Implementation method and device of intelligent business circulation engine, and storage medium

Abstract

The invention relates to the technical field of artificial intelligence data processing, in particular to an implementation method and device of an intelligent business circulation engine and a storage medium. The context-aware intelligent filling and keyword adaptation mechanism reduces manual intervention, and the working efficiency is remarkably improved. A multi-dimensional adaptive scoring system and multiple natural language processing technologies are integrated, so that the data processing speed is higher, and the business circulation is smoother. And the filling strategy and the data format conversion mechanism are dynamically adjusted according to different service requirements, so that the adaptive capacity of the system is enhanced. Through a flexible data format conversion mechanism, the system can be compatible with and process different types of service forms. A keyword adaptation and data cutting mechanism ensures that data fields are highly matched with actual requirements, and the accuracy of data filling and reporting is improved. By means of intelligent data extraction and processing, errors caused by manual operation are reduced, and the overall data quality is improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence data processing technology, and specifically to a method, device, and storage medium for implementing a business intelligence flow engine. Background Art

[0002] In traditional business process handling, users usually need to manually fill out forms or perform tedious operations to complete business flow. Such processes are inefficient, error-prone and time-consuming. Therefore, an engine that can intelligently process user needs and automatically flow business is needed to improve work efficiency.

[0003] Specific pain points include: Traditional business processes are cumbersome: Traditional business processes require users to manually fill out forms or perform multiple operations, which is time-consuming and prone to errors. Unclear expression of user needs: Users may have unclear language or inaccurate expressions when expressing business needs, resulting in difficulties in subsequent processing. Insecure identity authentication: Traditional identity authentication methods may have security risks and cannot effectively guarantee the legitimacy of the user's identity. Inefficient information processing: Information processing under traditional methods requires manual intervention, which is inefficient and prone to omissions.

[0004] Therefore, the business intelligent flow engine is proposed to solve the above pain points, process user needs through intelligent technology, and improve the efficiency and accuracy of business flow. Through 3D Depth live monitoring, speech to text, natural language processing, keyword retrieval, large model training and other technologies, intelligent processing and automatic flow of user needs are achieved, thereby improving work efficiency, reducing errors and improving user experience. Summary of the invention

[0005] The objects of the present invention are as follows: 1. Improve work efficiency: The business intelligent flow engine can intelligently process user needs and automatically circulate business, reducing tedious manual operations, saving time and labor costs, and improving work efficiency. 2. Reduce error rate: Through natural language processing, keyword extraction and other technologies, the engine can accurately understand user needs, reduce the incidence of human errors, and improve the accuracy of business processing. 3. Improve user experience: Users can record business needs through video, without filling out cumbersome forms, easy to operate, and improve user experience. 4. Enhance security: Identity authentication is completed through 3D Depth live monitoring technology to ensure the legitimacy of user identity and enhance the security of the system. 5. Data storage and analysis: The storage medium can save videos recorded by users and data generated during the processing process to provide support for subsequent data analysis and business optimization.

[0006] In summary, the implementation method, device and storage medium of this business intelligent flow engine can effectively improve work efficiency, reduce error rate, improve user experience, enhance security, and provide support for data storage and analysis, bringing many beneficial effects to the business flow process.

[0007] The present invention solves the existing technical problems by adopting the following technical solutions: A method for implementing a business intelligence flow engine comprises the following steps: S10, initializing the business intelligent flow engine, starting the business intelligent flow engine, and encapsulating the initial data program and algorithm information into the business intelligent flow engine; through the above implementation, the business intelligent flow engine is solidified into the business intelligent flow engine device as an initial application, so that it can be directly called when needed later, and the user triggers according to the required video acquisition; S11, the user uses the video recording / voice recording device to obtain personal information and perform security verification based on the 3D Depth live monitoring technology, and writes the verified user information into the context. Then, the Speech-to-Text technology is used to convert the input voice information into text form. Subsequently, the natural language processing (NLP) technology is applied to understand and process the text to generate highly readable language. Keywords are identified through the TextRank algorithm, and the LLM large model is used to complete the keyword and application adaptation, keyword and business form field data adaptation according to the keywords, and the adapted data is stored; S12, the LLM executes the corresponding application system through the OpenAPI according to the context, including: executing the corresponding application system through the OpenAPI according to the context, in combination with the keywords and the data adapted by the application. Through the above process, the LLM can interact with the target application system through the OpenAPI according to the context, in combination with the keywords and the data adapted by the application, so as to perform corresponding operations and provide results or feedback; S13, intelligently filling in the business form data according to the context of the LLM to generate a filled form, including: obtaining the form data fields of each task required by the business form; through the above implementation, the data between the at least one business and the business flow can be shared, and the at least one business form can interact with the business flow; S14, outputting in the calling application a manual confirmation whether the filled-in information meets the required expression information.

[0008] A device for implementing a method of a business intelligence flow engine includes a memory 100, a processor 200, an input interface 300, a display unit 400, and a network port 500; The memory 100 includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer-readable instructions. The internal memory provides an environment for the operation of the operating system and computer-readable instructions in the non-volatile storage medium. The processor 200 is a central processing unit CPU, a controller, a microcontroller, a microprocessor or other data processing chip, which provides computing and control capabilities for electronic devices. When executing the computer program stored in the memory 100, it executes the steps of the implementation method of a business intelligent flow engine described in claim 1. The input interface 300 is connected to the processor 200 and is used to obtain externally imported computer programs, parameters and instructions, and save them in the memory 100 under the control of the processor 200. The input interface 300 can be connected to an input device to receive parameters or instructions manually input by a user. The input device is a touch layer covering the display screen, a key, trackball or touchpad set on the terminal shell, a keyboard, a touchpad or a mouse; the display unit 400 is connected to the processor 200 and is used to display the data processed by the processor 200 and to display a visual user interface; the display unit 400 can be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display and an OLED touch device; the network port 500 is connected to the processor 200 and is used to communicate with various external terminal devices. The communication technology used in the communication connection is a wired communication technology or a wireless communication technology.

[0009] A storage medium for an implementation method of a business intelligent flow engine includes a computer-readable storage medium, and the computer-readable storage medium includes: a U disk, a mobile hard disk, a read-only memory ROM, a random access memory RAM, a disk or an optical disk, and other media that can store program codes; a computer program is stored on the storage medium, and when the computer program is executed by a processor, the steps of the implementation method of the business intelligent flow engine are implemented.

[0010] The present invention has the following beneficial effects: 1. Improve work efficiency: Context-aware intelligent reporting and keyword adaptation mechanisms reduce manual intervention and significantly improve work efficiency. The multi-dimensional adaptation scoring system and the integration of multiple natural language processing technologies make data processing faster and business flow smoother.

[0011] 2. Enhanced data accuracy: Keyword adaptation and data clipping mechanisms ensure that data fields are highly matched with actual needs, improving the accuracy of data reporting. Intelligent data extraction and processing reduce errors caused by manual operations and improve overall data quality.

[0012] 3. Improve system flexibility: Dynamically adjust reporting strategies and data format conversion mechanisms according to different business needs to enhance the system's adaptability. Through the flexible data format conversion mechanism, the system can be compatible with and process different types of business forms.

[0013] 4. Provide comprehensive adaptability evaluation: Comprehensively consider multiple dimensions such as keyword matching, data format compatibility and performance indicators to provide a comprehensive adaptability evaluation to ensure that business applications are highly compatible with the intelligent flow engine. Through the scoring system, deficiencies in the system can be better identified and targeted optimization can be performed to improve overall performance.

[0014] 5. Support complex business needs: Using a variety of natural language processing technologies such as text preprocessing, word segmentation, part-of-speech tagging, syntactic analysis and semantic analysis, it provides a solid technical foundation for handling complex business needs. Intelligent reporting based on contextual information can handle more complex and diverse business scenarios and meet the diverse needs of enterprises.

[0015] 6. Reduce labor costs: Reduce the workload of filling out forms manually and reduce labor costs. Through intelligent data processing and analysis, reduce the reliance on manual data processing and save costs.

[0016] 7. Improve user experience: The intelligent business flow engine simplifies the user operation process and improves the user experience.

[0017] 8. Enhanced decision support: High-quality data input and processing provide reliable data support for corporate decision-making. By comprehensively analyzing multi-dimensional scoring results, companies can better optimize their business and adjust their strategies.

[0018] The business intelligence flow engine approach has brought significant benefits by improving work efficiency, enhancing data accuracy, improving system flexibility, providing comprehensive adaptation evaluation, supporting complex business needs, reducing labor costs, improving user experience, and enhancing decision support. These advantages not only make the business processing of enterprises more efficient, accurate and intelligent, but also have important significance for the long-term development and competitiveness of enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of the method for implementing the circulation engine of the present invention; Figure 2 It is a schematic diagram of the flow engine device of the present invention. DETAILED DESCRIPTION

[0020] The implementation method of the present invention comprises the following steps: like Figure 1 Shown is a flow chart of a preferred embodiment of a method of a business intelligent flow engine of the present invention.

[0021] The method of a business intelligent flow engine is applied to one or more electronic devices, wherein the electronic device is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASIC), programmable gate arrays (FPGA), digital signal processors (DSP), embedded devices, etc.

[0022] The electronic device may be any electronic product that can perform human-computer interaction with a user, such as a personal computer, a tablet computer, a smart phone, a game console, an interactive Internet Protocol Television (IPTV), a smart wearable device, etc.

[0023] The electronic device may also include a network device and / or a user device, wherein the network device includes, but is not limited to, a single network server, a server group consisting of multiple network servers, or a cloud consisting of a large number of hosts or network servers based on cloud computing.

[0024] The network where the electronic device is located includes but is not limited to the Internet, a wide area network, a metropolitan area network, a local area network, a virtual private network (VPN), etc.

[0025] S10, initializing the business intelligent circulation engine, starting the business intelligent circulation engine, and encapsulating the initial digitization program and algorithm information into the business intelligent circulation engine.

[0026] Among them, the initial data refers to algorithm information: including algorithm information required for various links such as identity authentication, speech-to-text, natural language processing, keyword extraction, and large model training.

[0027] Initial model data: including initial model data for liveness monitoring, speech recognition, natural language processing, etc., used for the initial operation and processing of the engine.

[0028] User database: includes user identity information, historical records and other data, used for identity authentication and personalization.

[0029] Application data: includes application data related to business flow, which is used by the engine to execute applications and generate business forms.

[0030] Other necessary data: Depending on specific business needs, other necessary initial data may also be included to ensure that the engine can run normally and complete the business flow process.

[0031] For example: User leave business is relatively fixed, mainly including 3D Depth, Speech-to-Text, TextRank algorithm, user information, leave business application, application leave business module data, form data (leave days, leave type, approver, leave reason), etc.

[0032] Through the above implementation, the business intelligent flow engine is solidified into the business intelligent flow engine device as the initial application, so that it can be directly called when needed later. The user triggers the video acquisition according to the demand, thereby realizing the intelligence of business flow.

[0033] S11, the user uses the video recording / voice recording device to obtain personal information and perform security verification based on the 3D Depth live monitoring technology, and writes the verified user information into the context. Then, the Speech-to-Text technology is used to convert the input voice information into text form. Subsequently, the natural language processing (NLP) technology is applied to understand and process the text to generate readable language. Keywords are identified through the TextRank algorithm, such as annual leave, 3 days, and XXX approval. Then, the LLM large model is used to complete the keyword and application adaptation, keyword and business form field data adaptation according to the keywords, and the adapted data is stored.

[0034] In at least one embodiment of the present invention, authenticating the user includes: 1. Collect 3D depth images 1.1 Use a dedicated camera or sensor to collect facial images of the person being verified. These images can capture the three-dimensional depth information of the face.

[0035] 1.2 Ensure that the captured images are of high quality and can clearly display facial details and depth information. For example: high-resolution images (the camera should be able to capture high-resolution images to ensure that the details of facial features are clear), accurate depth resolution (the depth resolution should be high enough to ensure that subtle facial contours and bumps can be captured), wide viewing angle and depth range (the camera should have a wide enough viewing angle and depth range to ensure that the entire face can be covered), etc.

[0036] 2. Perform facial detection and key point positioning 2.1 Use computer vision technology to perform face detection on the collected 3D depth image and determine the facial area.

[0037] 2.2 Use the facial key point detection algorithm to locate the key points of the face, such as eyes, nose, mouth, etc.

[0038] 3: Liveness detection 3.1 Use deep learning algorithms to perform liveness detection on facial images to confirm that the person being verified is a real face rather than a photo or other deceptive means.

[0039] 3.2 A variety of liveness detection technologies can be used, such as detecting facial micro-expressions, blinking, head movements, etc.

[0040] 4: Feature extraction 4.1 Extract features from facial images that have passed liveness detection. These features can be facial geometric structures, texture features, etc.

[0041] 4.2 Using deep learning models, facial images are mapped into high-dimensional feature space for subsequent comparison and verification.

[0042] 5: Personal information matching 5.1 Match the extracted facial features with the personal information stored in the database to verify the identity of the person being verified.

[0043] 5.2 The personal information in the database can be a facial feature vector and its corresponding identity information, such as name, ID number, etc.

[0044] 6: Security Verification 6.1 Perform security verification based on the matching results to determine whether the person being verified is a legitimate user.

[0045] 6.2 Whether verification is passed can be determined based on certain thresholds or rules.

[0046] 7: Write to the large model context 7.1 If the security verification passes, the verified personal information is written into the context of the big model to update the cognitive state of the model.

[0047] 7.2 A large model can be a complex system consisting of multiple sub-models to handle various tasks and scenarios.

[0048] In at least one embodiment of the present invention, converting the video speech into text comprises: 1: Video and voice data extraction 1.1 Extract the speech part from the video file, which may involve audio signal processing technology to separate the audio stream from the video.

[0049] 1.2 If the video already contains an audio track, this step can be skipped.

[0050] 2: Preprocessing 2.1 Preprocess the recorded speech signal, including noise reduction, noise removal, and speech signal enhancement.

[0051] 2.2 Preprocessing can improve the accuracy of subsequent speech recognition.

[0052] 3: Speech feature extraction 3.1 Convert the preprocessed speech signal into a speech feature sequence.

[0053] 3.2 Commonly used speech features include Mel-frequency cepstral coefficients (MFCC), linear predictive coding (LPC) coefficients, etc.

[0054] 4: Speech Recognition Model 4.1 Use the speech recognition model to recognize the speech feature sequence and convert it into text form.

[0055] 4.2 Commonly used speech recognition models include end-to-end models based on deep learning (such as CTC and Attention-based models) and traditional models based on HMM-GMM, DNN-HMM, etc.

[0056] 5: Post-processing 5.1 Post-process the text obtained from speech recognition, including language model modification, spelling correction, etc.

[0057] 5.2 Post-processing can improve the accuracy and fluency of text recognition.

[0058] In at least one embodiment of the present invention, the using of natural language processing (NLP) technology to understand and process text to generate highly readable language includes: 5.2.1: Text Preprocessing 5.2.1.1 Preprocess the input text, including removing special characters and punctuation marks, converting to lowercase, etc.

[0059] 5.2.1.2 Standard text preprocessing techniques such as regular expressions, string operations, etc. can be used.

[0060] 5.2.2: Word segmentation 5.2.2.1 Divide the preprocessed text into sequences of words or phrases.

[0061] 5.2.2.2 You can use ready-made word segmentation tools or rule-based word segmentation algorithms, such as NLTK, spaCy, etc.

[0062] 5.2.3: Part-of-speech tagging 5.2.3.1 Perform part-of-speech tagging on the words or phrases after segmentation, that is, determine the part of speech of each word.

[0063] 5.2.3.2 You can use a pre-trained part-of-speech tagger, such as the one provided by NLTK.

[0064] 5.2.4: Syntactic Analysis 5.2.4.1 Analyze the structure and grammatical relationships of sentences and generate syntax trees or dependency graphs.

[0065] 5.2.4.2 You can use rule-based syntactic analysis methods or machine learning-based syntactic analyzers, such as Stanford Parser, spaCy, etc.

[0066] 5.2.5: Semantic Analysis 5.2.5.1 Understand the semantic meaning of sentences, including tasks such as word sense disambiguation and semantic role labeling.

[0067] 5.2.5.2 You can use semantic analysis models based on deep learning, such as BERT, GPT, etc.

[0068] 5.2.6: Logical Reasoning 5.2.6.1 Perform logical reasoning based on the content of the text and infer the relationships and logical chains between the texts.

[0069] 5.2.6.2 Technologies such as logical reasoning algorithms and knowledge graphs can be used, such as rule-based reasoning engines, reasoning models based on graph neural networks, etc.

[0070] 5.2.7: Generative Language 5.2.7.1 Generate readable language output based on the results of the previous steps.

[0071] 5.2.7.2 Template filling, language model generation and other techniques may be used to combine syntactic and semantic rules to generate natural language text.

[0072] In at least one embodiment of the present invention, identifying keywords by using the TextRank algorithm includes: 1: Text preprocessing 1.1 Preprocess the input text, including removing special characters, punctuation marks, converting to lowercase, etc.

[0073] 1.2 This step aims to convert the text into a standard form that can be processed.

[0074] 2: Word segmentation and part-of-speech tagging 2.1 Segment the preprocessed text and mark the part of speech for each word.

[0075] 2.2 Word segmentation can use ready-made word segmentation tools or rule-based word segmentation algorithms, and part-of-speech tagging can use trained part-of-speech taggers.

[0076] 3: Build the graph structure 3.1 Build a graph structure based on the results of word segmentation and part-of-speech tagging, where nodes represent words and edges represent the relationship between words.

[0077] 3.2 The association relationship can be determined by word co-occurrence and other methods.

[0078] 4: Calculate the weights between words 4.1 Based on the constructed graph structure, the weights between words are calculated, usually using the co-occurrence matrix or the similarity calculation method between words.

[0079] 4.2 The weights between words reflect their importance and relevance in the text.

[0080] 5: Iteratively calculate the TextRank score of the word 5.1 Initially, equal weight values ​​are assigned to each node.

[0081] 5.2 Calculate the TextRank score of the word by iteration until convergence.

[0082] 5.3 The TextRank score can be updated based on the association and weight between words, usually using an iterative calculation method similar to the PageRank algorithm.

[0083] 6: Extract keywords 6.1 Based on the TextRank scores of the words, extract the top-ranked words as keywords.

[0084] 6.2 The number of keywords to be extracted can be determined based on the set threshold.

[0085] Score(Si,Sj)=log(∣Si∣)+log(∣Sj∣))∣wk∣wkϵSi∩wkϵSj∣ in: Si, Sj represent two sentences respectively, and Wk represents the words in the sentences.

[0086] Numerator: The number of words in the two sentences Si and Sj that have the same word Denominator: The sum of the logarithms of the number of words in the sentence. The design of the denominator log can curb the advantage of longer sentences in similarity calculation.

[0087] In at least one embodiment of the present invention, the use of the LLM big model to complete the keyword and application adaptation, keyword and business form field data adaptation according to the keyword includes: Using a large language model (LLM) to adapt keywords to application and business form field data and store the adapted data involves a series of detailed steps.

[0088] Here is a basic overview of the process: 1. Data Preparation and Preprocessing 1.1 Collect data: First, collect relevant data of the application and field data of the business form. This may include the functional description of the application, user interface elements (such as buttons, input box labels, etc.) and all field names and descriptions of the business form.

[0089] 1.2 Keyword preparation: Define or collect a list of keywords that represent features, form fields, or specific tasks in your application that users might search for.

[0090] 2. Keyword and application adaptation 2.1 Input encoding: Preprocess keywords and application data, for example, optimize text data through stemming, removing stop words, tokenization, etc. to make it more suitable for model processing.

[0091] 2.2 Model query construction: Using LLM, build queries to match keywords with application features. This may involve generating sentences or queries that describe keyword features and asking the model to find the best matching features or interface elements based on application data.

[0092] 2.3 Fit and Scoring: Using the LLM output, score the fit of each keyword to possible application features or elements. The natural language understanding capabilities of the model can be used to assess the fit, such as by calculating semantic similarity.

[0093] Specific indicators and calculation methods may include the following aspects: 2.3.1. Keyword matching: Evaluate the matching degree of keywords between the application and the business intelligence flow engine. This can be done by calculating the overlap of keywords or using natural language processing technology to perform semantic matching. For example, the semantic similarity between keywords can be calculated using a word vector model.

[0094] Assuming a specific example: Suppose there is a leave management application. We want to evaluate the suitability of a leave management application named "LeaveManager" with the engine and calculate the suitability score.

[0095] 2.3.1.1. Keyword matching index: - Assume that the keywords required by "LeaveManager" are "leave application", "leave type", "application status", etc. The "LeaveFlow" engine provides keywords such as "leave process", "leave approval", "leave status update", etc.

[0096] - Keyword exact match: Assume that the keyword required by "LeaveManager" exactly matches 80% of the keywords provided by the engine.

[0097] - Keyword relevance: Through calculation, assume the relevance score is 0.85.

[0098] 2.3.1.2.Data format compatibility indicators: - The data format required by "LeaveManager" is leave application form, approval process data, etc., while the data format output by the "LeaveFlow" engine is approval results, leave balance, etc.

[0099] - Data format matching: Assume that the data format matching is 70%.

[0100] - Data conversion complexity: Assume that the data conversion complexity is 0.4.

[0101] 2.3.1.3. Performance indicators: - Assuming that the average response time of the "LeaveFlow" engine is 0.6 seconds and the processing speed is 800 leave applications per minute, the resource consumption is relatively low.

[0102] - Response time: Assume a score of 0.85.

[0103] - Processing Speed: Assume a score of 0.75.

[0104] - Resource consumption: Assume a score of 0.9.

[0105] 2.3.1.4. Calculation method of adaptation score: - Assume weights (w_1 = 0.4), (w_2 = 0.3), (w_3 = 0.3).

[0106] Based on the above data, we can calculate the adaptation score: {Fitness score} = 0.4 *(0.8*0.85) + 0.3*(0.7*0.4) + 0.3 * (0.85 * 0.75 +0.75*0.9) {Fitness score} = 0.4*(0.68) + 0.3 *(0.28) + 0.3 *(0.6375 + 0.675) {Fitness score} = 0.272 + 0.084 + 0.6045 {Fitness score} = 0.9605 Therefore, based on the above data and calculations, the fit score between the "LeaveManager" leave management application and the business intelligence flow engine is 0.9605.

[0107] 2.3.1.5. Comparative analysis Assume that there is another candidate engine "AlternativeFlow", whose keyword matching degree is 70%, data format matching degree is 60%, and response time, processing speed and resource consumption scores are 0.8, 0.7 and 0.8 respectively. Using the same calculation method, assume that its adaptation score is 0.83.

[0108] 2.3.1.6. Comparison results Keyword match: “LeaveFlow” (0.272) vs. “AlternativeFlow” (0.224) Data format compatibility: “LeaveFlow” (0.084) vs. “AlternativeFlow” (0.072) Performance: “LeaveFlow” (0.6045) vs. “AlternativeFlow” (0.534) The adaptation score of “LeaveFlow” is significantly higher than that of “AlternativeFlow”, indicating that “LeaveFlow” is more suitable for integration with “LeaveManager” in terms of keyword matching, data format compatibility, and performance.

[0109] 2.3.2. Data format compatibility: Inspect the compatibility of data formats between the application and the business intelligence flow engine, including input data format and output data format. Evaluate whether the application can correctly parse and process the output data of the business intelligence flow engine and apply it to the corresponding business scenario.

[0110] Specifically examine the compatibility of data formats between the application and the business intelligence flow engine, including input data format and output data format. The specific implementation process includes: 2.3.2.1 Input data format requirements: Clarify the input data format requirements of the business intelligence flow engine. This includes constraints such as field name, data type (such as integer, floating point, string, date, etc.), data structure (such as JSON, XML, CSV), required fields, and field length. Define the semantic requirements of input data, such as the business significance and dependencies of the fields.

[0111] 2.3.2.2 Output data format requirements: Clarify the data format output by the business intelligence flow engine, including the structure of the return value, field type, data precision (such as currency, percentage), date format, etc. Determine the response time requirements and error handling mechanism of the output data, such as error codes and message formats in abnormal situations.

[0112] 2.3.2.3 Format mapping: Create a mapping table for input and output data formats to clarify the correspondence between each field between the business intelligence flow engine and the application. This includes data type conversion rules, such as converting strings to date formats, converting floating point numbers to integers, etc. Perform format mapping for different data types required by the business, such as converting JSON format to XML format, or converting date strings to timestamps.

[0113] 2.3.2.4 Input data validation: Before data enters the business intelligence flow engine, implement format validation of input data to ensure that the data conforms to the predetermined format. If it does not conform, the system should immediately return detailed error information. Implement automatic data cleaning functions, such as deleting invalid characters and standardizing date formats, to ensure the correctness and consistency of data.

[0114] 2.3.2.5 Output data validation: Before the business intelligence flow engine generates output data, format validation is performed to ensure that the output data can be correctly identified and processed by the application.

[0115] 2.3.3. Performance indicators: including response time, processing speed, resource consumption and other performance indicators. Evaluate the performance of the application in the process of interaction with the business intelligence flow engine to ensure that the system can achieve the expected results in actual operation.

[0116] 2.3.4. Compatibility score calculation method: The weighted sum method can be used to combine the various indicators to obtain a comprehensive score. The weights can be adjusted according to actual needs and priorities to reflect the importance of different indicators to the overall adaptability of the system.

[0117] 3. Keywords and business form field adaptation 3.1 Input encoding: Similar to the above application adaptation steps, pre-process the business form field data.

[0118] 3.2 Model query construction: Build queries, requiring the model to match keywords with the most relevant business form fields.

[0119] 3.3 Adaptation and scoring: Score the matching of keywords and form fields to determine the best match.

[0120] 4. Data Storage 4.1 Matching result collation: The matching results of keywords, application elements and business form fields are collated to form a structured data set. Each entry should contain keywords, matching elements / fields, and matching scores.

[0121] 4.2 Database storage: Choose a suitable database system to store the adaptation results. This can be a relational database (such as MySQL, PostgreSQL) or a NoSQL database (such as MongoDB, DynamoDB), depending on the structure of the data and query requirements.

[0122] 4.3 Index optimization: In order to improve query efficiency, the data in the database is indexed, especially the keywords and matching items, to speed up the search.

[0123] 5. Continuous updates and optimizations 5.1 User feedback loop: Based on actual user feedback, continuously optimize the matching effect between keywords and applications and business form fields.

[0124] 5.2 Model training: If feasible, the model can be fine-tuned using user feedback data to improve the accuracy and relevance of matching. Pre-trained language models (such as GPT-3 and GPT-4) are mainly used and adapted to specific tasks through fine-tuning.

[0125] S12, further, for the LLM, according to the context, executing the corresponding application system through OpenAPI includes: Based on the context, keywords, and application-adapted data, the steps to execute the corresponding application system through OpenAPI usually include the following key links: 1. Contextual understanding and analysis 1.1 Text understanding: LLM understands the user's intentions and needs by parsing the text entered by the user, including keywords and contextual information.

[0126] 1.2 Context Modeling: Based on the context information provided by the user, LLM builds an internal model to represent the current situation and environment in order to better execute subsequent steps.

[0127] 2. Keyword and application adaptation data integration 2.1 Adaptation data acquisition: LLM uses the matching results of keywords and application elements obtained in the previous adaptation process to obtain corresponding adaptation data. This data may include keywords, matching application functions or interface elements, and matching scores.

[0128] 2.2 Data processing and conversion: LLM may need to further process and convert the adaptation data to match the interface format of the target application system. This may involve operations such as data structure adjustment and format conversion.

[0129] 3. OpenAPI call 3.1 API interface selection: Based on user requirements and context information, LLM determines the OpenAPI interface of the target application system that needs to be called. LLM uses GPT-4 or similar models for requirement parsing and context understanding. OpenAPI mainly includes standardized documentation of RESTful API.

[0130] 3.2 Build request: LLM builds the corresponding API request according to the interface specification and adaptation data of the target application system, including parameter settings, identity authentication information, etc.

[0131] 3.3 Execute request: LLM calls the OpenAPI of the target application system through the network connection and sends the constructed request to the target system.

[0132] 3.4 Processing responses: LLM receives and parses the response data from the target system, and parses, converts or extracts specific information from the response as needed.

[0133] 4. Data Processing and Interpretation 4.1 Response Parsing: LLM parses the data returned by the target system, understands the information contained in it, and converts it into an easily understandable format.

[0134] 4.2 Information extraction: LLM may extract specific information from the response data, such as the operation result status, the returned data content, etc.

[0135] 4.3 Data processing: LLM further processes the returned data based on the context and user needs, such as filtering, sorting, and summarizing.

[0136] 5. Results presentation and feedback 5.1 Result presentation: LLM presents the processed data to the user in an appropriate manner, and further executes the business application to start and initialize it to the business form interface.

[0137] 5.2 User Interaction: If necessary, LLM may interact with the user to provide more details, confirm or solicit feedback.

[0138] 5.3 Feedback Collection: LLM may collect user feedback to improve the performance of subsequent operations.

[0139] 6. Error handling and follow-up actions 6.1 Error Handling: If an error occurs during execution, LLM will try to identify the cause of the error and take appropriate countermeasures based on the situation, such as retrying, providing hints or suggestions, etc.

[0140] 6.2 Subsequent actions: Based on the execution results and user feedback, LLM may trigger further actions or operations to meet user needs or adjust the execution strategy.

[0141] 6.2.1 Data processing Data preprocessing Input data cleaning: Clean the data input by the user, including removing noise (such as special characters and punctuation marks), converting to lowercase, deleting redundant spaces, etc. Use regular expressions and string operations.

[0142] Tokenization and part-of-speech tagging: Use tools (such as NLTK, spaCy) to tokenize and tag the text to ensure a basic understanding of the text content.

[0143] Structured data extraction: Extract structured data such as time, place, quantity, etc. from text using named entity recognition (NER) technology.

[0144] Data format conversion: Convert data formats of different systems into a compatible unified format. Define standard interfaces and data exchange formats (such as JSON, XML), and use mapping and conversion tools to achieve format compatibility.

[0145] 6.2.1.2. Data processing flow Data input: Receive user requests and input data.

[0146] Data parsing: Parse input data to identify key content and intent.

[0147] Call API: Based on the analysis results, call the OpenAPI interface of the corresponding target application system and pass the necessary parameters.

[0148] Data integration: Collect data returned by various systems and integrate and format them.

[0149] Result generation: Generate readable results and feedback to users.

[0150] 6.2.2 Error Handling 6.2.2.1. Error type definition Input error: The data format or content entered by the user does not meet expectations, such as missing required parameters, incorrect data type, etc.

[0151] Parsing errors: problems that occur during text parsing, such as failure in word segmentation, part-of-speech tagging, or named entity recognition.

[0152] API call errors: problems that occur when calling the API of an external system, such as network errors, interface mismatch, insufficient permissions, etc.

[0153] Data conversion errors: problems that occur during the data format conversion process, such as format incompatibility, data loss, etc.

[0154] Logical reasoning errors: problems that occur when making logical inferences based on text content, such as rule conflicts, incomplete reasoning, etc.

[0155] Response generation errors: problems that occur when generating the final response, such as template filling failure, result formatting errors, etc.

[0156] 6.2.2.2. Error handling strategy Error detection: In each data processing step, add error detection mechanisms to capture and log error information.

[0157] Error classification: Classify errors according to their types and define different processing strategies.

[0158] Fallback mechanism: When a critical error occurs, enable a fallback mechanism to return to the previous stable state or provide a default response.

[0159] Logging: Record error information and processing steps in detail to facilitate subsequent analysis and optimization. Use log management tools (such as ELK stack) to centrally manage and analyze logs.

[0160] User prompts: Friendly prompts of error messages are displayed to users on the user interface, and they are guided on how to correct input or operation.

[0161] Automatic retry: For temporary errors (such as network errors), you can set up an automatic retry mechanism to try to re-execute the operation.

[0162] Error feedback: Feedback error information to the LLM system for continuous training and optimization of the model to improve the robustness and adaptability of the system.

[0163] Specifically, the error information is fed back to the LLM system for continuous training and optimization of the model: 6.2.2.2.1. Error Capture and Classification Real-time monitoring and logging: In each module of the business flow engine (such as API calls, rule execution, data processing, etc.), an error capture mechanism is integrated to record all abnormal situations and error information in the log system.

[0164] Error classification: Classify captured errors, including: API call failures (such as timeouts, invalid responses), rule execution errors (such as rule conflicts, mismatches), data processing anomalies (such as data format errors, missing values), user interaction errors (such as incomprehensible input, invalid requests) 6.2.2.2.2. Error data collection and cleaning Data collection: The error information recorded in the business flow engine is regularly collected into a centralized database to form an error data set.

[0165] Data cleaning: Clean the collected erroneous information, remove redundant data, repair or mark outliers, and ensure data quality.

[0166] 6.2.2.2.3. Error feedback generation and extraction Error summary extraction: Through natural language processing (NLP) technology, error information is converted into a concise error summary, and key error causes and context information are extracted.

[0167] User feedback integration: Allow users to provide additional feedback information on error situations, such as marking the error type and describing the error situation, to further enrich the error data.

[0168] 6.2.2.2.4. Error feedback to the LLM system Automated feedback mechanism: Establish an automated process to regularly pass error summaries and user feedback to the LLM system. Error information can be sent to the LLM training system through an API interface or a dedicated data channel.

[0169] Manual review and confirmation: For critical or complex errors, a manual review process can be set up to ensure that the error feedback delivered to the LLM system is accurate.

[0170] 6.2.2.2.5. Continuous model training and optimization Training data expansion: Error feedback is used as a new training sample and added to the LLM training data set. Through data enhancement technology, more data of similar scenarios is generated to improve the generalization ability of the model.

[0171] Model retraining: Regularly retrain the LLM, using newly added erroneous data to optimize the model's parameters and weights so that it can better adapt to and handle complex scenarios in the business flow engine.

[0172] 6.2.2.2.6. Model evaluation and testing Error repair verification: After the model is retrained, the LLM is evaluated using a historical error dataset to verify how well the model repairs specific error types.

[0173] A / B testing: In the actual business environment, A / B testing is used to compare the performance of the new and old models to ensure that the new model has significant improvements in handling errors and improving business flow efficiency.

[0174] 6.2.2.2.7. Feedback loops and continuous optimization Iterative optimization: The error feedback mechanism is embedded into the continuous optimization process of the business intelligence flow engine to form a closed-loop feedback system to ensure continuous optimization and improvement of the model.

[0175] Through the above process, LLM can interact with the target application system through OpenAPI according to the context, combined with keywords and application-adapted data, to perform corresponding operations and provide results or feedback.

[0176] S13, further, for the LLM, intelligently filling in business form data according to the context, generating a filled form, including: Obtaining form data fields for each of the tasks required by a business form; Determine the number of the data fields and trim the number of the data fields to adapt to the data stored in the LLM, including field type matching: ensure that the type of the form data field (such as text, number, date, etc.) is consistent with the data type stored in the LLM. Field quantity trimming: If the number of form data fields exceeds the number of data stored in the LLM, the fields need to be trimmed or merged to adapt to the LLM data. Field semantic matching: ensure that the semantics of the form fields (field name, description, etc.) are consistent with the semantics of the LLM stored data to ensure accurate data reporting. Contextual relevance: based on contextual information, intelligently select the most relevant data to fill in the corresponding form fields. Data format unification: ensure that the data format (such as date format, currency format, etc.) is consistent to meet the form requirements.

[0177] Determine the name of the data field, and convert the name of the data field according to the standard data format of the business form; Determine the format of the data field and convert the format of the data field according to the adaptation format of the business form; and obtain the operation data of the business form, encapsulate the operation data into standard format data, and feed the encapsulated operation data back to the business form.

[0178] The data field refers to the relevant business data of the business form, such as time data and the like.

[0179] The number of data fields is determined when each business form is initially configured, for example, based on LLM context interaction.

[0180] Furthermore, for the format of the data field, taking time data as an example, it can be a format including year, month and day, such as 20200518, which means May 18, 2020.

[0181] Through the above implementation, data commonality between the at least one business and the business flow can be achieved, which facilitates data interaction between the at least one business form and the business flow.

[0182] S14, outputting in the calling application a manual confirmation whether the filled-in information meets the required expression information.

[0183] The process of outputting manually confirmable filled-in information in the calling application to see whether it meets the required expression information can be divided into the following steps: 1. Data Preparation and Processing 1.1 Obtaining reported information: Obtaining user reported information from the business intelligence flow engine, which may include various data entered by the user in the business form.

[0184] 2. Comparison and analysis 2.1 Comparison between reported information and demand expression information: Compare the information reported by the user with the demand expression information, and compare their degree of compliance item by item.

[0185] 2.2 Data analysis and marking: Identify and mark the parts of the reported information that do not match the information expressed in the demand through algorithms or manual marking.

[0186] 3. Output results that can be confirmed by humans 3.1 Mark inconsistent items: Mark the specific items in the reported information that do not meet the required expression information in the generated report for manual confirmation and correction.

[0187] 3.2 Provide correction suggestions: Based on the analysis results, provide corresponding correction suggestions or suggest users to take corresponding actions.

[0188] 4. Manual confirmation and correction 4.1 User intervention: The generated reports or prompts are presented to relevant personnel or users for manual confirmation and correction.

[0189] 4.2 Adjustment of reported information: Based on the results of manual confirmation, users can make corresponding adjustments or corrections to the reported information to make it meet the required expression information.

[0190] Furthermore, the implementation method of the business intelligent flow engine also includes an execution device. Figure 2 The figure shows a functional module diagram of a preferred embodiment of the business intelligent flow engine execution device of the present invention.

[0191] A memory 100, used for storing computer programs; The processor 200 is used to implement the steps of the business processing method disclosed in any of the aforementioned embodiments when executing the computer program.

[0192] Specifically, the memory 100 includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer-readable instructions, and the internal memory provides an environment for the operation of the operating system and computer-readable instructions in the non-volatile storage medium. In some embodiments, the processor 200 can be a central processing unit (CPU), a controller, a microcontroller, a microprocessor or other data processing chip, which provides computing and control capabilities for electronic devices. When executing the computer program stored in the memory 100, the steps of the business processing method disclosed in any of the aforementioned embodiments can be implemented.

[0193] Based on the above embodiments, as a preferred embodiment, see Figure 2 As shown, the electronic device also includes: The input interface 300 is connected to the processor 200 and is used to obtain externally imported computer programs, parameters and instructions, and save them in the memory 100 under the control of the processor 200. The input interface 300 can be connected to an input device to receive parameters or instructions manually input by a user. The input device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the terminal housing, or a keyboard, touchpad or mouse.

[0194] The display unit 400 is connected to the processor 200 and is used to display data processed by the processor 200 and to display a visual user interface. The display unit 400 can be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc.

[0195] The network port 500 is connected to the processor 200 and is used to communicate with external terminal devices. The communication technology used for the communication connection can be a wired communication technology or a wireless communication technology, such as mobile high-definition link technology (MHL), universal serial bus (USB), high-definition multimedia interface (HDMI), wireless fidelity technology (WiFi), Bluetooth communication technology, low-power Bluetooth communication technology, communication technology based on IEEE802.11s, etc.

[0196] The electronic device may include a memory, a processor, and may also include a computer program stored in the memory and executable on the processor, such as an execution program of a business intelligence flow engine.

[0197] Specifically, the memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer-readable instructions, and the internal memory provides an environment for the operation of the operating system and computer-readable instructions in the non-volatile storage medium. In some embodiments, the processor can be a central processing unit (CPU), a controller, a microcontroller, a microprocessor or other data processing chip, which provides computing and control capabilities for electronic devices. When executing the computer program stored in the memory, the steps of any of the aforementioned business intelligence flow engine methods can be implemented.

[0198] The present application also provides a computer-readable storage medium, which may include: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes. The storage medium stores a computer program, and when the computer program is executed by a processor, the steps of any of the aforementioned business intelligent flow engine methods are implemented.

[0199] The business intelligent flow engine combines a variety of advanced technologies, including 3D Depth liveness monitoring, speech-to-text, natural language processing, keyword retrieval and large model training, to intelligently process the video input by the device, effectively solving the pain points in identity authentication, multimodal data processing, information search and understanding, intelligent decision support and business process automation, thereby improving the efficiency and accuracy of business processing, reducing the need for manual intervention and improving business processing efficiency.

[0200] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

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

[0202] Technical points of the present invention: 1. Context-aware intelligent reporting: Based on contextual information, LLM can intelligently identify and extract key data and automatically fill in business forms. This approach not only relies on fixed rules, but can also dynamically adapt to different business needs, improving flexibility and intelligence.

[0203] 2. Keyword adaptation and data clipping mechanism: Through precise keyword matching and data field clipping mechanisms, the data fields of business forms are ensured to be highly compatible with the data stored in LLM. This mechanism can be flexibly adjusted according to actual needs to improve the accuracy of data matching.

[0204] 3. Multi-dimensional adaptation scoring system: A multi-dimensional adaptation scoring system is introduced, including multiple dimensions such as keyword matching, data format compatibility and performance indicators, to comprehensively evaluate the adaptability of business applications and the intelligent circulation engine and provide a comprehensive adaptation evaluation.

[0205] 4. Integrate multiple natural language processing technologies: Comprehensive use of multiple NLP technologies such as text preprocessing, word segmentation, part-of-speech tagging, syntactic analysis, and semantic analysis ensures the accuracy and efficiency of data processing, providing a solid foundation for intelligent reporting.

[0206] 5. Dynamic data format conversion: According to the data format requirements of different business forms, a flexible data format conversion mechanism is provided to ensure the uniformity of data formats and meet the needs of different business forms.

[0207] This patent solution can achieve 1. Requirements analysis and understanding: Use natural language processing (NLP) and machine learning (ML) technologies to analyze and understand requirements from user input and extract key information and intent.

[0208] 2. Dynamic decision-making and process orchestration: Dynamically select and orchestrate business processes based on user needs and contextual information, and determine the API interfaces and services that need to be called.

[0209] 3. Exception handling and recovery: It has the ability to detect and handle exceptions, and can automatically take remedial measures when abnormal situations occur to ensure the continuity and stability of business processes.

[0210] 4. Data-driven optimization: Through data analysis and machine learning, we continuously optimize business processes and decision-making mechanisms to improve the intelligence and efficiency of the system.

[0211] 5. Cross-system integration: Support integration with different business systems and platforms to achieve cross-system data sharing and business collaboration, and improve overall business flow efficiency.

[0212] Possible future application scenarios 1. Intelligent customer service system: Provide 24 / 7 automated customer service, provide solutions by understanding user problems and calling corresponding API interfaces, reduce the workload of manual customer service, and improve customer satisfaction.

[0213] 2. Supply Chain Management: Dynamically allocate resources, optimize inventory management and logistics distribution. Automatically adjust the processes of each link in the supply chain according to real-time data and market demand to improve efficiency and reduce costs.

[0214] 3. Healthcare: Realize intelligent appointment, diagnosis and treatment plan recommendation. By integrating electronic health records (EHR), medical imaging and other data, the medical process is automated to improve the efficiency and accuracy of diagnosis and treatment.

[0215] 4. E-commerce: Provide personalized shopping recommendations, automated order processing and after-sales service. By analyzing user behavior and preferences, dynamically adjust product recommendations and promotion strategies to improve user experience and sales.

[0216] 5. Smart Manufacturing: Realize automation and intelligent management of production processes. By real-time monitoring of production data, dynamically adjust production plans and process parameters to improve production efficiency and product quality.

[0217] 6. Smart City: In the fields of traffic management, public safety, environmental monitoring, etc., through the intelligent business flow engine, it can dynamically respond to and handle various urban management affairs, improve urban operation efficiency and residents' quality of life.

[0218] Glossary LLM (Large Language Model): A large language model, such as GPT (Generative Pre-trained Transformer), is a natural language processing model based on deep learning that can generate and understand text data.

[0219] GPT (Generative Pre-trained Transformer): A large-scale language generation model that uses the Transformer architecture and can perform various natural language processing tasks after pre-training, such as text generation, summary generation, etc.

[0220] Mel-frequency cepstral coefficient (MFCC): is a feature extraction method commonly used in speech and audio signal processing, especially in speech recognition. MFCC is an acoustic feature based on the human ear perception characteristics, which describes the characteristics of the sound by converting the sound signal into cepstral coefficients in the Mel frequency domain.

[0221] Linear Predictive Coding (LPC) is a feature extraction method for speech and audio signal processing, commonly used in speech coding and speech recognition. The basic idea of ​​LPC is to assume that the speech signal is composed of the residual of the sound generated by the vocal tract system after passing through the prediction filter. LPC coefficients are used to describe the characteristics of these prediction filters.

[0222] HMM-GMM (Hidden Markov Model - Gaussian Mixture Model) is a classic speech recognition model. In this model, the Hidden Markov Model (HMM) is used to model the temporal structure of speech signals, while the Gaussian Mixture Model (GMM) is used to model the acoustic features of speech signals.

[0223] DNN-HMM (Deep Neural Network - Hidden Markov Model) is a model used in the field of speech recognition. In this model, the deep neural network (DNN) is used to model the acoustic features of speech signals, while the hidden Markov model (HMM) is used to model the temporal structure of speech signals.

[0224] Speech-to-Text is a technology or system that converts speech signals into text form. This technology has been widely used in various applications, including voice assistants, automatic subtitle generation, voice search, speech recognition, etc.

Claims

1. A method for implementing a business intelligence flow engine, characterized in that The steps include: S10, initializing the business intelligent circulation engine, starting the business intelligent circulation engine, and encapsulating the initial digitization program and algorithm information into the business intelligent circulation engine; The business intelligence flow engine is fixed as the initial application to the business intelligence flow engine device, and is directly called when needed, and the user triggers it according to the required video acquisition; S11, the user uses the video voice recording device to obtain personal information and perform security verification based on the 3D Depth live monitoring technology, and writes the verified user information into the context; then the Speech-to-Text technology is used to convert the input voice information into text form; then, the natural language processing (NLP) technology is used to understand and process the text to generate highly readable language; the TextRank algorithm is used to identify keywords, and the LLM large model is used to complete the keyword and application adaptation, keyword and business form field data adaptation according to the keywords, and the adapted data is stored; S12, the LLM executes the corresponding application system through the OpenAPI according to the context, including: executing the corresponding application system through the OpenAPI according to the context, in combination with the keywords and the data adapted by the application. Through the above process, the LLM can interact with the target application system through the OpenAPI according to the context, in combination with the keywords and the data adapted by the application, so as to perform corresponding operations and provide results or feedback; S13, intelligently filling in the business form data according to the context of the LLM to generate a filled form, including: obtaining the form data fields of each task required by the business form; through the above implementation, the data between the at least one business and the business flow can be shared, and the at least one business form can interact with the business flow; S14, outputting in the calling application a manual confirmation to determine whether the filled-in information meets the requirement expression information.

2. The device according to claim 1, characterized in that The device comprises a memory 100, a processor 200, an input interface 300, a display unit 400 and a network port 500; The memory 100 includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer-readable instructions. The internal memory provides an environment for the operation of the operating system and computer-readable instructions in the non-volatile storage medium. The processor 200 is a central processing unit CPU, a controller, a microcontroller, a microprocessor or other data processing chip, which provides computing and control capabilities for electronic devices. When executing the computer program stored in the memory 100, the steps of the implementation method of a business intelligent flow engine described in claim 1 are executed. The input interface 300 is connected to the processor 200 and is used to obtain externally imported computer programs, parameters and instructions, which are saved to the processor 200 under the control of the processor 200. The memory 100 includes an input interface 300 connected to an input device for receiving parameters or instructions manually input by a user; the input device is a touch layer covering the display screen, a key, a trackball or a touchpad provided on a terminal housing, a keyboard, a touchpad or a mouse; a display unit 400 is connected to the processor 200 for displaying data processed by the processor 200 and for displaying a visual user interface; the display unit 400 is an LED display, a liquid crystal display, a touch-sensitive liquid crystal display and an OLED touch device; a network port 500 is connected to the processor 200 for communication connection with external terminal devices; the communication technology adopted by the communication connection is a wired communication technology or a wireless communication technology.

3. The storage medium of the method for implementing the business intelligence flow engine according to claim 2 is characterized in that It includes a computer-readable storage medium, which includes: a USB flash drive, a mobile hard disk, a read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk, various media for storing program codes; a computer program is stored on the storage medium, and when the computer program is executed by a processor, the steps of the implementation method of the business intelligence flow engine are implemented.

Images