7 results about "Forward algorithm" patented technology

The invention discloses a household energy centralized management method and device, electronic equipment and a storage medium. The method comprises the following steps: collecting real-time observation data of a family environment to construct a multi-dimensional context feature vector; generating a dynamic state transition probability matrix and emission parameters of the hidden Markov model through a dynamic parameter mapping mechanism, and calculating initial posterior probability distribution of a family activity state by using a forward algorithm; when the information entropy of the probability distribution exceeds a threshold value, executing an active state exploration strategy, issuing an optimal perturbation instruction and correcting a posterior probability based on an active response observation evidence so as to determine a final family activity state; and based on the state loading control constraint set, solving a real-time optimization objective function containing energy efficiency and comfort penalty, and generating an equipment control instruction. According to the method, the uncertainty of state judgment is reduced by combining a dynamic model and an active exploration mechanism, and household energy cost optimization is realized on the premise of ensuring comfort.

The present application relates to the technical field of psychophysiological monitoring, and discloses a video psychophysiological index analysis method and system; the method comprises the following steps: positioning facial key points and partitioning from a video, extracting remote optoelectronic plethysmography pulse wave and action unit breathing characteristics, outputting heart rate and breathing signals through Kalman filtering; obtaining a heartbeat interval through peak value detection, calculating heart rate variability indicators and breathing frequency, and normalizing by establishing an individual baseline in an initial calm period; fusing emotional characteristics and normalized indicators to form an observation vector, inferring a stress state through an unsupervised model and a forward algorithm; detecting index mutation points to identify stimulus events, and adopting a gated recurrent unit network to predict excessive stress risk and grade early warning; storing multi-session data into a database, quantifying desensitization rate, tolerance and recovery capacity through linear regression, weighted scoring and model updating. The present application realizes contactless monitoring, real-time stress evaluation and early warning, and supports quantitative analysis of cross-training effects.

This invention discloses a method and apparatus for heart sound segmentation based on a Hidden Markov Model (HMM) cardiac cycle. The method involves: extracting the signal envelope from the heart sound signal using Hilbert's algorithm; obtaining the peak positions of the first heart sound (S1) and the second heart sound (S2) after Hilbert envelope extraction based on preset baselines for different leads; locating the peaks based on the cardiac cycle; extracting the diastolic phase based on an improved HMM; locating the optimal time span of heart sound signals S1 and S2 using an improved Viterbi algorithm; and segmenting the original heart sound signal using the time span and the peak positions of S1 and S2. This invention employs an improved HMM and an improved Viterbi forward algorithm to calculate the duration of heart sound intervals S1 and S2, combined with the cardiac cycle, to accurately locate the peak positions of heart sound intervals, thus improving the performance of heart sound segmentation.

The application provides a kind of based on attack chain modeling and serialization inference's exercise behavior detection method and system, this method includes: attack chain knowledge graph includes tactical node, technical node, the tactical node is used to describe tactical phase or tactical state, the technical node is used to describe the specific method or means for realizing a certain tactic;Attack chain knowledge graph is modeled, and the state transition probability matrix between different tactical nodes and the observation probability matrix observed to a certain event under a certain tactical node are obtained;The global confidence of current system under continuous time sequence in a certain tactical phase is dynamically inferred and calculated using HMM forward algorithm;The hierarchical risk value is determined according to the global confidence of current system under continuous time sequence in a certain tactical phase, attack chain integrity, asset risk level, and the corresponding response strategy is determined according to hierarchical risk value, effectively improve the reliability of network attack and defense exercise behavior detection.

The application belongs to the technical field of depth inversion, and provides a Rayleigh wave dispersion inversion method and system fusing physical information, which adaptively generates a multi-type layered geological model based on dynamic Markov decision according to the conditional probability relationship of the previous layer parameter; a fast vectorization Rayleigh wave forward algorithm is used to calculate the base order and first-order dispersion curve corresponding to the geological model, and construct dispersion sample data; a dispersion curve joint inversion model based on deep learning is constructed, the base order dispersion curve, the first-order dispersion curve and the corresponding mask information are taken as inputs, a loss function with physical information constraint is introduced to train the dispersion curve joint inversion model; the trained dispersion curve joint inversion model is used to process target geological data, and the layered transverse wave velocity and layer thickness parameters of the underground medium are obtained, and Rayleigh wave dispersion curve inversion is realized; the application can enhance the constraint on the real stratum relationship, and realize efficient and accurate stratum velocity and layer thickness inversion.

The application discloses a service state prediction method and device, electronic equipment and storage medium, and the system comprises: acquiring log historical data set and initial observation probability of each log of a system; constructing an initial hidden Markov model; calculating observation anomaly probability of any log observation sequence according to a forward algorithm module, initial observation probability and model initial parameters, and obtaining anomaly probability distribution result of the log; performing weighted fusion on the anomaly probability distribution result, determining service state sequence length of the system according to the weighted anomaly probability distribution result; acquiring initial service state of the system; performing backtracking analysis on the initial service state according to a Viterbi algorithm module, model initial parameters and service state sequence length, and obtaining target service state sequence of the system; and through a target hidden Markov model, fault recovery and intervention are advanced from post-repair to in-process or even pre-process, so that the maintenance capability and business continuity guarantee capability of the system are improved.

The invention relates to the technical field of psychological and physiological monitoring, and discloses a video type psychological and physiological index analysis method and system. The method comprises the following steps of: positioning facial key points from a video, partitioning, extracting remote photoelectric volume pulse waves and breathing characteristics of an action unit, and outputting heart rate and breathing signals through Kalman filtering; obtaining a heartbeat interval through peak detection, calculating a heart rate variability index and a respiratory rate, and establishing an individual baseline for normalization by using an initial calm period; fusing the emotional features and the normalized index to form an observation vector, and deducing a stress state through an unsupervised model and a forward algorithm; detecting an index mutation point to identify a stimulation event, and predicting an excessive stress risk and performing graded early warning by adopting a gating circulation unit network; and storing the multi-session data into a database, quantifying the desensitization rate, tolerance and recovery capability through linear regression, performing weighted scoring, and updating the model. According to the invention, non-contact monitoring and real-time stress evaluation and early warning are realized, and quantitative analysis of a cross-training effect is supported.