81 results about "Complexity index" patented technology

The invention relates to the technical field of platinum smelting, and discloses a platinum smelting impurity prediction and process optimization system based on deep learning. A feature extraction module of the system extracts a multi-scale smelting feature component through a deep convolutional neural network and calculates a complexity index; the key feature determination module identifies key smelting features most related to impurity sources; the optimization parameter generation module is used for generating personalized optimal regulation and control parameters in combination with historical smelting baseline characteristics, smelting state characteristics and complexity indexes; the sequence reconstruction module reconstructs the feature sequence and decomposes the feature sequence into a slow-varying component reflecting the steady-state characteristic of the impurity and a fast-varying component reflecting the dynamic change of the impurity; the impurity prediction module determines impurity types and generates a mapping relation between the impurity types and smelting process parameters; and the process optimization execution module generates a control instruction based on the mapping relation to realize smelting process optimization. According to the system, through cooperation of multiple modules, precise prediction of platinum smelting impurities and dynamic optimization of the process are achieved.

The invention discloses a method and device for determining the number of test cases and electronic equipment. The method comprises the following steps: generating an initial orthogonal table of a target software system; determining a target service operated by the target software system, obtaining a complexity index variable, and determining a service influence index variable of the target service; determining a complexity factor value according to the complexity index variable and the service influence index variable; acquiring historical operation information of the target software system, acquiring an error rate index variable from the historical operation information and the software code, and determining an error rate factor value of the target software system according to the error rate index variable; and adjusting the preset number according to the complexity factor value and the error rate factor value to obtain a target number of the test cases of the target software system. By means of the method and device, the problem that in the related technology, due to the fact that the accuracy of determining the number of supplementary cases of the orthogonal table is low, the accuracy of software testing according to the orthogonal table is low is solved.

The invention relates to the technical field of leakage detection, in particular to a drainage pipe network leakage area AI identification method and system, and the method comprises the following steps: collecting sound pressure changes through an acoustic collection device, extracting candidate leakage sections, collecting images through an inspection robot, screening training samples, analyzing brightness and profile changes, adjusting the training weight, and identifying the leakage position. And calculating a boundary coincidence area ratio, and adjusting model parameters in combination with a pipe section material. According to the method, leakage candidate sections are compared and screened based on continuous fluctuation of a sound pressure multi-frequency channel, the problem of frequent occurrence of misrecognition under noise interference is avoided, a complexity index is constructed by combining a texture direction and a span in an image, and sample screening is performed; the gray-scale curved surface constructed in the non-abnormal area in the image is compared with the profile trend of the abnormal area to identify the abnormal structure, the extraction capability of irregular leakage features is enhanced, the leakage boundary and material coding data are superposed to evaluate the identification interference degree of the mixed section, and the adaptability and stability of model parameters under different materials are guaranteed.

The invention relates to the technical field of point cloud data processing, and provides a ground filtering method and system based on cloth simulation parameter dynamic optimization, and the method comprises the following steps: obtaining point cloud data of a to-be-processed region, carrying out the preprocessing, constructing a complexity index, and calculating a complexity index value based on the preprocessed data; clustering the obtained complexity indexes by adopting a clustering method fusing an elbow rule and a spatial neighborhood constraint to obtain a terrain complexity category; and selecting distribution parameters of a distribution algorithm based on the terrain complexity category and the dynamic mapping of the constructed complexity category-optimal parameter group, and further performing ground point filtering and ground extraction by adopting the distribution algorithm. According to the invention, distribution simulation parameter configuration is guided through terrain complexity analysis, accurate adaptation of different terrain areas is realized, and the accuracy and global consistency of ground point extraction are improved.

The invention relates to a deep learning cloud particle target detection method based on a dynamic detection head. According to the method, firstly, complexity evaluation is carried out on a cloud particle image data set, the weighted information entropy, the weighted texture complexity and the weighted local standard deviation of each image are calculated, and the comprehensive complexity is calculated based on the complexity indexes. Then, according to the overall complexity grade of the data set and the image data complexity, a detection head selection standard is formulated, and a proper dynamic detection head combination is selected according to the standard; and finally, performing target detection on the cloud particle image data set by using the selected dynamic detection head combination. Through the method, the most suitable detection head combination can be adaptively selected, the target detection precision and efficiency are effectively improved, and the method is particularly suitable for a large-scale cloud particle image data set with complexity difference.

The invention relates to the technical field of PLC (Programmable Logic Controller) control parameter setting, and discloses a PID (Proportion Integration Differentiation) parameter intelligent setting method and system based on a PLC control system. According to the method, a process variable signal, a set value signal and an actuator output signal are collected in real time through a PLC communication protocol, feature quantities are extracted after smooth filtering and noise elimination processing, and a standardized control data set is formed; querying matched setting strategy entries by using the set, calculating complexity indexes, evaluating parameter configuration compliance conditions, and generating setting state judgment of a single control loop; detecting common entities among the loops and calculating an interaction strength index by combining a plurality of standardized control data sets, and generating a multi-loop interaction relation graph; and based on the map fusion setting state judgment, calculating the overall quality score and forming a setting recommendation scheme, and after field verification and closed-loop control simulation test, outputting a final PID parameter setting instruction.

The invention discloses an automatic processing method and system for graphic labeling, and particularly relates to the field of image clustering and automatic labeling, and the method comprises the steps: calculating an entropy value and a gradient variance of a pixel region of an input image to generate a complexity index, adjusting the region division scale according to the complexity index, and obtaining a divided region; calculating a color histogram, a texture direction and an edge curvature in the divided region to form a region feature, and combining a complexity index with the region feature to generate a feature matrix; according to the method, the feature matrix is generated by calculating the region complexity index and extracting the multi-dimensional features, the clustering model is input to identify the boundary cluster and the hopping region, and multi-category and single-category annotations are generated in combination with the weight map and segmented weighting, so that the problems of semantic confusion, boundary instability and error accumulation are solved.

The invention relates to the technical field of data processing, in particular to an AI-based intelligent question setting exercise method and device and a storage medium, and the method comprises the steps: collecting target identity data, behavior data and question data; determining a target behavior index according to the target behavior data, and determining a question complexity index according to the question data; determining a target type according to the target behavior index and the topic complexity index; performing intelligent question setting and pushing to the target according to the target historical exercise question and the target type in the monitoring period; and judging a target state based on the question difficulty, the target question answering accuracy and the target question answering code line number in the management period, and updating a process of performing intelligent question setting and pushing to the target in the next management period according to the target state. The exercise efficiency of the user is effectively improved, and the learning efficiency of the user is improved.

The specification provides a selection method, device and equipment of a well completion mode. The method comprises: obtaining membership degrees of wellbore complexity data corresponding to different well completion modes; processing the membership degrees of the wellbore complexity data corresponding to the different well completion modes by using a wellbore complexity index prediction model to obtain wellbore complexity indexes corresponding to the different well completion modes, the wellbore complexity index prediction model being established in advance by using membership degrees corresponding to wellbore complexity data samples; obtaining engineering data corresponding to the different well completion modes, determining risk levels corresponding to the different well completion modes according to the wellbore complexity indexes and the engineering data; and selecting a target well completion mode from the different well completion modes according to the risk levels corresponding to the different well completion modes. Based on the above method, the optimal well completion mode can be accurately determined.

The invention relates to the technical field of digital recognition and intelligent classification, in particular to an intelligent recognition and classification method for non-missing patterns, which comprises the following steps of: constructing a multi-dimensional embroidery pattern knowledge base, calculating a disparity map and a height map to obtain a stereoscopic strength coefficient, and calculating a stitch complexity index as a physical vector by combining a histogram of oriented gradients; meanwhile, the pattern recognition model is used for outputting the category probability of each independent pattern part as a semantic vector, then an embroidery method is determined, graph nodes are created based on physical and semantic vectors, a spatial relation edge and a cultural relation edge are established, then a multi-modal heterogeneous graph is generated, the graph is input into a graph neural network for message passing and global pooling, and the multi-modal heterogeneous graph is obtained. According to the method, a global feature vector is obtained, an embroidery theme is matched through a theme classifier, finally, the comprehensive credibility is calculated, and through multi-modal data fusion and heterogeneous graph analysis, the accuracy and automation level of pattern recognition are effectively improved, and the method is particularly suitable for intelligent classification of complex non-missing embroidery.

The invention discloses a dynamic potential function stability detection method and device of a material, equipment and a medium, and relates to the field of computational material science, and the method comprises the steps: building a molecular dynamics model of a target material to execute molecular dynamics simulation operation to obtain a simulation result; the molecular dynamics model comprises a to-be-detected dynamics potential function; determining the local stiffness of each material atom in the simulation result, and determining the performance index of the multi-dimensional performance corresponding to the to-be-detected dynamic potential function based on the local stiffness data set; the performance indexes comprise a structural uniformity index, a temperature stability index, an atomic configuration robustness index, a fitting curve complexity index and a physical rationality index; and calculating a multi-dimensional comprehensive stability index corresponding to the to-be-detected dynamic potential function based on the performance index and the performance weight of each performance, and determining a stability grade of the to-be-detected dynamic potential function of the target material based on the multi-dimensional comprehensive stability index and a preset grading rule. And the dynamic potential function stability can be detected.

The invention relates to the technical field of data processing, in particular to an outpatient service payment medical insurance data rationality analysis method and system based on clustering analysis, and the method comprises the steps: converting the multidimensional resource consumption of outpatient service records into standard comparable structure expression through constructing a unified resource structure vector, and providing unified input for subsequent analysis. On the basis, an improved clustering algorithm with resource sensitivity and an abnormal penalty term is introduced, different resource consumption modes are automatically recognized, resource cluster tags are generated, and it is guaranteed that a clustering result fits the actual features of medical insurance. Furthermore, a behavior path modeling mechanism is established in the cluster, and an operation type weight and a redundant path penalty function are introduced to quantify and form a complexity index reflecting structural rationality and behavior normalization. And finally, comparing the resource cluster with the medical insurance group, and proposing an intra-group structure consistency score and a payment deviation evaluation index.

The invention provides a cross-layer annotation correction and real-time rendering method and application thereof, and belongs to the technical field of digital pathological image processing. According to the method, the problems of labeling dislocation, editing distortion and large-point-set rendering lagging during switching of the multi-focal-plane slices are solved. The method is characterized by comprising the following steps: determining interlayer offset parameters of each focal plane, and carrying out associative storage with focal plane identifiers; anchoring an original focal plane and storing original geometric data when the label is created; when the target focal plane is switched, cross-layer mapping is carried out according to the relative offset parameters, and geometric semantic stability is kept for different labeling types; a rendering complexity index is calculated, hierarchical simplified rendering is executed in a continuous interaction stage, and high-precision display is recovered after interaction is stable; and after editing, reversely writing the target geometric data back to the original focal plane coordinate system. According to the method, cross-layer accurate correction and real-time smooth interaction of the labels are realized, and pathological diagnosis is effectively assisted.

The invention discloses a main network SVG graph model updating method, system and device based on an overlapping sliding window mechanism and dynamic correction and a medium, and the method comprises the steps: carrying out the sliding cutting of a main network SVG graph according to a preset window size and an overlapping rate, and generating a sub-graph set containing an overlapping region; electrical equipment primitives and coordinates in a sub-graph set are recognized through a target detection model, connecting lines between the equipment are extracted based on a straight line detection algorithm, the equipment primitives and the coordinates of the connecting lines are accurately positioned, and the problems of missed recognition and false recognition in a high-density graph are solved; the connection relation is dynamically corrected according to a power topology rule, reconnection operation is executed on the connecting lines crossing the bus to correct abnormal topology, a final to-be-rendered area is generated, and it is ensured that a graph model is consistent with an actual power grid structure; calculating a comprehensive complexity index of the region according to the attribute value of the to-be-rendered region; and on the basis of the comprehensive complexity index and the set threshold value judgment result, the judgment area is subjected to CIM model data rendering or updating operation, interface jamming is avoided, and the main network SVG graph model updating requirements of different complexities based on an overlapped sliding window mechanism and dynamic correction can be met.

The invention provides a course reinforcement learning method, device and equipment for low-resource dialect recognition, relates to the technical field of artificial intelligence, and aims to solve the problems that in a low-resource dialect speech recognition scene, existing reinforcement learning enables a model not to perform effective learning and performance improvement is limited. The method comprises the following steps: performing supervised fine tuning on an initial speech recognition model by adopting a low-resource dialect speech data set to obtain a preliminary optimization model; recognizing the voice data based on the preliminary optimization model, and calculating a complexity index for reflecting the recognition difficulty of the voice data according to a voice recognition result and the labeled reference answer; according to the complexity index, sorting the voice data from low to high according to the recognition difficulty to form a training data sequence of course learning; and applying reinforcement learning on the training data sequence, and optimizing, iteratively updating the preliminary optimization model through a dynamic reward function and a group relative strategy until a preset convergence condition is met, thereby obtaining a target dialect recognition model.

This invention discloses a computer-based method for evaluating the complexity of the working environment of unmanned ground vehicles (UGVs). First, it determines environmental complexity evaluation indicators and establishes an evaluation indicator system. Then, it acquires and normalizes multi-source environmental data using onboard sensors and a deep learning model. Next, it determines the weights of static environmental indicators using the Analytic Hierarchy Process (AHP) and calculates the static environmental complexity index using a nonlinear mapping function. Simultaneously, it establishes a potential energy function incorporating distance, speed, and predicted collision time, calculating the total environmental potential energy to output the dynamic environmental complexity index. For the state evolution complexity evaluation indicator, it constructs a state evolution complexity quantification model based on state change entropy, outputting the state evolution complexity index. Finally, it outputs the working environment complexity index by establishing a multi-dimensional coupled complexity model. This invention achieves real-time quantification and evaluation of working environment complexity, providing reliable data for autonomous decision-making, path planning, and safety control of UGVs.

The application relates to the technical field of PLC control parameter setting, and discloses a PID parameter intelligent setting method and system based on a PLC control system. The method collects process variable signals, set value signals and actuator output signals in real time through a PLC communication protocol, extracts characteristic quantities after smoothing filtering and noise elimination processing, and forms a standardized control data set; the set is used to query matched setting strategy entries, calculate a complexity index and evaluate parameter configuration compliance, and generate a setting state judgment of a single control loop; a plurality of standardized control data sets are combined to detect shared entities between loops and calculate an interaction intensity index, and a multi-loop interaction relationship graph is generated; based on the graph, the setting state judgment is fused, the overall quality score is calculated, and a setting recommendation scheme is formed; after field verification and closed-loop control simulation testing, a final PID parameter setting instruction is output.

The application relates to a motor intention detection method based on electroencephalogram microstate system automaton theory analysis, which comprises the following steps: acquiring multi-try and multi-category electroencephalogram signals based on a motor imagination experiment paradigm and converting the signals into discrete observation sequences; reconstructing an optimal prediction model for the discrete observation sequences based on a causal state splitting reconstruction algorithm; calculating multi-dimensional neural dynamics complexity indexes based on the optimal prediction model to form feature vectors for inter-group comparison; taking the feature vectors as inputs to train a classifier, evaluating the performance of the model in distinguishing different motor intention categories through cross-validation, and outputting the probability of an individual belonging to a certain motor intention category to realize the decoding of motor intention. The application combines formal automaton theory with neural time series analysis to provide a new way for realizing non-invasive and fine decoding of motor intention.

The invention belongs to the technical field of radar signal sorting, and particularly discloses an environment adaptive optimization method for radar signal sorting under a Bayesian framework, which comprises the following steps of: acquiring a plurality of characteristic parameters of a radar signal, grouping the characteristic parameters into a plurality of characteristic domains according to physical attributes of the characteristic parameters, and then calculating statistical dispersion of the characteristic parameters in each domain, obtaining a complexity index of each feature domain; performing cross-domain weighted fusion on the complexity indexes of the feature domains to obtain a comprehensive electromagnetic environment complexity quantitative score; and dynamically mapping the electromagnetic environment complexity quantization score into a probability parameter for controlling generation of a new cluster in the CRP model under the Bayesian framework, and realizing adaptive optimization of the CRP model parameter. According to the method, adaptive adjustment of the CRP model parameters can be realized, and the robustness and accuracy of a sorting algorithm in a dynamic electromagnetic environment are improved.

Systems and method for an AI-based calculation of a case complexity index, CCI. For training a neural network, NN, the method includes receiving training data comprising a medical image of a set of medical images and for example a related report for the medical image and a CCI for the medical image. The method may further include training the NN for providing a trained NN, that is configured for determining the CCI for a medical image by adjusting weights and biases of the NN such that a loss function is minimized.

The invention discloses a threat intelligence analysis method and system. The threat intelligence analysis method comprises the steps of obtaining a full-flow mirror image copy and log data; obtaining each PCAP file according to the full-flow mirror image copy; preprocessing the log data to obtain a structured log stream; extracting the flow metadata of each PCAP file; obtaining a flow data set after layered sampling according to the flow metadata; carrying out feature extraction on the flow data set and the log data after layered sampling so as to obtain a high-order fusion feature vector; and inputting the high-order fusion feature vector into a trained gating cycle unit network so as to obtain an anomaly detection probability value. According to the method, a comprehensive complexity index is constructed through a spatial frequency feature (IP change frequency), a port turnover rate feature (port switching rate) and a protocol switching entropy feature (protocol diversity), and the traffic behavior abnormal degree is quantified.

The invention relates to the technical field of material informatics and data science, in particular to a material performance sensitivity quantification and grading method based on multi-dimensional statistical analysis, which comprises the following steps: inputting a standardized identifier containing a material chemical formula and at least one column of performance index data; a diversity index and an element complexity index are formed; and calculating a variable coefficient, a quantile range ratio, a skewness absolute value, a kurtosis absolute value and an abnormal value proportion for the performance index data column. According to the method, the problems of dependence on expert experience and high subjectivity are solved through multi-dimensional statistical feature fusion and a quantitative model for forming variability correction; the analysis accuracy is improved through five complementary statistical characteristics; different material systems are adapted by forming variability coefficients, evaluation deviation is eliminated, and universality is enhanced; automatic grading and knowledge base construction significantly improve the material design optimization efficiency, provide precise support for process control and quality assurance, and promote intelligent transformation of material research and development.

The application provides an adaptive Bayesian identification method for long-term prediction of industrial chain electric load, comprising: aggregating active power data into daily frozen power time series; performing discrete Fourier transform on the daily frozen power time series, obtaining amplitude spectrum, and calculating spectral complexity index; constructing an adaptive basis function library containing sine and cosine functions according to the screened dominant frequency; establishing a sparse regression model containing regularization, representing the daily frozen power time series as a linear combination of basis functions in the adaptive basis function library, and obtaining a sparse coefficient vector by solving an optimization problem; and using the identified sparse coefficient vector and adaptive basis function library, calculating the daily frozen power prediction value of the future time step through time index extrapolation. The method can plan the characteristics of upstream and downstream industrial chains, accurately capture multi-scale periodic structures, and has low calculation cost and high physical interpretability.

The present invention relates to the technical field of computers, and disclosed are a data processing method and system, a device, and a storage medium. The method comprises: acquiring a plurality of sample prompts for model training; analyzing the correspondence between entities in the sample prompts and entity attributes, and extracting sample prompts in which the correspondence reaches a complexity index, so as to obtain a first data set; extracting, from the first data set, sample prompts satisfying a difference index, so as to obtain a second data set; and performing model training on the basis of the second data set.

The application belongs to the technical field of data processing, and particularly relates to a large model question and answer method and system based on a knowledge base, which comprises the following steps: receiving a user question and extracting a plurality of basic linguistic features; constructing a two-dimensional query essence space and mapping the user question into a query point in the space; in the query essence space, nonlinearly calculating dynamic weights of the basic linguistic features based on geometric distances between the query point and a plurality of preset characteristic ideal anchor points; calculating a dynamic question complexity index according to the dynamic weights and the basic linguistic features, and determining a retrieval quantity of the knowledge base; retrieving knowledge fragments from the knowledge base according to the retrieval quantity, and generating an answer in combination with the user question. The application realizes accurate self-adaptation of a retrieval range by dynamically evaluating a question essence, can provide appropriate context information for a large language model, and thus improves the accuracy and reliability of large model question and answer.

The application relates to the technical field of intelligent unmanned aerial vehicle manufacturing, and discloses a method for constructing and dynamically evaluating an unmanned aerial vehicle cluster comprehensive task complexity index system, which comprises the following steps: a target state parameter and a task-equipment matching function are proposed and improved; the unmanned aerial vehicle cluster task allocation is converted into a multi-target optimization problem; an improved NSGA-II algorithm is used for task allocation under the condition of multi-target optimization; an index system for evaluating task complexity is established by comprehensively considering the index composition of the task target dimension and the task allocation dimension; and multi-dimensional task complexity comprehensive evaluation is carried out. According to the application, personnel can quickly and accurately understand the change of a task situation and make relatively correct judgments accordingly. The application considers the dynamic change of a target state and a task, and can provide guidance for unmanned aerial vehicle task execution.

The application discloses a kind of graphic annotation automation processing method and system, specifically related to image clustering and automatic annotation field, including the entropy value and gradient variance of the pixel region of input image are calculated to generate complexity index, according to complexity index adjustment region division scale, obtain division region;Color histogram, texture direction and edge curvature are calculated in division region, form region feature, and complexity index and region feature are combined to generate feature matrix;The application generates feature matrix by calculating region complexity index and extracting multi-dimensional features, inputs clustering model to identify boundary cluster and jump region, combined with weight map and segmented weighting generates multi-class and single-class annotation, to solve the problem of semantic confusion, unstable boundary and error accumulation.

The invention relates to the technical field of machine vision and operation evaluation, and discloses a size measurement scoring system and method based on machine vision. The method includes parsing out a pause event and associating the pause event with a predefined standard measurement phase by synchronously capturing a continuous spatial position stream and a timestamp stream during a measurement process, forming a sequence of steps to check a logical sequence and generate a sequence deviation factor. Meanwhile, preset image features are dynamically detected, feature capture moments of the features are recorded, position point sets in time windows before and after the moments are extracted, and local trajectory complexity indexes are calculated. And integrating the total time consumption, the sequence deviation factor and the local trajectory complexity index, calculating an original operation efficiency value, carrying out normalized fusion on the original operation efficiency value and the reference size value, and finally outputting a comprehensive measurement score. According to the invention, automatic comprehensive evaluation of the standardization of the measurement operation flow and the stability of the operation process is realized.

The application discloses a lightweight convolutional neural network deployment method and system for smart glasses, a dynamic reasoning strategy is embedded in a feature extraction backbone network, the strategy is a reasoning strategy dynamically selected according to a texture complexity index of a feature map of the current input feature extraction backbone network, the dynamic difference of the input feature is fully considered, a feature extraction backbone network embedded with the dynamic reasoning strategy is trained using knowledge distillation, a lightweight convolutional neural network is obtained, a gradient balance factor inversely proportional to the loss gradient modulus of the training sample is introduced into the distillation loss function, the gradient balance factor is used to effectively balance the contribution of different difficulty samples to parameter updating, the lightweight convolutional neural network is subjected to mixed precision quantization, the quantized lightweight convolutional neural network is mapped to a special instruction set of a smart glasses processor, a binary execution file is obtained, and the binary execution file is deployed to the smart glasses processor, so that the reasoning efficiency and precision of the smart glasses are improved, and the power consumption is reduced.

The application relates to the technical field of machine vision and operation evaluation, and discloses a size measurement scoring system and method based on machine vision. The method comprises the following steps: synchronously capturing a continuous spatial position stream and a time stamp stream in a measurement process, analyzing a pause event and associating the pause event with a pre-defined standard measurement stage, forming a step sequence to check a logical sequence and generating a sequence deviation factor. Meanwhile, preset image features are dynamically detected and their feature capture moments are recorded, a position point set in a time window before and after the moment is extracted, and a local trajectory complexity index is calculated. The original operation efficiency value is calculated by comprehensively considering the overall time consumption, the sequence deviation factor and the local trajectory complexity index, and is normalized and fused with a reference size value, and finally a comprehensive measurement score is output. The application realizes automatic comprehensive evaluation of the standardization of a measurement operation process and the stability of an operation process.