, and it is also convenient to provide the processing results to the user for user judgment operations.

[0108] Analyze the text to be processed and divide the text content in the text to be processed into character groups and word groups;

[0109] The text groups and word groups are numbered separately; through the numbering method, the standardization and normalization of the text content is achieved. Through numbering, the same elements in the text can be represented uniformly, reducing repeated descriptions, simplifying the data structure, and helping to improve the efficiency of data processing.

[0110] According to the number, the text to be processed is processed to obtain repeated text segments and repeated word segments;

[0111] A repeated word deep learning model is obtained through data training;

[0112] Through the repeated word deep learning model, a first processing operation and a second processing operation are respectively performed on the repeated text segment and the repeated word segment;

[0113] Recording the to-be-processed text that has completed the first processing operation and the second processing operation as a standard text;

[0114] Analyze the audio to be processed, obtain the region to which the language in the audio to be processed belongs, and record it as the marked region;

[0115] Obtain all dialect databases existing in the marked area, recorded as marked database;

[0116] According to the marked database, the dialect text content existing in the standard text is obtained and recorded as the marked dialect; the dialect refers to the special expression meaning of some words in certain regions;

[0117] Get the number of marked dialects;

[0118] The marked dialects are combined according to the quantity, and the combined marked dialects are substituted into the marked text to obtain the text to be analyzed;

[0119] Perform semantic analysis on the text to be analyzed, and classify the text to be analyzed according to the results of the semantic analysis;

[0120] According to the classification results of the text to be analyzed, the meaning of the standard text expression is obtained.

[0121] The step of analyzing the text to be processed and dividing the text contents in the text to be processed into text groups includes:

[0122] Traverse the characters in the text to be processed;

[0123] Obtain different characters in the text to be processed and classify the same characters into the same character group;

[0124] Get the words in the text to be processed and group the same words into the same word group.

[0125] The step of numbering the character groups and the word groups respectively comprises:

[0126] Get the divided text groups, and number each text group in sequence;

[0127] The characters in the same character group are numbered uniformly, and the number is consistent with the number of the character group;

[0128] Obtain the divided word groups, and number each word group in alphabetical order;

[0129] The words in the same word group are numbered uniformly, and the number is consistent with the number of the word group to which the word belongs.

[0130] The text to be processed is processed according to the number to obtain repeated text segments and repeated word segments, including:

[0131] Traverse the text to be processed according to the text number;

[0132] Obtaining continuous text content with the same number in the text to be processed, dividing the text content into a text segment, and recording it as a repeated text segment;

[0133] Get all repeated text segments in the text to be processed;

[0134] Traverse the text to be processed according to the word number;

[0135] Obtaining continuous word contents with the same number in the text to be processed, dividing the word contents into a word segment, and recording it as a repeated word segment;

[0136] Get all repeated word segments in the text to be processed.

[0137] The method of performing a first processing operation and a second processing operation on a repeated text segment and a repeated word segment respectively through a repeated word deep learning model includes:

[0138] Input the text to be processed into the repeated word deep learning model to determine the meaning of the repeated word segments in the text to be processed;

[0139] If the repeated text segment does not affect the meaning of the text to be processed, the repeated text segment is retained;

[0140] If the repeated text segment does not have any expression meaning in the text to be processed, then mark the repeated text segment;

[0141] Traverse all repeated text segments and obtain all marked repeated text segments;

[0142] For each marked repeated text segment, one character is retained, and the remaining characters with the same number in the repeated text segment are deleted, thereby completing the first processing operation;

[0143] Input the text to be processed into the repeated word deep learning model to determine the meaning of the repeated word segments in the text to be processed;

[0144] If the repeated word segment does not affect the meaning of the text to be processed, the repeated word segment is retained;

[0145] If the repeated word segment does not have any expression meaning in the text to be processed, then mark the repeated word segment;

[0146] Traverse all repeated word segments and obtain all marked repeated word segments;

[0147] For each marked repeated word segment, one word is retained and the remaining words with the same number in the repeated word segment are deleted to complete the second processing operation; for repeated text segments and repeated word segments, deletion operations are performed based on whether they have normal expression meanings in the standard text, which can effectively remove redundant information in the text and retain the main expression meanings, thereby making the text more concise and easy to understand; and by deleting repeated content that does not affect the expression meaning, the possibility of misunderstanding and ambiguity can be reduced; in addition, this deletion operation also optimizes the length of the text, improves the processing efficiency and storage efficiency of the text, and saves computing resources and time.

[0148] The marking dialects are combined according to the quantity, and the combined marking dialects are substituted into the marking text to obtain the text to be analyzed, including:

[0149] If there is no marked dialect in the standard text, the standard text is used as the text to be analyzed;

[0150] If there are marked dialects in the standard text, the number of marked dialects is obtained, recorded as N;

[0151] Combine the marked dialects N times:

[0152] In the first combination, N marked dialects are selected for combination;

[0153] Among the marked dialects, select N marked dialects at random until all combinations of N marked dialects are traversed;

[0154] Record the selected marked dialect each time, obtain the expression meaning of each marked dialect, and substitute it into the standard text as the text to be analyzed;

[0155] In the second combination, N-1 marked dialects are selected for combination;

[0156] Among the marked dialects, select N-1 marked dialects at random until all combinations of N-1 marked dialects are traversed;

[0157] Record the selected marked dialect each time, obtain the expression meaning of each marked dialect, and substitute it into the standard text as the text to be analyzed;

[0158] In the third combination, N-2 marked dialects are selected for combination;

[0159] Among the marked dialects, select N-2 marked dialects at random until all combinations of N-2 marked dialects are traversed;

[0160] Record the selected marked dialect each time, obtain the expression meaning of each marked dialect, and substitute it into the standard text as the text to be analyzed;

[0161] For the Nth combination, one marked dialect is selected for combination;

[0162] In the marked dialects, select any one of them, until all marked dialects are traversed; assuming that the number of marked dialects in the marked text is 3, the marked dialects are combined 3 times;

[0163] In the first combination, the meanings of the expressions of the three marked dialects are substituted into the standard text as the text to be analyzed;

[0164] During the second combination, the first two marked dialects are selected first, and the expression meanings of the two marked dialects are obtained. The remaining third marked dialect is regarded as a standard common language, and the expression meaning as a dialect is not obtained. The first two marked dialects are substituted into the standard text as the text to be analyzed.

[0165] Afterwards, the last two marked dialects are selected, and the expression meanings of the two marked dialects are obtained. The remaining first marked dialect is regarded as a standard common language, and the expression meaning as a dialect is not obtained. The selected last two marked dialects are substituted into the standard text as the text to be analyzed;

[0166] When combining for the third time, the first marked dialect is selected first, and the expression meaning of the first marked dialect is obtained. The remaining second and third marked dialects are regarded as standard common languages, and the expression meanings as dialects are not obtained. The selected first marked dialect is substituted into the standard text as the text to be analyzed.

[0167] Then the second marked dialect is selected, and the expression meaning of the second marked dialect is obtained. The remaining first and third marked dialects are regarded as standard common languages, and the expression meaning as dialects is not obtained. The selected second marked dialect is substituted into the standard text as the text to be analyzed;

[0168] Finally, the third marked dialect is selected, and the expression meaning of the third marked dialect is obtained. The remaining first and second marked dialects are regarded as standard common languages, and the expression meaning as dialects is not obtained. The selected third marked dialect is substituted into the standard text as the text to be analyzed;

[0169] Record the selected marked dialect each time, obtain the expression meaning of each marked dialect, and substitute it into the standard text as the text to be analyzed; by combining the existing marked dialects and substituting them into the standard text, it is possible to cover and reveal the potential multiple interpretations caused by dialect differences in the text, and obtain all the expression meanings existing in the standard text. This process can ensure that no possible expression meaning is missed, making the final semantic analysis more comprehensive and accurate.

[0170] The performing semantic analysis on the text to be analyzed and classifying the text to be analyzed according to the semantic analysis result includes:

[0171] Get all the text to be analyzed;

[0172] Analyze the expression meaning of each text to be analyzed in turn;

[0173] If the text to be analyzed has an expression meaning, the text to be analyzed is marked and the expression meaning of the text to be analyzed is recorded;

[0174] If the text to be analyzed does not have any expression meaning, the text to be analyzed will be deleted.

[0175] The step of obtaining the meaning of the standard text expression according to the classification result of the text to be analyzed includes:

[0176] Obtain the expression meanings of all recorded texts to be analyzed, and form an expression meaning set of standard texts;

[0177] Selecting one expression meaning from the expression meaning set of the standard text as the first expression meaning, and sending it to the user through the user selection module;

[0178] Obtaining the user's judgment on the meaning of the first expression;

[0179] If the user makes a correct judgment on the first expression meaning, the first expression meaning is used as the expression meaning of the standard text, and other expression meanings in the expression meaning set are deleted;

[0180] If the user makes an incorrect judgment on the first expression meaning, the first expression meaning is deleted from the expression meaning set;

[0181] Selecting another expression meaning from the expression meaning set, recording it as a second expression meaning, and sending it to the user through the user selection module again;

[0182] Obtaining the user's judgment on the meaning of the second expression;

[0183] If the user makes a correct judgment on the second expression meaning, the second expression meaning is used as the expression meaning of the standard text, and other expression meanings in the expression meaning set are deleted;

[0184] If the user makes an incorrect judgment on the second expression meaning, the second expression meaning is deleted from the expression meaning set;

[0185] Selecting another expression meaning from the expression meaning set, recording it as a third expression meaning, and sending it to the user through the user selection module again;

[0186] Obtaining the user's judgment on the meaning of the third expression;

[0187] If the user makes a correct judgment on the third expression meaning, the third expression meaning is used as the expression meaning of the standard text, and other expression meanings in the expression meaning set are deleted;

[0188] If the user makes an incorrect judgment on the third expression meaning, the third expression meaning is deleted from the expression meaning set, and the fourth expression meaning is selected again and sent to the user. Based on the user's judgment, it is determined whether to continue to select expression meanings from the expression meaning set and send them to the user, until all expression meanings in the expression meaning set are traversed, and the expression meaning of the standard text is obtained based on the user's judgment. By introducing a user feedback mechanism, it is helpful to adjust the analysis results based on the user's judgment and opinions, so that the expression meaning of the standard text finally determined can better meet the user's personalized needs and make the analysis results more in line with the actual human expression.

[0189] Embodiment 2, a system for implementing the intelligent semantic analysis based on the data big model, comprising: referring to Figure 1 and Figure 2 :

[0190] Audio processing module: processes the audio to be processed and converts the audio to be processed into text to be processed;

[0191] Text segmentation module: divides the text into character groups and word groups according to the text content;

[0192] Text processing module: number the texts and process them according to the numbers to generate standard texts;

[0193] Audio analysis module: analyzes the region to which the language in the audio to be processed belongs based on the audio to be processed;

[0194] Data acquisition module: obtains the dialect database of the marked area through the big data model;

[0195] Dialect marking module: obtains the dialect text content in the standard text and marks it;

[0196] Combination module: record any combination of marked dialects;

[0197] Meaning analysis module: determines whether the text to be analyzed has expression meaning;

[0198] Classification module: classifies the text to be analyzed according to the semantic analysis results of the text to be analyzed;

[0199] User selection module: provides the user with the expression meaning of the recorded standard text for the user to choose;

[0200] Data recording module: stores standard texts that express the meaning of records.

[0201] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0202] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. An intelligent semantic analysis method based on a large data model, characterized in that: include: Obtain the audio file to be analyzed and record it as the audio to be processed; According to the content of the audio to be processed, the audio to be processed is converted into a text form and recorded as the text to be processed; Analyze the text to be processed and divide the text content in the text to be processed into character groups and word groups; Number the word groups and phrase groups separately; According to the number, the text to be processed is processed to obtain repeated text segments and repeated word segments; A repeated word deep learning model is obtained through data training; Through the repeated word deep learning model, a first processing operation and a second processing operation are respectively performed on the repeated text segment and the repeated word segment; Recording the to-be-processed text that has completed the first processing operation and the second processing operation as a standard text; Analyze the audio to be processed, obtain the region to which the language in the audio to be processed belongs, and record it as the marked region; Obtain all dialect databases existing in the marked area, recorded as marked database; According to the marked database, the dialect text content existing in the standard text is obtained and recorded as the marked dialect; Get the number of marked dialects; The marked dialects are combined according to the quantity, and the combined marked dialects are substituted into the marked text to obtain the text to be analyzed; Perform semantic analysis on the text to be analyzed, and classify the text to be analyzed according to the results of the semantic analysis; According to the classification results of the text to be analyzed, the meaning of the standard text expression is obtained.

2. According to claim 1, the intelligent semantic analysis method based on the data big model is characterized in that: The step of analyzing the text to be processed and dividing the text contents in the text to be processed into text groups includes: Traverse the characters in the text to be processed; Obtain different characters in the text to be processed and classify the same characters into the same character group; Get the words in the text to be processed and group the same words into the same word group.

3. According to claim 2, the intelligent semantic analysis method based on the data big model is characterized in that: The step of numbering the character groups and the word groups respectively comprises: Get the divided text groups, and number each text group in sequence; The characters in the same character group are numbered uniformly, and the number is consistent with the number of the character group; Obtain the divided word groups, and number each word group in alphabetical order; The words in the same word group are numbered uniformly, and the number is consistent with the number of the word group to which the word belongs.

4. The intelligent semantic analysis method based on a data big model according to claim 3 is characterized in that: The text to be processed is processed according to the number to obtain repeated text segments and repeated word segments, including: Traverse the text to be processed according to the text number; Obtaining continuous text content with the same number in the text to be processed, dividing the text content into a text segment, and recording it as a repeated text segment; Get all repeated text segments in the text to be processed; Traverse the text to be processed according to the word number; Obtaining continuous word contents with the same number in the text to be processed, dividing the word contents into a word segment, and recording it as a repeated word segment; Get all repeated word segments in the text to be processed.

5. The intelligent semantic analysis method based on a data big model according to claim 4 is characterized in that: The method of performing a first processing operation and a second processing operation on a repeated text segment and a repeated word segment respectively through a repeated word deep learning model includes: Input the text to be processed into the repeated word deep learning model to determine the meaning of the repeated word segments in the text to be processed; If the repeated text segment does not affect the meaning of the text to be processed, the repeated text segment is retained; If the repeated text segment does not have any expression meaning in the text to be processed, then mark the repeated text segment; Traverse all repeated text segments and obtain all marked repeated text segments; For each marked repeated text segment, one character is retained, and the remaining characters with the same number in the repeated text segment are deleted, thereby completing the first processing operation; Input the text to be processed into the repeated word deep learning model to determine the meaning of the repeated word segments in the text to be processed; If the repeated word segment does not affect the meaning of the text to be processed, the repeated word segment is retained; If the repeated word segment does not have any expression meaning in the text to be processed, then mark the repeated word segment; Traverse all repeated word segments and obtain all marked repeated word segments; For each marked repeated word segment, one word is retained, and the remaining words with the same number in the repeated word segment are deleted, completing the second processing operation.

6. The intelligent semantic analysis method based on a data big model according to claim 5 is characterized in that: The marking dialects are combined according to the quantity, and the combined marking dialects are substituted into the marking text to obtain the text to be analyzed, including: If there is no marked dialect in the standard text, the standard text is used as the text to be analyzed; If there are marked dialects in the standard text, the number of marked dialects is obtained, recorded as N; Combine the marked dialects N times: In the first combination, N marked dialects are selected for combination; Among the marked dialects, select N marked dialects at random until all combinations of N marked dialects are traversed; Record the selected marked dialect each time, obtain the expression meaning of each marked dialect, and substitute it into the standard text as the text to be analyzed; In the second combination, N-1 marked dialects are selected for combination; Among the marked dialects, select N-1 marked dialects at random until all combinations of N-1 marked dialects are traversed; Record the selected marked dialect each time, obtain the expression meaning of each marked dialect, and substitute it into the standard text as the text to be analyzed; In the third combination, N-2 marked dialects are selected for combination; Among the marked dialects, select N-2 marked dialects at random until all combinations of N-2 marked dialects are traversed; Record the selected marked dialect each time, obtain the expression meaning of each marked dialect, and substitute it into the standard text as the text to be analyzed; For the Nth combination, one marked dialect is selected for combination; Among the marked dialects, select any one of them, until all marked dialects are traversed; The selected marked dialect each time is recorded, the expression meaning of each marked dialect is obtained, and it is substituted into the standard text as the text to be analyzed.

7. The intelligent semantic analysis method based on a data big model according to claim 6 is characterized in that: The performing semantic analysis on the text to be analyzed and classifying the text to be analyzed according to the semantic analysis result includes: Get all the text to be analyzed; Analyze the expression meaning of each text to be analyzed in turn; If the text to be analyzed has an expression meaning, the text to be analyzed is marked and the expression meaning of the text to be analyzed is recorded; If the text to be analyzed does not have any expression meaning, the text to be analyzed will be deleted.

8. The intelligent semantic analysis method based on a data big model according to claim 7 is characterized in that: The step of obtaining the meaning of the standard text expression according to the classification result of the text to be analyzed includes: Obtain the expression meanings of all recorded texts to be analyzed, and form an expression meaning set of standard texts; Selecting one expression meaning from the expression meaning set of the standard text as the first expression meaning, and sending it to the user through the user selection module; Obtaining the user's judgment on the meaning of the first expression; If the user makes a correct judgment on the first expression meaning, the first expression meaning is used as the expression meaning of the standard text, and other expression meanings in the expression meaning set are deleted; If the user makes an incorrect judgment on the first expression meaning, the first expression meaning is deleted from the expression meaning set; Selecting another expression meaning from the expression meaning set, recording it as a second expression meaning, and sending it to the user through the user selection module again; Obtaining the user's judgment on the meaning of the second expression; If the user makes a correct judgment on the second expression meaning, the second expression meaning is used as the expression meaning of the standard text, and other expression meanings in the expression meaning set are deleted; If the user makes an incorrect judgment on the second expression meaning, the second expression meaning is deleted from the expression meaning set; Selecting another expression meaning from the expression meaning set, recording it as a third expression meaning, and sending it to the user through the user selection module again; Obtaining the user's judgment on the meaning of the third expression; If the user makes a correct judgment on the third expression meaning, the third expression meaning is used as the expression meaning of the standard text, and other expression meanings in the expression meaning set are deleted; If the user makes an incorrect judgment on the third expression meaning, the third expression meaning will be deleted from the expression meaning set, and the fourth expression meaning will be selected again and sent to the user. Based on the user's judgment, it will be determined whether to continue selecting expression meanings from the expression meaning set and sending them to the user, until all expression meanings in the expression meaning set are traversed, and the expression meaning of the standard text is obtained based on the user's judgment.

9. A system using the intelligent semantic analysis method based on a large data model as claimed in claim 1, characterized in that: include: Audio processing module: processes the audio to be processed and converts the audio to be processed into text to be processed; Text segmentation module: divides the text into character groups and word groups according to the text content; Text processing module: number the text, process it according to the number, and generate standard text; Audio analysis module: analyzes the region to which the language in the audio to be processed belongs based on the audio to be processed; Data acquisition module: obtains the dialect database of the marked area through the big data model; Dialect marking module: obtains the dialect text content in the standard text and marks it; Combination module: record any combination of marked dialects; Meaning analysis module: determines whether the text to be analyzed has expression meaning; Classification module: classifies the text to be analyzed according to the semantic analysis results of the text to be analyzed; User selection module: provides the user with the expression meaning of the recorded standard text for the user to choose; Data recording module: stores standard texts that express the meaning of records.

Images