An application system integration method and system based on machine intelligence

Abstract

The present invention discloses a method and system for integrating application systems based on machine intelligence, which relates to the field of artificial intelligence technology. The method comprises capturing Internet fusion technology through multi-source data within the application system, processing the Internet fusion technology using a distributed computing framework, constructing a machine intelligence retrieval model based on the information knowledge graph of machine intelligence, and optimizing the parameters of the retrieval model through machine learning; and completing information retrieval services based on application system integration by matching the partitioning results through a user feedback mechanism. The present invention can significantly improve the efficiency and accuracy of information retrieval services, effectively respond to the challenges of massive data on the Internet, achieve fast and accurate information retrieval, ensure the continuous improvement of retrieval service quality, possess high flexibility, support changing information retrieval requirements and rapid iteration of services, and provide users with a more personalized and intelligent information retrieval experience.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an application system integration method and system based on machine intelligence. Background Art

[0002] With the rapid development of Internet technology, the amount of global information has exploded, and people's demand for information retrieval and management has also increased. Against this background, information services such as Internet search services, online news services, and website navigation have become an important part of modern life. Users not only hope to find the information they need quickly and accurately, but also expect to obtain a more personalized and intelligent service experience. This requires continuous advancement of related technologies to adapt to the development trend of the information society.

[0003] However, existing Internet information service technologies still have many shortcomings. For example, in terms of the accuracy and relevance of information retrieval, existing technologies are difficult to meet the high requirements of users. In terms of personalized recommendations, although there has been some progress, the accuracy and timeliness of recommendations still need to be improved. In addition, the ability to integrate information from different sources is weak, which requires users to switch between multiple platforms to obtain complete information. To address these problems, the present invention proposes an application system integration method based on machine intelligence. By introducing more advanced algorithms and models, the efficiency and quality of information retrieval are significantly improved, providing users with a more intelligent and convenient service experience. Summary of the Invention

[0004] In view of the problems existing in the existing application system integration methods based on machine intelligence, the present invention proposes an application system integration method and system based on machine intelligence.

[0005] Therefore, the purpose of the present invention is to provide an application system integration method and system based on machine intelligence. In order to solve the problems of low efficiency and insufficient accuracy of information retrieval services in the existing technology, the present invention adopts technical means such as multi-source data capture and processing based on machine intelligence, optimization of retrieval model parameters and microservice architecture integration to improve the efficiency and accuracy of information retrieval.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, an embodiment of the present invention provides an application system integration method based on machine intelligence, which includes: capturing Internet fusion technology through multi-source data in the application system, and processing the Internet fusion technology using a distributed computing framework; constructing a machine intelligence retrieval model based on the information knowledge graph of machine intelligence, and optimizing the parameters of the retrieval model through machine learning; integrating the optimized retrieval model parameters using a microservice architecture, dividing the integrated retrieval model, matching the division results through a user feedback mechanism, and completing information retrieval services based on application system integration.

[0008] As a preferred solution of the machine intelligence-based application system integration method of the present invention, the Internet fusion technology includes collecting massive information data through the Internet, and the Internet fusion technology includes crawling massive information data through multi-source data in the application system. The crawling steps are as follows:

[0009] Define pre-processing rules for multi-source data within the application system;

[0010] Add exception handling mechanism to verify the defined preprocessing rules;

[0011] Use automated tools to manage pre-processing rules that do not show any anomalies, triggering the application system display page that captures massive amounts of information data;

[0012] Defining the pre-processing rules for multi-source data in the application system includes introducing a character encoding conversion function. In the stage of defining the pre-processing rules for multi-source data in the application system, the field positions of the multi-source data are recorded and the field positions of the multi-source data are identified;

[0013] Use machine intelligence learning algorithms to train field locations, and use the sample data obtained during the training process to train the model and complete the location of the field.

[0014] The adding of the exception handling mechanism includes setting a defined request timeout in the application system display page;

[0015] When the definition process captures an abnormal character encoding in the application system display page, the pre-processing rule stage of defining the multi-source data in the application system times out, and the abnormal information is recorded at this time.

[0016] As a preferred solution of the machine intelligence-based application system integration method of the present invention, wherein: the verification of the defined preprocessing rules includes verifying the preprocessing rules by test execution, the scenarios of the preprocessing rules include abnormal information and normal information, and the test execution includes automated testing and manual testing;

[0017] When abnormal information is recorded, manual testing is used to verify the preprocessing rules;

[0018] When no abnormal information is recorded, normal information is recorded, and the pre-processing rules are verified by automated testing;

[0019] The verification of the pre-processing rules by manual testing includes dynamically adjusting the pre-processing rules according to changes in abnormal information and information from a user feedback mechanism;

[0020] The method of using the test execution to verify the pre-processing rules includes dynamically adjusting the number of retries and the time interval according to the number of recorded exception information and the time interval, generating an automated test based on the dynamically adjusted number of retries and the time interval, and completing the function of automatically feeding back the exception information.

[0021] As a preferred solution of the machine intelligence-based application system integration method of the present invention, the processing of the Internet fusion technology includes allocating weights of abnormal information based on a character encoding conversion function and a task scheduling algorithm of a distributed computing framework;

[0022] The weight allocation of the abnormal information includes calculating a dynamic adjustment factor based on the adjustment task queue length and the resource utilization rate of the application system during the dynamic adjustment process. The specific formula for calculating the dynamic adjustment factor is:

[0023]

[0024] Among them, γ and δ represent weight coefficients, represents the ratio of the task queue length, (1-U) represents the remaining amount of resource utilization of the application system, U represents the resource utilization of the application system, and D i represents the dynamic adjustment factor;

[0025] By dynamically adjusting factors, a machine intelligence information knowledge graph is constructed based on the severity and frequency of abnormal information, where the severity includes fatal errors, errors, warnings, and notifications, and the frequency includes analyzing the number of occurrences of the severity;

[0026] According to the changes in the dynamic adjustment factor, the node positions of the information knowledge graph are dynamically updated to complete the construction of the machine intelligent retrieval model.

[0027] As a preferred solution of the machine intelligence-based application system integration method of the present invention, wherein: the optimization of the parameters of the retrieval model includes deriving abnormal information data from the information knowledge graph for training, and annotating the training data;

[0028] The training of the abnormal information data derived from the information knowledge graph includes standardizing the character encoding conversion function, wherein the standardization includes converting the character encoding into a numerical variable using one-hot encoding, obtaining a numerical variable rule by iterating the numerical variable N times, traversing all numerical variable rules, evaluating the performance of each numerical variable rule, and performing data segmentation on the performance, wherein the data segmentation includes a training set, a validation set, and a test set;

[0029] The training set includes a probability model for constructing an objective function according to the next numerical variable rule of N iterations;

[0030] The validation set includes sampling points for searching the probability model in all numerical variable rules;

[0031] The test set includes optimizing sampling points based on information from a user feedback mechanism.

[0032] As a preferred solution of the machine intelligence-based application system integration method of the present invention, the integration of the optimized retrieval model parameters includes integrating the optimized sampling points into the microservice architecture of the application system, wherein the microservice architecture includes data acquisition service, data preprocessing service, model training service, model deployment service, user feedback service and information retrieval service;

[0033] The division of the integrated retrieval model includes functional division and regional division, wherein the functional division includes data preprocessing service, model training service and user feedback service;

[0034] The said area division includes data acquisition services, model deployment services and information retrieval services;

[0035] According to the division results of the retrieval model, the user feedback mechanism is traced;

[0036] Verify the information of the user feedback mechanism and the division results. The verification steps are as follows:

[0037] When the information of the user feedback mechanism is inconsistent with the segmentation results, the exported abnormal information data is trained again;

[0038] When the information of the user feedback mechanism is classified into any of the aforementioned severity levels, the user feedback mechanism information is recorded and the reasons for the classification of the information of the user feedback mechanism are analyzed;

[0039] When the information from the user feedback mechanism is not relevant to the partitioning results, the microservice architecture is adjusted.

[0040] As a preferred solution of the machine intelligence-based application system integration method described in the present invention, the adjustment of the microservice architecture includes combining automated testing to automatically collect information of the user feedback mechanism that is irrelevant to the division results, changing the information knowledge graph of machine intelligence, adding the construction of a machine intelligence retrieval model, and completing a variety of information retrieval services.

[0041] In the second aspect, an embodiment of the present invention provides an application system integration system based on machine intelligence, which includes: a processing module, which captures Internet fusion technology through multi-source data in the application system and processes the Internet fusion technology using a distributed computing framework; an optimization module, which constructs a machine intelligence retrieval model based on the information knowledge graph of machine intelligence, and optimizes the parameters of the retrieval model through machine learning; an integration module, which uses a microservice architecture to integrate the optimized retrieval model parameters, divides the integrated retrieval model, matches the division results through a user feedback mechanism, and completes information retrieval services based on application system integration.

[0042] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the processor executes the computer program, it implements any step of the above-mentioned machine intelligence-based application system integration method.

[0043] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the above-mentioned machine intelligence-based application system integration method is implemented.

[0044] The beneficial effects of the present invention are: the present invention can significantly improve the efficiency and accuracy of information retrieval services by adopting advanced means such as multi-source data capture and processing based on machine intelligence, optimization of retrieval model parameters, and microservice architecture integration. The present invention can not only effectively respond to the challenges of massive data on the Internet and achieve fast and accurate information retrieval, but also continuously optimize the retrieval model through a continuous user feedback mechanism to ensure the continuous improvement of the quality of retrieval services. In addition, the present invention also has a high degree of flexibility and scalability, supports changing information retrieval needs and rapid iteration of services, and provides users with a more personalized and intelligent information retrieval experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0046] Figure 1 A specific flow chart of a machine intelligence-based application system integration method and system provided in one embodiment of the present invention.

[0047] Figure 2 A schematic diagram of an application system integration method based on machine intelligence and a system integration method is provided for one embodiment of the present invention. DETAILED DESCRIPTION

[0048] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only 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 ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0049] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0050] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0051] The present invention is described in detail with reference to schematic diagrams. For ease of illustration, cross-sectional views of device structures may be partially enlarged and not to scale when describing embodiments of the present invention. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0052] In the description of the present invention, it should be noted that the terms "upper, lower, inner, and outer" and other references to orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first, second, or third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0053] In this disclosure, unless otherwise specified or limited, the terms "mounted," "connected," and "connected" should be interpreted broadly. For example, they may refer to fixed, removable, or integral connections. They may also refer to mechanical, electrical, or direct connections, indirect connections through an intermediary, or internal communication between two components. Those skilled in the art will understand the specific meanings of these terms in this disclosure.

[0054] Example 1

[0055] Reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides an application system integration method based on machine intelligence, including:

[0056] S1: Capture Internet fusion technology through multi-source data within the application system and process it using a distributed computing framework.

[0057] Among them, Internet fusion technology includes using the Internet to collect massive information data and crawling through multi-source data within the application system. The crawling steps are as follows:

[0058] Define pre-processing rules for multi-source data within the application system;

[0059] Add exception handling mechanism to verify the defined preprocessing rules;

[0060] Use automated tools to manage pre-processing rules that do not show any anomalies, triggering the application system display page that captures massive amounts of information data;

[0061] Defining pre-processing rules for multi-source data within the application system includes introducing a character encoding conversion function. During the stage of defining pre-processing rules for multi-source data within the application system, the field positions of the multi-source data are recorded and identified.

[0062] Use machine intelligence learning algorithms to train field locations, and use the sample data obtained during the training process to train the model and complete the location of the field.

[0063] Adding an exception handling mechanism includes setting a defined request timeout in the application system display page;

[0064] When the definition process captures an abnormal character encoding in the application system display page, the pre-processing rule stage of defining the multi-source data in the application system times out, and the abnormal information is recorded at this time.

[0065] Furthermore, through the multi-source data interface within the application system, massive information data is collected from various channels on the Internet, such as social media, news websites, etc. The average daily data volume can reach tens of TB. First, the pre-processing rules of the multi-source data in the application system are defined. This step includes introducing a character encoding conversion function to solve the problem of character encoding differences between different data sources. At the same time, the field positions of the multi-source data are recorded. Identifying the positions of these fields is crucial for subsequent data processing. Experimental results show that this method can achieve a field position recognition rate of more than 98%. Then, the field position is trained using a machine intelligence learning algorithm, and a large amount of sample data accumulated during the training process is used to build a model to achieve field Accurately identify the location. To ensure the effectiveness and stability of the preprocessing rules, an exception handling mechanism is added, including setting the request timeout to 30 seconds in the application system display page. When encountering abnormal character encoding during the preprocessing rule definition process, the system will automatically time out and record the exception information, thereby reducing invalid waiting time and improving processing efficiency. Statistics show that this exception handling mechanism can reduce approximately 80% of invalid waiting time. Finally, an automated tool is used to manage those preprocessing rules that do not have exceptions, triggering the data crawling process of the application system display page to ensure the efficiency and stability of data crawling. The entire process fully reflects the advantages of the present invention in processing large-scale Internet data.

[0066] S1.1: Verifying the defined preprocessing rules involves executing tests to verify the preprocessing rules. The preprocessing rule scenarios include both abnormal and normal information. Test execution includes both automated and manual testing.

[0067] When abnormal information is recorded, manual testing is used to verify the preprocessing rules;

[0068] When no abnormal information is recorded, normal information is recorded, and the pre-processing rules are verified by automated testing;

[0069] Utilize manual testing to verify pre-processing rules, including dynamically adjusting pre-processing rules based on changes in exception information and information from user feedback mechanisms;

[0070] Utilizing the test execution to verify the pre-processing rules includes dynamically adjusting the number of retries and the time interval according to the number of recorded exception information and the time interval, generating automated tests based on the dynamically adjusted number of retries and the time interval, and completing the function of automatically feeding back exception information.

[0071] Furthermore, verifying the defined preprocessing rules includes using test execution to test the effectiveness of the preprocessing rules. The test scenarios include both abnormal information and normal information. The test execution methods are divided into automated testing and manual testing. When the system records abnormal information, manual testing is used to verify the preprocessing rules. Based on the changing trend of the abnormal information and the information provided by the user feedback mechanism, the manual test dynamically adjusts the preprocessing rules to adapt to changes in the data source and improve the accuracy of data processing. When no abnormal information is recorded, that is, the system records normal information, automated testing is used to verify the preprocessing rules. Automated testing can automatically adjust the number of retries and time intervals based on the number of recorded abnormal information and time intervals. For example, if the system detects that a certain abnormal information has appeared more than 5 times in the past 24 hours, it will automatically increase the frequency of checking for the abnormality and reduce the retry time interval, thereby identifying and handling abnormal situations more quickly. Based on the dynamically adjusted number of retries and time intervals, the system can automatically generate a new automated test process to achieve automatic feedback and processing of abnormal information. The entire process ensures the effectiveness and reliability of the data preprocessing rules and improves the efficiency and quality of data processing.

[0072] S1.2: Processing of Internet integration technology includes assigning weights to abnormal information using a task scheduling algorithm within a distributed computing framework based on character encoding conversion functions;

[0073] The weight allocation of abnormal information includes calculating the dynamic adjustment factor based on the adjustment task queue length and the resource utilization rate of the application system during the dynamic adjustment process. The specific formula for calculating the dynamic adjustment factor is:

[0074]

[0075] Among them, γ and δ represent weight coefficients, represents the ratio of the task queue length, (1-U) represents the remaining amount of resource utilization of the application system, U represents the resource utilization of the application system, and D i represents the dynamic adjustment factor;

[0076] By dynamically adjusting factors, we build a machine intelligence information knowledge graph based on the severity and frequency of abnormal information. Severity includes fatal errors, errors, warnings, and notifications, and frequency includes the number of occurrences of the analyzed severity.

[0077] According to the changes in the dynamic adjustment factors, the node positions of the information knowledge graph are dynamically updated to complete the construction of the machine intelligent retrieval model.

[0078] Furthermore, the task scheduling algorithm of the distributed computing framework is used to intelligently allocate the weight of abnormal information. The dynamic adjustment factor is calculated by dynamically adjusting the task queue length and the resource utilization of the application system. For example, when the task queue length ratio increases by 20% and the remaining resource utilization of the application system drops to 10%, the dynamic adjustment factor will increase accordingly to optimize task allocation and improve processing efficiency. This process takes into account the severity of abnormal information (such as fatal errors, errors, warnings and notifications) and its frequency of occurrence. By analyzing the number of occurrences of abnormal information of different severities within a specific time period, for example, no more than 5 fatal errors are recorded per month, while warning information may be as high as hundreds of times, a machine intelligence information knowledge graph is constructed based on this. The change of the dynamic adjustment factor directly affects the update of the node position in the information knowledge graph, ensuring that the information graph can reflect the latest abnormal information distribution status in real time, and finally completing the construction of the machine intelligence retrieval model. This process not only improves the pertinence and effectiveness of abnormal information processing, but also provides accurate data support for subsequent information retrieval services.

[0079] S2: Based on the information knowledge graph of machine intelligence, a machine intelligence retrieval model is constructed, and the parameters of the retrieval model are optimized through machine learning.

[0080] Among them, optimizing the parameters of the retrieval model includes deriving abnormal information data from the information knowledge graph for training and labeling the training data;

[0081] Extracting abnormal information data from the information knowledge graph for training includes standardizing the character encoding conversion function. Standardization includes using one-hot encoding to convert character encodings into numerical variables, iterating the numerical variables N times to obtain numerical variable rules, traversing all numerical variable rules, evaluating the performance of each numerical variable rule, and performing data segmentation based on the performance. The data segmentation includes training sets, validation sets, and test sets.

[0082] The training set includes a probability model of the target function constructed according to the next numerical variable rule of N iterations;

[0083] The validation set includes sampling points for searching the probability model in all numerical variable rules;

[0084] The test set includes optimization of sampling points based on information from the user feedback mechanism.

[0085] Preferably, the process of constructing a machine intelligence retrieval model based on the information knowledge graph of machine intelligence includes deriving abnormal information data from the information knowledge graph for training and annotating the training data. Specifically, after deriving the abnormal information data, the character encoding conversion function is first standardized, and the character encoding is converted into a numerical variable using a one-hot encoding method. The numerical variable is iterated multiple times (such as 1000 times) to extract the numerical variable rules, and then all the numerical variable rules are traversed to evaluate the performance of each rule. After the performance evaluation, the data is divided into three parts: a training set, a validation set, and a test set. The training set is used to extract the numerical variable rules according to the performance of each iteration. Next, a numerical variable rule is used to construct a probabilistic model of the objective function. The validation set is used to search for the optimal sampling points of the probabilistic model among all numerical variable rules. The test set is used to further optimize these sampling points based on information from the user feedback mechanism to ensure the accuracy and practicality of the model. For example, in one practice, the training set accounts for 70% of the total data volume, the validation set accounts for 15%, and the test set also accounts for 15%. Through this data segmentation strategy, the model training process can fully utilize the existing data while maintaining the model's good predictive ability for future data, ultimately achieving effective optimization of the retrieval model parameters. The key parameters in the retrieval model process are shown in Table 1 below:

[0086] Table 1 Key parameters in the retrieval model process

[0087]

[0088] Table 1 shows the key parameter settings for building machine intelligence retrieval models in five different practical cases, including the number of iterations, dataset split ratio (training set, validation set, test set), character encoding conversion method, data preprocessing method, and model optimization goal. Comparing these parameters reveals that model training details were adjusted according to specific needs in different cases. For example, Cases A and D both used a 70% training set ratio and one-hot encoding, but the number of iterations was 1000 and 1200, respectively, with the optimization goals of improving model prediction accuracy and increasing model training speed. This data helps us understand how to optimize model performance through flexible parameter adjustments.

[0089] S3: Use the microservice architecture to integrate the optimized retrieval model parameters, divide the integrated retrieval model, match the division results through the user feedback mechanism, and complete the information retrieval service based on application system integration.

[0090] Integrating the optimized retrieval model parameters includes integrating the optimized sampling points into the microservice architecture of the application system. The microservice architecture includes data acquisition services, data preprocessing services, model training services, model deployment services, user feedback services, and information retrieval services.

[0091] The integrated retrieval model is divided into functional divisions and regional divisions. The functional divisions include data preprocessing services, model training services, and user feedback services.

[0092] The regional division includes data acquisition services, model deployment services, and information retrieval services;

[0093] According to the division results of the retrieval model, the user feedback mechanism is traced;

[0094] Verify the information of the user feedback mechanism and the division results. The verification steps are as follows:

[0095] When the information of the user feedback mechanism is inconsistent with the segmentation results, the exported abnormal information data is trained again;

[0096] When the information of the user feedback mechanism is classified into any of the serious conditions, the user feedback mechanism information is recorded and the reasons for the classification of the information of the user feedback mechanism are analyzed;

[0097] When the information from the user feedback mechanism is not relevant to the partitioning results, the microservice architecture is adjusted.

[0098] Furthermore, the optimized sampling points are integrated into the microservice architecture of the application system, including data acquisition service, data preprocessing service, model training service, model deployment service, user feedback service and information retrieval service. The integrated retrieval model is divided according to function and service area. Functional division covers data preprocessing service, model training service and user feedback service, while regional division involves data acquisition service, model deployment service and information retrieval service. This division method ensures the independence and flexibility of each service, and is easy to manage and maintain. According to the division results of the retrieval model, the system can complete the traceability of the user feedback mechanism. When the user feedback information is inconsistent with the division results, the system will re- The exported abnormal information data is trained to improve the accuracy of the model. For example, in one practice, the system detected that 10% of user feedback did not match the model classification results, and then started the retraining process. When the user feedback information belongs to any category of severity, such as fatal errors or warnings, the system will record these feedback information and conduct in-depth analysis of the reasons for such feedback for subsequent improvements. In the case that the user feedback information is completely irrelevant to the model classification results, the system will make appropriate adjustments to the microservice architecture to ensure service efficiency and user satisfaction. For example, if it is found that the correlation between user feedback and information retrieval services is lower than the expected 80%, the system will adjust the service configuration to better meet user needs.

[0099] S3.1: Adjustments to the microservice architecture include combining automated testing to automatically collect information from user feedback mechanisms that are irrelevant to the segmentation results, changing the information knowledge graph of machine intelligence, adding the construction of machine intelligence retrieval models, and completing diverse information retrieval services.

[0100] In a preferred embodiment, an application system integration system based on machine intelligence includes a processing module, which captures Internet fusion technology through multi-source data in the application system and processes the Internet fusion technology using a distributed computing framework; an optimization module, which builds a machine intelligence retrieval model based on the information knowledge graph of machine intelligence and optimizes the parameters of the retrieval model through machine learning; an integration module, which integrates the optimized retrieval model parameters using a microservice architecture, divides the integrated retrieval model, matches the division results through a user feedback mechanism, and completes information retrieval services based on application system integration.

[0101] The above-mentioned unit modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of the above-mentioned modules.

[0102] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0103] In summary, the present invention can significantly improve the efficiency and accuracy of information retrieval services by adopting advanced means such as multi-source data capture and processing based on machine intelligence, optimization of retrieval model parameters, and microservice architecture integration. The present invention can not only effectively cope with the challenges of massive data on the Internet and achieve fast and accurate information retrieval, but also continuously optimize the retrieval model through a continuous user feedback mechanism to ensure the continuous improvement of the quality of retrieval services. In addition, the present invention also has a high degree of flexibility and scalability, supports changing information retrieval needs and rapid iteration of services, and provides users with a more personalized and intelligent information retrieval experience.

[0104] Example 2

[0105] Reference Figure 1 and Figure 2 , which is the second embodiment of the present invention, provides an application system integration method based on machine intelligence. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0106] During the experiment of this technical solution, a large amount of information data was first collected from Internet channels through the multi-source data interface within the application system. For example, during the experiment, an average of about 15TB of data was captured from more than 1,200 different websites or platforms every day. The preprocessing rules for multi-source data were defined, including the introduction of character encoding conversion function and recording the field positions of multi-source data. The recognition rate reached 99%. The field positions were trained through the machine intelligence learning algorithm. After 1,200 iterations, the model was able to accurately identify and locate the key fields in the multi-source data. In order to ensure the effectiveness of the preprocessing rules, a 30-second request timeout was set in the application system display page, and the exception information was recorded through the exception handling mechanism. The experimental results showed that the exception handling mechanism reduced the invalid waiting time by about 85%.

[0107] Abnormal information data was extracted from the information knowledge graph for training, and the training data was annotated. The training data was converted into numerical variables using the one-hot encoding method, and 1,500 iterations were performed to refine the rules of the numerical variables. The data was split into a training set (72%), a validation set (18%), and a test set (10%) to ensure the comprehensiveness and accuracy of the model training. Using these data sets, a probabilistic model of the objective function was constructed, and the optimal sampling points were searched in the validation set. In the test set, these sampling points were further optimized based on information from the user feedback mechanism to ensure the accuracy and practicality of the model.

[0108] The optimized retrieval model parameters are integrated into the microservice architecture of the application system. The architecture includes data acquisition service, data preprocessing service, model training service, model deployment service, user feedback service and information retrieval service. The model is divided according to function and service area, and feedback is obtained through users. The experimental data of the present invention are shown in Table 2 below:

[0109] Table 2 Experimental data table of the present invention

[0110] Data Type Specific values Number of daily data sources More than 1,200 websites or platforms Average daily crawled data volume About 15TB Field position recognition rate 99％ Field position training iterations 1200 times Request timeout 30 seconds Exception handling mechanism reduces invalid waiting time About 85% Number of iterations of numerical variable rules 1500 times Training set ratio 72％ Validation set ratio 18％ Test set ratio 10％

[0111] Table 2 shows the specific parameter settings and experimental results for each key step in the process of building an application system integrating information retrieval services based on machine intelligence. During the experiment, an average of approximately 15TB of data was crawled from more than 1,200 websites or platforms every day, with a field location recognition rate of 99%. The model was trained through 1,200 iterations to accurately identify key fields. A 30-second request timeout was set on the application system display page. The exception handling mechanism reduced invalid waiting time by approximately 85%. The training data was one-hot encoded through 1,500 iterations and divided into a training set of 72%, a validation set of 18%, and a test set of 10%. These data helped optimize the model parameters and ensure the accuracy and practicality of the model. A comparison of the present invention with the prior art is shown in Table 3 below:

[0112] Table 3 Comparison between the present invention and the prior art

[0113]

[0114]

[0115] Table 3 demonstrates the significant advantages of the technical solution of the present invention in terms of data capture capability, field location recognition rate, exception handling mechanism, number of model training iterations, dataset segmentation ratio, user feedback mechanism, and microservice architecture, effectively improving the overall performance and user experience of information retrieval services.

[0116] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for integrating an application system based on machine intelligence, characterized by: include: Capture Internet fusion technology through multi-source data within the application system and process it using a distributed computing framework; Based on the machine intelligence information knowledge graph, a machine intelligence retrieval model is constructed, and the parameters of the retrieval model are optimized through machine learning; Adopting a microservice architecture, the optimized retrieval model parameters are integrated, the integrated retrieval model is divided, and the division results are matched through a user feedback mechanism to complete the information retrieval service based on application system integration. The Internet fusion technology includes collecting massive amounts of information data through the Internet. The Internet fusion technology includes capturing massive amounts of information data through the multi-source data in the application system. The capturing steps are as follows: Define pre-processing rules for multi-source data within the application system; Add exception handling mechanism to verify the defined preprocessing rules; Use automated tools to manage pre-processing rules that do not show any anomalies, triggering the application system display page that captures massive amounts of information data; Defining the pre-processing rules for multi-source data in the application system includes introducing a character encoding conversion function. In the stage of defining the pre-processing rules for multi-source data in the application system, the field positions of the multi-source data are recorded and the field positions of the multi-source data are identified; Use machine intelligence learning algorithms to train field locations, and use the sample data obtained during the training process to train the model and complete the location of the field. The adding of the exception handling mechanism includes setting a defined request timeout in the application system display page; When the definition process captures an abnormal character encoding in the application system display page, the pre-processing rule phase of defining the multi-source data in the application system times out, and the abnormal information is recorded. The verification of the defined pre-processing rules includes verifying the pre-processing rules by using test execution, the scenarios of the pre-processing rules include abnormal information and normal information, and the test execution includes automated testing and manual testing; When abnormal information is recorded, manual testing is used to verify the preprocessing rules; When no abnormal information is recorded, normal information is recorded, and the pre-processing rules are verified by automated testing; The verification of the pre-processing rules by manual testing includes dynamically adjusting the pre-processing rules according to changes in abnormal information and information from a user feedback mechanism; The use of test execution to verify the pre-processing rules includes dynamically adjusting the number of retries and the time interval according to the number of recorded exception information and the time interval, and generating an automated test based on the dynamically adjusted number of retries and the time interval to complete the function of automatically feeding back the exception information; The processing of the Internet fusion technology includes allocating weights of abnormal information using a task scheduling algorithm of a distributed computing framework based on a character encoding conversion function; The weight allocation of the abnormal information includes calculating a dynamic adjustment factor based on the adjustment task queue length and the resource utilization rate of the application system during the dynamic adjustment process. The specific formula for calculating the dynamic adjustment factor is: ， in, and represents the weight coefficient, represents the ratio of the task queue length, Indicates the remaining amount of resource utilization of the application system. Indicates the resource utilization of the application system. represents the dynamic adjustment factor; By dynamically adjusting factors, a machine intelligence information knowledge graph is constructed based on the severity and frequency of abnormal information, where the severity includes fatal errors, errors, warnings, and notifications, and the frequency includes analyzing the number of occurrences of the severity; According to the changes in the dynamic adjustment factor, the node positions of the information knowledge graph are dynamically updated to complete the construction of the machine intelligent retrieval model; Optimizing the parameters of the retrieval model includes deriving abnormal information data from the information knowledge graph for training, and labeling the training data; The training of the abnormal information data derived from the information knowledge graph includes standardizing the character encoding conversion function, wherein the standardization includes converting the character encoding into a numerical variable using one-hot encoding, obtaining a numerical variable rule by iterating the numerical variable N times, traversing all numerical variable rules, evaluating the performance of each numerical variable rule, and performing data segmentation on the performance, wherein the data segmentation includes a training set, a validation set, and a test set; The training set includes a probability model for constructing an objective function according to the next numerical variable rule of N iterations; The validation set includes sampling points for searching the probability model in all numerical variable rules; The test set includes optimizing sampling points based on information from a user feedback mechanism.

2. The machine intelligence-based application system integration method according to claim 1, wherein: Integrating the optimized retrieval model parameters includes integrating the optimized sampling points into the microservice architecture of the application system, wherein the microservice architecture includes data acquisition service, data preprocessing service, model training service, model deployment service, user feedback service and information retrieval service; The division of the integrated retrieval model includes functional division and regional division, wherein the functional division includes data preprocessing service, model training service and user feedback service; The said area division includes data acquisition services, model deployment services and information retrieval services; According to the division results of the retrieval model, the user feedback mechanism is traced; Verify the information of the user feedback mechanism and the division results. The verification steps are as follows: When the information of the user feedback mechanism is inconsistent with the segmentation results, the exported abnormal information data is trained again; When the information of the user feedback mechanism is classified into any of the aforementioned severity levels, the user feedback mechanism information is recorded and the reasons for the classification of the information of the user feedback mechanism are analyzed; When the information from the user feedback mechanism is not relevant to the partitioning results, the microservice architecture is adjusted.

3. The machine intelligence-based application system integration method according to claim 2, wherein: The adjustment of the microservice architecture includes combining automated testing to automatically collect information of the user feedback mechanism that is not related to the division results, changing the information knowledge graph of machine intelligence, adding the construction of a machine intelligence retrieval model, and completing a variety of information retrieval services.

4. A machine intelligence-based application system integration system, based on the machine intelligence-based application system integration method according to any one of claims 1 to 3, characterized in that: include, A processing module, which captures Internet fusion technology through multi-source data within the application system and processes the Internet fusion technology using a distributed computing framework; The optimization module builds a machine intelligence retrieval model based on the machine intelligence information knowledge graph and optimizes the parameters of the retrieval model through machine learning; The integration module uses a microservice architecture to integrate the optimized retrieval model parameters, divides the integrated retrieval model, matches the division results through a user feedback mechanism, and completes the information retrieval service based on application system integration.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the machine intelligence-based application system integration method according to any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the machine intelligence-based application system integration method according to any one of claims 1 to 3 are implemented.

Images