684 results about "Baseline model" patented technology

The invention provides an HBase client main-standby switching method and system based on fault perception, and relates to the technical field of database management, and the method comprises the steps: collecting system operation data through a distributed fault monitoring module, building an anomaly detection baseline model, carrying out fault diagnosis based on a system health assessment distribution matrix, and generating a fault characteristic spectrum; and switching is triggered according to a risk level evaluation result, and the optimal standby node is selected for sequential switching by using a deep strategy learning algorithm, so that rapid and accurate fault detection, diagnosis and intelligent switching can be realized, and the reliability and performance of the system are improved.

The invention relates to the technical field of intelligent building energy-saving control, and discloses a green building equipment intelligent energy-saving control method and system based on artificial intelligence, and the method comprises the steps: collecting building environment data, and carrying out the preprocessing; and simulating the energy consumption of the building equipment by using the digital twin model, and training the baseline neural network model to obtain an initial control strategy. And adopting a genetic algorithm to optimize hyper-parameters of the baseline model, and generating a plurality of candidate energy-saving control strategies. And inputting the optimized strategy into the deep reinforcement learning model, carrying out strategy iteration, calculating the energy-saving efficiency and comfort score after each iteration, screening the optimal strategy through a reward mechanism, and improving the accuracy of energy consumption prediction of the building equipment. The hyper-parameter of the neural network is optimized by using the genetic algorithm, and the intelligent degree of the energy-saving control strategy is improved. On the basis of deep reinforcement learning, the intelligent agent can adaptively adjust an energy-saving strategy under different environmental conditions, and the user comfort is guaranteed while the optimal building energy consumption is ensured.

The invention provides an enterprise employee behavior analysis and safety risk early warning monitoring method and system, and relates to the technical field of enterprise risk management, and the method comprises the steps: collecting employee terminal operation behavior data, building an activity thermal distribution diagram based on office area grid behavior association intensity, and training a behavior evaluation model. And inputting the behavior scoring matrix into a deep neural network to extract target behavior characteristics, calculating an abnormal behavior risk weight coefficient in combination with a department security policy, performing classification and analyzing a risk diffusion probability, and generating a risk situation index to determine an early warning level. Starting a response strategy according to the early warning level, blocking high-risk early warning in real time, tracking associated accounts, establishing a risk association map, identifying potential risk propagation nodes and performing active protection, finally generating an early warning report and feeding back effective protection rules to a behavior baseline model, and realizing continuous optimization of a risk early warning mechanism. And the accuracy and effectiveness of safety risk early warning in the enterprise are improved.

The invention belongs to the technical field of applets, and discloses an applet abnormal behavior management system fused with artificial intelligence. Comprising the steps of collecting a multi-dimensional data stream in real time, performing space-time slicing on the multi-dimensional data stream, extracting a behavior feature vector in each time window, and constructing a dynamic behavior baseline model bound with a user identity; calculating a deviation degree between the behavior feature vector and a dynamic behavior baseline model, and generating a real-time anomaly confidence coefficient in combination with a similarity matching degree with a preset historical anomaly event library; the hidden relevance of a cross-interface calling link is recognized, then an abnormal propagation probability graph is constructed, the weights of the nodes in the abnormal propagation probability graph are dynamically corrected through a real-time abnormal confidence coefficient, and the corrected weights of the nodes in the abnormal propagation probability graph are obtained; and calculating a comprehensive risk entropy value, and when the comprehensive risk entropy value exceeds a risk threshold value, dynamically triggering a multi-level management and control strategy, so that complex applet abnormal behaviors can be comprehensively detected and managed.

The invention discloses a model training method and device, electronic equipment and a computer storage medium. The model training method comprises the following steps: performing supervision fine tuning on an initial baseline model based on a first training sample set in a preset training sample set to obtain a supervision fine tuning model and a preference data sample; performing optimization training on the initial reward model based on the preference data sample to obtain a target reward model, and performing optimization training on the supervision fine tuning model based on the preference data sample to obtain a preference optimization model; generating a first reward signal based on the preference optimization model and the first training sample set, and generating a second reward signal based on the preference optimization model, the target reward model and a second training sample set in the preset training sample set; and performing optimization training on the preference optimization model based on the first reward signal, the second reward signal and the second training sample set to obtain a target object generation type pre-training model. By adopting the method, the generalization ability of the model can be improved.

The invention relates to the technical field of audio and video analysis and data processing, in particular to an AI-based multimedia equipment operation and maintenance management system, which comprises a data acquisition module, an analysis module, an equipment state modeling module, a fault prediction module, a maintenance strategy optimization module, a resource allocation module and a visual platform. The data acquisition module acquires temperature, vibration, audio and video feature data of the equipment through the sensor and the video equipment, and the data are preprocessed and then transmitted to the cloud; the equipment state modeling module constructs a dynamic baseline model based on historical data, abnormal signals are generated through real-time comparison, and the visual platform realizes multi-dimensional state monitoring through 3D topology rendering, AR auxiliary diagnosis and early warning linkage. The system solves the problems of response lag and low resource utilization rate caused by dependence on manual operation and maintenance in the prior art through a closed-loop management mechanism driven by a data stream, and improves the fault diagnosis precision and the operation and maintenance efficiency.

The embodiment of the invention provides a multi-factor security authentication method and device based on AI adaptive recognition and a block chain, and the method comprises the steps: collecting user biological features, behavior data and equipment information, and extracting static and dynamic features through employing a convolutional neural network and a long-short term memory network; a feature fusion network is innovatively constructed, an attention mechanism is introduced to carry out dynamic weight distribution and fusion on different features, and feature correction is carried out based on a user behavior baseline model. And meanwhile, a block chain network and an intelligent contract are deployed, so that credible storage and risk level evaluation of authentication data are realized. According to the method, the limitation of traditional single authentication is broken through, the effective balance between safety and user experience is realized, and an advanced technical scheme is provided for the field of identity authentication.

The invention relates to the technical field of line arc abnormity monitoring, in particular to a data center operation and maintenance service environment monitoring system which comprises a multi-mode sensing unit, a dynamic baseline modeling unit, a transition state abnormity extraction unit and a grading early warning unit. The dynamic baseline modeling unit builds a three-layer safety baseline model, the base layer builds harmonic references in a segmented mode according to the load rate, the environment layer generates voiceprint feature template libraries of different temperature and humidity intervals, the time layer fits a 24-hour change trend envelope line, and the transition state anomaly extraction unit separates and decouples an arc voiceprint feature frequency band through a blind source and calculates the energy ratio. The harmonic abrupt change phase deviation amplitude is analyzed in combination with window sliding correlation, the abnormal evolution index is generated through fusion, the graded early warning unit triggers response according to the index, transition state abnormity is accurately recognized, the perspectiveness and reliability of operation and maintenance early warning are improved, and the method is suitable for safe operation and maintenance of data center equipment.

The invention relates to the technical field of network security and data protection, and discloses a data leakage prevention method and system based on user behavior perception, and the method comprises the steps: obtaining historical operation data, and constructing a personalized behavior reference library; monitoring a current access behavior in real time by sliding a time window, calculating a deviation degree and triggering anomaly detection; performing multi-level feature analysis on the abnormal behavior and calculating a comprehensive abnormal score; dynamically adjusting the access authority according to the score, recording an abnormal behavior and carrying out relevance matching; optimizing the reference model through feedback learning; and evaluating the credibility and the risk level based on an accurate modeling result, and adaptively adjusting permission configuration. According to the method, the user behavior rule can be accurately captured, the data leakage risk can be efficiently identified, the access permission can be dynamically adjusted, the false alarm rate can be reduced, adaptive protection can be realized, and data security and business smoothness can be guaranteed.

The invention relates to a point cloud semantic segmentation method and model based on density perception and feature enhancement, and the method comprises the steps: firstly, carrying out the dynamic density perception sampling in an input point cloud, and carrying out the grouping according to the semantic feature distance; then, after multi-level extraction of deep semantic information, enhancement of relative position information is carried out, and local features of different areas are enhanced; and finally, guiding to recover sampling points according to semantic homogeneity in the up-sampling of the decoder, and fusing the point-level semantic features and the channel-level semantic features by using a bidirectional self-attention mechanism. The model comprises a point cloud input module, a down-sampling module, a cross attention module, a position feature enhancement module, an up-sampling module, a point cloud output module and a trigger threshold judgment module. The method is tested on public data sets S3DIS and Scannet, mIoU reaches 73.0% and 74.8% respectively and is increased by 2.6% and 4.2% respectively relative to a baseline model, and the effectiveness of the method is proved.

The invention relates to a data delivery risk early warning and intelligent management and control method and system. The method comprises the following steps: collecting flow characteristic data, user behavior data and environment characteristic data of data delivery in real time; establishing a dynamic baseline model, and generating reference range data of normal delivery behaviors based on historical data and industrial standards; performing anomaly detection on the user behavior data and the environment characteristic data by using a dynamic baseline model, and generating a risk early warning signal; when a risk early warning signal is received, identifying a risk type, and generating a corresponding risk level; and matching management and control actions in a preset strategy library based on the risk type and the risk level, executing the management and control actions, monitoring a feedback effect, and continuously optimizing dynamic baseline model parameters. According to the method, through dynamic reference generation, multi-dimensional data fusion and a continuous optimization mechanism, the problem of pain points of a traditional model in adaptability, comprehensiveness and maintenance cost is solved.

The invention provides an AI-based carbon footprint and supply chain carbon emission collaborative management method and platform, and the method comprises the steps: collecting various types of carbon-related data from each link of a supply chain in real time, and carrying out the standardization processing to form a unified data flow. An intelligent sensing technology is combined with a dynamic reference model, and sudden carbon emission fluctuation generated in production operation is automatically recognized. The propagation influence and potential risk of these carbon emission fluctuations in the entire supply chain are analyzed. The risk propagation condition and the processing capacity of each node are comprehensively considered, and an optimal carbon emission distribution processing scheme is formulated through an optimization algorithm. A distributed account book technology and an automatic execution mechanism are utilized, and an excitation system is combined to ensure that each participant completes a corresponding carbon processing task according to an optimal scheme. According to the method, the intelligent level and the overall effect of supply chain carbon management can be effectively improved, and a technical support is provided for an enterprise to realize a carbon emission reduction target.

The invention provides an efficient and accurate small target detection method aiming at the problems of low detection precision, high false detection rate and poor real-time efficiency of small target objects in the field of automatic driving. According to the method, a YOLOv8 network is used as a baseline model, a dynamic enhancement convolution module is introduced into a backbone feature extraction network, and more detailed target detail information is extracted. And an adaptive weighted receptive field fusion module is adopted, so that image feature enhancement of diversified receptive fields is realized, and the richness and adaptability of feature representation are improved. And a large kernel channel space attention module is utilized to highlight a small target area and suppress background noise, so that the distinguishing capability and the detection performance of the whole model are remarkably improved. According to the technical scheme, the precision of small target detection is effectively improved, the omission ratio and the false detection ratio are reduced, and the perception capability and the safety of the automatic driving system in a complex traffic environment are enhanced.

The invention provides a learning data accurate mining method based on cognitive calculation driving. The learning data accurate mining method comprises the following steps of S1, performing multi-modal learning behavior data acquisition and heterogeneous integration; s2, a dynamic feature weight optimization step based on calculus; s3, performing cognitive state differential equation modeling; s4, a cognitive diagnosis hybrid model based on statistics; s5, incremental construction of the dynamic knowledge graph is carried out; s6, constructing a federated learning framework for privacy protection; s7, a cognitive intervention strategy is generated; s8, constructing a multi-granularity effect evaluation system; s9, a step of constructing an interpretability enhancement module; s10, a step of carrying out adaptive iterative optimization; the learning data accurate mining method based on cognitive calculation driving has the following advantages that the data utilization rate breaks through the limitation of a traditional method through federated learning and heterogeneous graph fusion; the attention prediction error is reduced through differential equation modeling, and the method is superior to all existing ARIMA / LSTM baseline models.

The invention discloses a high-current accurate measurement method, which relates to the technical field of current measurement, and comprises the following steps: obtaining multi-dimensional measurement parameters, carrying out preprocessing, constructing an error model to obtain a comprehensive error, and carrying out calculation based on the comprehensive error to obtain a comprehensive feedback value; based on the comprehensive feedback value, utilizing a closed-loop feedback mechanism to dynamically optimize a compensation parameter, and further utilizing the compensation parameter and the comprehensive feedback value to carry out error correction on the measurement current; real-time measurement parameters are obtained through a built-in sensor and a factory calibration reference model, an error model is constructed, the dynamic closed-loop self-calibration function is achieved, dependence on external calibration equipment is not needed, the problems that a traditional measurement method is high in complexity and high in cost are solved, and the method is suitable for integrated application of a high-precision electric energy meter and industrial measurement equipment and has wide application prospects. And meanwhile, the real-time performance and the convenience of measurement are remarkably improved.

The invention relates to the field of traffic transportation information engineering, and discloses a group division method and system based on individual movement mode dynamic hypergraph clustering, and the method comprises the steps: constructing an individual movement mode feature matrix under each time slice; constructing an individual movement mode hypergraph under each time slice, and modeling a'many-to-many 'complex high-order association relationship among individual movement modes under the whole time period; proposing a dynamic hypergraph clustering model based on a moving mode, and performing category division of individual associated groups; the provided model is applied to a series of empirical data sets for training and testing, and it is proved that the model can obtain good group division performance superior to that of a baseline model.

The invention provides a patient vital sign anomaly detection method based on an artificial intelligence technology, and relates to the technical field of data processing, and the method comprises the steps: collecting original sign data of a patient; calculating multi-dimensional characteristic parameters; establishing an individual baseline model, and determining a comprehensive reference interval in the model; the vital sign features monitored in real time are constructed into multi-dimensional feature vectors, the multi-dimensional feature vectors are input into the individual baseline model, and the deviation degree of the real-time multi-dimensional feature vectors in the comprehensive reference interval is calculated; performing a clustering analysis to identify an anomalous aggregation region; performing trend analysis, calculating change direction consistency and continuous change amplitude of the multi-dimensional feature vector, and generating a trend analysis result; calculating an accumulated change index in a continuous time window according to a trend analysis result to obtain a dynamic confidence score; when the dynamic confidence score continuously exceeds an adaptive threshold value, determining that a vital sign abnormal event exists; the autonomy and accuracy of the method for detecting the vital sign abnormity of the patient are improved.

The invention relates to the technical field of deformation measurement, in particular to a building deformation monitoring system based on fusion and intersection of Beidou and machine vision technologies, and the system comprises a pose calculation module which is used for setting anchoring nodes, matching Beidou three-dimensional coordinates and inclination angle data of the anchoring nodes obtained in real time with two-dimensional pixel coordinates of the anchoring nodes in a visual image, and calculating the pose of the anchoring nodes; resolving a real-time pose, and converting the real-time pose into a deformation measurement result; the three-dimensional deformation field generation module is used for processing the topological graph of the building structure by adopting a graph calculation model, fusing multi-source heterogeneous information by aggregating and updating node features, and decoding to generate a three-dimensional deformation field of the whole building; and the measurement early warning module is used for constructing a normal behavior baseline model to predict a normal deformation state, carrying out deformation state evaluation and generating deformation early warning information by comparing deformation deviation degrees and combining context analysis and a deformation trend, so that the precision and comprehensiveness of building deformation monitoring are improved.

The invention discloses a power grid abnormal flow detection and identification method based on a multi-mode machine learning algorithm. The method comprises the following steps: step 1, acquiring industrial control protocol message data: capturing a communication message of network flow of an electric power industrial control system through a data collector; step 2, training a multi-modal auto-encoder: training the auto-encoder based on the network traffic characteristics of the normal service; 3, abnormal traffic monitoring: detecting and collecting abnormal traffic by taking the decoding reduction degree under the normal service as a baseline; step 4, abnormal data analysis modeling: training a multi-modal neural network model in combination with the abnormal traffic and the coding result of the baseline model for classifying the abnormal traffic; according to the method, some abnormal traffic existing in the network traffic of the power industrial control system can be detected, and the abnormal traffic in the industrial control system is classified by analyzing the data features and the service scenes of the abnormal traffic.

The invention is suitable for the field of abnormal data monitoring, and provides an abnormal data monitoring method and system based on artificial intelligence, and the system comprises a real-time data flow collection module, a cross-level correlation analysis module, a dynamic baseline model building module, an abnormal behavior comparison module, and an abnormal traceability module. According to the method, through real-time data acquisition, cross-hierarchy analysis, dynamic modeling, anomaly detection and deep threat tracing, an efficient and intelligent anomaly monitoring closed loop is constructed, and remarkable beneficial effects are presented. According to the method, firstly, real-time data flow capture and cross-level correlation analysis are utilized to ensure that interaction behaviors of multiple protocol levels in a network are comprehensively covered, and a complex abnormal mode is recognized in combination with context information; according to the multi-dimensional analysis, the detection accuracy is remarkably improved, false alarm and missing alarm are effectively reduced, meanwhile, an abnormal mark data set with weights is generated, and high-quality data support is provided for follow-up processes.

The invention discloses a cross-platform job resource integrated management system and method based on the Internet, and belongs to the technical field of job resource management. A baseline model is constructed, and whether resource consumption data is abnormal or not is judged; extracting an abnormal resource type, obtaining a corresponding resource demand, identifying tasks having resource competition with the current job, and obtaining an abnormal competition task set; defining job attributes, determining a resource association relationship between tasks, and constructing an abnormal competition task relationship graph; determining the latest execution time of each competition task; analyzing the maximum bearing capacity of the current resource meeting all task execution under the conflict-free condition; calculating the quantity of resources needing to be added; subtracting the actually added resource quantity from the resource quantity needing to be added to obtain a vacant resource quantity; when personnel resources are vacant, the competition tasks are sorted, and the competition tasks are preferentially allocated to high-priority exclusive jobs in combination with the maximum bearing capacity; when the device resources are vacant, idle special devices are preferentially allocated for exclusive jobs.

The invention discloses a dynamic electrocardiogram early warning system based on artificial intelligence, which relates to the technical field of medical health monitoring, and comprises a model building module, a threshold setting module, an early warning threshold setting module based on the baseline model, a threshold adjusting module, and a dynamic electrocardiogram early warning module based on the real-time activity state of the user. And dynamically adjusting the early warning threshold. By dynamically adjusting the early warning threshold value, daily activity changes of the user can be better adapted, unnecessary false alarm and missing alarm are reduced, the early warning reliability and the user experience are improved, in addition, a dynamic threshold value adjusting mechanism can also capture the influence of user emotion fluctuation on physiological indexes, and the user experience is improved for high-intensity motion and static states. The early warning threshold can be relaxed and tightened respectively, the sensitivity and accuracy of early warning are ensured, and the intelligent level of early warning monitoring is further improved.

The invention discloses a line aging monitoring system applied to a safe electric device, and the system comprises the steps: collecting time series data of line state parameters (such as multi-point temperature, key node current and voltage, and leakage current), and training a dynamic baseline model through employing historical data collected under a normal operation condition; the model can learn and predict an expected parameter value after considering the influence of factors such as load and environment in a normal operation state of the line; then, comparing a line state parameter time sequence acquired in real time with a parameter baseline value time sequence predicted by the dynamic baseline model, and calculating a residual vector sequence between the two time sequences; furthermore, the abnormal state of the line is accurately judged by analyzing the mode characteristics of the residual error sequence in a period of time, so that the problem that normal fluctuation and real aging abnormity are difficult to distinguish by a traditional method is effectively solved, and the accuracy and reliability of line aging monitoring are remarkably improved.

The invention relates to the technical field of air conditioner simulation optimization, in particular to an efficient machine room multi-subsystem cooperative control method based on intelligent simulation, and the method comprises the steps: building a digital twinborn model and a virtual simulation environment of central air conditioner key equipment; performing simulation analysis on the health state and the operation performance of key equipment in the digital twin model through a virtual simulation environment, and establishing a performance reference model; acquiring an external data source to obtain equipment operation parameters; in the virtual simulation environment, the equipment operation parameters are analyzed based on the performance reference model, and health state parameters and operation performance parameters are obtained; dynamically updating the digital twin model based on the health state parameters and the operation performance parameters to obtain an equipment state twin model; and based on the equipment state twin model, performing simulation design in a virtual simulation environment to obtain equipment compensatory cooperative control parameters. According to the invention, through simulation analysis of the digital twin model, cooperative control of the central air-conditioning subsystem is realized.

The invention discloses a joint multi-task table semantic analysis method based on a pre-training model, and relates to the technical field of natural language processing and databases. According to the method, SQL statements are crawled from a specified website, SQL and tables are converted into natural language texts through a large language model, columns and tables are extracted to form positive and negative samples, experimental data is converted into a Spider data set format, and natural language problems and database mode tasks are completed through a cue word template and a few-sample frame; then, an MLNaT model of a 12-layer relation perception Transform architecture is constructed, statements and column names spliced according to a specific format are input, three tasks of mask language setting, column prediction and SQL generation are set, and pre-training is carried out; and finally, evaluating on a Spider data set according to an accurate set matching rate, and taking an RAT-SQL as a baseline model. It is verified that the MLNaT model is superior to the reference model in the aspects of column prediction and SQL generation.

The invention provides a leasing equipment full life cycle monitoring method and system based on the Internet of Things, and belongs to the technical field of equipment operation and maintenance monitoring, and the method comprises the steps: obtaining real-time state data of leasing equipment in an operation process through a multi-source sensor; performing feature extraction on the real-time state data to obtain various feature vectors, and constructing a multi-dimensional feature vector set; inputting the multi-dimensional feature vector set into a random forest model, and judging whether the leasing equipment is in an abnormal state or not in combination with a space-time clustering algorithm; if yes, fault prediction is carried out on the leasing equipment in combination with an abnormal detection result, historical operation data and a baseline model, and a fault probability value of the leasing equipment is obtained; the future position of the leasing equipment is predicted based on the position data, an electronic fence boundary is set for the leasing equipment, if the future position exceeds the electronic fence boundary, an early warning notification of trajectory deviation is generated, and the system finally realizes trajectory and fault visual presentation in a thermodynamic diagram mode, so that the equipment operation and maintenance intelligence and management precision are improved.

The invention discloses an urban occupational health risk real-time supervision method and system based on data analysis, and the method comprises the steps: continuously collecting the behavior data flow of a target employee and the physical environment dynamic data flow of an associated working region, and generating an individual-environment dynamic data sequence; detecting a behavior abnormal signal based on the historical behavior mode of the target employee; in combination with the physical environment dynamic data flow at the current moment, the association strength between the physical environment dynamic data flow and the occupational health risk is evaluated, and a personalized micro-intervention instruction is generated and triggered; capturing the instant feedback response of the employee to the micro-intervention instruction in real time, updating the personal behavior baseline model of the target employee in real time, and calibrating the evaluation rule of the association strength; and dynamically generating a real-time occupational health risk portrait of the employee based on the individual behavior baseline model, the association strength evaluation rule and the individual-environment dynamic data sequence. According to the embodiment of the invention, accurate early warning and personalized intervention of occupational health risks can be realized.

The invention relates to the technical field of heat transmission, in particular to a heat transmission data storage and management system based on an industrial big data platform, and the system comprises a self-adaptive acoustic baseline modeling module which generates a self-adaptive baseline model library for storing the mapping relation between a working condition area and a model; the abnormal deviation degree calculation module is used for calculating and generating an abnormal deviation degree; the health state evaluation module is used for generating a comprehensive health index representing the long-term service performance of the pipe network; and the closed-loop correction and scheduling module is used for determining a comprehensive risk level according to the comprehensive health index and the change trend thereof, generating an operation and maintenance scheduling instruction for dynamically adjusting an abnormal deviation degree calculation process and pipe network operation parameters, and realizing closed-loop feedback control. According to the invention, accurate identification and positioning of abnormal events such as leakage, third-party damage and the like are realized.

The invention relates to the technical field of seismic station network operation and maintenance management, and particularly discloses a seismic station network operation and maintenance edge computing system and a seismic station network operation and maintenance edge computing device, which efficiently access a multi-source original data stream of seismic station network terminal equipment by using edge computing nodes, construct a data-equipment-environment ternary association graph, realize comprehensive integration and dynamic weaving of data, and improve the efficiency of seismic station network operation and maintenance. The integrity and the real-time performance of the data are ensured; by adopting a cross-modal attention mechanism, multi-source features such as earthquakes, equipment and environments are effectively fused, and the effectiveness and relevance of feature expression are enhanced; by constructing a multi-factor dynamic baseline model, the system can accurately quantify the abnormal risk level, and a basis is provided for operation and maintenance decision making; and automatic scheduling and effect evaluation of operation and maintenance tasks are realized through the intelligent response module, and the operation and maintenance efficiency and the response speed are improved.

The invention belongs to the technical field of graph classification, discloses a graph classification method based on subgraph integration and position sensing, and relates to a graph classification method SIPA combining substructure embedding and node position sensing. The SIPA firstly extracts sub-graphs through two different strategies so as to capture multi-sample sub-structures in the graphs, the structural features of the sub-structures are coded by adopting a graph convolutional network, and information of different sub-graphs is effectively fused through an attention mechanism. Then, anchor nodes are introduced to calculate relative position information of nodes, and the information is embedded into node representation to better capture global position features. And finally, a graph information bottleneck mechanism is used for optimizing node representation and removing redundant information irrelevant to a classification task. The method not only effectively learns the local structure information, but also enhances the perception capability of the relative position information of the nodes. Experimental results show that SIPA is superior to an existing baseline model in five data sets, and the superiority of SIPA in a graph classification task is verified.