25 results about "Probability matching" patented technology

A method of determining whether a first record and a second record are a match may include performing probabilistic matching, including assigning weights to record attributes to create weighted attributes and computing a probabilistic matching score using the weighted attributes, and performing rule based deterministic matching. The method may also include, returning a result that indicates a match based on the probabilistic matching score and the rule based deterministic matching; performing a modeled analysis of the first record and the second record based on a combined result of the rule based deterministic matching and the probabilistic matching score; returning a result that indicates a match based on a determination, via the modeled analysis, that the first record and the second record are a match; and returning a result that indicates that manual review is needed based on an inconclusive result via the modeled analysis.

The present disclosure provides a resource recommendation method, a recommendation prediction model training method, an apparatus and a device, and belongs to the technical field of multimedia. The resource recommendation method comprises: obtaining historical recommendation information and historical search information of an object; obtaining recommendation interest features and search interest features of the object based on a resource to be recommended, the historical recommendation information and the historical search information; performing dimension reduction processing on object features of the object, resource features of the resource, the recommendation interest features and the search interest features to obtain a recommendation probability of the resource; and recommending a target resource in a plurality of resources to the object based on the recommendation probability of the plurality of resources to be recommended. In the scheme provided by the present disclosure, it is ensured that the indication of the recommendation probability matches the interest of the object, the accuracy of the recommendation probability is ensured, and then the resource recommended to the object is the resource of interest to the object, so as to ensure the accuracy of the resource recommendation and then ensure the resource recommendation effect.

A method of determining whether a first record and a second record are a match may include performing probabilistic matching, including assigning weights to record attributes to create weighted attributes and computing a probabilistic matching score using the weighted attributes, and performing rule based deterministic matching. The method may also include returning a result that indicates a match based on the probabilistic matching score and the rule based deterministic matching; performing a modeled analysis of the first record and the second record based on a combined result of the rule based deterministic matching and the probabilistic matching score; returning a result that indicates a match based on a determination, via the modeled analysis, that the first record and the second record are a match; and returning a result that indicates that manual review is needed based on an inconclusive result via the modeled analysis.

The application discloses a data processing method and device, equipment and a storage medium, and belongs to the technical field of artificial intelligence. The method comprises the following steps: obtaining a noise measurement index of a first candidate task based on an initial neural network model; obtaining a sampling probability of the first candidate task based on the noise measurement index of the first candidate task; selecting a reference task meeting a first selection condition from the first candidate task according to the sampling probability; updating the parameters of the initial neural network model based on the reference task; learning a target task based on the neural network model after the parameter update, obtaining a target neural network model, and using the target neural network model for data processing. The sampling probability obtained by considering the noise measurement index is a sampling probability matched with the interference degree of the candidate task on the parameter update of the model. The reliability of the reference task selected from the candidate task based on the sampling probability is high, which is beneficial to improving the update quality of the parameters of the model, thereby improving the accuracy of data processing.

The invention discloses a rainfall prediction method based on dynamic downscaling, which relates to the field of meteorological numerical weather forecast, and comprises the following steps: acquiring an initial field of regional forecast, accessing real-time rainfall data, radar reflectivity data and digital elevation model data of a ground meteorological station, and performing quality control and spatial interpolation processing to obtain a rainfall prediction result; generating an initial field with continuous space; based on the spatial continuous initial field, obtaining cloud-water conversion by adopting fractional calculus, and dynamically adjusting the cloud drop collision efficiency according to the terrain gradient field to obtain a cloud microphysical variable; and based on the cloud micro-physical variables, performing regional space encryption calculation by using a WRF model power framework. According to the method, a power frame of a triple nested grid and a Thompson / WSM6 / Morrison multi-scheme probability matching fusion technology are adopted, so that the rainfall forecast error of a complex terrain area is reduced, the strong rainfall forecast accuracy is improved in a mountain area with the gradient larger than 15 degrees, and an innovative technical system from initial field downscaling processing to final high-precision rainfall forecast generation is achieved.

A maneuver detection method based on model probability matching, the method comprising: analyzing different characteristic parameters when a hypersonic reentry vehicle has a maneuver mutation, constructing a maneuver model description matrix, and determining a maneuver model; defining an influence factor and designing a model matching degree function; setting thresholds for the characteristic parameters of different maneuver models, and calculating a matching vector between the current system and each preset model based on the actual characteristic parameters of the current system; substituting the prior influence factor and the calculated matching vector into the model matching degree function to deduce the model matching probability; calculating a model probability correction sequence according to the probabilities of different preset maneuver models corresponding to the true state at the current moment, and finally correcting the model probability vector; and judging whether the preset model switching condition is satisfied according to the model probability vector. According to the present invention, different maneuver models are set according to the characteristics of different maneuver mutations, adapting to the characteristics of the hypersonic reentry vehicle having extremely strong unknown maneuver characteristics.

The invention discloses a sales agent voice conversation method. The method comprises the following steps: firstly, carrying out time sequence context frame association analysis and emotion dynamic spectrum quantitative modeling on a real-time voice data stream to generate a user demand dynamic feature conformation; then, through hierarchical denoising reconstruction and intention posterior probability matching, a core demand feature base vector is determined, and a scenarized demand positioning feature topology is generated; calling the commodity attribute knowledge graph to execute weighted association mapping and recommendation semantic logic generation to obtain a customized recommendation response text; and finally, fusing context and emotion features to execute text-to-speech conversion and emotion speech adaptation modulation, and generating a real-time response speech stream. According to the invention, through context-emotion two-dimensional cooperative processing, accurate capture of user demands and emotion adaptive response are realized, and the demand identification accuracy and the user satisfaction are improved.

The application provides a structure probability risk assessment method and system based on SRSMBRA and Copula-MCM, and the method comprises the following steps: improving the probability matching point of a classical random response surface method, obtaining undetermined coefficients of a Hermite expansion by a random response surface method SRSMBRA based on regression analysis, taking the probability matching point method as a breakthrough point, and constructing a limit state equation of a structure; based on the limit state equation of the structure, constructing a joint probability density function of multi-dimensional random variables of the structure by means of a Copula function and a Vine model, obtaining a failure probability when there is a correlation between the multi-dimensional random variables of the structure based on a structure probability risk assessment method of Copula-MCM; and integrating the structure probability risk assessment of SRSMBRA and Copula-MCM by means of Rosenblatt transformation, obtaining the failure probability by a Copula-SRSMBRA risk assessment method by using a moment method or a Monte Carlo method, and performing efficient and feasible structure probability risk assessment of an engine life-limited part by the above method.

The application discloses a vehicle trajectory map matching method based on a hidden Markov model, which comprises the following steps: (1) preprocessing a trajectory T to form a trajectory T', and extracting each trajectory point from the trajectory T' to obtain a candidate point of each trajectory point; (2) matching k optimal paths of the trajectory T' based on a hidden Markov probability matching model; and (3) calculating path scores of the corresponding paths based on a steering cost of the k optimal paths, and taking a path with the highest path score as an optimal matching path of the trajectory T. The driver tends to select a path with less steering points, which means that they are more likely to follow a straight path or a road with less turns. When map matching, the path preference of the driver is considered, so that the matching result is more consistent with the actual road driving, the quality of path reconstruction is improved, and the accuracy of the matching result is improved.

The application provides an adaptive abnormality identification method and system for a tunneling machine operating condition, and relates to the technical field of mining engineering tunneling. The method comprises the following steps: when a lithology mutation event is captured, a plurality of associated devices are driven to perform condition collection based on a preset cooperative sampling rule, the lithology mutation event is matched and traced back in time sequence, a real-time lithology transition stage is dynamically called to determine a device associated judgment model, a correlation number matrix is calculated, and is loaded into the device associated judgment model to make a space-time associated coupling decision; when the decision result is associated coupling failure, a failure characteristic vector is analyzed to perform fault direction probability matching, and a real-time fault device group is output. The technical problems that the prior art has single or fixed parameter monitoring of the tunneling machine state, resulting in insufficient timeliness and accuracy of abnormality identification are solved. The technical effects of improving the comprehensiveness, real-time performance and accuracy of tunneling machine abnormality identification, and ensuring efficient and safe operation of the tunneling machine are achieved.

The invention relates to the technical field of lossless video coding, and provides a lossless compression method for eliminating inter-frame redundancy based on partitioning and probability matching, which comprises the following steps of: dividing a current frame into a plurality of sub-blocks, and classifying each sub-block by calculating the similarity between a current block and a matching block; the classified sub-blocks are subjected to matched coding operation through a differentiation processing strategy, and compressed data are generated; and carrying out entropy coding on the compressed data, and outputting a lossless compressed code stream. The video frame is divided into a plurality of sub-blocks through the technical means of combining block processing, probability matching, multi-level classification and reversible integer predictive coding, the sub-blocks are classified according to the inter-frame similarity, differentiated processing strategies are adopted for different types, and finally residual and motion information is compressed through entropy coding. The inter-frame redundancy is effectively eliminated, the compression efficiency is remarkably improved, and meanwhile, the robustness of fast motion and scene switching is kept.

The invention discloses a task-oriented equipment support resource regional configuration model establishment method, and aims to solve the problem that a support resource configuration model is insufficient in generality in the aspects of an open environment, a combined target and hierarchical constraints in the task execution process of regional equipment. In order to solve the problems of the prior art, such as the lack of a support strategy and a configuration scheme on system analysis of an influence mechanism level, on the basis of the definition of a node set of a support object and a support mechanism, the number matching of node resources is converted into probability matching, an equipment support resource region configuration model is established based on an optimal transmission theory, and a basic algorithm for model solution is given. Case analysis shows that the configuration model provided by the invention is better in performance in intermittent demand, and shows a local zigzag structure for different resource types.

The invention discloses a visual multi-target tracking method based on a panoramic image, and belongs to the technical field of computer vision. The method comprises the following steps: inputting a panoramic RGB image into a feature extraction network, extracting image features and generating a learnable instance, and obtaining a bounding box and attribute information of a target through a decoding network; constructing a trajectory instance based on a historical trajectory to extract prediction information of a current frame, and in combination with a trajectory pool management strategy, realizing initialization, updating or deletion of the trajectory through confidence threshold judgment; designing a distance calculation and probability matching algorithm, fusing observation data of a learnable instance and a track instance, and optimizing a joint probability model associated with a target; an association mode is dynamically selected according to trajectory stability and environmental noise, and continuous online tracking and unique number management of multiple targets are ensured to be realized in a full scene. According to the method, the precision and robustness of multi-target tracking in a complex scene are effectively improved through high coverage of the panoramic image and an efficient trajectory association mechanism.

The invention provides a self-adaptive anomaly identification method and system for the operation condition of a heading machine, and relates to the technical field of mining engineering tunneling, and the method comprises the steps: driving a plurality of associated devices to execute working condition collection based on a preset cooperative sampling rule when a lithology sudden change event is captured, carrying out the time sequence feature matching backtracking of the lithology sudden change event, and obtaining a lithology sudden change event; and dynamically calling an equipment association judgment model according to a real-time lithologic transition stage, calculating a correlation coefficient matrix, loading the correlation coefficient matrix to the equipment association judgment model to carry out space-time association coupling judgment, and when a judgment result is association coupling failure, analyzing a failure feature vector to carry out fault pointing probability matching, and outputting a real-time fault equipment group. The technical problem of insufficient timeliness and accuracy of abnormal recognition caused by monitoring the state of the heading machine through single or fixed parameters in the prior art is solved. The technical effects of improving the comprehensiveness, real-time performance and accuracy of abnormal recognition of the heading machine and guaranteeing efficient and safe operation of the heading machine are achieved.

Embodiments of the present disclosure disclose a device association method and device based on behavior blocking and probability matching. A specific implementation of the method comprises: obtaining a set of device behavior records; combining each device behavior record in the set of device behavior records that meets a preset similarity condition into a similar device record group to obtain a set of similar device record groups; for each similar device record group in the set of similar device record groups, determining a matching probability value between each two similar device records in the similar device record group to obtain a group of matching probability values; generating a device behavior association graph according to each group of matching probability values obtained; identifying connected nodes of the device behavior association graph to obtain a set of connected node groups; and for each connected node group in the set of connected node groups, assigning each connected node in the connected node group with a same identifier to complete device association. The implementation can improve the accuracy of device association.

The present invention discloses an abnormal event early warning method and system for multimodal perception of a traffic road network, which relates to the technical field of data processing. The method includes: obtaining a set of frequently-occurring path accident data of a target traffic road network, constructing a set of abnormal event prototypes, and obtaining vehicle status information of a target vehicle based on V2X communication. Probability matching is performed in combination with multimodal perception data to determine a target abnormal event prototype; a set of historical abnormal event triggering features is called to perform feature enhancement on it, and the enhanced abnormal event triggering features are used as early warning conditions to perform real-time abnormal event early warning on the driving process of the target vehicle. The present invention solves the technical problems of the prior art that rely on a single data source and lack real-time dynamic response capabilities, resulting in incomplete detection of traffic abnormal events and early warning delays, and achieves the technical effect of realizing comprehensive and real-time detection and dynamic early warning of traffic abnormal events through multimodal perception and V2X communication technologies.

The present invention proposes a method, system, device and medium for quantitative precipitation forecasting for an extended period, and relates to the field of weather forecasting. The method comprises: based on the current forecast time validity of a certain grid point, taking the ensemble average of the real-time forecast as the object, in the re-forecast data sample library with the same forecast time validity and subjected to ensemble average processing, sorting according to the similarity between the ensemble average processed re-forecast data and the real-time forecast, taking the first predefined number of re-forecast data in the sorting as the model forecast similarity sample set; based on the model forecast similarity sample set, corresponding to the QPE data, obtaining the historical actual precipitation similarity set; based on the precipitation similarity set, using the probability matching average technology and the bilinear interpolation technology, calculating the extended period precipitation grid forecast product for a predefined number of days. The present invention conducts in-depth research on the problems of long extended period forecast time validity, small number of samples required for the same period, small number of extended period models, and large forecast uncertainty, and establishes a new generation of extended period precipitation forecast model technology.

The invention provides an intelligent design optimization system and method for a pure methanol power collecting and distributing dual-purpose ship, and relates to the technical field of ship design, and the system comprises a probability matching module which carries out the probability matching in a ship design prototype library; the structure optimization module is used for performing fine-grained structure optimization on the initial ship design prototype; the effect verification module is used for verifying the fine-grained structure layout scheme; and the interference analysis module is used for extracting a signal equipment-layout position set to perform signal interference simulation analysis when the design effect verification result is passed, adding a local signal conditioning circuit, and integrating an optimized signal equipment-layout position set and a fine-grained structure layout scheme into an intelligent design optimization scheme. By means of the method and device, the technical problem that in the prior art, due to the fact that different structures of ship design are designed independently, potential conflict problems are found in the later period, and the ship design efficiency is further affected can be solved, potential design conflicts are found and resolved in the early period, and the ship design efficiency is improved.

A computer-implemented method for improving entity matching in a probabilistic matching engine may train a graph neural network (GNN) model according to an output of the probabilistic matching engine to perform entity matching and determine an anti-fact interpretation for non-matching of entities. A data transformation list may be identified by operable traces using a GNN model. The data transformation lists may be ranked based on computational overhead and an improvement in entity matching within an estimated probability matching engine using a GNN model.

The invention relates to the technical field of hydrology and water resource prediction, in particular to a cross-border river short and medium hydrology meteorological numerical forecasting method and system based on an ensemble forecasting system and multi-model collaboration and a medium. Multi-mode weighted fusion is realized by adopting probability matching correction and combining with improved geometric averaging, and regression-based threshold probability quantitative rainfall forecast is constructed; making a set average and set probability product according to the basin boundary, performing precision verification and subjective correction on the result, and generating a basin precipitation distribution diagram; the corrected rainfall is input into a subarea hydrological model, a three-water-source Xinanjiang model is used for calculating runoff and confluence, a Muskinggen method is adopted for river channel confluence, reservoir flood regulation is completed in combination with static reservoir capacity flood regulation calculation, and the runoff process, the flood peak and the occurrence time are obtained; and selecting a typical flood process to carry out inspection.

The application discloses a kind of visual multi-target tracking methods based on panoramic image, belong to computer vision technical field.The method is by inputting panoramic RGB image to feature extraction network, extracts image feature and generates learnable instance, and obtains the boundary box and attribute information of target by decoding network;Based on historical trajectory, trajectory instance is constructed to extract the prediction information of current frame, in combination with trajectory pool management strategy, the initialization, update or deletion of trajectory is realized by confidence threshold judgment;Design distance calculation and probability matching algorithm, the observation data of learnable instance and trajectory instance are fused, and the joint probability model of target association is optimized;According to the stability of trajectory and environmental noise, dynamically select association mode, ensure that continuous online tracking and unique numbering management of multi-target are realized in full scene.The method improves the precision and robustness of multi-target tracking in complex scene by the high coverage of panoramic image and efficient trajectory association mechanism.

The present application relates to the technical field of precipitation forecast, in particular to a kind of hourly time lag ensemble precipitation forecast correction method based on weight coefficient, it is based on the time lag ensemble forecast method of Taylor test result and the forecast-observation probability matching method based on GAMMA cumulative probability analysis, and the fine gridding display of precipitation correction product is realized, aims at reducing precipitation forecast error, improve hourly precipitation forecast and 24 hours precipitation forecast accuracy, provide scientific reference for hourly precipitation, especially short-time heavy precipitation forecast;It includes the following steps: using new type of ensemble weight coefficient Wi based on Taylor test result is developed;Using the time lag ensemble forecast method based on new type of weight coefficient, the forecast member of different time starting is collected;Using the forecast-observation probability matching method based on GAMMA cumulative probability analysis, corrects hourly precipitation change and 24 hours precipitation order of magnitude;Precipitation forecast test;Its beneficial effects are: the present application improves hourly, 24 hours sunny and heavy rain forecast capability.

The present application belongs to the technical field of microseismic source positioning, and particularly relates to a microseismic source positioning method without accurate time picking, which comprises the following steps: obtaining microseismic monitoring data of a plurality of sensors arranged in a monitoring area, extracting a microseismic wave arrival time segment of each sensor, and calculating an observed time difference interval of a sensor pair. Then, a theoretical time difference calculation model of the sensor pair is constructed based on a source position coordinate and a wave velocity. A time uncertainty parameter is introduced to establish a probability matching relationship between the theoretical time difference and the observed time difference interval, and a likelihood function is constructed. Finally, the source position and the wave velocity that maximize the likelihood function in a search space are searched as the final positioning result. The present application replaces the traditional accurate time point, relaxes the requirement for microseismic data picking accuracy, and effectively reduces the picking error caused by wave field complexity, equipment error and artificial factors.

The present invention relates to a sparsity-aware neural processing unit that performs index matching of a constant probability regardless of IA and weighted density, and a processing method thereof. According to an embodiment of the present invention, the sparsity-aware neural processing unit processing method may include: receiving a plurality of input activations (IA); obtaining weights with non-zero values from each weighted output channel; obtaining an input channel index, wherein the input channel index stores the weights and IA in a memory and includes a memory address location where the weights and IA are stored; and in an index matching unit (IMU) including a buffer memory storing the input channel index, arranging the weights with non-zero values of each weighted output channel according to the row size of the IMU and matching the weights and IA.

The application provides a data processing method for improving the efficiency of answering questions. The method comprises: obtaining a first input sent by a user; determining a plurality of target drafts matching the first input from a draft library according to the first input, the target drafts comprising at least one draft token; calling a pre-trained model to obtain a matching probability of each draft token in each target draft, the matching probability being a probability of the pre-trained model outputting the draft token when the first input is input; determining a verified draft token, the verified draft token being a draft token with a matching probability greater than a target threshold; and determining an available target draft in the plurality of target drafts, the verified draft token in the target available draft being used to generate an answer to the first input. The application also provides corresponding devices, computing device clusters, computer-readable storage media, and computer program products.