283 results about "Classification rule" patented technology

The invention relates to the technical field of databases, in particular to an intelligent mapping and classifying method based on heterogeneous data sources, which comprises the following steps: collecting heterogeneous data streams through an API (Application Program Interface) gateway and converting the heterogeneous data streams into structured data packets; using a semantic topology engine to fuse BERT semantic extraction, a graph convolutional network and a dynamic time warping technology to generate a cross-source association graph; constructing a field type clustering center by adopting a meta-learning framework based on the atlas, generating an initial classification rule through mode compatibility measurement, and dynamically updating the rule by means of adversarial training; outputting a DSL configuration script in combination with a template engine and an AST compiling technology; and dynamically adjusting a graph convolution weight and classifier parameters by using a strategy gradient algorithm through a reinforcement learning agent, and establishing a mapping-classification-verification collaborative optimization mechanism. According to the method, cross-source data semantic association accuracy is improved, small sample adaptive classification is realized, and system robustness and efficiency are improved.

A method and system for controlling a heating, ventilation, and air conditioning system (HVAC) in a building based on adaptive model predictive control (MPC) is disclosed, and belongs to the technical field of intelligent building adaptive control. The method may include: acquiring operation data of the HVAC under different working conditions, and performing preprocessing on the acquired data; calculating a marginal fitness measure (MAD) of each data point for each class based on the preprocessed data, and calculating a fitness measure of each data point for each class by combining MADs of all descriptors to obtain a global fitness measure (GAD); setting a GAD threshold, and classifying current operation states of the system according to classification rules; and adopting adaptive control strategies for different operation states since model parameters of each class are fixed after classification.

The invention discloses an AI-based archive digitization full-process management and control method, which comprises the following steps: acquiring archive image data and text data, constructing an original archive data set, and generating an initial directory set; constructing an archive directory template set, executing template matching and structure adaptation, and generating an initial structure recognition result; setting a classification rule of directory deviation and structural contradiction, constructing a deviation rule base, executing rule scanning and node comparison, and generating a deviation contradiction label sequence; constructing an evolution state vector, inputting the evolution state vector into an improved GMN model, and generating a structure repair state vector through a structure guiding unit, a deviation sensing unit, a path fusion unit and a node evolution unit in sequence; and executing result updating and conflict rearrangement, completing standard template alignment and semantic coverage verification, and generating a final structure recognition result. According to the method, the structural integrity, the semantic consistency and the automatic repair capability of directory recognition are improved, and the method has relatively high adaptability and intelligent level.

The invention relates to an automatic driving risk identification model training method and system, and a risk assessment method and system, and belongs to the technical field of vehicle automatic driving. The automatic driving risk identification model training method comprises the steps of performing time synchronization processing on multi-source sensing data to form a driving data set, and performing category labeling on the driving data set according to a set classification rule; the categories comprise traffic infrastructures, traffic auxiliary facilities, traffic participants and traffic environments; performing physical state information labeling on the driving data set subjected to category labeling to obtain a preprocessed data set; generating character risk cues describing different driving scenes based on the preprocessed data set; risk assessment is carried out on different driving scenes, risk states of the driving scenes are marked based on risk assessment results, and a set number of risk targets are marked for the driving scenes which are confirmed to be risk scenes; and forming a sample set by using the character risk prompt words, the risk states and the risk targets corresponding to different driving scenes.

Described embodiments provide systems and methods for mapping high-priority WiFi packets to low latency data pipe. An access point can receive the one or more low latency classification rules from a device monitoring one or more low latency classification rules at a cable modem termination system (CMTS). The access point can be in communication with the CMTS and provide wireless access to one or more devices in a local network. The access point can receive packets corresponding to the one or more low latency classification rules from the CMTS. The access point can communicate, based at least on the one or more low latency classification rules, the packets to a queue used for transmission of low latency traffic to the one or more devices in the local network.

The invention relates to the technical field of non-road mechanical vehicle monitoring, in particular to a non-road mechanical vehicle fuel consumption real-time estimation method and system. According to the method, multi-source data are collected and preprocessed, and the current working state of a vehicle is recognized in real time based on classification rules; dynamically setting Kalman filtering parameters according to the identified working state of the vehicle; predicting a predicted fuel state vector and an error covariance matrix at the next sampling moment based on Kalman filtering; meanwhile, the refueling event is detected in parallel, and if the refueling event is detected, the refueling event is recorded, and the fuel oil state and the error covariance matrix are reset; and carrying out fusion processing on the actual observation information and the predicted fuel state vector based on Kalman filtering to obtain optimal fuel state estimation. According to the method, the multi-dimensional observation model is constructed, and the fuel consumption estimation precision is greatly improved through multi-source information complementation; and a dynamic mapping relation between working condition identification and filter parameters is established, so that the filtering algorithm keeps excellent performance under different working conditions.

The present disclosure provides a method for selectively encrypting and decrypting portions of a document. The method includes receiving a document containing plaintext, analyzing the document using an artificial intelligence system to identify portions containing sensitive information, assigning predefined security levels or attributes to the identified portions, sending an encrypt file content request to a cryptographic engine, receiving an encrypted file content response, building an encrypted document by replacing portions with masking symbols and storing encrypted portions as metadata, and sending the encrypted document to a client device. The artificial intelligence system uses machine learning models trained on domain-specific data and employs pattern recognition and contextual analysis to identify sensitive content based on semantic understanding and classification rules.

The invention discloses a data modeling method and device, computer equipment and a storage medium, relates to the technical field of artificial intelligence, and is applied to data modeling of a financial big data platform. The original data elements are processed in a standardized mode, and the type labels are automatically marked according to the preset classification rules; in combination with a feature coding technology, converting multi-dimensional data elements into uniform characterization representation; calculating by using a data field standard map and semantic similarity, automatically identifying an abnormal mode, and performing quantitative scoring through a pre-training review model; and finally, performing iterative optimization according to the review index score, and generating a detailed feedback report. According to the method, the modeling efficiency is greatly improved, automation and intelligentization of the field creation, type judgment and standard binding processes are realized, the data standard coverage rate and the modeling quality are effectively improved, the risk of manual review and script deployment errors is reduced, and the modeling efficiency is improved. And the modeling reliability and the operation and maintenance efficiency of the financial big data platform in a complex business scene are enhanced.

The application relates to the technical field of network flow management, and discloses a network flow detection method and device. The method comprises the following steps: a gateway acquires to-be-detected network flow corresponding to a specified port on a server; the gateway classifies and processes to-be-detected flow data according to a preset rule to obtain flow data corresponding to each flow type, wherein the preset rule is a classification rule set for the flow type; the gateway performs abnormality detection on the flow data of each flow type to obtain to-be-distributed flow data; the gateway distributes the to-be-distributed flow data to a service node associated with the corresponding flow type; and the service node performs abnormality detection on the flow data received from the gateway when the micro service of the service node is started. Through implementation of the application, the gateway and each service node can directly perform independent abnormality detection on the flow data, and the detection time of malicious flow is saved, and the detection efficiency of the malicious flow is improved.

The invention provides a dynamic self-adaptive data stream accelerated transmission method and device oriented to a credential environment. The method comprises the following steps: carrying out capability detection on a central processing unit, an operating system, a network interface and a password acceleration unit in a credential computing node to obtain environment capability information, and organizing the environment capability information into a credential environment capability portrait; analyzing the service type, the data scale, the timeliness level and the security level of the data stream, and generating a data stream portrait corresponding to the data stream in combination with a preset classification rule; acquiring running state information in a cross-layer manner and organizing the running state information into running state vectors; based on the credential environment capability portrait, the data stream portrait and the operation state vector, executing a preset strategy decision process, and determining a transmission strategy parameter set for the data stream; and sending the data fragments according to the transmission strategy parameter set. The resource utilization rate of the creative environment can be improved, the multi-type data stream transmission efficiency is improved, and the manual tuning overhead of transmission parameters is reduced.

The invention relates to the technical field of intelligent management, and discloses an after-sales service management method based on intelligent scheduling and automatic order dispatching, and the method comprises the steps: obtaining the maintenance and after-sales work order of a customer, and extracting the work order information; classifying the work order information according to a preset classification rule, generating different types of work orders, and distributing the work orders to a to-be-processed queue; matching a plurality of engineers for the work order to be processed by adopting an intelligent matching algorithm, and calculating an engineer score based on factors such as skill matching, load, response time efficiency and historical evaluation; assigning engineer processing work orders according to the scores, marking the work orders as to-be-processed work orders if the work orders cannot be matched, and transferring to manual auditing; if the distribution is successful, tracking the position of the engineer and the service progress in real time, and pushing information to the client; after the work order is completed, client evaluation is collected, and engineer scores are updated in real time. According to the method, the problem that a multi-dimensional scoring judgment, real-time position information tracking and service process visualization feedback mechanism is lacked can be solved.

The invention discloses a medical data grading, classifying and desensitizing system and method based on role authority, and relates to the technical field of medical data security, and the system comprises a classification rule making module which maps a CRF template to a classification space and carries out dynamic sensitive weight calibration; the desensitization information identification engine integrates BERT, BiLSTM and CRF deep learning technologies, carries out sensitive entity identification and label labeling, and executes differential desensitization according to a mapping relation between a sensitive level and a desensitization strategy; the real-time data quality detection module is used for realizing real-time evaluation and early warning of data quality; the role authority management module is used for constructing an attribute encryption strategy space and realizing authority management; according to the method and the device, full-process automatic processing of the medical data from classification, identification, desensitization and quality detection to authority management and feedback optimization is formed, the sensitivity level of the medical data field can be accurately identified, a scientific desensitization strategy basis is provided, and meanwhile, the quality and privacy security of the data in the desensitization process are ensured.

The invention provides a manufacturing resource multi-modal information real-time sensing method based on cloud edge collaboration, and relates to the technical field of cloud edge collaboration computation.The method comprises the steps that tasks are preliminarily divided according to preset task classification rules, and the task classification rules comprise real-time requirements and processing characteristics; different execution strategies are triggered according to different task types obtained through preliminary division; according to the real-time computing resource state and the task dynamic characteristics in the task execution process, the task type is dynamically adjusted, and according to the dynamically adjusted task type, a corresponding execution strategy is triggered; and according to historical execution data and a global resource state after task execution is completed, performing long-term optimization division on the dynamically adjusted task type through a reinforcement learning algorithm to obtain a global optimal scheduling strategy.

The invention discloses an anomaly detection method, device and equipment for a maglev train suspension system and a medium, and relates to the technical field of maglev train detection. The method comprises the following steps: acquiring multi-source operation time sequence data of a maglev train suspension system in real time; carrying out automatic feature extraction on the suspension system through a first neural network model to obtain feature vectors used for representing dynamic response behaviors of the suspension system under different working conditions; inputting the feature vector into a trained second neural network model to obtain an anomaly detection score; the second neural network model is a time sequence neural network and is used for outputting a quantitative score value representing the state abnormity possibility of the suspension system according to the input vector; and when the anomaly detection score indicates that the suspension system is in an abnormal state, performing fault mode classification on the abnormal state according to a preset fault classification rule. According to the technical scheme of the invention, the method breaks through the limitation that a conventional method is insufficient in coupling feature extraction, and achieves the precise anomaly detection and fault classification of the suspension system.

The invention relates to the technical field of embedded equipment control, and discloses a multi-event cooperative control method and device for audio equipment, equipment and a storage medium, and the method comprises the steps: constructing a queue structure body of multi-event circulation according to hardware parameters of the audio equipment, and achieving the efficient cooperative processing of events through a classification rule and a time scheduling method. The method has the advantages of effectively managing event queues, reducing high-priority event delay, avoiding equipment downtime and improving user experience and equipment performance.

The invention relates to the technical field of cooling tower energy-saving control, and discloses a closed cooling tower energy-saving control system, which comprises a multi-dimensional sensing module, an intelligent decision module, a precise execution module and a self-diagnosis optimization module, and is characterized in that the multi-dimensional sensing module acquires target parameters in real time and transmits the target parameters to the intelligent decision module; the intelligent decision-making module analyzes and processes the collected parameters and outputs a control instruction based on a preset working condition classification rule in combination with the dynamic energy efficiency model and the dual-threshold judgment logic; the precise execution module adjusts the operation state of a core component of the closed cooling tower according to the control instruction and feeds back real-time operation data to the intelligent decision module; the self-diagnosis optimization module monitors the equipment state based on feedback data, the system can flexibly adjust the operation state according to environment parameters and heat exchange loads through cooperative frequency conversion adjustment of a draught fan and a spraying pump and a dynamic adaptation strategy according to different working conditions, and compared with a traditional fixed rotating speed control mode, the comprehensive energy consumption of the closed cooling tower is effectively reduced.

The invention discloses a multi-modal data classification method and device and a storage medium, and belongs to the technical field of multi-modal data processing. The method comprises the following steps: extracting feature vectors of multi-modal energy data, fusing the feature vectors into a comprehensive feature vector, and determining an association relationship between the multi-modal energy data and existing energy field knowledge; according to the association relationship and the data characteristics of the multi-modal energy data, the classification rule of the existing energy field knowledge is updated; and according to the comprehensive feature vector and the updated classification rule, classifying the multi-modal energy data to obtain a classification result, that is, generating the comprehensive feature vector from the multi-modal energy data, and determining an association relationship between the multi-modal energy data and the existing knowledge, so that the multi-modal energy data can be classified according to the association relationship and the data characteristics of the multi-modal energy data. And updating the classification rule to meet the effect of classifying the multi-modal energy data.

The embodiment of the invention provides an intelligent file management method and system. According to the method, when a target file uploaded by a user is received, metadata, user behavior data, file content and scene information of the target file can be extracted by using a data acquisition layer, and a foundation is laid for subsequent accurate processing; then a classification rule is optimized by means of multi-modal semantic analysis of a large model and a reinforcement learning algorithm, and a final target classification label is generated in combination with user behavior data and scene information, so that a traditional manual labeling mode is eliminated, and intelligent, rapid and accurate classification can be realized; and finally, a corresponding encryption algorithm, a key updating period and an access permission are matched based on the target classification label, so that the problems that a traditional static encryption mode carries out one-cut on files with different sensitivities and the matching degree of the encryption algorithm and the file sensitiveness is low are avoided, and automatic classification and dynamic matching of encryption strategies according to file contents, use scenes and user behavior data are realized. Therefore, the working efficiency can be obviously improved.

The described embodiments provide systems and methods for mapping high priority WiFi packets to low latency data pipelines. An access point may receive one or more low latency classification rules from a device that monitors the one or more low latency classification rules at a cable modem terminal system (CMTS). The access point may communicate with the CMTS and provide wireless access to one or more devices in a local network. The access point may receive packets from the CMTS that correspond to the one or more low latency classification rules. The access point may communicate the packet to a queue for transmitting low latency traffic to the one or more devices in the local network based at least on the one or more low latency classification rules.

The invention relates to the technical field of government affair services, in particular to a rural land two-round package-extension registration method and system, and the method comprises the following steps: obtaining a first character and a length of an identification number field, writing the first character and the length into an affiliation field, extracting a delay period record classification cycle type, combining a purpose classification operation mode to generate adaptive information, and matching a permission field to establish field correspondence; the additional path content hooks the unit code and the period field. According to the method, identification number affiliation classification is completed by extracting first characters and lengths of fields, plot stage types are divided in combination with periodic direction information, a time sequence structure corresponding to a packet extension state is constructed, purpose fields are combined according to a standard classification rule to generate adaptation items, and permission field contents are subjected to inter-field value matching and content connection according to a purpose structure. A unit code and periodic information are written into a path field to form a data hooking chain, so that the collaborative linkage capability of a field structure, the node connection stability and the extension level of information registration are enhanced, and a registration system of a data structure under multiple scenes is promoted.

The invention provides a user file management method and device, electronic equipment and a storage medium, and relates to the technical field of data processing. Target files uploaded by a user are analyzed and converted into a unified first semantic vector; storing the semantic vector and the metadata into a dynamic knowledge base together, synchronizing newly added contents in real time through an incremental updating mechanism, constructing a cross-modal knowledge graph based on a tag classification rule and association logic, and generating a second semantic vector after receiving natural language query to perform joint retrieval of the semantic vector and the metadata. By means of cross-modal knowledge graph optimization result relevance, multi-modal answers containing texts, chart links and the like are finally generated through a pre-training model, and the multi-modal answers are returned through a standardized interface. The problems that in the prior art, due to the fact that a unified semantic processing mechanism is lacked, heterogeneous files are difficult to retrieve, newly-added content cannot be contained in time in a knowledge base staticizing mode, and retrieval results are fragmented due to the fact that cross-modal knowledge association is lacked can be solved.

The invention relates to the technical field of reservoir flood control early warning, and discloses a reservoir flood control risk assessment and early warning system based on big data. A data aggregation unit of the system continuously collects historical hydrology, real-time meteorology, reservoir area geographic information and surrounding social economic data; the data preprocessing unit performs multi-dimensional cleaning and standardization on the data to generate an assessment input data set with a unified format; the risk dynamic analysis unit constructs a flood occurrence probability prediction model by applying a machine learning algorithm on the basis of the data set, constructs a submerging range simulation model in combination with geographic space characteristics, and calculates a potential economic loss quantized value; the early warning grade mapping unit inputs the result into an early warning grading rule base to generate an overall flood control risk early warning grade; a plan matching decision-making unit retrieves a flood control coping plan set with the highest matching degree according to the early warning level; and the early warning information issuing unit converts the pre-arranged plan into visual guidance information, and broadcasts and pushes the visual guidance information to an emergency management department and downstream resident terminals.

The invention belongs to the technical field of artificial intelligence, and particularly relates to an emergency perception method containing semantic extraction, which comprises the following steps of: S1, constructing a historical word library about emergencies, and forming an association weight between a vocabulary group and a user through historical data training; s2, constructing an entity variant word expansion model based on a double-layer model to realize generation of a variant word bank, wherein the double-layer model comprises a proper noun expansion layer and a historical word bank expansion layer; s3, collecting text data on a network, performing classification processing on the text data based on a text classification rule, and preliminarily extracting entity words; s4, inputting the extracted entity words into the variant word expansion model to obtain a corresponding variant word group, and completing identification and extraction of the entity words; and S5, carrying out situation analysis based on the knowledge graph, and predicting the occurrence and development situation of the emergencies. According to the method, the variant word bank suitable for the whole network is mined on the basis of the incidence relation of different users to different similar words, and the reliability of vocabulary expansion is improved.

The invention relates to a multi-dimensional index model-oriented index system automatic construction method and system, and the method comprises the steps: firstly obtaining a concept index input by a business person, converting the concept index into an entity index, and carrying out the logic verification of the entity index through a metadata verification method; identifying and matching the entity indexes passing the verification with the business classification rules to generate a reference business theme, establishing a derived individual business theme according to the reference business theme, and constructing a dimension bus architecture based on the reference business theme and the individual business theme; generating a multi-dimensional index model based on the dimension bus architecture and performing consistency verification; then, through layer-by-layer tracing of a service system, a service entity and service attributes, positioning data sources of basic indexes and associated dimensions in the verified multi-dimensional index model, and correcting the verified multi-dimensional index model according to the data sources, so as to generate a physical model containing a mapping relationship between the indexes and the physical model; and finally, configuring operation parameters of the physical model, and automatically generating an index system.

The invention discloses a pet motion state judgment method and system, and relates to the technical field of intelligent pet keeping. Comprising the following steps: continuously collecting original sensor data of a pet, wherein the original sensor data at least comprises three-axis acceleration and three-axis angular velocity; data preprocessing and motion parameter resolving are carried out on the original sensor data, multi-dimensional motion data used for motion state judgment are obtained, and the multi-dimensional motion data comprise at least one piece of continuous motion data in a time domain; acquiring a preset classification rule base corresponding to the body type category according to the body type category of the pet; and obtaining a target motion state based on the multi-dimensional motion data and the classification rule base. Through the multi-dimensional data and the targeted classification rule base, the technical problem that in the prior art, a unified model or a fixed threshold value cannot adapt to motion feature differences of pets with different body types is solved, and accurate judgment of various fine-grained motion states of the pets is achieved; and reliable technical support is provided for pet health monitoring and behavior analysis.

The invention provides a business decision-making method and device, equipment, a medium and a program product, relates to the technical field of artificial intelligence, and can solve the problem that accurate business decision-making results cannot be made for different customer groups. According to the specific technical scheme, N first group feature vectors are determined based on to-be-processed customer data of multiple customers, and the N first group feature vectors comprise a customer credit evaluation feature vector, a customer relationship feature vector, a business processing feature vector and other business factor feature vectors; based on M preset classification rules and the N first group feature vectors, dividing the to-be-processed customer data into M first customer group data sets; and for each first customer group data set, processing the first customer group data set by adopting the decision tree model of the business scene corresponding to the first customer group data set to obtain a business decision result of the customer corresponding to the first customer group data set. The method is used for making accurate business decision results for different customer groups.

The application relates to the technical field of distributed data processing, in particular to a data collection and storage method and system, an organic light-emitting diode inkjet printing equipment and a storage medium. The method comprises the following steps: a collection client obtains a collection task of the organic light-emitting diode inkjet printing equipment; based on a preset data type classification rule, the collection task is split into multiple independent subtasks, wherein the number of the independent subtasks is adapted to the number of the collection clients; each collection client collects original data of the organic light-emitting diode inkjet printing equipment in parallel according to each independent subtask, and sends the original data to a master node of a distributed cache cluster; each master node receives original data of a corresponding type, divides a storage partition according to a type corresponding to the preset data type classification rule, and directly writes the original data into a local memory to complete real-time storage, so that the data storage delay of the organic light-emitting diode inkjet printing equipment multi-type data collection is low.

This application provides an object segmentation method and an object segmentation model training method, which can be applied to the field of artificial intelligence technology. The object segmentation method includes: determining whether the current video frame in a video frame sequence is a keyframe, a regular frame, or a frame to be inspected based on a preset classification rule; if the current video frame is a keyframe, extracting features from the current video frame to obtain the current frame image features; if the current video frame is a regular frame or a frame to be inspected, determining the motion change features of the current video frame relative to a reference keyframe associated with the current video frame, and determining the current frame image features based on the motion change features and the reference image features; and performing object segmentation on the current frame image features based on the target mask features representing the target object to obtain the object segmentation result of the current video frame, wherein the object segmentation result indicates the position and shape of the target object.

The application discloses an online checking method and system for an SF6 gas detector, and the method comprises the following steps: controlling a standard gas source to output an asymmetric triangular concentration waveform, wherein the rising rate of the waveform is not equal to the falling rate; collecting real-time concentration of the detector; drawing rising and falling response curves with standard concentration as the horizontal coordinate and real-time concentration as the vertical coordinate; calculating the deviation area between the two curves and an ideal straight line; if the area ratio is not greater than a dynamic threshold, extracting all intersection points of the two curves and sorting them; calculating the algebraic area of the region surrounded by adjacent intersection points to generate a surrounding area symbol string; and determining the fault type according to the intersection point number parity and zero value, intersection point horizontal coordinate interval dispersion degree and symbol string distribution according to the topological classification rule. The asymmetric excitation fully exposes the adsorption-desorption asymmetric characteristics of the sensor, and the response curve topological characteristics realize online identification and classification of baseline separation, translation drift, response lag, nonlinear superposition and noise jamming and the like.

The invention relates to the technical field of information, in particular to a composite stepped event text classification marking method, which comprises the following steps of: defining event classification tags, and constructing a tree-shaped hierarchical tag structure according to the event classification tags; formulating classification rules for the event classification labels and constructing a rule base; rule matching classification is carried out on the input production data, if rule matching classification fails or rule matching classification confidence is lower than a set threshold value, first model classification is carried out, and if first model classification fails or first model classification confidence is lower than the set threshold value, second model classification is carried out; performing multi-dimensional evaluation and cross validation on the classification result of the second model; outputting classification result data, and outputting the classification result data according to at least one of the rule classification result, the first model classification result and the second model classification result. A traditional rule matching technology and a large-scale and small-scale language pre-training model are organically integrated to construct a classification system with high efficiency, accuracy and adaptability.