Enterprise conversation data storage and analysis method and system based on cloud computing

Through the cloud-based enterprise session data storage and analysis method, the problem of difficult to extract effective information in large amounts of enterprise session data is solved, efficient storage and in-depth analysis are achieved, manual analysis costs are reduced, and understanding of customer needs is enhanced.

CN118278420BActive Publication Date: 2024-11-12WESHINE CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410465975.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-18
Publication Date
2024-11-12
Estimated Expiration
2044-04-18

AI Technical Summary

Technical Problem

It is difficult to extract effective information from massive texts of enterprise session data, and manual analysis is costly and storage efficiency is inefficient.

Method used

Adopt cloud-based enterprise session data storage analysis methods, including session data collection, elastic storage, format cleaning, word segmentation, part-of-speech annotation, entity recognition, entity link extraction and deep semantic extraction, and extracted through multimodal fusion analysis and semantic analysis Key information.

Benefits of technology

It realizes efficient storage and in-depth analysis of enterprise session data, reduces data noise and redundancy, improves information quality and availability, reduces manual analysis costs, and enhances understanding of customer needs and intentions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118278420B_ABST
    Figure CN118278420B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of data processing technology, and in particular to a method and system for storing and analyzing enterprise conversation data based on cloud computing. The method comprises the following steps: performing conversation data collection and elastic storage processing on an enterprise's internal communication tool to obtain enterprise conversation elastically stored text information data; performing formatting and cleaning processing and conversation entity recognition analysis on the enterprise conversation elastically stored text information data to obtain enterprise conversation text entity information data; performing entity link extraction processing and deep semantic extraction analysis on the enterprise conversation text entity information data to obtain enterprise conversation text deep semantic information data; performing conversation suggestion analysis on enterprise conversation text information standard data according to the enterprise conversation text deep semantic information data to obtain enterprise conversation text content reply suggestion data. The present invention can provide data guarantee for the storage and analysis of enterprise conversation data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for storing and analyzing enterprise conversation data based on cloud computing. Background Art

[0002] With the advent of the digital age, the way companies, organizations and individuals communicate has undergone tremendous changes. Many communication and exchange activities have changed from traditional face-to-face meetings and telephone exchanges to conversations based on the Internet and digital platforms. These conversations generate a large amount of text and data, which contains a lot of valuable information, such as user preferences, market trends, customer feedback, etc. However, it is often difficult to directly extract effective information from these massive amounts of text data, and the cost of manual analysis is high, while the storage efficiency is low. Summary of the invention

[0003] Based on this, it is necessary for the present invention to provide a cloud computing-based enterprise conversation data storage and analysis method and system to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a cloud computing-based enterprise conversation data storage and analysis method includes the following steps:

[0005] Step S1: collecting and processing conversation data of the enterprise's internal communication tools to obtain enterprise multimodal conversation text information data; elastically storing the enterprise multimodal conversation text information data through a cloud computing platform to obtain enterprise conversation elastically stored text information data;

[0006] Step S2: Formatting and cleaning the enterprise conversation elastically stored text information data to obtain enterprise conversation text information standard data; performing text segmentation processing on the enterprise conversation text information standard data to obtain an enterprise conversation text information segmentation set; performing part-of-speech tagging processing on the enterprise conversation text information segmentation set to obtain an enterprise conversation text segmentation part-of-speech set;

[0007] Step S3: performing conversation entity recognition analysis on the enterprise conversation text information standard data to obtain enterprise conversation text entity information data; performing entity link extraction processing on the enterprise conversation text entity information data based on the enterprise conversation text information segmentation set and the enterprise conversation text segmentation part-of-speech set to obtain enterprise conversation text entity link relationship data;

[0008] Step S4: Based on the enterprise conversation text entity link relationship data, deep semantic extraction analysis is performed on the enterprise conversation text information standard data to obtain the enterprise conversation text deep semantic information data; based on the enterprise conversation text deep semantic information data, conversation suggestion analysis is performed on the enterprise conversation text information standard data to obtain the enterprise conversation text content reply suggestion data.

[0009] The present invention firstly collects and processes the conversation data of the internal communication tools of the enterprise, and can integrate and analyze the collected information in different forms, so as to obtain more comprehensive and in-depth enterprise multimodal conversation text information data. This analysis and processing process can help enterprises discover potential communication problems, trends and patterns, so as to provide more accurate and comprehensive data support for subsequent processing processes. At the same time, by using the cloud computing platform to perform elastic storage processing on the enterprise multimodal conversation text information data, the enterprise's conversation data can be safely stored in the cloud to achieve high availability and elastic expansion of the conversation data. Such a storage method can not only ensure the security and stability of the conversation data, but also improve the storage efficiency of the conversation data, so as to provide reliable technical support and guarantee for the enterprise information construction. Secondly, by formatting and cleaning the elastic storage text information data of the enterprise conversation, the consistency and standardization of the data can be ensured to make it conform to the unified data format and specification, which includes removing special characters, punctuation marks, HTML tags, etc. in the text, uniformly converting the uppercase and lowercase formats of the text, and processing duplicate data, noise and content format errors, etc. Through formatting and cleaning processing, the noise and redundant information of the data can be reduced, the quality and availability of the data can be improved, so as to provide a reliable data basis for subsequent text analysis and processing. By performing text segmentation on the standard data of enterprise conversation text information, the text can be split into independent vocabulary units for subsequent semantic analysis and information extraction. Through word segmentation, complex text data can be converted into a structured information set, thus providing a basis for subsequent text analysis and mining. In addition, by performing part-of-speech tagging on the enterprise conversation text information word set, the purpose is to add corresponding part-of-speech tags to each vocabulary unit, such as nouns, verbs, adjectives, etc. This part-of-speech tagging process can help understand the grammatical and semantic relationships of words in the text and analyze the meaning and structure of the text more deeply. Through part-of-speech tagging, the ability to understand and mine text data can be further improved, thereby providing support for enterprises to obtain more valuable information and insights from text data. Then, by performing conversation entity recognition analysis on the standard data of enterprise conversation text information, entities can be identified from enterprise conversation texts, including people, places, organizations, time, etc., and relevant information can be extracted, which helps enterprises better understand the content and context of the conversation, thereby providing a basis for subsequent analysis and decision-making. By identifying entities in the conversation, enterprises can have a deeper understanding of the key topics and participants involved in the conversation, which helps to discover potential business insights and opportunities. By performing entity link extraction on the enterprise conversation text entity information data based on the enterprise conversation text information segmentation set and the enterprise conversation text segmentation part-of-speech set, that is, identifying and establishing the association relationship between named entities, the same entities mentioned in the conversation can be linked, thereby establishing a relationship network between entities.Through entity link extraction processing, enterprises can better understand the relationship and connection between various entities involved in the conversation, thereby providing a more comprehensive perspective and information support for further analysis and application. Finally, deep semantic extraction and analysis of enterprise conversation text information standard data is performed to reveal the deep semantics hidden behind the conversation text. Through deep semantic extraction and analysis, enterprises can better understand the implicit information and potential meanings in the conversation, provide support for deeper data mining and intelligent applications, and thus extract key effective information from enterprise conversation data. In addition, by performing conversation suggestion analysis on enterprise conversation text information standard data based on enterprise conversation text deep semantic information data, enterprises can provide suggestions and response plans for conversation content, which helps enterprises better respond to user needs and intentions, improve customer satisfaction and experience quality, and provide enterprises with more intelligent customer service and communication support, thereby reducing manual analysis costs.

[0010] Preferably, the present invention further provides a cloud computing-based enterprise conversation data storage and analysis system, which is used to execute the cloud computing-based enterprise conversation data storage and analysis method as described above. The cloud computing-based enterprise conversation data storage and analysis system comprises:

[0011] The enterprise internal conversation data elastic storage processing module is used to collect and process the conversation data of the enterprise internal communication tools to obtain the enterprise multimodal conversation text information data; the enterprise multimodal conversation text information data is elastically stored through the cloud computing platform to obtain the enterprise conversation elastic storage text information data;

[0012] The enterprise conversation text information word segmentation processing module is used to format and clean the enterprise conversation elastic storage text information data to obtain the enterprise conversation text information standard data; perform text segmentation processing on the enterprise conversation text information standard data to obtain the enterprise conversation text information word segmentation set; perform part-of-speech tagging processing on the enterprise conversation text information word segmentation set to obtain the enterprise conversation text word segmentation part-of-speech set;

[0013] The enterprise conversation text entity relationship extraction module is used to perform conversation entity recognition analysis on the enterprise conversation text information standard data to obtain the enterprise conversation text entity information data; perform entity link extraction processing on the enterprise conversation text entity information data based on the enterprise conversation text information segmentation set and the enterprise conversation text segmentation part of speech set, thereby obtaining the enterprise conversation text entity link relationship data;

[0014] The enterprise conversation text semantic suggestion analysis module is used to perform deep semantic extraction and analysis on the enterprise conversation text information standard data based on the enterprise conversation text entity link relationship data to obtain the enterprise conversation text deep semantic information data; and perform conversation suggestion analysis on the enterprise conversation text information standard data based on the enterprise conversation text deep semantic information data to obtain the enterprise conversation text content reply suggestion data.

[0015] In summary, the present invention provides an enterprise conversation data storage and analysis system based on cloud computing. The enterprise conversation data storage and analysis system based on cloud computing consists of an enterprise internal conversation data elastic storage processing module, an enterprise conversation text information word segmentation processing module, an enterprise conversation text entity relationship extraction module, and an enterprise conversation text semantic suggestion analysis module. It can implement any one of the enterprise conversation data storage and analysis methods based on cloud computing described in the present invention, and is used to combine the operations between computer programs running on each module to implement any one of the enterprise conversation data storage and analysis methods based on cloud computing. The internal structures of the system cooperate with each other, which can greatly reduce duplication of work and manpower investment, and can quickly and effectively provide a more accurate and efficient enterprise conversation data storage and analysis process, thereby simplifying the operation process of the enterprise conversation data storage and analysis system based on cloud computing. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0017] Figure 1 A schematic diagram of the steps of the enterprise session data storage and analysis method based on cloud computing of the present invention;

[0018] Figure 2 for Figure 1 Detailed step flow diagram of step S1;

[0019] Figure 3 for Figure 2 Detailed step flow chart of step S14 in FIG. DETAILED DESCRIPTION

[0020] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0021] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0022] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0023] To achieve this, please refer to Figures 1 to 3 The present invention provides a method for storing and analyzing enterprise conversation data based on cloud computing, comprising the following steps:

[0024] Step S1: collecting and processing conversation data of the enterprise's internal communication tools to obtain enterprise multimodal conversation text information data; elastically storing the enterprise multimodal conversation text information data through a cloud computing platform to obtain enterprise conversation elastically stored text information data;

[0025] Step S2: Formatting and cleaning the enterprise conversation elastically stored text information data to obtain enterprise conversation text information standard data; performing text segmentation processing on the enterprise conversation text information standard data to obtain an enterprise conversation text information segmentation set; performing part-of-speech tagging processing on the enterprise conversation text information segmentation set to obtain an enterprise conversation text segmentation part-of-speech set;

[0026] Step S3: performing conversation entity recognition analysis on the enterprise conversation text information standard data to obtain enterprise conversation text entity information data; performing entity link extraction processing on the enterprise conversation text entity information data based on the enterprise conversation text information segmentation set and the enterprise conversation text segmentation part-of-speech set to obtain enterprise conversation text entity link relationship data;

[0027] Step S4: Based on the enterprise conversation text entity link relationship data, deep semantic extraction analysis is performed on the enterprise conversation text information standard data to obtain the enterprise conversation text deep semantic information data; based on the enterprise conversation text deep semantic information data, conversation suggestion analysis is performed on the enterprise conversation text information standard data to obtain the enterprise conversation text content reply suggestion data.

[0028] In the embodiment of the present invention, please refer to Figure 1 FIG. 1 is a schematic diagram of the steps of the enterprise conversation data storage and analysis method based on cloud computing of the present invention. In this example, the enterprise conversation data storage and analysis method based on cloud computing includes the following steps:

[0029] Step S1: collecting and processing conversation data of the enterprise's internal communication tools to obtain enterprise multimodal conversation text information data; elastically storing the enterprise multimodal conversation text information data through a cloud computing platform to obtain enterprise conversation elastically stored text information data;

[0030] The embodiment of the present invention determines the communication tools used within the enterprise (such as email systems, instant messaging software (such as WeChat, etc.), video conferencing systems, etc.), and automatically obtains and records the text, image and audio communication content between employees in the communication tools by setting a suitable data collection program or interface, and converts the analyzed text, image and audio communication content into text content, and uses corresponding analysis methods (such as text feature extraction, similarity calculation, association rule mining algorithm and other technical means) to perform association mining analysis on the analyzed text content, image text content and audio text content to explore the association relationship between different modal data, such as between text and image, between text and audio, between image and audio, and between text, image and audio, etc. Secondly, by combining the text content association relationship obtained by analysis, multimodal fusion technology (such as ensemble learning, deep fusion network and other algorithms) is used to integrate and fuse the text text content, image text content and audio text content corresponding to the association relationship, so as to effectively integrate the text, image and audio data, and provide more comprehensive and accurate conversation text information, thereby obtaining enterprise multimodal conversation text information data. Then, by analyzing the corresponding storage requirements and access patterns of the enterprise's multimodal conversation text information data, and elastically expanding the storage capacity space on the cloud computing platform, the corresponding enterprise multimodal conversation text information data can meet its cloud storage service needs, thereby safely storing the enterprise's conversation data in the cloud, and achieving high availability and elastic expansion of the conversation data, and finally obtaining the enterprise's elastically stored text information data.

[0031] Step S2: Formatting and cleaning the enterprise conversation elastically stored text information data to obtain enterprise conversation text information standard data; performing text segmentation processing on the enterprise conversation text information standard data to obtain an enterprise conversation text information segmentation set; performing part-of-speech tagging processing on the enterprise conversation text information segmentation set to obtain an enterprise conversation text segmentation part-of-speech set;

[0032] The embodiment of the present invention first identifies and extracts sensitive information in the elastically stored text information data of the enterprise conversation, such as personal identity information, sensitive words, etc., and replaces the identified sensitive information with a virtual identifier by using data desensitization technology to ensure that the data reduces its sensitivity while maintaining availability. At the same time, the desensitized text information data is deduplicated by using a corresponding deduplication algorithm to eliminate repeated content in the text information data and retain unique data entries, and the deduplication text information data is de-noised to identify and remove noise data in the text, such as errors, redundant or irrelevant information. The semantic content is checked and grammatical errors are corrected by using natural language processing technology, including identifying and correcting grammatical errors, logical errors and other semantic errors in the text, so as to further improve the accuracy and comprehensibility of the data and ensure that errors and incomplete information in the text are corrected. The corrected text information data is processed by using formatting processing technology, including removing special characters, punctuation marks, HTML tags in the text, adjusting the structure, typesetting and marking of the text, etc., to meet specific data format requirements and make it conform to the standards of enterprise conversation text information, thereby obtaining enterprise conversation text information standard data. Secondly, by constructing a suitable enterprise conversation text segmenter, the corresponding enterprise conversation text information standard data is segmented to divide the text information into independent words or phrases, and the complex text data is converted into a structured information set, thereby obtaining the enterprise conversation text information segmentation set. Then, by using the part-of-speech tagging method to process each segmentation vocabulary in the enterprise conversation text information segmentation set, the part of speech or grammatical role of each vocabulary in the enterprise conversation text information segmentation set in the sentence is determined, such as noun, verb, adjective, etc., and the grammatical structure and meaning of the text are further analyzed and understood, and finally the enterprise conversation text segmentation part-of-speech set is obtained.

[0033] Step S3: performing conversation entity recognition analysis on the enterprise conversation text information standard data to obtain enterprise conversation text entity information data; performing entity link extraction processing on the enterprise conversation text entity information data based on the enterprise conversation text information segmentation set and the enterprise conversation text segmentation part-of-speech set to obtain enterprise conversation text entity link relationship data;

[0034] The embodiment of the present invention analyzes the processed enterprise conversation text information standard data by using an entity recognition algorithm (such as NER) to identify entities from the enterprise conversation text, including people, places, organizations, time, etc., and extracts important information units from the conversation text, namely, entity information data, thereby obtaining enterprise conversation text entity information data. At the same time, the enterprise conversation text entity information data obtained by analysis is screened and processed by using pre-defined named entity types (such as personal names, organization names, place names, etc.) to remove those entities that do not meet the analysis objectives, and further focus on the key information in the conversation, so as to screen out the corresponding named entities. Secondly, each named entity analyzed is analyzed by combining the enterprise conversation text information word set and the enterprise conversation text word segmentation part-of-speech set to analyze, identify and understand the surrounding word segmentation semantic information associated with these named entities, and by using the association rule mining algorithm to perform association analysis on the surrounding word segmentation semantic information and each named entity that appears, so as to analyze the semantic relevance between the named entity and its surrounding text. Then, by combining the analyzed semantic relevance, the link relationship extraction technology is used to analyze the association link relationship between entities in the enterprise conversation text entity information data, so as to analyze, identify and establish direct or indirect association link relationships between entities, such as co-occurrence relationships, semantic similarity relationships, etc., and finally obtain the enterprise conversation text entity link relationship data.

[0035] Step S4: Based on the enterprise conversation text entity link relationship data, deep semantic extraction analysis is performed on the enterprise conversation text information standard data to obtain the enterprise conversation text deep semantic information data; based on the enterprise conversation text deep semantic information data, conversation suggestion analysis is performed on the enterprise conversation text information standard data to obtain the enterprise conversation text content reply suggestion data.

[0036] The embodiment of the present invention analyzes the enterprise conversation text information standard data by using a semantic relationship extraction algorithm to identify the semantic relationships between different entities from the conversation text, such as the relationship between characters, the relationship between events and places, etc., and constructs a knowledge graph for the semantic information in the enterprise conversation text information standard data by combining the enterprise conversation text entity link relationship data and the enterprise conversation text semantic relationship data obtained by the analysis, so as to represent the entities and corresponding relationships in the enterprise conversation text information standard data in the form of a graph, thereby constructing a structured semantic knowledge graph. At the same time, by using a semantic understanding model based on deep learning (such as BERT, GPT, etc.), the text semantic information in the constructed semantic knowledge graph is analyzed to further mine and understand the hidden information and semantic associations in the text, and reveal the deep semantics hidden behind the conversation text, thereby obtaining deep semantic information data of the enterprise conversation text. Then, the deep semantic information data of the enterprise conversation text obtained by analysis is analyzed by using machine learning (such as convolutional neural networks, etc.) and data mining technology to predict the user's behavior and intention based on the enterprise conversation text content and semantic information, thereby predicting the response, operation or decision that the user may need, and performing suggestion analysis on the corresponding enterprise conversation text information standard data by combining the user behavior prediction data obtained by analysis, so as to provide appropriate reply suggestions for user conversation messages according to the predicted user behavior and intention, so as to optimize the conversation experience and meet the needs of users, and finally obtain enterprise conversation text content reply suggestion data.

[0037] The present invention firstly collects and processes the conversation data of the internal communication tools of the enterprise, and can integrate and analyze the collected information in different forms, so as to obtain more comprehensive and in-depth enterprise multimodal conversation text information data. This analysis and processing process can help enterprises discover potential communication problems, trends and patterns, so as to provide more accurate and comprehensive data support for subsequent processing processes. At the same time, by using the cloud computing platform to perform elastic storage processing on the enterprise multimodal conversation text information data, the enterprise's conversation data can be safely stored in the cloud to achieve high availability and elastic expansion of the conversation data. Such a storage method can not only ensure the security and stability of the conversation data, but also improve the storage efficiency of the conversation data, so as to provide reliable technical support and guarantee for the enterprise information construction. Secondly, by formatting and cleaning the elastic storage text information data of the enterprise conversation, the consistency and standardization of the data can be ensured to make it conform to the unified data format and specification, which includes removing special characters, punctuation marks, HTML tags, etc. in the text, uniformly converting the uppercase and lowercase formats of the text, and processing duplicate data, noise and content format errors, etc. Through formatting and cleaning processing, the noise and redundant information of the data can be reduced, the quality and availability of the data can be improved, so as to provide a reliable data basis for subsequent text analysis and processing. By performing text segmentation on the standard data of enterprise conversation text information, the text can be split into independent vocabulary units for subsequent semantic analysis and information extraction. Through word segmentation, complex text data can be converted into a structured information set, thus providing a basis for subsequent text analysis and mining. In addition, by performing part-of-speech tagging on the enterprise conversation text information word set, the purpose is to add corresponding part-of-speech tags to each vocabulary unit, such as nouns, verbs, adjectives, etc. This part-of-speech tagging process can help understand the grammatical and semantic relationships of words in the text and analyze the meaning and structure of the text more deeply. Through part-of-speech tagging, the ability to understand and mine text data can be further improved, thereby providing support for enterprises to obtain more valuable information and insights from text data. Then, by performing conversation entity recognition analysis on the standard data of enterprise conversation text information, entities can be identified from enterprise conversation texts, including people, places, organizations, time, etc., and relevant information can be extracted, which helps enterprises better understand the content and context of the conversation, thereby providing a basis for subsequent analysis and decision-making. By identifying entities in the conversation, enterprises can have a deeper understanding of the key topics and participants involved in the conversation, which helps to discover potential business insights and opportunities. By performing entity link extraction on the enterprise conversation text entity information data based on the enterprise conversation text information segmentation set and the enterprise conversation text segmentation part-of-speech set, that is, identifying and establishing the association relationship between named entities, the same entities mentioned in the conversation can be linked, thereby establishing a relationship network between entities.Through entity link extraction processing, enterprises can better understand the relationship and connection between various entities involved in the conversation, thereby providing a more comprehensive perspective and information support for further analysis and application. Finally, deep semantic extraction and analysis of enterprise conversation text information standard data is performed to reveal the deep semantics hidden behind the conversation text. Through deep semantic extraction and analysis, enterprises can better understand the implicit information and potential meanings in the conversation, provide support for deeper data mining and intelligent applications, and thus extract key effective information from enterprise conversation data. In addition, by performing conversation suggestion analysis on enterprise conversation text information standard data based on enterprise conversation text deep semantic information data, enterprises can provide suggestions and response plans for conversation content, which helps enterprises better respond to user needs and intentions, improve customer satisfaction and experience quality, and provide enterprises with more intelligent customer service and communication support, thereby reducing manual analysis costs.

[0038] Preferably, step S1 comprises the following steps:

[0039] Step S11: collecting and processing conversation text data from the enterprise's internal communication tool to obtain text information data of the enterprise's internal conversation records;

[0040] Step S12: collecting and processing conversation image data of the enterprise's internal communication tool to obtain the enterprise's internal conversation record image information data;

[0041] Step S13: collecting and processing the conversation audio data of the enterprise internal communication tool to obtain the enterprise internal conversation record audio information data;

[0042] Step S14: performing multimodal fusion analysis on the text information data of the internal conversation record of the enterprise, the image information data of the internal conversation record of the enterprise, and the audio information data of the internal conversation record of the enterprise to obtain the enterprise multimodal conversation text information data;

[0043] Step S15: elastically storing the enterprise multimodal conversation text information data through the cloud computing platform to obtain elastically stored enterprise conversation text information data.

[0044] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed step flow diagram of step S1 in FIG. 1 , in this embodiment, step S1 includes the following steps:

[0045] Step S11: collecting and processing conversation text data from the enterprise's internal communication tool to obtain text information data of the enterprise's internal conversation records;

[0046] The embodiment of the present invention determines the communication tools used within the enterprise (such as email systems, instant messaging software (such as WeChat, etc.), video conferencing systems, etc.), and automatically obtains and records the text communication content between employees in the communication tools, including text messages, emails, etc., by setting up appropriate data collection programs or interfaces, and finally obtains text information data of internal conversation records of the enterprise.

[0047] Step S12: collecting and processing conversation image data of the enterprise's internal communication tool to obtain the enterprise's internal conversation record image information data;

[0048] The embodiment of the present invention uses an image capture interface to collect images uploaded by users on internal enterprise communication tools in real time, so as to capture and record image exchanges between employees, such as pictures, emoticons, meeting screenshots, screen shots, avatars, etc., and finally obtain image information data of internal enterprise conversation records.

[0049] Step S13: collecting and processing the conversation audio data of the enterprise internal communication tool to obtain the enterprise internal conversation record audio information data;

[0050] The embodiment of the present invention uses audio recording equipment or software to collect voice exchanges between users on internal communication tools of an enterprise in real time, so as to record voice audio streams between employees in real time, such as telephone calls, voice messages, voice chats, etc., and finally obtain audio information data of internal conversation records of the enterprise.

[0051] Step S14: performing multimodal fusion analysis on the text information data of the internal conversation record of the enterprise, the image information data of the internal conversation record of the enterprise, and the audio information data of the internal conversation record of the enterprise to obtain the enterprise multimodal conversation text information data;

[0052] The embodiment of the present invention mines and analyzes the text information data of the enterprise internal conversation record by using natural language processing technology, including keyword extraction, sentiment analysis, theme recognition and other technologies, to help obtain the key information content in the conversation communication, and extracts the key features of the image information data of the enterprise internal conversation record by using computer vision technology (such as color histogram, texture features, shape features, etc.), so as to extract the key features of the image, such as color, shape, texture, etc. At the same time, the corresponding enterprise internal conversation record image information data is analyzed by using an image content recognition method to identify the text content that the user wants to express in the image information data, such as identifying conference room photos, electronic whiteboard content, etc. Secondly, the feature extraction of the audio information data of the enterprise internal conversation record is performed by using signal processing technology to extract the feature information such as spectrum features and sound intensity in the audio information, and the corresponding enterprise internal conversation record audio information is recognized by using speech recognition technology by combining the audio features obtained by analysis. The information data is analyzed to convert these features into recognizable audio text content information data, such as identifying the specific text content of conference recordings or telephone calls. Then, the text content, image text content and audio text content obtained by the analysis are subjected to association mining analysis by using corresponding analysis methods (such as text feature extraction, similarity calculation, association rule mining algorithm and other technical means) to explore the association relationship between different modal data, such as between text and image, between text and audio, between image and audio, and between text, image and audio. Finally, by combining the text content association relationship obtained by the analysis, multimodal fusion technology (such as ensemble learning, deep fusion network and other algorithms) is used to integrate and fuse the text content, image text content and audio text content with corresponding association relationship, so as to effectively integrate the text, image and audio data, and provide more comprehensive and accurate conversation text information, so as to finally obtain the enterprise multimodal conversation text information data.

[0053] Step S15: elastically storing the enterprise multimodal conversation text information data through the cloud computing platform to obtain elastically stored enterprise conversation text information data.

[0054] The embodiment of the present invention first analyzes the corresponding storage requirements and access patterns of enterprise multimodal conversation text information data, and then elastically expands the storage capacity space on the cloud computing platform by combining the storage requirements and access patterns obtained by the analysis, so that the corresponding enterprise multimodal conversation text information data can meet its cloud storage service requirements, thereby safely storing the enterprise's conversation data in the cloud, and achieving high availability and elastic expansion of the conversation data, and finally obtaining the enterprise conversation elastically stored text information data.

[0055] The present invention firstly collects and processes the conversation text data of the internal communication tools of the enterprise, which can help the enterprise to obtain and record the text communication content between employees in the communication tools, including text messages, emails, etc., so as to obtain complete internal text information of the enterprise. This conversation text data collection and processing process is helpful to supervise and audit the communication activities within the enterprise, improve the communication efficiency, and provide important decision-making basis and historical records for the subsequent processing process. Secondly, by collecting and processing the conversation image data of the internal communication tools of the enterprise, the image communication between employees, such as pictures, emoticons, etc., can be captured and recorded, thereby forming an image information data set within the enterprise. These image information can cover the content of meeting screenshots, screenshots, avatars, etc., providing a more vivid and intuitive communication record within the enterprise, which helps to enhance the understanding and accuracy of communication. Then, by collecting and processing the conversation audio data of the internal communication tools of the enterprise, the voice communication between employees, such as telephone calls, voice messages, etc., can be recorded, thereby forming an audio information data set within the enterprise. These audio recording data include conference recordings, telephone call recordings, etc., which help to trace back and audit the voice communication content within the enterprise, thereby providing basic data guarantee for the subsequent processing process. Next, by performing multimodal fusion analysis on the text information data, image information data, and audio information data of the internal conversation records of the enterprise, different forms of information can be integrated and analyzed to obtain more comprehensive and in-depth enterprise multimodal conversation text information data. This analysis and processing process can help enterprises discover potential communication problems, trends, and patterns, and provide more accurate and comprehensive data support for subsequent processing. Finally, by using the cloud computing platform to perform elastic storage processing on the enterprise's multimodal conversation text information data, the enterprise's conversation data can be securely stored in the cloud to achieve high availability and elastic expansion of conversation data. This storage method can not only ensure the security and stability of conversation data, but also facilitate enterprise data access and analysis, thereby providing reliable technical support and guarantee for enterprise information construction.

[0056] Preferably, step S14 comprises the following steps:

[0057] Step S141: performing text content mining analysis on the text information data of the internal conversation records of the enterprise to obtain the text content information data of the internal conversation records of the enterprise;

[0058] Step S142: performing image feature analysis on the image information data of the internal conversation record of the enterprise to obtain image feature data of the internal conversation record of the enterprise; performing image content recognition analysis on the image information data of the internal conversation record of the enterprise according to the image feature data of the internal conversation record of the enterprise to obtain image text content information data of the internal conversation record of the enterprise;

[0059] Step S143: performing audio feature analysis on the audio information data of the internal conversation record of the enterprise to obtain audio feature data of the internal conversation record of the enterprise; performing audio content recognition analysis on the audio information data of the internal conversation record of the enterprise according to the audio feature data of the internal conversation record of the enterprise to obtain audio text content information data of the internal conversation of the enterprise;

[0060] Step S144: performing multimodal association mining analysis on the text content information data of the internal enterprise conversation, the image text content information data of the internal enterprise conversation, and the audio text content information data of the internal enterprise conversation to obtain multimodal text content association relationship data of the enterprise conversation;

[0061] Step S145: Based on the enterprise conversation multimodal text content association relationship data, a multimodal fusion analysis is performed on the enterprise internal conversation text content information data, the enterprise internal conversation image text content information data, and the enterprise internal conversation audio text content information data to obtain the enterprise multimodal conversation text information data.

[0062] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 2 Detailed step flow diagram of step S14 in the embodiment, step S14 includes the following steps:

[0063] Step S141: performing text content mining analysis on the text information data of the internal conversation records of the enterprise to obtain the text content information data of the internal conversation records of the enterprise;

[0064] The embodiment of the present invention mines and analyzes the text information data of internal conversation records of an enterprise by using natural language processing technology, including keyword extraction, sentiment analysis, topic identification and other technologies, to help the enterprise discover the key information content in conversation communication and ultimately obtain the text content information data of the internal conversation of the enterprise.

[0065] Step S142: performing image feature analysis on the image information data of the internal conversation record of the enterprise to obtain image feature data of the internal conversation record of the enterprise; performing image content recognition analysis on the image information data of the internal conversation record of the enterprise according to the image feature data of the internal conversation record of the enterprise to obtain image text content information data of the internal conversation record of the enterprise;

[0066] The embodiment of the present invention uses computer vision technology (such as color histogram, texture features, shape features, etc.) to extract key features of the image, such as color, shape, texture, etc., to obtain the image feature data of the internal conversation record of the enterprise. Then, the image content recognition method is used to analyze the corresponding image information data of the internal conversation record of the enterprise by combining the image feature data of the internal conversation record of the enterprise obtained by analysis, so as to identify the text content that the user wants to express in the image information data, such as identifying the conference room photo, the content of the electronic whiteboard, etc., and finally obtain the text content information data of the internal conversation image of the enterprise.

[0067] Step S143: performing audio feature analysis on the audio information data of the internal conversation record of the enterprise to obtain audio feature data of the internal conversation record of the enterprise; performing audio content recognition analysis on the audio information data of the internal conversation record of the enterprise according to the audio feature data of the internal conversation record of the enterprise to obtain audio text content information data of the internal conversation of the enterprise;

[0068] The embodiment of the present invention uses signal processing technology to extract features from audio information data of internal enterprise conversation records, so as to extract feature information such as spectrum features and sound intensity in the audio information, thereby obtaining audio feature data of internal enterprise conversation records. Then, by combining the obtained audio feature data of internal enterprise conversation records with speech recognition technology, the corresponding audio information data of internal enterprise conversation records are analyzed to convert these features into recognizable audio text content information data, such as identifying the specific text content of conference recordings or telephone call content, and finally obtaining audio text content information data of internal enterprise conversations.

[0069] Step S144: performing multimodal association mining analysis on the text content information data of the internal enterprise conversation, the image text content information data of the internal enterprise conversation, and the audio text content information data of the internal enterprise conversation to obtain multimodal text content association relationship data of the enterprise conversation;

[0070] The embodiment of the present invention uses corresponding analysis methods (such as text feature extraction, similarity calculation, association rule mining algorithm and other technical means) to perform association mining analysis on the text content information data of internal enterprise conversations, the image text content information data of internal enterprise conversations, and the audio text content information data of internal enterprise conversations to explore the association relationship between different modal data, such as between text and image, between text and audio, between image and audio, and between text, image and audio, etc., and finally obtains multi-modal text content association relationship data of enterprise conversations.

[0071] Step S145: Based on the enterprise conversation multimodal text content association relationship data, a multimodal fusion analysis is performed on the enterprise internal conversation text content information data, the enterprise internal conversation image text content information data, and the enterprise internal conversation audio text content information data to obtain the enterprise multimodal conversation text information data.

[0072] The embodiment of the present invention uses multimodal fusion technology (such as integrated learning, deep fusion network and other algorithms) to integrate and fuse the internal enterprise conversation text content information data, internal enterprise conversation image text content information data and internal enterprise conversation audio text content information data with corresponding association relationships by combining the multimodal text content association relationship data of enterprise conversations obtained by analysis, so as to effectively integrate the text, image and audio data, and provide more comprehensive and accurate conversation text information, and finally obtain the enterprise multimodal conversation text information data.

[0073] The present invention firstly performs text content mining and analysis on the text information data of the internal conversation records of the enterprise, and can realize the deep understanding and mining of the text data, including the technologies of keyword extraction, sentiment analysis, theme identification, etc. This analysis can help the enterprise discover the key information content in the communication of employees, understand the needs and concerns of employees, and then optimize the internal communication mode and management strategy of the enterprise, thereby providing basic data guarantee for the subsequent association mining and analysis process. Secondly, by performing image feature analysis on the image information data of the internal conversation records of the enterprise, the key features of the image, such as color, shape, texture, etc., can be extracted. In addition, by performing image content recognition and analysis on the image information data of the internal conversation records of the enterprise according to the image feature data of the internal conversation records of the enterprise, the text content that the user wants to express specifically in the image information data is identified, such as identifying the conference room photos, the content of the electronic whiteboard, etc. This recognition and analysis process helps the enterprise understand the meaning and context of the image information between internal employees, thereby further strengthening the understanding and recording of the communication content. Then, by performing audio feature analysis on the audio information data of the internal conversation records of the enterprise, the sound features of the audio, such as spectrum, rhythm, intonation, etc., can be extracted. Subsequently, audio content recognition analysis is performed on the audio information data of the enterprise's internal conversation records based on the audio feature data of the enterprise's internal conversation records to convert these features into recognizable audio text content information data, such as identifying the specific text content of conference recordings or telephone calls. This analysis process helps the enterprise obtain a full picture of the communication content and further improve the information grasp and management of conversation records. Next, by performing multimodal association mining analysis on the enterprise's internal conversation text content information data, the enterprise's internal conversation image text content information data, and the enterprise's internal conversation audio text content information data, different types of information can be associated and integrated to form multimodal text content association relationship data. This association analysis can reveal the inherent connections and patterns between different data types, providing enterprises with a more comprehensive and three-dimensional communication analysis perspective, thereby optimizing decision-making and management processes. Finally, by conducting multimodal fusion analysis on the text content information data of internal enterprise conversations, the image text content information data of internal enterprise conversations, and the audio text content information data of internal enterprise conversations based on the multimodal text content association data of enterprise conversations, comprehensive conversation content understanding and comprehensive analysis can be achieved. This multimodal fusion analysis helps to deeply explore the deep information of communication content, discover potential problems and opportunities, and provide enterprises with more intelligent and comprehensive management support, thereby providing more accurate and comprehensive data support for subsequent processing processes.

[0074] Preferably, step S144 includes the following steps:

[0075] Perform text-image association mining analysis on the enterprise internal conversation text content information data and the enterprise internal conversation image text content information data to obtain the enterprise internal conversation text-image text content association relationship data;

[0076] The embodiments of the present invention analyze the text content information data of internal enterprise conversations and the image text content information data of internal enterprise conversations by using technical means such as text feature extraction, similarity calculation, and association rule mining algorithm, so as to deeply explore the association relationship between text information and image information in internal enterprise conversations. For example, keywords mentioned by employees in chat records are identified and these keywords are associated with photos, pictures, or emoticon animations used during the chat, thereby revealing the intrinsic association between text descriptions and images, and finally obtaining internal enterprise conversation text-image text content association relationship data.

[0077] Preferably, the text content information data of the internal conversations in the enterprise and the audio text content information data of the internal conversations in the enterprise are subjected to text-audio association mining analysis to obtain the text-audio text content association relationship data of the internal conversations in the enterprise;

[0078] The embodiment of the present invention analyzes the text content information data of internal enterprise conversations and the audio text content information data of internal enterprise conversations by using technical means such as text feature extraction, similarity calculation, and association rule mining algorithm, so as to deeply explore the correlation between text information and audio information in internal enterprise conversations. For example, the text content in the chat record is associated with the chat voice, thereby revealing the intrinsic correlation between the text description and the audio, and finally obtaining the text-audio text content correlation relationship data of internal enterprise conversations.

[0079] Preferably, image-audio association mining analysis is performed on the enterprise internal conversation image text content information data and the enterprise internal conversation audio text content information data to obtain the enterprise internal conversation image-audio text content association relationship data;

[0080] The embodiments of the present invention analyze the image text content information data of internal enterprise conversations and the audio text content information data of internal enterprise conversations by using technical means such as text feature extraction, similarity calculation, and association rule mining algorithm, so as to deeply explore the correlation between image information and audio information in internal enterprise conversations. For example, photos, pictures, or emoticon animations used during the chat are associated with the chat voice, thereby revealing the intrinsic correlation between the image and the audio, and finally obtaining the image-audio text content correlation relationship data of the internal enterprise conversations.

[0081] Preferably, the text content information data of the internal enterprise conversation, the image text content information data of the internal enterprise conversation and the audio text content information data of the internal enterprise conversation are subjected to fusion association mining analysis to obtain the internal enterprise conversation fusion text content association relationship data;

[0082] The embodiment of the present invention analyzes the text content information data of internal enterprise conversations, the image text content information data of internal enterprise conversations, and the audio text content information data of internal enterprise conversations by using technical means such as text feature extraction, similarity calculation, and association rule mining algorithm, so as to comprehensively analyze the text, image and audio information, and deeply understand the association patterns and rules between different types of information, and finally obtain the internal enterprise conversation fusion text content association relationship data.

[0083] Preferably, the internal enterprise conversation text-image text content association relationship data, the internal enterprise conversation text-audio text content association relationship data, the internal enterprise conversation image-audio text content association relationship data and the internal enterprise conversation fusion text content association relationship data are merged to obtain enterprise conversation multimodal text content association relationship data.

[0084] The embodiment of the present invention processes the text-image text content association relationship data of internal enterprise conversations, the text-audio text content association relationship data of internal enterprise conversations, the image-audio text content association relationship data of internal enterprise conversations, and the fused text content association relationship data of internal enterprise conversations by using a data merging method, so as to merge the association relationship data between text-image, text-audio, image-audio and fused text content into a unified data set, and finally obtain the multimodal text content association relationship data of enterprise conversations.

[0085] The present invention firstly conducts text-image association mining analysis on the text content information data of the enterprise internal conversation and the text content information data of the enterprise internal conversation image, and can deeply explore the association relationship between the text information and the image information in the enterprise internal conversation, which includes extracting keywords, themes and emotions from the text description and associating them with the image content, thereby revealing the intrinsic connection between the text description and the image. For example, the keywords mentioned by the employees in the chat records can be identified, and these keywords can be associated with the photos, pictures or emoticon package animations used during the chat to better understand the relationship between the conversation content and the participants. This analysis process helps to understand the conversation content more comprehensively, thereby improving information understanding and decision-making efficiency. Secondly, by conducting text-audio association mining analysis on the text content information data of the enterprise internal conversation and the audio text content information data of the enterprise internal conversation, the relationship between the text information and the audio information in the enterprise internal conversation can be explored. This analysis method can help enterprises associate text records with audio content, such as associating the text content in the chat record with the chat voice, so as to more comprehensively understand the content and background of the conversation. This association analysis helps enterprises to deeply understand the details and context of the communication process and provide more information support for subsequent processing. Then, by performing image-audio association mining analysis on the enterprise internal conversation image text content information data and the enterprise internal conversation audio text content information data, the connection between image information and audio information in the enterprise internal conversation can be studied. This analysis method can associate image content with audio content, such as associating photos, pictures or emoticons used during the chat with the chat voice to gain a deeper understanding of the context and details of the conversation. This association analysis helps enterprises better understand the scenes in the conversation and the interactions between participants, thereby improving communication efficiency and decision-making quality. Next, by performing fusion association mining analysis on the enterprise internal conversation text content information data, the enterprise internal conversation image text content information data and the enterprise internal conversation audio text content information data, the text, image and audio information can be comprehensively analyzed to deeply explore the text content association relationship of the enterprise internal conversation. This analysis method can integrate and associate different types of information, thereby forming a more comprehensive and three-dimensional conversation understanding, providing enterprises with more information support and decision-making basis. Through fusion association analysis, enterprises can gain a deeper understanding of the connection and pattern between conversation content and provide more powerful guidance for business development. Finally, by merging various relational data, this merging and integration of comprehensive data can help companies fully understand the relationship between conversation content and provide more comprehensive information support for decision-making. Through multimodal text content relational data, companies can better understand the overall picture and characteristics of internal communication, thereby optimizing management decisions and improving work efficiency and employee satisfaction.

[0086] Preferably, step S15 comprises the following steps:

[0087] Step S151: performing data volume statistical analysis on enterprise multimodal conversation text information data to obtain enterprise multimodal conversation text data volume statistical data; performing growth trend forecast analysis on enterprise multimodal conversation text data volume statistical data to obtain enterprise multimodal conversation text data volume growth trend data;

[0088] The embodiment of the present invention uses a mathematical statistical analysis method to perform statistical analysis on the data volume of corresponding text information in the enterprise multimodal conversation text information data, so as to gain an in-depth understanding of the distribution of the text data volume of the enterprise in different time periods, including the total amount of data, the data volume of each mode, and the proportional relationship between different modes, thereby obtaining the enterprise multimodal conversation text data volume statistical data. Then, by using methods such as time series analysis and trend prediction models and combining historical data and external influencing factors, the growth trend of the enterprise multimodal conversation text data volume statistical data is analyzed to predict the growth trend of the enterprise multimodal conversation text data volume in the future period, and finally obtain the enterprise multimodal conversation text data volume growth trend data.

[0089] Step S152: performing a cloud computing storage scale demand analysis on the enterprise multimodal conversation text information data according to the enterprise multimodal conversation text data volume growth trend data, and obtaining the enterprise multimodal conversation cloud computing storage scale demand data;

[0090] The embodiment of the present invention analyzes the storage scale requirements of the corresponding enterprise multimodal conversation text information data on the cloud computing platform in combination with the growth trend data of the enterprise multimodal conversation text data volume, so as to analyze and determine the required storage capacity, storage type, data backup strategy, etc., and enable it to effectively plan and configure the enterprise's cloud computing resources, and finally obtain the enterprise multimodal conversation cloud computing storage scale requirement data.

[0091] Step S153: performing access frequency statistical analysis on the enterprise multimodal conversation text information data to obtain enterprise multimodal conversation text access frequency data; performing access pattern recognition analysis on the enterprise multimodal conversation text information data based on the enterprise multimodal conversation text access frequency data to obtain enterprise multimodal conversation text access pattern data;

[0092] The embodiment of the present invention analyzes enterprise multimodal conversation text information data by using a frequency statistical analysis method to analyze and understand the access conditions of various types of text data of the enterprise, including which data is frequently accessed and which data is rarely accessed, thereby obtaining enterprise multimodal conversation text access frequency data. Then, by combining the enterprise multimodal conversation text access frequency data obtained by the analysis, the corresponding enterprise multimodal conversation text information data is analyzed using an access pattern analysis method to identify access patterns of different types of data, such as periodic access, sudden access, etc., and finally obtain enterprise multimodal conversation text access pattern data.

[0093] Step S154: Designing an elastic storage architecture for the cloud computing platform according to the enterprise multimodal session cloud computing storage scale demand data and the enterprise multimodal session text access mode data to obtain an elastic storage architecture for the cloud computing platform;

[0094] The embodiment of the present invention optimizes the design of the storage architecture on the cloud computing platform by combining the enterprise multimodal session cloud computing storage scale demand data and the enterprise multimodal session text access pattern data obtained by analysis, so as to determine the number and location of storage nodes according to the corresponding storage requirements and access patterns, select appropriate storage technology and architecture, design data backup and disaster recovery solutions, etc., thereby designing a flexible storage architecture with strong applicability and high resource utilization to meet the storage and access requirements of different text data, and finally obtain a flexible storage architecture for the cloud computing platform.

[0095] Step S155: elastically storing the enterprise multimodal conversation text information data according to the elastic storage architecture of the cloud computing platform to obtain elastically stored enterprise conversation text information data.

[0096] The embodiment of the present invention uses a designed cloud computing platform elastic storage architecture to perform elastic storage processing on the corresponding enterprise multimodal conversation text information data, so as to store the enterprise multimodal conversation text information data on the corresponding storage nodes of the cloud computing platform, and dynamically adjust and optimize the storage capacity of its storage nodes as needed, and finally obtain the enterprise conversation elastically stored text information data.

[0097] The present invention firstly performs data volume statistics analysis on enterprise multimodal conversation text information data, and can deeply understand the distribution of text data volume of enterprises in different time periods, including the total amount of data, the amount of data of each mode, and the proportional relationship between different modes, etc., which helps enterprises understand the scale and characteristics of data and provides a basis for subsequent data processing and management. At the same time, by performing growth trend forecast analysis on the statistical data of enterprise multimodal conversation text data volume, the growth trend of enterprise multimodal conversation text data volume in the future can be predicted, thereby providing a reference basis for future resource planning and decision-making of enterprises. Secondly, by performing cloud computing storage scale demand analysis on enterprise multimodal conversation text information data according to the growth trend data of enterprise multimodal conversation text data volume, this includes determining the required storage capacity, storage type, data backup strategy, etc., to meet the data storage needs of enterprises in different time periods. By performing in-depth analysis on the demand for storage scale, the cloud computing resources of enterprises can be effectively planned and configured, thereby improving the efficiency and flexibility of data storage. Then, by performing access frequency statistics analysis on enterprise multimodal conversation text information data, the access situation of various types of text data of enterprises can be understood, including which data is frequently accessed and which data is rarely accessed. On this basis, by performing access pattern recognition and analysis on enterprise multimodal conversation text information data based on enterprise multimodal conversation text access frequency data, we can identify access patterns of different types of data, such as hot data, unpopular data, etc., which helps enterprises better understand the use of data and provide guidance for subsequent data management and optimization. Next, we design an elastic storage architecture for the cloud computing platform based on the enterprise multimodal conversation cloud computing storage scale demand data and enterprise multimodal conversation text access pattern data, which includes determining the number and location of storage nodes, selecting appropriate storage technology and architecture, and designing data backup and disaster recovery solutions. Through the design of elastic storage architecture, the cloud computing platform can have better flexibility and scalability, and can be dynamically adjusted and optimized according to actual data requirements and access patterns. Finally, by performing elastic storage processing on enterprise multimodal conversation text information data based on the elastic storage architecture of the cloud computing platform, this includes data storage, backup, migration and other operations to ensure data security, reliability and efficiency. Through elastic storage processing, enterprises can better manage and utilize multimodal conversation text information data to meet the needs of different business scenarios, thereby improving the efficiency and quality of data processing.

[0098] Preferably, step S2 comprises the following steps:

[0099] Step S21: Formatting and cleaning the enterprise conversation flexibly stored text information data to obtain enterprise conversation text information standard data;

[0100] In the embodiments of the present invention, first, sensitive information in the elastic storage text information data of enterprise conversations is identified and extracted, such as personal identity information, sensitive words, etc., and the identified sensitive information is replaced with virtual identifiers by using data desensitization technology to ensure that the data reduces its sensitivity while maintaining its availability. At the same time, the deduplication algorithm is used to deduplicate the desensitized text information data to eliminate the content that appears repeatedly in the text information data and retain the unique data entries. Secondly, the deduplicated text information data is denoised to identify and remove the noise data in the text, such as errors, redundancies, or irrelevant information. Then, natural language processing technology is used to perform operations such as semantic content checking and grammar correction, including identifying and correcting grammar errors, logical errors, and other semantic errors in the text to further improve the accuracy and comprehensibility of the data and ensure the correction of errors and incomplete information in the text. And the corrected text information data is processed by using formatting processing technology, including removing special characters, punctuation marks, HTML tags, adjusting the structure, layout, and markup of the text, etc., to meet specific data format requirements and make it conform to the standards of enterprise conversation text information, and finally the standard data of enterprise conversation text information is obtained.

[0101] Step S22: Analyze the characteristics of the standard data of enterprise conversation text information to obtain the characteristic information data of enterprise conversation text;

[0102] In the embodiments of the present invention, the characteristic analysis method is used to analyze the standard data of enterprise conversation text information obtained after processing to identify the characteristics such as the language style, theme content, and emotional tendency of the text, and finally the characteristic information data of enterprise conversation text is obtained.

[0103] Step S23: Identify and analyze the stop words in the standard data of enterprise conversation text information to obtain the stop word data of enterprise conversation text information;

[0104] In the embodiments of the present invention, the stop words in the text information of the standard data of enterprise conversation text information are identified and analyzed to identify the common stop words in the text, where the stop words refer to the high-frequency words that are often ignored in text analysis, such as "de", "shi", "le", etc., and finally the stop word data of enterprise conversation text information is obtained.

[0105] Step S24: Build a tokenizer adaptation according to the characteristic information data of enterprise conversation text and the stop word data of enterprise conversation text information to obtain an enterprise conversation text-adapted tokenizer; perform text tokenization on the standard data of enterprise conversation text information according to the enterprise conversation text-adapted tokenizer to obtain an enterprise conversation text information token set;

[0106] The embodiment of the present invention generates a corresponding stop word list by combining the enterprise conversation text characteristic information data and the enterprise conversation text information stop word data to construct a word segmenter suitable for the current text data, and identifies the key words in the text to ensure the accuracy and effectiveness of the text segmentation, thereby obtaining an enterprise conversation text adaptation word segmenter. Then, by using the constructed enterprise conversation text adaptation word segmenter, the corresponding enterprise conversation text information standard data is segmented to divide the text information into independent words or words, and the complex text data is converted into a structured information set, and finally the enterprise conversation text information word segmentation set is obtained.

[0107] Step S25: performing part-of-speech tagging on the enterprise conversation text information word set to obtain the enterprise conversation text word set.

[0108] The embodiment of the present invention processes each segmentation vocabulary in the enterprise conversation text information segmentation set by using a part-of-speech tagging method to determine the part of speech or grammatical role of each vocabulary in the enterprise conversation text information segmentation set in the sentence, such as noun, verb, adjective, etc., and further analyzes and understands the grammatical structure and meaning of the text, and finally obtains the enterprise conversation text segmentation part-of-speech set.

[0109] First, by performing formatting and cleaning on the text information data stored in enterprise conversations in a flexible manner, the consistency and standardization of the data can be ensured, making it conform to a unified data format and specification. This includes removing special characters, punctuation marks, HTML tags, etc. from the text, uniformly converting the case format of the text, and handling duplicate data, noise, and content format errors. Through formatting and cleaning, the noise and redundant information in the data can be reduced, improving the quality and usability of the data, thereby providing a reliable data foundation for subsequent text analysis and processing. Second, by analyzing the characteristics of the enterprise conversation text information standard data, the aim is to deeply understand the characteristics and content of the text data, which includes analyzing keywords, themes, sentiment tendencies, etc. in the text, as well as identifying key information and relevance in the text. Through text characteristic analysis, enterprises can better understand and grasp the meaning and value of the text data, thereby providing guidance and support for subsequent text mining and information extraction. Then, by performing stop word identification and analysis on the enterprise conversation text information standard data to identify common stop words in the text, such as "de", "shi", "le", etc., which usually have no practical meaning in text analysis but have a high frequency of occurrence. By identifying stop words, the noise interference in the text processing and analysis process can be reduced, improving the accuracy and efficiency of the analysis, and more accurately grasping the key information and characteristics of the text. Next, by constructing a tokenizer adaptation based on the enterprise conversation text characteristic information data and the enterprise conversation text information stop word data to construct a text tokenizer suitable for the needs of enterprise conversations, the tokenizer can identify key vocabulary in the text and perform filtering processing according to the stop word list. Subsequently, by performing text tokenization on the enterprise conversation text information standard data using the enterprise conversation text-adapted tokenizer, the text can be split into independent lexical units for subsequent semantic analysis and information extraction. Through tokenization, complex text data can be converted into a structured information set, thereby providing a basis for subsequent text analysis and mining. Finally, by performing part-of-speech tagging on the enterprise conversation text information token set, the aim is to add corresponding part-of-speech tags, such as nouns, verbs, adjectives, etc., to each lexical unit. This part-of-speech tagging process can help understand the grammatical and semantic relationships of the words in the text and analyze the meaning and structure of the text more deeply. Through part-of-speech tagging, the ability to understand and mine text data can be further improved, thereby providing support for enterprises to obtain more valuable information and insights from the text data.

[0110] Preferably, step S21 includes the following steps:

[0111] Step S211: Perform data desensitization processing on the text information data stored in enterprise conversations in a flexible manner to obtain text desensitization information data for enterprise conversation storage;

[0112] The embodiment of the present invention first identifies and extracts sensitive information in the enterprise session elastically stored text information data, such as personal identity information, sensitive words, etc., and then replaces the identified sensitive information with a virtual identifier by using data desensitization technology to ensure that the data is reduced in sensitivity while maintaining its availability, and finally obtains the enterprise session stored text desensitized information data.

[0113] Step S212: performing deduplication processing on the enterprise conversation stored text desensitized information data to obtain the enterprise conversation stored text deduplication information data;

[0114] The embodiment of the present invention performs duplicate content identification and analysis on the enterprise conversation stored text desensitized information data by using a text content comparison method to identify duplicate information in the text data, and then deduplicates the identified duplicate information by using a corresponding deduplication algorithm to eliminate repeated content in the text data and retain unique data entries, ultimately obtaining the enterprise conversation stored text deduplication information data.

[0115] Step S213: performing context association analysis on the enterprise conversation stored text deduplication information data to obtain the enterprise conversation stored text semantic association information data; performing semantic perception denoising processing on the enterprise conversation stored text deduplication information data according to the enterprise conversation stored text semantic association information data to obtain the enterprise conversation stored text semantic denoising information data;

[0116] The embodiment of the present invention uses an association statistical analysis method to analyze the relationship between text information and its context in enterprise conversation stored text deduplication information data, so as to analyze and identify the semantic relationship between text information and its context, thereby identifying and extracting semantic related information between texts, thereby obtaining enterprise conversation stored text semantic association information data. Then, by combining the enterprise conversation stored text semantic association information data obtained by the analysis, the corresponding enterprise conversation stored text deduplication information data is used for denoising, so as to identify and remove noise data in the text, such as erroneous, redundant or irrelevant information, and remove interference items of the text data, thereby retaining valid information related to semantics, and finally obtaining enterprise conversation stored text semantic denoising information data.

[0117] Step S214: performing semantic content error correction on the enterprise conversation stored text semantic denoising information data to obtain the enterprise conversation stored text semantic correction information data;

[0118] The embodiment of the present invention uses natural language processing technology to perform semantic content checking, grammatical error correction and other operations on the semantic denoising information data of enterprise conversation storage text, including identifying and correcting grammatical errors, logical errors and other semantic errors in the text, so as to further improve the accuracy and comprehensibility of the data, and ensure that errors and incomplete information in the text are corrected, and finally obtain the semantic correction information data of enterprise conversation storage text.

[0119] Step S215: Formatting the enterprise conversation stored text semantic correction information data to obtain enterprise conversation text information standard data.

[0120] The embodiment of the present invention processes the corrected enterprise conversation stored text semantic correction information data by using formatting processing technology, including adjusting the structure, layout and marking of the text, so as to meet specific data format requirements and make it conform to the enterprise conversation text information standards, and finally obtain the enterprise conversation text information standard data.

[0121] The present invention firstly performs data desensitization processing on the text information data of enterprise conversation elastic storage to protect user privacy and sensitive information, while ensuring the security of data. Data desensitization processing includes replacing, deleting or encrypting personal identity information, sensitive words, etc. in the text to eliminate potential privacy leakage risks. Through data desensitization processing, user privacy can be effectively protected, in compliance with the requirements of relevant laws and regulations, while ensuring that the text data processed by the enterprise conversation is safe and reliable. Secondly, by performing deduplication processing on the desensitized information data of the enterprise conversation storage text, repeated information in the text data can be eliminated, data redundancy can be reduced, and the overall quality of the data can be improved. The deduplication processing process can identify and remove repeated content in the text, retain unique data entries, thereby reducing the storage space occupied and improving data processing efficiency. Through deduplication processing, the enterprise conversation text data can be made clearer and manageable, thereby providing a more reliable basis for subsequent data analysis and mining. Then, by performing context association analysis on the deduplication information data of the enterprise conversation storage text, the semantic association in the text data can be found, and the semantic related information between the texts can be identified and extracted. Subsequently, semantic-aware denoising is performed on the enterprise conversation storage text deduplication information data according to the enterprise conversation storage text semantic association information data, that is, identifying and removing noise data in the text, such as errors, redundant or irrelevant information, so as to remove interference items in the data, retain valid information related to semantics, and improve the quality and availability of the data. Through semantic-aware denoising, the meaning and structure of the text data can be grasped more accurately, thus providing a more reliable basis for subsequent text analysis and mining. Next, semantic content correction is performed on the enterprise conversation storage text semantic denoising information data, aiming to improve the accuracy and coherence of the text data, including identifying and correcting grammatical errors, logical errors and other semantic errors in the text, making the text clearer and easier to understand. By correcting errors and incomplete information in the text, the quality and readability of the text data can be improved, thus providing a more reliable basis for subsequent text analysis and application. Finally, the enterprise conversation storage text data after semantic correction is formatted to ensure that the text data conforms to a unified format and specification. Formatting includes adjusting the structure, layout and tags of the text to make it neater and easier to manage. Through formatting, the presentation of text data can be unified, and the consistency and maintainability of data can be improved, thereby providing better support and convenience for the company's subsequent data analysis and application.

[0122] Preferably, step S3 comprises the following steps:

[0123] Step S31: performing conversation entity recognition analysis on the enterprise conversation text information standard data to obtain enterprise conversation text entity information data;

[0124] The embodiment of the present invention analyzes the processed enterprise conversation text information standard data by using an entity recognition algorithm (such as NER) to identify entities from the enterprise conversation text, including people, places, organizations, time, etc., and extracts important information units from the conversation text, namely, entity information data, and finally obtains the enterprise conversation text entity information data.

[0125] Step S32: performing named entity screening analysis on the enterprise conversation text entity information data to obtain an enterprise conversation text named entity data set;

[0126] The embodiment of the present invention uses predefined named entity types (such as personal names, organization names, place names, etc.) to filter and process the analyzed enterprise conversation text entity information data to remove entities that do not meet the analysis objectives, and further focus on the key information in the conversation, reduce the interference of irrelevant content, and integrate the screened named entities at the same time, and finally obtain an enterprise conversation text named entity dataset.

[0127] Step S33: performing an occurrence position recognition analysis on the enterprise conversation text named entity dataset to obtain the enterprise conversation text named entity occurrence position information data; performing an occurrence frequency statistical analysis on the enterprise conversation text named entity dataset based on the enterprise conversation text named entity occurrence position information data to obtain the enterprise conversation text named entity occurrence frequency data;

[0128] The embodiment of the present invention uses a text analysis method to perform recognition analysis on each named entity appearing in the enterprise conversation text named entity data set to identify the specific appearance position of each named entity in the text, including chapter, paragraph, sentence number, etc., thereby obtaining the enterprise conversation text named entity appearance position information data. Then, by combining the enterprise conversation text named entity appearance position information data obtained by the analysis, a frequency statistical analysis method is used to perform statistical analysis on each named entity appearing in the enterprise conversation text named entity data set to calculate the appearance frequency of each named entity in the conversation text, and understand which named entities are more important or more concerned in the conversation, and finally obtain the enterprise conversation text named entity appearance frequency data.

[0129] Step S34: performing text surrounding semantic association analysis on the enterprise conversation text named entity dataset based on the enterprise conversation text information segmentation set and the enterprise conversation text segmentation part-of-speech set to obtain enterprise conversation text named entity surrounding semantic association data;

[0130] The embodiment of the present invention analyzes each named entity appearing in an enterprise conversation text named entity data set by combining an enterprise conversation text information word segmentation set and an enterprise conversation text word segmentation part-of-speech set to analyze, identify and understand surrounding word segmentation semantic information associated with these named entities. Then, an association analysis is performed on the surrounding word segmentation semantic information and each named entity appearing in the enterprise conversation text named entity data set by using an association rule mining algorithm to analyze the semantic relevance between the named entity and its surrounding text, and finally obtains the surrounding semantic association data of the enterprise conversation text named entities.

[0131] Step S35: performing entity link extraction processing on the enterprise conversation text entity information data according to the occurrence frequency data of the enterprise conversation text named entities and the semantic association data around the enterprise conversation text named entities to obtain the enterprise conversation text entity link relationship data.

[0132] The embodiment of the present invention uses link relationship extraction technology to analyze the associated link relationships between entities in the enterprise conversation text entity information data by combining the enterprise conversation text named entity occurrence frequency data and the semantic association data around the enterprise conversation text named entities obtained by analysis, so as to analyze, identify and establish direct or indirect associated link relationships between entities, such as co-occurrence relationships, semantic similarity relationships, etc., and finally obtain enterprise conversation text entity link relationship data.

[0133] The present invention firstly performs conversation entity recognition analysis on the standard data of enterprise conversation text information, and can identify entities from enterprise conversation text, including people, places, organizations, time, etc., and extract relevant information, which helps enterprises better understand the content and context of the conversation, thereby providing a basis for subsequent analysis and decision-making. By identifying entities in the conversation, enterprises can have a deeper understanding of the key topics and participants involved in the conversation, which helps to discover potential business insights and opportunities. Secondly, named entities are filtered out by performing named entity screening analysis on the enterprise conversation text entity information data, and named entities refer to entities with specific names or identifiers, such as company names, product names, names, etc. By screening out named entities, key information in the conversation can be further focused on, irrelevant content interference can be reduced, and the efficiency and accuracy of subsequent processing can be improved. The enterprise conversation text named entity data set obtained in this way can be used as important basic data for subsequent analysis and application. Then, the enterprise conversation text named entity data set is subjected to occurrence position recognition analysis to determine the specific location information of each named entity in the conversation text. Subsequently, the frequency of occurrence of the named entity dataset of the enterprise conversation text is statistically analyzed based on the location information data of the named entity in the enterprise conversation text to calculate the frequency of occurrence of each named entity in the conversation text. In this way, the named entities that appear frequently in the conversation can be identified, reflecting the focus and important topics of the conversation, thereby providing enterprises with guidance for in-depth understanding of the content of the conversation. Next, the semantic association analysis of the text surrounding the named entity dataset of the enterprise conversation text is performed based on the enterprise conversation text information segmentation set and the enterprise conversation text segmentation part of speech set. This can capture the semantic association between the named entity and its surrounding text, and help understand the context and meaning of the entity in the text. By analyzing the semantic association between the named entity and the surrounding text, the role and meaning of the entity in the conversation can be more accurately grasped, thereby providing a more reliable basis for subsequent entity linking. Finally, the entity information data of the enterprise conversation text is subjected to entity linking extraction processing based on the frequency of occurrence data of the named entity in the enterprise conversation text and the semantic association data around the named entity in the enterprise conversation text, that is, the association relationship between the named entities is identified and established, so that the same entity mentioned in the conversation can be linked, thereby establishing a relationship network between entities. Through entity link extraction processing, enterprises can better understand the relationships and connections between the various entities involved in the conversation, thereby providing a more comprehensive perspective and information support for further analysis and application.

[0134] Preferably, step S4 comprises the following steps:

[0135] Step S41: performing semantic relationship extraction and analysis on the enterprise conversation text information standard data to obtain enterprise conversation text semantic relationship data;

[0136] The embodiment of the present invention analyzes the standard data of enterprise conversation text information by using a semantic relationship extraction algorithm to identify the semantic relationships between different entities from the conversation text, such as the relationship between characters, the relationship between events and places, etc., and finally obtains the semantic relationship data of the enterprise conversation text.

[0137] Step S42: constructing a semantic knowledge graph for the enterprise conversation text information standard data based on the enterprise conversation text entity link relationship data and the enterprise conversation text semantic relationship data to obtain the enterprise conversation text semantic knowledge graph;

[0138] The embodiment of the present invention constructs a knowledge graph for the semantic information in the enterprise conversation text information standard data by combining the enterprise conversation text entity link relationship data and the enterprise conversation text semantic relationship data obtained by analysis, so as to represent the entities and corresponding relationships in the enterprise conversation text information standard data in the form of a graph, thereby constructing a structured semantic knowledge graph, which can clearly display the relationship between entities in the enterprise conversation text, and finally obtain the enterprise conversation text semantic knowledge graph.

[0139] Step S43: performing deep semantic reasoning analysis on the text semantic information in the enterprise conversation text semantic knowledge graph to obtain deep semantic information data of the enterprise conversation text;

[0140] The embodiments of the present invention analyze the text semantic information in the semantic knowledge graph of enterprise conversation text by using a semantic understanding model based on deep learning (such as BERT, GPT, etc.) to further mine and understand the hidden information and semantic associations in the text, and reveal the deep semantics hidden behind the conversation text, and finally obtain the deep semantic information data of the enterprise conversation text.

[0141] Step S44: performing conversation behavior prediction analysis on the enterprise conversation text deep semantic information data to obtain enterprise conversation text user behavior prediction data;

[0142] The embodiment of the present invention analyzes the deep semantic information data of enterprise conversation texts obtained by analysis by using machine learning (such as convolutional neural networks, etc.) and data mining technology to predict the user's behavior and intention based on the enterprise conversation text content and semantic information, thereby predicting the user's possible response, operation or decision, and finally obtaining the enterprise conversation text user behavior prediction data.

[0143] Step S45: performing conversation suggestion analysis on the enterprise conversation text information standard data according to the enterprise conversation text user behavior prediction data to obtain enterprise conversation text content reply suggestion data.

[0144] The embodiment of the present invention performs suggestion analysis on the corresponding enterprise conversation text information standard data by combining the enterprise conversation text user behavior prediction data obtained by analysis, so as to provide appropriate reply suggestions for user conversation messages according to the predicted user behavior and intention, so as to optimize the conversation experience and meet the needs of users, and finally obtain enterprise conversation text content reply suggestion data.

[0145] The present invention firstly extracts and analyzes the semantic relationship of enterprise conversation text information standard data, which can help enterprises identify the semantic relationship between different entities from the conversation text, such as the relationship between characters, the relationship between events and places, etc. This helps enterprises to understand the meaning and internal association of conversation text more deeply, and provides an important basis for subsequent analysis and application. Through semantic relationship extraction and analysis, enterprises can discover potential association patterns and laws, help optimize decision-making and improve efficiency. Secondly, by constructing a semantic knowledge graph for enterprise conversation text information standard data based on enterprise conversation text entity link relationship data and enterprise conversation text semantic relationship data, the enterprise conversation text information can be constructed into a structured semantic knowledge graph, which can clearly show the relationship between entities, help enterprises organize and manage conversation information, and discover new associations and insights. The establishment of a semantic knowledge graph can provide enterprises with a higher level of semantic understanding and knowledge expression, thereby laying the foundation for subsequent in-depth analysis and application. Then, by performing deep semantic reasoning analysis on the text semantic information in the enterprise conversation text semantic knowledge graph, the deep semantics hidden behind the conversation text can be revealed. Through deep semantic reasoning analysis, enterprises can better understand the implicit information and potential meanings in conversations, thereby providing support for deeper data mining and intelligent applications. Next, by performing conversation behavior prediction analysis on the deep semantic information data of enterprise conversation texts, which includes predicting users' possible replies, operations or decisions, it helps enterprises better understand user needs and behavior patterns, so that they can make corresponding responses and decisions in advance. Finally, by performing conversation suggestion analysis on the enterprise conversation text information standard data based on the enterprise conversation text user behavior prediction data, it can provide enterprises with suggestions and response plans for conversation content, which helps enterprises better respond to user needs and intentions, improve customer satisfaction and experience quality, and provide enterprises with more intelligent customer service and communication support, thereby enhancing corporate image and competitiveness.

[0146] Preferably, the present invention further provides a cloud computing-based enterprise conversation data storage and analysis system, which is used to execute the cloud computing-based enterprise conversation data storage and analysis method as described above. The cloud computing-based enterprise conversation data storage and analysis system comprises:

[0147] The enterprise internal conversation data elastic storage processing module is used to collect and process the conversation data of the enterprise internal communication tools to obtain the enterprise multimodal conversation text information data; the enterprise multimodal conversation text information data is elastically stored through the cloud computing platform to obtain the enterprise conversation elastic storage text information data;

[0148] The enterprise conversation text information word segmentation processing module is used to format and clean the enterprise conversation elastic storage text information data to obtain the enterprise conversation text information standard data; perform text segmentation processing on the enterprise conversation text information standard data to obtain the enterprise conversation text information word segmentation set; perform part-of-speech tagging processing on the enterprise conversation text information word segmentation set to obtain the enterprise conversation text word segmentation part-of-speech set;

[0149] The enterprise conversation text entity relationship extraction module is used to perform conversation entity recognition analysis on the enterprise conversation text information standard data to obtain the enterprise conversation text entity information data; perform entity link extraction processing on the enterprise conversation text entity information data based on the enterprise conversation text information segmentation set and the enterprise conversation text segmentation part of speech set, thereby obtaining the enterprise conversation text entity link relationship data;

[0150] The enterprise conversation text semantic suggestion analysis module is used to perform deep semantic extraction and analysis on the enterprise conversation text information standard data based on the enterprise conversation text entity link relationship data to obtain the enterprise conversation text deep semantic information data; and perform conversation suggestion analysis on the enterprise conversation text information standard data based on the enterprise conversation text deep semantic information data to obtain the enterprise conversation text content reply suggestion data.

[0151] In summary, the present invention provides an enterprise conversation data storage and analysis system based on cloud computing. The enterprise conversation data storage and analysis system based on cloud computing consists of an enterprise internal conversation data elastic storage processing module, an enterprise conversation text information word segmentation processing module, an enterprise conversation text entity relationship extraction module, and an enterprise conversation text semantic suggestion analysis module. It can implement any one of the enterprise conversation data storage and analysis methods based on cloud computing described in the present invention, and is used to combine the operations between computer programs running on each module to implement any one of the enterprise conversation data storage and analysis methods based on cloud computing. The internal structures of the system cooperate with each other, which can greatly reduce duplication of work and manpower investment, and can quickly and effectively provide a more accurate and efficient enterprise conversation data storage and analysis process, thereby simplifying the operation process of the enterprise conversation data storage and analysis system based on cloud computing.

[0152] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0153] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for storing and analyzing enterprise conversation data based on cloud computing, characterized in that: The following steps are involved: Step S1: collecting and processing conversation data of the enterprise's internal communication tools to obtain enterprise multimodal conversation text information data; The enterprise multimodal conversation text information data is elastically stored through the cloud computing platform to obtain the enterprise conversation elastically stored text information data; wherein step S1 includes the following steps: Step S11: collecting and processing conversation text data from the enterprise's internal communication tool to obtain text information data of the enterprise's internal conversation records; Step S12: collecting and processing conversation image data of the enterprise's internal communication tool to obtain the enterprise's internal conversation record image information data; Step S13: collecting and processing the conversation audio data of the enterprise internal communication tool to obtain the enterprise internal conversation record audio information data; Step S14: performing multimodal fusion analysis on the text information data of the internal conversation record of the enterprise, the image information data of the internal conversation record of the enterprise, and the audio information data of the internal conversation record of the enterprise to obtain the enterprise multimodal conversation text information data; wherein step S14 includes the following steps: Step S141: performing text content mining analysis on the text information data of the internal conversation records of the enterprise to obtain the text content information data of the internal conversation records of the enterprise; Step S142: performing image feature analysis on the image information data of the internal conversation record of the enterprise to obtain image feature data of the internal conversation record of the enterprise; performing image content recognition analysis on the image information data of the internal conversation record of the enterprise according to the image feature data of the internal conversation record of the enterprise to obtain image text content information data of the internal conversation record of the enterprise; Step S143: performing audio feature analysis on the audio information data of the internal conversation record of the enterprise to obtain audio feature data of the internal conversation record of the enterprise; performing audio content recognition analysis on the audio information data of the internal conversation record of the enterprise according to the audio feature data of the internal conversation record of the enterprise to obtain audio text content information data of the internal conversation of the enterprise; Step S144: performing multimodal association mining analysis on the text content information data of the internal enterprise conversation, the image text content information data of the internal enterprise conversation, and the audio text content information data of the internal enterprise conversation to obtain multimodal text content association relationship data of the enterprise conversation; wherein step S144 includes the following steps: Perform text-image association mining analysis on the enterprise internal conversation text content information data and the enterprise internal conversation image text content information data to obtain the enterprise internal conversation text-image text content association relationship data; Perform text-audio association mining and analysis on the text content information data of internal enterprise conversations and the audio text content information data of internal enterprise conversations to obtain the text-audio text content association relationship data of internal enterprise conversations; Performing image-audio association mining analysis on the enterprise internal conversation image text content information data and the enterprise internal conversation audio text content information data to obtain the enterprise internal conversation image-audio text content association relationship data; Performing fusion association mining analysis on the text content information data of internal enterprise conversations, the image text content information data of internal enterprise conversations, and the audio text content information data of internal enterprise conversations to obtain the fusion text content association relationship data of internal enterprise conversations; Merging the enterprise internal conversation text-image text content association relationship data, the enterprise internal conversation text-audio text content association relationship data, the enterprise internal conversation image-audio text content association relationship data, and the enterprise internal conversation fusion text content association relationship data to obtain enterprise conversation multimodal text content association relationship data; Step S145: Based on the enterprise conversation multimodal text content association relationship data, a multimodal fusion analysis is performed on the enterprise internal conversation text content information data, the enterprise internal conversation image text content information data, and the enterprise internal conversation audio text content information data to obtain the enterprise multimodal conversation text information data; Step S15: elastically storing the enterprise multimodal conversation text information data through the cloud computing platform to obtain elastically stored enterprise conversation text information data; wherein step S15 includes the following steps: Step S151: performing data volume statistical analysis on enterprise multimodal conversation text information data to obtain enterprise multimodal conversation text data volume statistical data; performing growth trend forecast analysis on enterprise multimodal conversation text data volume statistical data to obtain enterprise multimodal conversation text data volume growth trend data; Step S152: performing a cloud computing storage scale demand analysis on the enterprise multimodal conversation text information data according to the enterprise multimodal conversation text data volume growth trend data, and obtaining the enterprise multimodal conversation cloud computing storage scale demand data; Step S153: performing access frequency statistical analysis on the enterprise multimodal conversation text information data to obtain enterprise multimodal conversation text access frequency data; performing access pattern recognition analysis on the enterprise multimodal conversation text information data based on the enterprise multimodal conversation text access frequency data to obtain enterprise multimodal conversation text access pattern data; Step S154: Designing an elastic storage architecture for the cloud computing platform according to the enterprise multimodal session cloud computing storage scale demand data and the enterprise multimodal session text access mode data to obtain an elastic storage architecture for the cloud computing platform; Step S155: elastically storing the enterprise multimodal conversation text information data according to the elastic storage architecture of the cloud computing platform to obtain the enterprise conversation elastically stored text information data; Step S2: Formatting and cleaning the enterprise conversation elastically stored text information data to obtain enterprise conversation text information standard data; performing text segmentation processing on the enterprise conversation text information standard data to obtain an enterprise conversation text information segmentation set; performing part-of-speech tagging processing on the enterprise conversation text information segmentation set to obtain an enterprise conversation text segmentation part-of-speech set; Step S3: performing conversation entity recognition analysis on the enterprise conversation text information standard data to obtain enterprise conversation text entity information data; performing entity link extraction processing on the enterprise conversation text entity information data based on the enterprise conversation text information segmentation set and the enterprise conversation text segmentation part-of-speech set to obtain enterprise conversation text entity link relationship data; wherein step S3 includes the following steps: Step S31: performing conversation entity recognition analysis on the enterprise conversation text information standard data to obtain enterprise conversation text entity information data; Step S32: performing named entity screening analysis on the enterprise conversation text entity information data to obtain an enterprise conversation text named entity data set; Step S33: performing an occurrence position recognition analysis on the enterprise conversation text named entity dataset to obtain the enterprise conversation text named entity occurrence position information data; performing an occurrence frequency statistical analysis on the enterprise conversation text named entity dataset based on the enterprise conversation text named entity occurrence position information data to obtain the enterprise conversation text named entity occurrence frequency data; Step S34: performing text surrounding semantic association analysis on the enterprise conversation text named entity dataset based on the enterprise conversation text information segmentation set and the enterprise conversation text segmentation part-of-speech set to obtain enterprise conversation text named entity surrounding semantic association data; Step S35: performing entity link extraction processing on the enterprise conversation text entity information data according to the occurrence frequency data of the enterprise conversation text named entity and the semantic association data around the enterprise conversation text named entity to obtain the enterprise conversation text entity link relationship data; Step S4: Based on the enterprise conversation text entity link relationship data, deep semantic extraction analysis is performed on the enterprise conversation text information standard data to obtain the enterprise conversation text deep semantic information data; based on the enterprise conversation text deep semantic information data, conversation suggestion analysis is performed on the enterprise conversation text information standard data to obtain the enterprise conversation text content reply suggestion data.

2. The method for storing and analyzing enterprise conversation data based on cloud computing according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: Formatting and cleaning the enterprise conversation flexibly stored text information data to obtain enterprise conversation text information standard data; Step S22: Analyze the text characteristics of the enterprise conversation text information standard data to obtain enterprise conversation text characteristic information data; Step S23: performing stop word recognition analysis on the enterprise conversation text information standard data to obtain stop word data of the enterprise conversation text information; Step S24: constructing a word segmenter based on the enterprise conversation text characteristic information data and the enterprise conversation text information stop word data to obtain an enterprise conversation text adaptation word segmenter; performing text word segmentation processing on the enterprise conversation text information standard data based on the enterprise conversation text adaptation word segmenter to obtain an enterprise conversation text information word segmentation set; Step S25: performing part-of-speech tagging on the enterprise conversation text information word set to obtain the enterprise conversation text word set.

3. The enterprise conversation data storage and analysis method based on cloud computing according to claim 2, characterized in that: Step S21 includes the following steps: Step S211: performing data desensitization processing on the enterprise conversation elastically stored text information data to obtain the enterprise conversation stored text desensitized information data; Step S212: performing deduplication processing on the enterprise conversation stored text desensitized information data to obtain the enterprise conversation stored text deduplication information data; Step S213: performing context association analysis on the enterprise conversation stored text deduplication information data to obtain the enterprise conversation stored text semantic association information data; performing semantic perception denoising processing on the enterprise conversation stored text deduplication information data according to the enterprise conversation stored text semantic association information data to obtain the enterprise conversation stored text semantic denoising information data; Step S214: performing semantic content error correction on the enterprise conversation stored text semantic denoising information data to obtain the enterprise conversation stored text semantic correction information data; Step S215: Formatting the enterprise conversation stored text semantic correction information data to obtain enterprise conversation text information standard data.

4. The enterprise conversation data storage and analysis method based on cloud computing according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: performing semantic relationship extraction and analysis on the enterprise conversation text information standard data to obtain enterprise conversation text semantic relationship data; Step S42: constructing a semantic knowledge graph for the enterprise conversation text information standard data based on the enterprise conversation text entity link relationship data and the enterprise conversation text semantic relationship data to obtain the enterprise conversation text semantic knowledge graph; Step S43: performing deep semantic reasoning analysis on the text semantic information in the enterprise conversation text semantic knowledge graph to obtain deep semantic information data of the enterprise conversation text; Step S44: performing conversation behavior prediction analysis on the enterprise conversation text deep semantic information data to obtain enterprise conversation text user behavior prediction data; Step S45: performing conversation suggestion analysis on the enterprise conversation text information standard data according to the enterprise conversation text user behavior prediction data to obtain enterprise conversation text content reply suggestion data.

5. A cloud computing-based enterprise conversation data storage and analysis system, characterized in that: For executing the enterprise conversation data storage and analysis method based on cloud computing as claimed in claim 1, the enterprise conversation data storage and analysis system based on cloud computing comprises: The enterprise internal conversation data elastic storage processing module is used to collect and process the conversation data of the enterprise internal communication tools to obtain the enterprise multimodal conversation text information data; the enterprise multimodal conversation text information data is elastically stored through the cloud computing platform to obtain the enterprise conversation elastic storage text information data; The enterprise conversation text information word segmentation processing module is used to format and clean the enterprise conversation elastic storage text information data to obtain the enterprise conversation text information standard data; perform text segmentation processing on the enterprise conversation text information standard data to obtain the enterprise conversation text information word segmentation set; perform part-of-speech tagging processing on the enterprise conversation text information word segmentation set to obtain the enterprise conversation text word segmentation part-of-speech set; The enterprise conversation text entity relationship extraction module is used to perform conversation entity recognition analysis on the enterprise conversation text information standard data to obtain the enterprise conversation text entity information data; perform entity link extraction processing on the enterprise conversation text entity information data based on the enterprise conversation text information segmentation set and the enterprise conversation text segmentation part of speech set, thereby obtaining the enterprise conversation text entity link relationship data; The enterprise conversation text semantic suggestion analysis module is used to perform deep semantic extraction and analysis on the enterprise conversation text information standard data based on the enterprise conversation text entity link relationship data to obtain the enterprise conversation text deep semantic information data; and perform conversation suggestion analysis on the enterprise conversation text information standard data based on the enterprise conversation text deep semantic information data to obtain the enterprise conversation text content reply suggestion data.