Operation and maintenance method and device of business system, storage medium and electronic equipment

Abstract

The invention discloses an operation and maintenance method and device of a service system, a storage medium and electronic equipment, the method is applied to the field of artificial intelligence, and the method comprises the following steps: collecting original time sequence data of the service system, and carrying out data noise reduction processing on the original time sequence data to obtain noise-reduced time sequence data; performing data preprocessing and data recombination operation on the time sequence data after noise reduction to obtain target time sequence data; combining the target time sequence data and the user intention of the target user to obtain reference information of the user intention; and inputting the reference information of the user intention into the large language model, generating an operation and maintenance result of the service system, and playing the operation and maintenance result to a target user, so that the target user performs operation and maintenance operation on the service system based on the operation and maintenance result. Through the operation and maintenance method and device, the problem that operation and maintenance personnel of the service system need to perform operation and maintenance operation on the service system based on the operation and maintenance report after complex analysis in the related technology, so that the real-time performance of the operation and maintenance operation of the service system is poor is solved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and more specifically, to a method, device, storage medium, and electronic device for operating and maintaining a business system. Background Art

[0002] With the rapid development of business, the amount of data and complexity of business systems are growing exponentially, placing higher demands on operations and maintenance inspections. Traditional operations and maintenance inspection methods primarily involve periodic manual inspections and automatic monitoring based on fixed thresholds. The main problems with traditional operations and maintenance inspection methods include: inspection content is limited by the experience and knowledge of operations and maintenance personnel, which can easily lead to incomplete inspection coverage; inspection content relies on fixed thresholds, which make it difficult to detect potential risks when system-related indicators do not meet thresholds; inspection results are inconvenient to view, requiring inspection reports to view results, making it impossible to quickly perceive system status; configuration is cumbersome, requiring manual creation and addition of inspection rules, which has high learning and labor costs.

[0003] In related technologies, business system operation and maintenance personnel need to perform operation and maintenance operations on the business system based on operation and maintenance reports after complex analysis, which leads to poor real-time performance of the operation and maintenance operations of the business system. No effective solution has been proposed so far. Summary of the Invention

[0004] The main purpose of this application is to provide a business system operation and maintenance method, device, storage medium and electronic equipment to solve the problem in related technologies that business system operation and maintenance personnel need to perform operation and maintenance operations on the business system based on operation and maintenance reports after complex analysis, resulting in poor real-time performance of operation and maintenance operations on the business system.

[0005] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a business system operation and maintenance method is provided, which includes: collecting original time series data of the business system, and performing data noise reduction processing on the original time series data to obtain noise-reduced time series data; performing data preprocessing operations and data reorganization operations on the noise-reduced time series data to obtain target time series data; merging the target time series data and the user intention of the target user to obtain reference information of the user intention, wherein the user intention is obtained by performing intent recognition on first text information through a large language model, and the first text information is obtained after processing the output speech of the target user; inputting the reference information of the user intention into the large language model to generate an operation and maintenance result of the business system, and playing the operation and maintenance result to the target user, so that the target user performs operation and maintenance operations on the business system based on the operation and maintenance result.

[0006] Furthermore, data preprocessing operations and data reorganization operations are performed on the denoised time series data to obtain target time series data, including: performing feature extraction operations on multiple time series data of any indicator in the denoised time series data to obtain a first feature vector set; calculating the similarity between each first feature vector in the first feature vector set and the vector corresponding to the preset indicator threshold based on the cosine similarity algorithm to obtain a similarity set; sorting the similarities in the similarity set in descending order, and determining a target similarity whose sorting order is lower than the first preset threshold; determining a target feature vector corresponding to the target similarity, and determining the target time series data based on the time series data corresponding to the target feature vector.

[0007] Furthermore, a feature extraction operation is performed on the multiple time series data to obtain a first feature vector set, including: data sampling is performed on the multiple time series data based on a preset time window to obtain multiple time series data after sampling; statistical features are extracted from the multiple time series data after sampling based on a preset statistical algorithm, and the statistical features are used as vector index values to obtain the first feature vector set.

[0008] Furthermore, the target time series data and the user intent of the target user are merged to obtain reference information of the user intent, including: determining operation and maintenance rules related to the user intent, wherein the operation and maintenance rules are pre-input rules for running the business system; filtering data in the target time series data according to the operation and maintenance rules to obtain filtered target time series data; converting the filtered target time series data into text data to obtain reference information of the user intent.

[0009] Furthermore, the original time series data is subjected to data denoising processing to obtain the denoised time series data, including: calculating the average value of multiple time series data in the original time series data, wherein the multiple time series data refer to the time series data in the original time series data whose collection time is earlier than the preset time; using a time series data analysis algorithm to calculate the average value and the index data at each moment in the original time series data to obtain the first time series data; calculating the difference between the index data at each moment in the first time series data and the index data at each moment in the original time series data one by one to obtain the difference at each moment; deleting noise data from the original time series data to obtain the denoised time series data, wherein the noise data is the index data corresponding to the target moment where the difference is greater than the second preset threshold.

[0010] Furthermore, before merging the target time series data and the user intention of the target user to obtain reference information of the user intention, the method also includes: deploying a target program in the target device, controlling the microphone in the target device through the target program to receive the first voice information output by the target user; transmitting the first voice information received by the microphone to a voice conversion tool through the target program based on a wireless communication network; and controlling the voice conversion tool through the target program to convert the first voice information into text information to obtain the first text information.

[0011] Furthermore, the reference information of the user intention is input into the large language model to generate the operation and maintenance results of the business system, including: generating a target template based on the user intention and preset prompt words, wherein the target template is used to instruct the large language model to output operation and maintenance operation suggestions based on the current operation status of the business system; using the reference information of the user intention to fill the target template and input it into the large language model, and generating the operation and maintenance results of the business system through the large language model.

[0012] In order to achieve the above-mentioned purpose, according to another aspect of the present application, an operation and maintenance device for a business system is provided, which includes: an acquisition unit, which is used to acquire original time series data of the business system, and perform data noise reduction processing on the original time series data to obtain noise-reduced time series data; a reorganization unit, which is used to perform data preprocessing operations and data reorganization operations on the noise-reduced time series data to obtain target time series data; a merging unit, which is used to merge the target time series data and the user intention of the target user to obtain reference information of the user intention, wherein the user intention is obtained by performing intent recognition on first text information through a large language model, and the first text information is obtained after processing the output voice of the target user; a playback unit, which is used to input the reference information of the user intention into the large language model, generate an operation and maintenance result of the business system, and play the operation and maintenance result to the target user, so that the target user can perform operation and maintenance operations on the business system based on the operation and maintenance result.

[0013] Furthermore, the reorganization unit includes: an operation subunit, which is used to perform feature extraction operations on multiple time series data of any indicator in the denoised time series data to obtain a first feature vector set; a first calculation subunit, which is used to calculate the similarity between each first feature vector in the first feature vector set and the vector corresponding to the preset indicator threshold based on the cosine similarity algorithm to obtain a similarity set; a sorting subunit, which is used to sort the similarities in the similarity set in descending order, and determine a target similarity whose sorting order is lower than the first preset threshold; a determination subunit, which is used to determine a target feature vector corresponding to the target similarity, and determine the target time series data based on the time series data corresponding to the target feature vector.

[0014] Furthermore, the operation sub-unit includes: a sampling module, which is used to sample the multiple time series data based on a preset time window to obtain multiple time series data after sampling; an extraction module, which is used to extract statistical features from the multiple time series data after sampling based on a preset statistical algorithm, and use the statistical features as vector index values to obtain the first feature vector set.

[0015] Furthermore, the merging unit includes: a determination subunit, used to determine the operation and maintenance rules related to the user intention, wherein the operation and maintenance rules are pre-input rules for running the business system; a screening subunit, used to screen data in the target time series data according to the operation and maintenance rules to obtain screened target time series data; a conversion subunit, used to convert the screened target time series data into text data to obtain reference information of the user intention.

[0016] Furthermore, the acquisition unit includes: a second calculation subunit, used to calculate the average value of multiple time series data in the original time series data, wherein the multiple time series data refer to the time series data in the original time series data whose collection time is earlier than the preset time; a third calculation subunit, used to use a time series data analysis algorithm to calculate the average value and the indicator data at each moment in the original time series data to obtain the first time series data; a fourth calculation subunit, used to calculate the difference between the indicator data at each moment in the first time series data and the indicator data at each moment in the original time series data one by one to obtain the difference at each moment; a deletion subunit, used to delete noise data in the original time series data to obtain the denoised time series data, wherein the noise data is the indicator data corresponding to the target moment where the difference is greater than the second preset threshold.

[0017] Furthermore, the device also includes: a receiving unit, which is used to deploy a target program in the target device before merging the target time series data and the user intention of the target user to obtain reference information of the user intention, and control the microphone in the target device through the target program to receive the first voice information output by the target user; a transmission unit, which is used to transmit the first voice information received by the microphone to a voice conversion tool through the target program based on a wireless communication network; and a conversion unit, which is used to control the voice conversion tool through the target program to convert the first voice information into text information to obtain the first text information.

[0018] Furthermore, the playback unit includes: a first generation sub-unit, used to generate a target template based on the user intention and a preset prompt word, wherein the target template is used to instruct the large language model to output operation and maintenance suggestions based on the current operation status of the business system; a second generation sub-unit, used to fill the target template with reference information of the user intention, and input it into the large language model, and generate the operation and maintenance results of the business system through the large language model.

[0019] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a computer program product is provided, including a computer program, which, when executed by a processor, implements the operation and maintenance method of the business system described in any one of the above-mentioned items, and which, when executed by a processor, implements the steps of the operation and maintenance method of the business system described in each embodiment of the present application.

[0020] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes stored computer instructions, wherein when the computer instructions are executed by a processor, an operation and maintenance method of the business system described in any one of the above-mentioned items is implemented.

[0021] In order to achieve the above-mentioned purpose, according to one aspect of the present application, an electronic device is provided, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors implement the operation and maintenance method of the business system described in any one of the above-mentioned items.

[0022] Through the present application, the following steps are adopted: collecting original time series data of the business system, and performing data noise reduction processing on the original time series data to obtain noise-reduced time series data; performing data preprocessing operations and data reorganization operations on the noise-reduced time series data to obtain target time series data; merging the target time series data and the user intention of the target user to obtain reference information of the user intention, wherein the user intention is obtained by performing intent recognition on first text information through a large language model, and the first text information is obtained after processing the output voice of the target user; inputting the reference information of the user intention into the large language model to generate the operation and maintenance results of the business system, and playing the operation and maintenance results to the target user, so that the target user can perform operation and maintenance operations on the business system based on the operation and maintenance results, thereby solving the problem in the related technology that the operation and maintenance personnel of the business system need to perform operation and maintenance operations on the business system based on operation and maintenance reports after complex analysis, resulting in poor real-time performance of operation and maintenance operations on the business system.

[0023] By collecting the original time series data of the business system and performing data noise reduction, data preprocessing, and data reorganization operations on it, the target time series data can be extracted to obtain more concise and noise-reduced time series data that accurately reflects the system's operating status. This provides an accurate data foundation for subsequent AI big model analysis, thereby significantly improving the accuracy of subsequent analysis. At the same time, by using the AI big model to recognize the intent of the text converted from the operation and maintenance personnel's voice query, reference information of the user's intent is generated, the target time series data is integrated with the user's intent, and this integrated reference information is input into the AI big model to generate business system operation and maintenance results that meet user needs. The results are then immediately conveyed to the operation and maintenance personnel through voice broadcast. This not only ensures the pertinence and efficiency of the operation and maintenance results generated by the AI big model, but also makes the operation and maintenance operations more intuitive and rapid, greatly reducing the burden on operation and maintenance personnel to analyze complex data, realizing real-time operation and maintenance operations based on the operating status of the business system, and providing real-time operation and maintenance decision support, achieving the technical effect of significantly improving operation and maintenance efficiency and accuracy, further optimizing the work experience of operation and maintenance personnel and ensuring the stable and efficient operation of the business system. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0025] Figure 1 This is a flowchart of the operation and maintenance method of the business system provided in Example 1 of the present application;

[0026] Figure 2This is a schematic diagram of the structure of an optional intelligent inspection tool based on voice interaction technology provided in accordance with the first embodiment of the present application;

[0027] Figure 3 This is a schematic diagram of the processing flow of time series data by the optional intelligent analysis system provided in Example 1 of the present application;

[0028] Figure 4 This is a schematic diagram of an optional process for outputting voice-based operation and maintenance inspection results through an intelligent inspection tool according to the first embodiment of the present application;

[0029] Figure 5 This is a schematic diagram of an operation and maintenance device of a business system provided in accordance with the second embodiment of the present application;

[0030] Figure 6 This is a schematic diagram of the operation and maintenance electronic equipment of the business system provided in Example 5 of the present application. DETAILED DESCRIPTION

[0031] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0032] It should be noted that the user information (including but not limited to user device information, user personal information, collected data, used data, generated data, processed data, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, collected information, used information, generated information, processed information, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data comply with the relevant laws, regulations and standards of relevant countries and regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse. For example, an interface is set up between this system and relevant users or institutions. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or institution through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or institution.

[0033] It should be noted that this application provides users with corresponding operation entrances for them to choose to agree or reject the automated decision-making results; if the user chooses to reject, the expert decision-making process will be entered.

[0034] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0035] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0036] Example 1

[0037] The present invention will be described below in conjunction with preferred implementation steps. Figure 1 This is a flow chart of the operation and maintenance method of the business system provided in Example 1 of the present application. Figure 1 As shown, the method includes the following steps:

[0038] Step S101: collect original time series data of the business system, and perform data noise reduction processing on the original time series data to obtain noise-reduced time series data.

[0039] In the first embodiment of the present invention, in order to monitor the operating status of the business system in real time and perform operation and maintenance operations on the business system in a timely manner, it is necessary to continuously collect original time series data reflecting its operating status from the business system, and perform noise reduction processing on these original time series data, so as to obtain more concise, smoother and more accurate time series data that reflects the actual operating status of the system, that is, the above-mentioned noise-reduced time series data, which is conducive to providing high-quality data for subsequent operation and maintenance processing, thereby achieving the effect of improving the accuracy and efficiency of operation and maintenance inspections.

[0040] The raw time series data mentioned above refers to raw metric data collected from business systems that reflects the system's operating status and is arranged in chronological order. Time series data may include, but is not limited to, server CPU usage, memory usage, disk I / O, network traffic, application response time, error rate, and other metrics. Time series data is the foundation of monitoring systems and is used to determine the health of business systems.

[0041] Step S102 , performing data preprocessing and data reorganization operations on the denoised time series data to obtain target time series data.

[0042] In this first embodiment, to further improve the response efficiency of operations and maintenance personnel, it is necessary to extract valid data from the time series data. This involves preprocessing the noise-reduced data (e.g., data segmentation, feature extraction, etc.), then selecting the data most relevant to the abnormal state of the business system from the preprocessed data, reorganizing the data, and ultimately generating the target time series data described above. This step allows for the refinement of data in the noise-reduced time series data, highlighting key information, and providing more accurate analytical data for the AI large model, thereby effectively improving the accuracy and efficiency of operations and maintenance decisions.

[0043] In step S103, the target time series data and the user intent of the target user are merged to obtain reference information of the user intent, wherein the user intent is obtained by performing intent recognition on the first text information through a large language model, and the first text information is obtained by processing the output speech of the target user.

[0044] In this first embodiment, to facilitate maintenance of the business system by operation and maintenance personnel (i.e., the target users mentioned above, and other staff such as managers) or to quickly understand the operating status of the business system, the voice query of the operation and maintenance personnel is converted into text (i.e., the first text information mentioned above). The AI large model is used to perform intent recognition, understand the specific needs of the operation and maintenance personnel, and obtain the user intent mentioned above. Next, the preprocessed and reorganized target time series data is combined with the user intent to generate reference information of the user intent.

[0045] The above steps can not only accurately understand the operation and maintenance needs of operation and maintenance personnel, but also provide the most relevant and timely feedback information to operation and maintenance personnel based on the current operating status of the business system, so that operation and maintenance personnel can quickly grasp the system status and respond quickly to potential problems, thereby effectively improving operation and maintenance efficiency and system stability.

[0046] In step S104, the reference information of the user's intention is input into the large language model to generate an operation and maintenance result of the business system, and the operation and maintenance result is played to the target user so that the target user can perform operation and maintenance operations on the business system based on the operation and maintenance result.

[0047] In this first embodiment, to quickly respond to the operator's query intent (i.e., the aforementioned user intent), the merged user intent reference information is used as input. A comprehensive analysis is performed using the AI model to generate an operational result tailored to the operator's query, including system health status and explanations of potential issues. This result is then communicated to the operator via voice broadcast, enabling them to intuitively and quickly understand the current operational status of the business system.

[0048] The above steps not only provide immediate feedback but also reduce the difficulty for operation and maintenance personnel to parse complex data, thereby enabling more rapid and accurate operation and maintenance operations to ensure the stable operation of the business system.

[0049] To summarize, the operation and maintenance method of the business system provided in the first embodiment of the present application collects the original time series data of the business system, and performs data noise reduction processing on the original time series data to obtain the noise-reduced time series data; performs data preprocessing operations and data reorganization operations on the noise-reduced time series data to obtain target time series data; merges the target time series data and the user intention of the target user to obtain reference information of the user intention, wherein the user intention is obtained by performing intent recognition on the first text information through a large language model, and the first text information is obtained after processing the output voice of the target user; the reference information of the user intention is input into the large language model to generate the operation and maintenance results of the business system, and the operation and maintenance results are played to the target user, so that the target user can perform operation and maintenance operations on the business system based on the operation and maintenance results, thereby solving the problem in the related technology that the operation and maintenance personnel of the business system need to perform operation and maintenance operations on the business system based on the operation and maintenance reports after complex analysis, resulting in poor real-time performance of the operation and maintenance operations on the business system.

[0050] By collecting the original time series data of the business system and performing data noise reduction, data preprocessing, and data reorganization operations on it, the target time series data can be extracted to obtain more concise and noise-reduced time series data that accurately reflects the system's operating status. This provides an accurate data foundation for subsequent AI big model analysis, thereby significantly improving the accuracy of subsequent analysis. At the same time, by using the AI big model to recognize the intent of the text converted from the operation and maintenance personnel's voice query, reference information of the user's intent is generated, the target time series data is integrated with the user's intent, and this integrated reference information is input into the AI big model to generate business system operation and maintenance results that meet user needs. The results are then immediately conveyed to the operation and maintenance personnel through voice broadcast. This not only ensures the pertinence and efficiency of the operation and maintenance results generated by the AI big model, but also makes the operation and maintenance operations more intuitive and rapid, greatly reducing the burden on operation and maintenance personnel to analyze complex data, realizing real-time operation and maintenance operations based on the operating status of the business system, and providing real-time operation and maintenance decision support, achieving the technical effect of significantly improving operation and maintenance efficiency and accuracy, further optimizing the work experience of operation and maintenance personnel and ensuring the stable and efficient operation of the business system.

[0051] Optionally, in the operation and maintenance method of the business system provided in Example 1 of the present application, data preprocessing operations and data reorganization operations are performed on the denoised time series data to obtain target time series data, including: performing feature extraction operations on multiple time series data of any indicator in the denoised time series data to obtain a first feature vector set; calculating the similarity between each first feature vector in the first feature vector set and the vector corresponding to the preset indicator threshold based on the cosine similarity algorithm to obtain a similarity set; sorting the similarities in the similarity set in descending order, and determining a target similarity whose sorting order is lower than the first preset threshold; determining a target feature vector corresponding to the target similarity, and determining the target time series data based on the time series data corresponding to the target feature vector.

[0052] In the first embodiment of the present invention, in order to extract key data from time series data, a feature extraction operation can be performed on multiple time series data of each indicator in the denoised time series data, that is, key features that can represent their changing trends and states are extracted from these data, and the first feature vector corresponding to each time series data under each indicator is generated, thereby forming a first feature vector set.

[0053] Then, based on the cosine similarity algorithm, the similarity between each first eigenvector in the first eigenvector set and the vector corresponding to the preset indicator threshold is evaluated to generate a similarity set that describes the angular difference between these vectors. Among them, the preset indicator threshold refers to the indicator limit value when monitoring and analyzing the operating status of the business system, which is an important reference for judging whether the performance of the business system (or application) is within the normal range. It may include but is not limited to the upper or lower limits of key indicators such as CPU utilization, memory usage, network traffic, disk I / O, and response time. When the actual monitored data deviates from these preset thresholds, it is considered that there are potential performance problems or abnormalities in the business system, thereby triggering the corresponding early warning mechanism or further intelligent analysis through the AI big model to help operation and maintenance personnel discover and deal with problems in a timely manner to maintain the stable operation of the business system and optimize performance.

[0054] Secondly, by arranging the similarities in the similarity set in descending order, we can identify target similarities that are significantly different from the normal operating state, further determine the target feature vectors associated with these target similarities, and based on the original time series data corresponding to these feature vectors, locate the specific time series data that exhibits abnormalities and determine them as the above-mentioned target time series data.

[0055] The above steps can accurately filter out anomalies from multi-dimensional data, providing more precise data support for subsequent intelligent analysis and user intent matching, thereby effectively improving the accuracy and timeliness of operation and maintenance decisions. At the same time, it can filter out the time series data that requires the most attention during the operation of the business system from massive time series data, improving the efficiency and accuracy of data processing, so that operation and maintenance problems that may be faced by the business system can be discovered in a timely manner, and the query intentions of the operation and maintenance personnel can be responded to in real time, thereby achieving the effect of improving the operation and maintenance efficiency and stability of the business system.

[0056] Optionally, in the operation and maintenance method of the business system provided in Example 1 of the present application, a feature extraction operation is performed on multiple time series data to obtain a first feature vector set, including: data sampling of multiple time series data based on a preset time window to obtain multiple time series data after sampling; extracting statistical features from the multiple time series data after sampling based on a preset statistical algorithm, and using the statistical features as vector index values to obtain a first feature vector set.

[0057] In the first embodiment of the present invention, the denoised monitoring data sequence can be periodically and quantitatively intercepted by means of a sliding window, that is, the above-mentioned data sampling of multiple time series data based on a preset time window is intended to capture the changes in the operating status of the system within a specific period, and ensure that the analyzed data is both representative and timely. For example, if the time window is set to every minute, the system will extract a fixed-length data segment from the original time series data every minute, and the sampling point can be set to 40 time points to reflect the recent dynamic performance of the system. It should be noted that the above-mentioned preset time window and sampling point can be adjusted according to actual conditions, and no specific restrictions are made in the first embodiment of the present invention.

[0058] After acquiring multiple sampled time series data, you can apply preset statistical algorithms, such as calculating variance, to extract statistical features from these data segments. Variance, an important indicator of data dispersion, reveals the fluctuation of data points relative to the mean and is particularly critical for identifying the stability of system operations.

[0059] For example, the calculated variance value is used as the index value of the vector, that is, the variance feature of each time series data is converted into a numerical vector element, thereby constructing a vector set composed of the features of multiple time series data under any indicator in the denoised time series data, namely the first feature vector set. This conversion process converts the original time series data into a feature vector form that is convenient for subsequent processing and analysis, laying the data foundation for identifying anomalies using the cosine similarity algorithm.

[0060] Through the above steps, we can more efficiently identify time series data segments that may be significantly different from the preset indicator threshold based on variance, that is, potential operation and maintenance problem points, and then provide users with accurate operation and maintenance inspection results and suggestions.

[0061] Optionally, in the operation and maintenance method of the business system provided in Example 1 of the present application, the target time series data and the user intent of the target user are merged to obtain reference information of the user intent, including: determining operation and maintenance rules related to the user intent, wherein the operation and maintenance rules are pre-input rules for operating the business system; performing data filtering in the target time series data according to the operation and maintenance rules to obtain filtered target time series data; converting the filtered target time series data into text data to obtain reference information of the user intent.

[0062] In the first embodiment of the present invention, in order to accurately respond to the query questions of the operation and maintenance personnel, it is necessary to determine the operation and maintenance rules related to the user's intention after the AI large model platform receives the query questions of the operation and maintenance personnel (i.e., the first text information mentioned above) and identifies the user's intention. The operation and maintenance rules are a series of guidelines pre-entered by the operation and maintenance experts or system administrators based on the characteristics and operation requirements of the business system, including but not limited to: the normal range of performance indicators, troubleshooting procedures, abnormal response strategies, etc. When the user's intention is directed to a specific operation and maintenance query or question, it will filter out the operation and maintenance rules that are most relevant to the user's intention from these preset operation and maintenance rules to facilitate subsequent data processing and operation and maintenance decisions.

[0063] Then, based on the filtered operation and maintenance rules, data filtering operations are performed on the target time series data. For example, the target time series data can be filtered for indicator data related to the operation and maintenance rules. The filtered target time series data is converted into text data to generate reference information tailored to the user's intent. For example, based on the characteristics of the business system and the operation and maintenance rules, an AI big model (or other model) converts the numerical and time information in the target time series data into descriptive text, such as "CPU usage abnormally increased to 85% between 2:00 PM and 2:10 PM" or "Memory usage suddenly increased beyond the normal range in the last hour." This text information not only contains specific data details but also incorporates the analysis results of the operation and maintenance rules, providing users with intuitive and easy-to-understand feedback. This allows even non-technical personnel to quickly understand the system's operation and maintenance status and take appropriate measures in a timely manner. This series of operations, from identifying user intent to filtering data based on operation and maintenance rules, and then converting the data into text information, achieves the effect of improving operation and maintenance efficiency and user experience.

[0064] By applying operation and maintenance rules, we can further narrow the scope of attention, focusing on analyzing the time series data that operators are actually concerned about, and eliminating anomalies that may be caused by false alarms or irrelevant factors. This process helps to improve the accuracy and pertinence of inspection results, ensuring that the information subsequently fed back to users has been strictly screened and verified, further achieving the technical effect of improving the efficiency and accuracy of operations and maintenance personnel.

[0065] Optionally, in the operation and maintenance method of the business system provided in Example 1 of the present application, data noise reduction processing is performed on the original time series data to obtain the noise-reduced time series data, including: calculating the average value of multiple time series data in the original time series data, wherein the multiple time series data refers to the time series data in the original time series data whose collection time is earlier than the preset time; using the time series data analysis algorithm to calculate the average value and the indicator data at each moment in the original time series data to obtain the first time series data; calculating the difference between the indicator data at each moment in the first time series data and the indicator data at each moment in the original time series data one by one to obtain the difference at each moment; deleting the noise data in the original time series data to obtain the noise-reduced time series data, wherein the noise data is the indicator data corresponding to the target moment whose difference is greater than the second preset threshold.

[0066] In the first embodiment, in order to improve data processing efficiency and enhance real-time operation and maintenance, it is necessary to identify and remove noise data from the original time series data to achieve the purpose of data noise reduction.

[0067] First, calculate the initial value for the noise reduction process. This is the average of the first few observations in the original time series data, i.e., the time series data collected before the preset time. The purpose of calculating the average value is to provide a reference baseline, ensuring strong resistance to noise while maintaining maximum sensitivity to new data.

[0068] Then, using a time series data analysis algorithm, the calculated average value is compared with the indicator data at each moment in the original time series data to generate the first time series data. This algorithm may use an exponentially weighted moving average (EWMA) or other statistical methods suitable for time series data anomaly detection. Its purpose is to measure the degree of deviation of the data at each collection moment from the historical average state. The calculation formula of the time series data analysis algorithm can be shown as Formula 1.

[0069] EMA(t)=α*X(t)+(1-α)*EMA(t-1) (1)

[0070] Among them, X(t) represents the indicator data at time t, X(0) represents the above average value, and α represents the balance factor, which can be set to 0.4. The balance factor can be flexibly adjusted according to time, indicator type or other factors in order to improve the accuracy of time series data.

[0071] Next, the difference between the first time series data and the indicator data at each moment in the original time series data is calculated one by one. The difference is used to indicate the degree of deviation between each data point and the average state at that moment, which helps improve the accuracy of subsequent noise identification.

[0072] Finally, the difference is compared with a second preset threshold, and noise data with a difference greater than the second preset threshold is identified and deleted, thereby obtaining the denoised time series data. The second preset threshold is used to determine whether the difference exceeds the normal range of variation. When the difference at a certain moment is greater than the second preset threshold, the indicator data at that moment is considered noise data. Noise data may be caused by factors such as system failures, external interference, or data collection errors, and is inconsistent with the normal operating status of the business system. By deleting this noise data, random fluctuations and anomalies in the data can be effectively reduced, the accuracy and reliability of the time series data can be improved, and the overall efficiency and quality of operation and maintenance inspections can be improved.

[0073] Through the construction and processing of the above-mentioned time series data, it is possible to ensure that the time series data used for operation and maintenance inspection analysis is more accurate and reliable, which helps to make accurate operation and maintenance decisions in the future, thereby achieving the effect of improving operation and maintenance accuracy.

[0074] Optionally, in the operation and maintenance method of the business system provided in Example 1 of the present application, before merging the target time series data and the user intent of the target user to obtain reference information of the user intent, the above method also includes: deploying a target program in the target device, controlling the microphone in the target device through the target program to receive the first voice information output by the target user; transmitting the first voice information received by the microphone to a voice conversion tool through the target program based on a wireless communication network; and controlling the voice conversion tool through the target program to convert the first voice information into text information to obtain first text information.

[0075] In this first embodiment, in order to quickly respond to queries from maintenance personnel, a specific program or software, namely the target program described above, needs to be installed and run on the target device. This program is responsible for voice input and output control and wireless communication capabilities of the target device. The target program activates the microphone in the target device to receive the first voice message from the maintenance personnel, namely their inquiries or instructions regarding system status. The target device can be a handheld device or a device deployed nearby the maintenance personnel.

[0076] Then, based on the wireless communication network, the target program encodes the first voice information captured by the microphone and transmits it to the speech conversion tool via a wireless network (such as Wi-Fi or Bluetooth). This transmission process ensures the real-time and integrity of the voice information, laying the foundation for network communication for subsequent speech recognition and text conversion.

[0077] Through the target program, the speech conversion tool processes the received first speech information and converts it into text information, namely, the first text information, using automatic speech recognition technology (ASR technology). ASR technology can recognize and analyze the speech content in the speech signal and convert it into a readable text format to facilitate further analysis and understanding by the AI large model platform.

[0078] In an optional embodiment, the target device may be a voice box for voice interaction with the user. The hardware portion of the voice box includes main components including a microprocessor, memory, audio codec, microphone, speaker, wireless communication module, input / output interface, power management unit, housing and structural design, as well as a control panel and indicator lights. The core part of the voice box is an IoT program developed based on Python, which mainly implements two functions: completing the drive and control of the microphone, speaker, and wireless communication module; and integrating with the TTS module and ASR module of the voice interaction platform to jointly complete the conversion between voice and text of questions and results. Among them, the TTS module is responsible for converting text data into spoken speech through text-to-speech technology. The ASR module is responsible for converting spoken speech into text through automatic speech recognition technology.

[0079] In an optional embodiment, the voice conversion tool can be an API service developed based on a python code framework, which can also be called a voice interaction platform. The voice interaction platform acts as a bridge between voice devices (such as the voice box mentioned above) and AI large models. The voice interaction platform mainly provides two core functions: providing ASR capabilities, that is, converting the received voice information into text for processing by the AI large model; providing TTS capabilities, that is, converting the text results processed by the AI large model into voice feedback to the operation and maintenance personnel.

[0080] Through the above steps, a seamless connection between the voice box device and the voice interaction platform is achieved, ensuring the accurate capture and transmission of user voice commands, as well as the effective conversion of voice information into text information, thereby providing operation and maintenance personnel with a convenient way to conduct operation and maintenance inspections of business systems through natural language, greatly improving operation and maintenance efficiency and user experience.

[0081] Optionally, in the operation and maintenance method of the business system provided in Example 1 of the present application, reference information of the user's intention is input into the large language model to generate the operation and maintenance results of the business system, including: generating a target template based on the user's intention and preset prompt words, wherein the target template is used to instruct the large language model to output operation and maintenance operation suggestions based on the current operation status of the business system; using the reference information of the user's intention to fill the target template and input it into the large language model, and generating the operation and maintenance results of the business system through the large language model.

[0082] In this first embodiment, to guide the AI big model in generating corresponding O&M action suggestions based on the user intent of the O&M personnel, a target template can be generated based on user intent and preset prompt words. Preset prompt words are guiding texts designed in advance based on common O&M problems and scenarios. They help the big model understand the problem context and ensure that the generated suggestions are highly relevant to actual O&M needs. The target template combines user intent and prompt words to form a structured input format, guiding the big model to generate O&M action suggestions that are consistent with the current operating conditions of the business system.

[0083] Then, the reference information of the user's intent is filled into the corresponding position of the target template. This filling process ensures that the large model can receive specific problem descriptions and relevant system operation data when generating operation and maintenance results, thereby providing more accurate and targeted operation and maintenance suggestions or diagnostic results. For example, if the user's intention is to inquire about the reason for the high server CPU utilization, the target template may include "The server CPU utilization rate has risen to XX% in the past hour. Please analyze the possible causes of the failure and provide solutions." This form of template filling provides clear guidance for the AI large model, enabling it to accurately analyze problems and provide suggestions based on actual data and operation and maintenance rules.

[0084] Finally, the populated target template is fed into the AI master model. Based on the populated template, the master model analyzes reference information about the user's intent, including the specific circumstances of the abnormal data and relevant operation and maintenance rules, and generates operation and maintenance recommendations tailored to the current operational situation. These recommendations may include explanations of abnormal indicators, possible causes, emergency response measures, and long-term optimization strategies, helping operations personnel quickly understand the problem, locate the fault, and take effective action to improve the stability and efficiency of the business system.

[0085] Through the above steps, the intelligent analysis and text generation capabilities of the AI big model can be fully utilized, combined with the user's specific operation and maintenance needs and the actual operation data of the system, to generate highly targeted and practical operation and maintenance results, providing operation and maintenance personnel with an efficient and intelligent operation and maintenance auxiliary means, thereby improving the operation and maintenance effects and accuracy of operation and maintenance personnel.

[0086] Optionally, in this embodiment 1, Figure 2It is a structural diagram of an intelligent inspection tool based on voice interaction technology. The intelligent inspection tool includes a voice box, a voice interaction platform, an AI big model, an intelligent analysis system and a monitoring system. The intelligent inspection tool uses intent recognition technology to convert the analysis results into conclusions generated by the big model, and then quickly conveys them to the user through the voice broadcast device. Among them, the voice box and the voice interaction platform are as described above. The AI big model is integrated with the intelligent analysis system through the software development kit (SDK) to process the retrieved data to identify the user's intention. After the user's voice is converted into text by ASR (automatic speech recognition), the AI big model can understand and respond to these questions and feedback the user's current operation and maintenance inspection status. The executor of the operation and maintenance method of a business system provided in Example 1 of the present application can be the intelligent analysis system in the intelligent inspection tool, as described above.

[0087] Optionally, in this embodiment 1, Figure 3 This is a schematic diagram of the intelligent analysis system's processing flow for time series data. The data denoising stage primarily involves periodically extracting and filtering time series data to eliminate data points or spikes that significantly deviate from normal conditions. This prevents these extreme data points from being assigned higher weights in subsequent calculations, which could affect the accuracy of the final results. Data segmentation involves processing the denoised data based on specific calculation formulas and generating corresponding structured data according to specific rules. Vectorization involves vectorizing the segmented data, using the healthiness of the structured data as vector elements of numerical features for subsequent calculations. Data reorganization involves performing mathematical operations on vector data to measure the similarity between different vectors. Based on these similarity scores, time series data indicating abnormal conditions in the business system can be retrieved. Optimized merging involves processing the recalled data into independent text files and merging them according to a specific logic to form the target time series data described above.

[0088] Optionally, in this embodiment 1, Figure 4This is a flow chart illustrating the output of voice-based O&M inspection results through intelligent inspection tools. Users use the voice box's microphone to verbally query information regarding the operational status of business systems. The IoT program within the voice box quickly responds, receiving and encoding the voice information. Its wireless communication module then converts the voice signal into a data packet and sends it to the ASR (Automatic Speech Recognition) module of the voice interaction platform. After receiving the data packet, the ASR module parses the voice signal and converts it into text format. This text information is then delivered to the AI ​​model on the backend. After receiving the text information, the AI ​​model analyzes the user's query intent to accurately grasp O&M requirements. Based on the analyzed user intent, the intelligent analysis system calls on O&M rules and historical time series data related to the system indicators of interest to O&M personnel. Simultaneously, the monitoring system continuously collects real-time time series data from the business system and transmits it to the intelligent analysis system to ensure the real-time and accuracy of the AI ​​model's analysis results.

[0089] The intelligent analysis system extracts key information from massive amounts of time-series data through data noise reduction, segmentation, reorganization, and vectorization, constructing vector index values for subsequent rapid retrieval and analysis. These processed data are then optimized and merged with the user intent analyzed by the AI big model platform to generate a comprehensive result that integrates user needs and system data, namely the reference information of user intent mentioned above. The reference information of user intent is then input into the AI big model to form a structured, easy-to-understand operation and maintenance inspection report. The TTS (text-to-speech) module receives this report and immediately converts it into a voice signal. Through the TTS module of the voice interaction platform, the voice information is again transmitted to the voice box via the wireless network. Finally, the speaker of the voice box broadcasts the inspection results to the user, completing a complete closed loop from voice query to voice feedback, ensuring that operation and maintenance personnel can obtain system status information instantly and accurately and respond quickly.

[0090] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0091] Example 2

[0092] The second embodiment of the present application further provides an operation and maintenance device for a business system. It should be noted that the operation and maintenance device for the business system in the second embodiment of the present application can be used to execute the operation and maintenance method for the business system provided in the first embodiment of the present application. The operation and maintenance device for the business system provided in the second embodiment of the present application is introduced below.

[0093] Figure 5Schematic diagram of the operation and maintenance device of the business system according to the second embodiment of the present application. Figure 5 As shown, the device includes: a collection unit 501, a reorganization unit 502, a merging unit 503 and a playing unit 504.

[0094] Specifically, the collection unit 501 is used to collect original time series data of the business system, and perform data noise reduction processing on the original time series data to obtain noise-reduced time series data.

[0095] The reorganization unit 502 is used to perform data preprocessing and data reorganization on the denoised time series data to obtain target time series data.

[0096] The merging unit 503 is used to merge the target time series data and the user intention of the target user to obtain reference information of the user intention, wherein the user intention is obtained by performing intent recognition on the first text information through a large language model, and the first text information is obtained after processing the output speech of the target user.

[0097] The playback unit 504 is used to input the reference information of the user's intention into the large language model, generate the operation and maintenance results of the business system, and play the operation and maintenance results to the target user so that the target user can perform operation and maintenance operations on the business system based on the operation and maintenance results.

[0098] The operation and maintenance device of the business system provided in the second embodiment of the present application collects the original time series data of the business system through the collection unit 501, and performs data noise reduction processing on the original time series data to obtain the noise-reduced time series data; the reorganization unit 502 performs data preprocessing operations and data reorganization operations on the noise-reduced time series data to obtain target time series data; the merging unit 503 merges the target time series data and the user intention of the target user to obtain reference information of the user intention, wherein the user intention is obtained by performing intent recognition on the first text information through a large language model, and the first text information is obtained after processing the output voice of the target user; the playback unit 504 inputs the reference information of the user intention into the large language model, generates the operation and maintenance results of the business system, and plays the operation and maintenance results to the target user, so that the target user can perform operation and maintenance operations on the business system based on the operation and maintenance results, thereby solving the problem in the related technology that the business system operation and maintenance personnel need to perform operation and maintenance operations on the business system based on the operation and maintenance reports after complex analysis, resulting in poor real-time performance of the operation and maintenance operations on the business system.

[0099] By collecting the original time series data of the business system and performing data noise reduction, data preprocessing, and data reorganization operations on it, the target time series data can be extracted to obtain more concise and noise-reduced time series data that accurately reflects the system's operating status. This provides an accurate data foundation for subsequent AI big model analysis, thereby significantly improving the accuracy of subsequent analysis. At the same time, by using the AI big model to recognize the intent of the text converted from the operation and maintenance personnel's voice query, reference information of the user's intent is generated, the target time series data is integrated with the user's intent, and this integrated reference information is input into the AI big model to generate business system operation and maintenance results that meet user needs. The results are then immediately conveyed to the operation and maintenance personnel through voice broadcast. This not only ensures the pertinence and efficiency of the operation and maintenance results generated by the AI big model, but also makes the operation and maintenance operations more intuitive and rapid, greatly reducing the burden on operation and maintenance personnel to analyze complex data, realizing real-time operation and maintenance operations based on the operating status of the business system, and providing real-time operation and maintenance decision support, achieving the technical effect of significantly improving operation and maintenance efficiency and accuracy, further optimizing the work experience of operation and maintenance personnel and ensuring the stable and efficient operation of the business system.

[0100] Optionally, in the operation and maintenance device of the business system provided in Example 2 of the present application, the above-mentioned reorganization unit 502 includes: an operation sub-unit, which is used to perform feature extraction operations on multiple time series data of any indicator in the denoised time series data to obtain a first feature vector set; a first calculation sub-unit, which is used to calculate the similarity between each first feature vector in the first feature vector set and the vector corresponding to the preset indicator threshold based on the cosine similarity algorithm to obtain a similarity set; a sorting sub-unit, which is used to sort the similarities in the similarity set in descending order, and determine a target similarity whose sorting order is lower than the first preset threshold; a determination sub-unit, which is used to determine a target feature vector corresponding to the target similarity, and determine the target time series data based on the time series data corresponding to the target feature vector.

[0101] Optionally, in the operation and maintenance device of the business system provided in Example 2 of the present application, the above-mentioned operation sub-unit includes: a sampling module, which is used to sample multiple time series data based on a preset time window to obtain multiple time series data after sampling; an extraction module, which is used to extract statistical features from the multiple time series data after sampling based on a preset statistical algorithm, and use the statistical features as vector index values to obtain a first feature vector set.

[0102] Optionally, in the operation and maintenance device of the business system provided in Example 2 of the present application, the above-mentioned merging unit 503 includes: a determination subunit, used to determine the operation and maintenance rules related to the user intention, wherein the operation and maintenance rules are pre-input rules for operating the business system; a screening subunit, used to screen data in the target time series data according to the operation and maintenance rules to obtain the screened target time series data; a conversion subunit, used to convert the screened target time series data into text data to obtain reference information of the user intention.

[0103] Optionally, in the operation and maintenance device of the business system provided in Example 2 of the present application, the above-mentioned collection unit 501 includes: a second calculation subunit, used to calculate the average value of multiple time series data in the original time series data, wherein the multiple time series data refers to the time series data in the original time series data whose collection time is earlier than the preset time; a third calculation subunit, used to use the time series data analysis algorithm to calculate the average value and the indicator data at each moment in the original time series data to obtain the first time series data; a fourth calculation subunit, used to calculate the difference between the indicator data at each moment in the first time series data and the indicator data at each moment in the original time series data one by one to obtain the difference at each moment; a deletion subunit, used to delete noise data in the original time series data to obtain the denoised time series data, wherein the noise data is the indicator data corresponding to the target moment whose difference is greater than the second preset threshold.

[0104] Optionally, in the operation and maintenance device of the business system provided in Example 2 of the present application, the above-mentioned device also includes: a receiving unit, which is used to deploy a target program in the target device before merging the target time series data and the user intention of the target user to obtain reference information of the user intention, and control the microphone in the target device through the target program to receive the first voice information output by the target user; a transmission unit, which is used to transmit the first voice information received by the microphone to the voice conversion tool through the target program based on the wireless communication network; and a conversion unit, which is used to control the voice conversion tool through the target program to convert the first voice information into text information to obtain first text information.

[0105] Optionally, in the operation and maintenance device of the business system provided in Example 2 of the present application, the above-mentioned playback unit 504 includes: a first generation sub-unit, used to generate a target template based on user intention and preset prompt words, wherein the target template is used to instruct the large language model to output operation and maintenance operation suggestions based on the current operation status of the business system; a second generation sub-unit, used to fill the target template with reference information of the user intention, and input it into the large language model, and generate the operation and maintenance results of the business system through the large language model.

[0106] The operation and maintenance device of the business system includes a processor and a memory. The above-mentioned collection unit 501, reorganization unit 502, merging unit 503 and playback unit 504 are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.

[0107] The processor contains a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be configured, and kernel parameters can be adjusted to improve the real-time operation and maintenance of the business system.

[0108] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0109] A third embodiment of the present invention provides a computer-readable storage medium having a program stored thereon, which implements an operation and maintenance method of a business system when executed by a processor.

[0110] A fourth embodiment of the present invention provides a processor, which is used to run a program, wherein the program executes an operation and maintenance method of a business system when the program is running.

[0111] like Figure 6 As shown, embodiment five of the present invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and runnable on the processor. When the processor executes the program, the following steps are implemented: collecting original time series data of the business system, and performing data noise reduction processing on the original time series data to obtain noise-reduced time series data; performing data preprocessing operations and data reorganization operations on the noise-reduced time series data to obtain target time series data; merging the target time series data and the user intention of the target user to obtain reference information of the user intention, wherein the user intention is obtained by performing intent recognition on the first text information through a large language model, and the first text information is obtained after processing the output speech of the target user; the reference information of the user intention is input into the large language model to generate an operation and maintenance result of the business system, and the operation and maintenance result is played to the target user, so that the target user can perform operation and maintenance operations on the business system based on the operation and maintenance result.

[0112] When the processor executes the program, the following steps are also implemented: performing data preprocessing operations and data reorganization operations on the denoised time series data to obtain target time series data, including: performing feature extraction operations on multiple time series data of any indicator in the denoised time series data to obtain a first feature vector set; calculating the similarity between each first feature vector in the first feature vector set and the vector corresponding to the preset indicator threshold based on the cosine similarity algorithm to obtain a similarity set; sorting the similarities in the similarity set in descending order, and determining the target similarity whose sorting order is lower than the first preset threshold; determining the target feature vector corresponding to the target similarity, and determining the target time series data based on the time series data corresponding to the target feature vector.

[0113] When the processor executes the program, it also implements the following steps: performing feature extraction operations on multiple time series data to obtain a first feature vector set, including: performing data sampling on multiple time series data based on a preset time window to obtain multiple time series data after sampling; extracting statistical features from the multiple time series data after sampling based on a preset statistical algorithm, and using the statistical features as vector index values to obtain a first feature vector set.

[0114] When the processor executes the program, it also implements the following steps: merging the target time series data and the user intent of the target user to obtain reference information of the user intent, including: determining operation and maintenance rules related to the user intent, wherein the operation and maintenance rules are pre-entered rules for running the business system; filtering data in the target time series data according to the operation and maintenance rules to obtain filtered target time series data; converting the filtered target time series data into text data to obtain reference information of the user intent.

[0115] When the processor executes the program, the following steps are also implemented: performing data noise reduction processing on the original time series data to obtain the noise-reduced time series data, including: calculating the average value of multiple time series data in the original time series data, wherein the multiple time series data refers to the time series data in the original time series data whose collection time is earlier than the preset time; using the time series data analysis algorithm to calculate the average value and the indicator data at each moment in the original time series data to obtain the first time series data; calculating the difference between the indicator data at each moment in the first time series data and the indicator data at each moment in the original time series data one by one to obtain the difference at each moment; deleting the noise data in the original time series data to obtain the noise-reduced time series data, wherein the noise data is the indicator data corresponding to the target moment whose difference is greater than the second preset threshold.

[0116] When the processor executes the program, the following steps are also implemented: before merging the target time series data and the user intention of the target user to obtain reference information of the user intention, the above method also includes: deploying a target program in the target device, and controlling the microphone in the target device through the target program to receive the first voice information output by the target user; transmitting the first voice information received by the microphone to the voice conversion tool through the target program based on the wireless communication network; and controlling the voice conversion tool through the target program to convert the first voice information into text information to obtain the first text information.

[0117] When the processor executes the program, it also implements the following steps: inputting reference information of the user's intention into the large language model to generate the operation and maintenance results of the business system, including: generating a target template based on the user's intention and preset prompt words, wherein the target template is used to instruct the large language model to output operation and maintenance operation suggestions based on the current operation status of the business system; using the reference information of the user's intention to fill the target template and input it into the large language model, and generating the operation and maintenance results of the business system through the large language model.

[0118] The devices in this article can be servers, PCs, PADs, mobile phones, etc.

[0119] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program initialized with the following method steps: collecting original time series data of a business system, and performing data noise reduction processing on the original time series data to obtain noise-reduced time series data; performing data preprocessing operations and data reorganization operations on the noise-reduced time series data to obtain target time series data; merging the target time series data and the user intention of the target user to obtain reference information of the user intention, wherein the user intention is obtained by performing intent recognition on first text information through a large language model, and the first text information is obtained after processing the output speech of the target user; inputting the reference information of the user intention into the large language model to generate operation and maintenance results of the business system, and playing the operation and maintenance results to the target user, so that the target user can perform operation and maintenance operations on the business system based on the operation and maintenance results.

[0120] When executed on a data processing device, it is also suitable for executing a program initialized with the following method steps: performing data preprocessing operations and data reorganization operations on the denoised time series data to obtain target time series data, including: performing feature extraction operations on multiple time series data of any indicator in the denoised time series data to obtain a first feature vector set; calculating the similarity between each first feature vector in the first feature vector set and the vector corresponding to the preset indicator threshold based on the cosine similarity algorithm to obtain a similarity set; sorting the similarities in the similarity set in descending order, and determining the target similarity whose sorting order is lower than the first preset threshold; determining the target feature vector corresponding to the target similarity, and determining the target time series data based on the time series data corresponding to the target feature vector.

[0121] When executed on a data processing device, it is also suitable for executing a program initialized with the following method steps: performing feature extraction operations on multiple time series data to obtain a first feature vector set, including: performing data sampling on multiple time series data based on a preset time window to obtain multiple time series data after sampling; extracting statistical features from the multiple time series data after sampling based on a preset statistical algorithm, and using the statistical features as vector index values to obtain a first feature vector set.

[0122] When executed on a data processing device, it is also suitable for executing a program initialized with the following method steps: merging the target time series data and the user intention of the target user to obtain reference information of the user intention, including: determining operation and maintenance rules related to the user intention, wherein the operation and maintenance rules are pre-input rules for running the business system; filtering data in the target time series data according to the operation and maintenance rules to obtain filtered target time series data; converting the filtered target time series data into text data to obtain reference information of the user intention.

[0123] When executed on a data processing device, it is also suitable for executing a program initialized with the following method steps: performing data noise reduction processing on the original time series data to obtain the noise-reduced time series data, including: calculating the average value of multiple time series data in the original time series data, wherein the multiple time series data refers to the time series data in the original time series data whose collection time is earlier than the preset time; using the time series data analysis algorithm to calculate the average value and the indicator data at each moment in the original time series data to obtain the first time series data; calculating the difference between the indicator data at each moment in the first time series data and the indicator data at each moment in the original time series data one by one to obtain the difference at each moment; deleting the noise data in the original time series data to obtain the noise-reduced time series data, wherein the noise data is the indicator data corresponding to the target moment whose difference is greater than the second preset threshold.

[0124] When executed on a data processing device, it is also suitable for executing an initialized program having the following method steps: before merging the target time series data and the user intention of the target user to obtain reference information of the user intention, the above method also includes: deploying a target program in the target device, controlling the microphone in the target device through the target program to receive the first voice information output by the target user; transmitting the first voice information received by the microphone to a voice conversion tool through the target program based on a wireless communication network; and controlling the voice conversion tool through the target program to convert the first voice information into text information to obtain first text information.

[0125] When executed on a data processing device, it is also suitable for executing an initialization program having the following method steps: inputting reference information of user intention into a large language model to generate operation and maintenance results of the business system, including: generating a target template based on user intention and preset prompt words, wherein the target template is used to instruct the large language model to output operation and maintenance operation suggestions based on the current operation status of the business system; using reference information of user intention to fill the target template and input it into the large language model, and generating the operation and maintenance results of the business system through the large language model.

[0126] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0127] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0128] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0129] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0130] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0131] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0132] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0133] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0134] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0135] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method for operating and maintaining a business system, characterized in that: include: Collecting original time series data from the business system and performing data noise reduction processing on the original time series data to obtain noise-reduced time series data; Performing data preprocessing and data reorganization on the noise-reduced time series data to obtain target time series data; Merging the target time series data and the user intent of the target user to obtain reference information of the user intent, wherein the user intent is obtained by performing intent recognition on first text information using a large language model, and the first text information is obtained by processing the output speech of the target user; The reference information of the user intention is input into the large language model to generate the operation and maintenance results of the business system, and the operation and maintenance results are played to the target user so that the target user can perform operation and maintenance operations on the business system based on the operation and maintenance results.

2. The method according to claim 1, characterized in that Performing data preprocessing and data reorganization on the noise-reduced time series data to obtain target time series data includes: For multiple time series data of any indicator in the noise-reduced time series data, perform a feature extraction operation on the multiple time series data to obtain a first feature vector set; Calculating the similarity between each first feature vector in the first feature vector set and a vector corresponding to a preset index threshold based on a cosine similarity algorithm to obtain a similarity set; Sorting the similarities in the similarity set in descending order, and determining a target similarity whose sorting order is lower than a first preset threshold; A target feature vector corresponding to the target similarity is determined, and the target time series data is determined based on the time series data corresponding to the target feature vector.

3. The method according to claim 2, characterized in that Performing a feature extraction operation on the plurality of time series data to obtain a first feature vector set includes: Sampling the plurality of time series data based on a preset time window to obtain a plurality of sampled time series data; Statistical features are extracted from the sampled multiple time series data based on a preset statistical algorithm, and the statistical features are used as vector index values to obtain the first feature vector set.

4. The method according to claim 1, wherein Merging the target time series data and the user intent of the target user to obtain reference information of the user intent includes: Determining operation and maintenance rules related to the user intention, wherein the operation and maintenance rules are pre-input rules for operating the business system; Filtering the target time series data according to the operation and maintenance rules to obtain filtered target time series data; The filtered target time series data is converted into text data to obtain reference information of the user's intention.

5. The method according to claim 1, wherein Performing data noise reduction processing on the original time series data to obtain noise-reduced time series data includes: Calculating an average value of multiple time series data in the original time series data, wherein the multiple time series data refer to time series data in the original time series data whose collection time is earlier than a preset time; Using a time series data analysis algorithm to calculate the average value and the indicator data at each moment in the original time series data to obtain first time series data; Calculating the difference between the indicator data at each moment in the first time series data and the indicator data at each moment in the original time series data one by one to obtain the difference at each moment; Noise data is deleted from the original time series data to obtain the noise-reduced time series data, wherein the noise data is indicator data corresponding to the target moment when the difference is greater than a second preset threshold.

6. The method according to claim 1, characterized in that Before merging the target time series data and the user intent of the target user to obtain reference information of the user intent, the method further includes: Deploying a target program in a target device, and controlling a microphone in the target device through the target program to receive first voice information output by the target user; transmitting the first voice information received by the microphone to a voice conversion tool via the target program based on a wireless communication network; The target program controls the voice conversion tool to convert the first voice information into text information to obtain the first text information.

7. The method according to claim 1, characterized in that Inputting the reference information of the user intention into the large language model to generate the operation and maintenance results of the business system includes: generating a target template based on the user intention and a preset prompt word, wherein the target template is used to instruct the large language model to output operation and maintenance suggestions based on the current operation status of the business system; The target template is filled with reference information of the user's intention and input into the large language model, and the operation and maintenance results of the business system are generated through the large language model.

8. An operation and maintenance device for a business system, characterized in that: include: The acquisition unit is used to collect the original time series data of the business system and perform data noise reduction processing on the original time series data to obtain the noise-reduced time series data; a reorganization unit, configured to perform data preprocessing and data reorganization on the noise-reduced time series data to obtain target time series data; a merging unit, configured to merge the target time series data and the user intent of the target user to obtain reference information of the user intent, wherein the user intent is obtained by performing intent recognition on first text information using a large language model, and the first text information is obtained by processing the output speech of the target user; A playback unit is used to input the reference information of the user's intention into the large language model, generate the operation and maintenance results of the business system, and play the operation and maintenance results to the target user so that the target user can perform operation and maintenance operations on the business system based on the operation and maintenance results.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium includes stored computer instructions, wherein when the computer instructions are executed by a processor, the operation and maintenance method of the business system according to any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: It includes one or more processors and a memory, the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the operation and maintenance method of the business system described in any one of claims 1 to 7.

Images