38 results about "State transition probability" patented technology

The invention relates to the technical field of encrypted communication, and discloses an intelligent analysis method and system based on multi-source data, and the method comprises the steps: obtaining multi-dimensional metadata of an encrypted communication flow, and generating a feature vector set; calculating a state transition probability of a communication session by using a Markov chain model, dividing the communication session according to a time window and mapping the communication session to a discrete state, and generating a session state evolution sequence; analyzing a session state evolution sequence based on a hierarchical topic model, mapping the state sequence to a behavior topic through potential Dirichlet allocation, mapping the behavior topic to an intention category through a hierarchical Dirichlet process, and outputting a communication intention category and a confidence score; according to the method, the limitation that static feature analysis cannot reflect the complete life cycle of the session is overcome, the problem of semantic gaps caused by limited metadata information dimensions is solved, and the accuracy and robustness of intention recognition are improved.

The invention discloses a gateway load balancing method, a gateway load balancing device and electronic equipment. The method comprises the following steps: determining an initial load state of a network element instance in a service communication proxy (SCP) gateway; the state transition probability of the network element instance is determined by adopting a Markov chain model according to the initial load state, the future load state of the network element instance is determined according to the state transition probability, and the state transition probability introduces a time decay factor; the time attenuation factor is used for adjusting the weight of the real-time network element performance data in the preset time window relative to the historical network element performance data; and receiving a network element service request, and executing network element instance allocation according to the importance degree of the network element service request and the future load state of the network element instance. The technical problems that the delay of important services is increased and the system performance is reduced due to the fact that the static load balancing strategy in the related technology cannot effectively predict and respond to the load state change of the network element instance are solved.

The invention belongs to the field of big data, and relates to a content recommendation method and device, equipment and a medium, and the method comprises the steps: obtaining a contact behavior sequence of a target user, the contact behavior sequence comprising an interaction behavior between the target user and a vehicle service platform in a target scene; modeling the contact behavior sequence through a Markov process algorithm, mapping interaction behaviors into state nodes, and generating a state sequence; acquiring a state transition record of a historical user in the target scene; based on the state sequence and the state transition record, a preset deep neural network is adopted for processing, and a state transition probability matrix of the target user is generated; and determining to-be-recommended contents based on the state transition probability matrix, and recommending the contents to the target user. The method can be applied to the business field of finance and the like, and content recommendation accuracy can be improved.

The invention provides a Tor network data parallel node state prediction method and system based on a Markov chain, and relates to the technical field of Tor network security analysis, and the method comprises the steps: S1, obtaining a historical state data sequence of a Tor network relay node; s2, establishing a Tor network relay node state prediction model based on a Markov chain, and predicting a Tor network relay node state transition probability; s3, performing Markov property test on the Tor network relay node state transition probability matrix; and S4, establishing a parallel Tor network relay node state prediction model by using a distributed computing framework Spark, so as to improve the prediction efficiency. Through the time-sharing state data sampling technology, the Markov property test passing rate of the node state sequence is improved, so that the effectiveness of model prediction is improved; a parallelization model is established to process large-scale Tor network data, the prediction speed is remarkably improved, the target range is narrowed for node routing analysis, and support is provided for real-time data analysis.

The invention discloses a water-wind-light complementary space-time scene generation method and device, and the method comprises the steps: discretizing a continuous state space of a multi-dimensional water-wind-light resource into a self-adaptive grid, and constructing a state transition probability matrix in an offline manner; in the online stage, a discrete state sequence is rapidly generated through Markov chain migration, and a continuous scene is reconstructed through kernel density estimation. A layered Markov switching model is introduced to capture long-period features, and smooth interpolation transition is adopted during macroscopic state switching. The device comprises a gridding module, a matrix construction module and a scene generation module, and supports a Z-order curve index and a compressed sparse row format to realize microsecond access. Compared with a traditional VAR iteration method, single-time long-period scene generation time is reduced from an hour level to a millisecond level, and calculation efficiency and scene quality are both considered.

The application provides a behavior control method and device; the method comprises the following steps: determining a current behavior state of a virtual object and a state transition matrix of the virtual object in a current time period; determining a first state transition probability distribution corresponding to the current behavior state from the state transition matrix; in response to the first state transition probability distribution, transferring the behavior state of the virtual object from the current behavior state to a first target behavior state in a behavior state set, generating a first behavior event queue based on a Poisson rate corresponding to the first target behavior state; and controlling the virtual object to execute behavior events in the first behavior event queue in sequence. Through the application, accurate control of the behavior of the virtual object can be realized, thereby improving the fidelity of the behavior of the virtual object.

A data estimation device includes an acquisition unit, a stratification unit, an estimation unit, and an output unit. The acquisition unit acquires data sets including pieces of data indicating mutually different probability distributions and an attribute used for stratification of the data sets. The stratification unit stratifies the data sets based on the attribute. The estimation unit estimates a state transition probability between the data sets after the stratification based on a difference in distribution between a state transition probability between the data sets before the stratification and a state transition probability between the data sets stratified for each of the attributes. The output unit outputs the state transition probability between the data sets after the stratification. The use of the state transition probability estimated in this manner enables the data estimation device to support decision making based on an estimation result of a transition destination of data.

The invention discloses a service state identification method and device and electronic equipment. Relates to the big data field. The method comprises the steps that a trusteeship fund allocation instruction sequence sent by a user is received, a classification variable sequence of the trusteeship fund allocation instruction sequence is generated, and the trusteeship fund allocation instruction sequence comprises M trusteeship fund allocation instructions; target parameters in the probability prediction model are obtained, and the target parameters comprise an initial probability vector representing an initial state probability, a state transition probability matrix representing a state transition probability and an observation probability matrix representing an observation probability; and calculating the target parameter and the classification variable sequence through a Viterbi algorithm to obtain a service state sequence corresponding to the trusteeship fund allocation instruction sequence. According to the method and the device, the problem of relatively low accuracy and efficiency of manually determining the operation state of the business indicated by the instruction in the related technology is solved.

The invention relates to the technical field of flow intelligent guidance and control, and discloses an engine training guidance and control flow control method and system. The method comprises the steps that a training process is divided into continuous stages, and execution tracks are extracted to construct state snapshots; analyzing the features to generate a flow operation state mapping table; reconstructing a track and establishing a node state transition sequence, and describing a flow rule through a state transition probability matrix; constructing a flow anomaly propagation network to calculate an anomaly diffusion path; establishing a dynamic weight mechanism to determine key nodes and risk paths; generating a health degree index and constructing a multi-dimensional evaluation system; designing a self-adaptive intervention strategy library to match a guide and control instruction; and dynamically adjusting an execution path and resource allocation according to real-time data to realize flow closed-loop control. According to the method, the operation rule of the modeling process can be quantified, the abnormal propagation can be actively predicted, and the intelligent level of process control and the system robustness are improved.

The invention provides an industrial equipment preventive maintenance decision-making method based on a discrete time Markov chain, and relates to the technical field of industrial equipment reliability management and maintenance strategy optimization. The method comprises the following steps: based on historical maintenance record statistics, establishing a discrete time Markov chain maintenance effect state transition model and determining a state transition probability matrix thereof, carrying out eigenvalue decomposition on the state transition probability matrix to obtain state probability distribution after n times of maintenance, and determining each state probability; carrying out weighted fusion on the random intensity functions of the three maintenance processes through corresponding state probabilities, and constructing a mixed fault rate function; calculating an expected failure frequency based on the mixed failure rate function; and based on the expected failure times, establishing a long-term cost rate function, and solving to determine an optimal replacement period. According to the method, three basic maintenance effect types are integrated through the dynamic probability, and accurate prediction of the number of failure times of equipment and calculation of the optimal maintenance period are achieved.

This disclosure provides an autonomous decision-making method and apparatus for a nuclear reactor system based on interpretable artificial intelligence. The autonomous decision-making method for the nuclear reactor system includes: acquiring the state space, action space, and failure probability of the nuclear reactor system; inputting the state space, action space, and failure probability into a state transition model, and outputting the state transition probability of the components transitioning to the corresponding second state after taking different target actions in the first state; inputting the benefits and costs of the nuclear reactor system corresponding to the second state into a reward model, and outputting the reward function corresponding to the second state; aiming to maximize the power generation efficiency of the nuclear reactor system, solving for the optimal operating decision of the nuclear reactor system based on the state transition probability and the reward function; thereby tracing the reason for executing each action based on the relationship between the reward function and the state transition probability, so that operators can understand the reason for recommending the optimal operating decision before taking any action.

The invention discloses a turbo fsk receiving and processing method and system based on a vector DSP, and relates to the field of signal processing, and the method comprises the steps: carrying out the parallel calculation of a state transition probability through a vector after the fusion of channel information and external information, and efficiently generating a probability matrix through batch matrix multiplication and addition, data rearrangement and parallel comparison operation; then, forward and backward probabilities are calculated through vector recursion; and finally calculating the logarithmic posterior probability through vector addition and comparison. According to the invention, the architecture of the vector DSP is fully utilized, traditional serial calculation is converted into batch parallel processing, the calculation complexity and time complexity of a decoding algorithm are remarkably reduced, high-throughput and low-power-consumption Turbo-FSK signal decoding is realized, and the method is particularly suitable for communication equipment with strict requirements on power consumption and real-time performance, such as the Internet of Things.

The invention provides a hydraulic cooler dual protection linkage method and system based on oil temperature and oil pressure, and relates to the technical field of cooling control, and the method comprises the steps: building a running state parameter matrix and a system deviation matrix, building a state transition probability matrix and a stress time sequence matrix, calculating a temperature-pressure coupling stress value, and determining output cooling power according to the temperature-pressure coupling stress value. Accurate control of the hydraulic cooler is achieved. The temperature and pressure parameters of the hydraulic system are monitored and controlled cooperatively, the cooling efficiency is improved, the service life of the system is prolonged, and energy consumption is reduced.

The invention discloses a polymer formula analysis method and device based on a graph Q learning network, and the method comprises the steps: building a Markov decision process of formula generation through defining a state space, an action space, a state transition probability and a reward function; constructing a graph Q learning network model based on the Markov decision process; training the graph Q learning network model through a double-Q learning algorithm and a strategy function to obtain a trained graph Q learning network model; and performing state transition and action selection processing through the trained graph Q learning network model to generate a polymer formula. According to the method, the problems of low learning speed, high computing resource consumption, difficulty in accurately matching multi-target performance and the like in the prior art are solved.

The invention relates to an offshore wind turbine generator dynamic maintenance decision-making method considering state prediction uncertainty, and the method comprises the steps: obtaining the historical state transition prediction data of a plurality of prediction methods, and considering the local density information of a nearest neighbor set, based on a K-nearest neighbor method, basic probability distribution of historical state prediction results under multiple prediction methods of the unit is dynamically updated through dynamic Bayesian, and a unit component state transition probability interval is obtained through a combined reliability model of unit component state prediction based on a prediction result conflict degree; and constructing a stochastic programming model of a maintenance dynamic decision considering the uncertainty of component state prediction, and optimizing and solving to obtain a unit dynamic maintenance strategy by taking the minimum total maintenance cost as a target. Compared with the prior art, the method can reduce the influence of the uncertainty of state prediction on the maintenance decision, reduces the maintenance cost, and improves the maintenance efficiency.

The present application relates to the technical field of intelligent breeding, and discloses a feed dynamic management method, device and equipment for intelligent breeding and a storage medium, which constructs key time sequence data and performs preprocessing, calculates a heat stress strategy switching action index based on the time sequence data, constructs an input-driven three-state switching state space model based on the action feature sequence and an average feeding residual error sequence, determines an output hidden state estimate and a state probability, recursively generates a future feed demand prediction value and a prediction interval, performs online bias updating on a prediction error based on the action index, and outputs a corrected feed demand prediction. Thus, the present application introduces a strategy switching action index, simultaneously injects continuous hidden state evolution and discrete state transition probability calculation, effectively identifies action intensity from a small amount of key data, quickly completes prediction mechanism switching, simultaneously performs rapid correction, and realizes feed dynamic management for intelligent breeding in a special stress scenario.

The application discloses an electrical appliance state derivation method based on electrical change amount state conversion probability matching, combines a method based on a graph model and a method based on deep learning, derives the state change of the electrical appliance by the method based on deep learning, derives the specific state change condition of the electrical appliance by the method based on the graph model, and can better handle the derivation problem of the low-frequency electrical signal electrical appliance. Because the advantages of the two methods are combined, the required data amount is greatly reduced. In addition, the application adopts a low-frequency non-invasive signal processing technology based on 1-minute sampling once, does not need to make hardware transformation on the electric meter, greatly saves the cost, and has wide application and popularization prospects.

The application discloses a user interaction log-oriented internal and external noise filtering method, which firstly carries out screening and processing of an event log, and extracts useful column information; secondly, a general template in the form of a natural language text is automatically generated; then, the general log template and an operation line are combined, a Markov state transition graph of different operation types and target element combinations is constructed, and corresponding state transition probabilities are output; finally, internal and external noises are marked out through transition probabilities between nodes, and an event log is marked in detail by using a mean value and a total standard deviation variance adaptive threshold of a normal distribution, so that effective detection of internal and external noises in robot process automation is realized, and the accuracy of task segmentation, routine segmentation and process model generation is optimized.

The application discloses a resource aggregation willingness modeling and scheduling method based on a Markov jump system, and relates to the technical field of load scheduling.The application collects multiple basic preference parameters of each user in real time, constructs a response willingness function, maps the response willingness function to a discrete state set, establishes a state transition probability matrix family in time periods based on the discrete state set, and thus can accurately depict the complex behavior characteristics of the user based on the state transition probability matrix family, determines the available resource capacity of the system at each moment based on the state transition probability matrix family, and determines the scheduling decision at each moment according to the minimum solution of a comprehensive scheduling target function, so that the scheduling method can respond to the uncertainty of the user in real time.

The invention provides a malicious software classification method and system based on feature space transformation and a storage medium, and relates to the technical field of malicious software classification.The method comprises the steps that a binary byte sequence of to-be-classified malicious software is obtained, the binary byte sequence is modeled into a Markov chain, and the state of the Markov chain corresponds to the value of a binary byte; constructing a state transition probability matrix based on the transition relation of each state in the Markov chain, wherein elements in the state transition probability matrix are used for representing the transition probability between bytes; applying a mapped feature space transform to the state transition probability matrix to generate an enhanced feature map, the feature space transform using a power law function to non-linearly map each element in the matrix; and inputting the enhanced feature map into a pre-trained convolutional neural network model for feature extraction and classification to obtain a family classification result of the to-be-classified malicious software.

The invention discloses an abnormal data detection and analysis method for an electric power artificial intelligence platform, and relates to the field of electric power system abnormal data analysis, and the method comprises the steps: clustering K operation state clusters, constructing N * K state transition probability matrixes under K operation scenes, constructing real-time operation state vectors of N real-time parameters, determining the most similar state cluster in the current operation scene in the K operation state clusters, anchoring a state transition probability matrix corresponding to the target real-time parameter, constructing M-1 real-time state transition events of the target real-time parameter in the M current time steps, and determining the state transition probability matrix corresponding to the most similar state cluster in the state transition probability matrix corresponding to the M-1 real-time state transition events. According to the method, the state change deviating from a conventional evolution path in a specific operation scene can be identified, and the detection and identification capability of abnormal operation state behaviors in a complex operation scene is improved.

The invention provides a behavior control method and device. The method comprises the steps of determining a current behavior state of a virtual object and a state transition matrix of the virtual object in a current time period; determining first state transition probability distribution corresponding to the current behavior state from a state transition matrix; in response to the first state transition probability distribution, the behavior state of the virtual object is transferred from the current behavior state to a first target behavior state in a behavior state set, and a first behavior event queue is generated based on the Poisson rate corresponding to the first target behavior state; and controlling the virtual object to sequentially execute the behavior events in the first behavior event queue. By means of the method and device, accurate control over the virtual object behavior can be achieved, and therefore the fidelity of the virtual object behavior is improved.

The invention discloses a double basic service set system throughput evaluation method based on a Markov chain model, and the method comprises the steps: representing a state in a double basic service set system as a state in a Markov chain, defining the transition probability between the states, and obtaining a state transition probability matrix of the Markov chain; obtaining the probability distribution of each state when the system is in a steady state by solving the feature vector of the state transition probability matrix; and calculating the total throughput of the system according to the average throughput and the steady-state probability of the system under different competition window combinations. According to the method, the steady-state probability of the system under different competition window combinations is calculated by establishing the Markov chain model, so that the average throughput and the total throughput of the system can be estimated, and theoretical support is provided for throughput evaluation and system performance optimization of the WLAN system.

The invention discloses a power system operation state evolution analysis method, device and equipment and a medium, and the method comprises the steps: obtaining the output data of various generator sets in a plurality of geographic regions and the power load data corresponding to each geographic region, so as to generate a power system operation time sequence; outputting a state transition probability matrix through a pre-trained hidden Markov model according to the operation time sequence of the power system; wherein the hidden state number of the hidden Markov model is determined according to an akaike information criterion and a Bayesian information criterion so as to ensure the optimality of the model; the state transition probability matrix is used for identifying an implicit operation state in the operation process of the power system; and acquiring real-time operation data of a target power system, and predicting an operation state corresponding to a preset future time sequence according to the real-time operation data and the state transition probability matrix. Compared with the prior art, the dynamic property and the accuracy of the evolution analysis of the operating state of the power system can be improved.

The invention discloses a diffusion reinforcement learning channel access method based on entropy driving step length self-adaption, which comprises the following steps of: establishing a channel state transition probability matrix based on a channel state between an air base station and a ground edge aggregation server; grouping the channels, and randomly selecting a channel group for activation from the channel groups corresponding to the current macroscopic environment state according to preset weight probability distribution; establishing a channel selection strategy and a target function; an LADSAC model is deployed on an air base station side and is in butt joint with an existing ARQ protocol stack, observation collection, environment perception, step number decision making, diffusion denoising, action issuing, result receipt and learning updating are sequentially carried out on the LADSAC model in each time slot, and training and updating of the LADSAC model are completed; according to the method, the problem that decision performance and calculation efficiency are difficult to consider in channel access in a non-stationary emergency communication scene is solved, and reasoning time delay and energy consumption are reduced while decision quality is kept.

An estimation device and the like are provided that can output state transition probabilities that take into account the tendency of data change in data transitions between distributions. [Solution] An estimation device according to one aspect of the present disclosure includes a receiving means for receiving input of a data set for each stage, an estimation means for estimating a state transition probability based on a transition from a first distribution, which is the distribution of data in a first stage, to a second distribution, which is the distribution of data in a second stage, based on a cost required for the transition to the destination, which is determined for each data in the first distribution, and an output means for outputting the estimated state transition probability.

The application provides a malware classification method and system based on feature space transformation and a storage medium, relates to the technical field of malware classification, and comprises the following steps: obtaining a binary byte sequence of malware to be classified, modeling the binary byte sequence as a Markov chain, and taking the value of a binary byte as the state of the Markov chain; constructing a state transition probability matrix based on the transition relationship of each state in the Markov chain, wherein the elements in the state transition probability matrix are used to represent the transition probability between bytes; applying a mapped feature space transformation to the state transition probability matrix to generate an enhanced feature map, wherein the feature space transformation adopts a power-law function to perform nonlinear mapping on each element in the matrix; and inputting the enhanced feature map into a pre-trained convolutional neural network model for feature extraction and classification to obtain the family classification result of the malware to be classified.

The invention belongs to the field of discrete event system control, intelligent manufacturing and reinforcement learning, and discloses a state feedback control strategy generation method and system of a time delay Petri network system in a non-deterministic environment, and the method comprises the steps: constructing a time delay Petri network model containing controllable / uncontrollable transition and minimum transmission delay constraint, defining a time extension state formed by the identification and the enable transition residual delay; a two-stage random judgment mechanism is introduced, two types of non-deterministic factors of control execution errors and uncontrollable event preemption are modeled, and the state transition probability is deduced; a control problem is formalized into a Markov decision process, and a reward function with time consumption as a negative reward and deadlock and an ultra-long path as strong punishment is designed; and adopting a table type Q learning algorithm based on dynamic learning rate scheduling, iteratively updating an action value function in interaction with a simulation environment, and exporting a state feedback control strategy. The method can be widely deployed in discrete event system control scenes such as flexible manufacturing, logistics scheduling and multi-robot cooperation.

To adequately capture a probabilistic characteristic of a process behavior with a minimum requirement of an amount of information, and generate a learning model which is highly accurate, safe, and suitable for controlling a target process without instability due to missing data.SOLUTION: A learning model generation device that combines plural signals to generate a state transition probability matrix for a defined multidimensional state includes an arithmetic processing unit for: setting an initial value for a lower limit value, an upper limit value, and a division number for each dimension of the multidimensional state; calculating a learning data sufficiency rate, which is a total value of a transition probability to a state near a distribution center of a transition destination, and is defined as one or more pre-transition states within a predetermined distance from a target pre-transition state or a transition destination state with a maximum transition probability; and decreasing the division number when the calculated learning data sufficiency rate is below a defined range, and increasing the division number when the calculated learning data sufficiency rate exceeds the defined range.SELECTED DRAWING: Figure 2

The application discloses a timing failure system periodic inspection maintenance method based on a state transition Monte Carlo method. According to a state transition process of the timing failure system, a state set and a decision set of the system are given; a calculation method of state transition probability of different actions in a periodic inspection interval is proposed; based on the state transition Monte Carlo method, each maintenance strategy is simulated to obtain an average maintenance cost rate and an average safety level of the system in a life cycle under a specified maintenance strategy; and according to the average safety level requirement, an optimal maintenance strategy of the timing failure system is determined. The application can be applied to the maintenance strategy formulation of the timing failure system in a civil aircraft, and has important theoretical significance and application value for improving the safety and maintenance efficiency of the timing failure system.