15 results about "Selection operator" patented technology

A global navigation satellite system (GNSS) spoofing detection and classification technique is provided. An optimization problem is formulated at the baseband correlator domain by using an optimization algorithm such as the Least Absolute Shrinkage and Selection Operator (LASSO) algorithm, for example. A model of correlator tap outputs of the intended received signal is created to form a dictionary of pre-computed waveform functions (e.g., triangle-like-shaped functions). Sparse signal processing can be leveraged to choose a decomposition of pre-computed waveform functions from the dictionary. The optimal solution of this minimization problem can discriminate the presence of a potential spoofing attack peak by observing the decomposition of two different code-phase values (authentic and spoofed) in a sparse vector output. A threshold can be used to mitigate false alarms. Furthermore, a variation of the minimization problem can be provided that enhances the dictionary to a higher resolution.

The invention provides a method for establishing an OPC model, electronic equipment, a storage medium and a program product. The method includes constructing a loss function of minimum absolute value contraction and selection operator (LASSO) linear regression based on an original loss function and an L1 regularized loss function for the OPC model, where the original loss function is constructed based on an OPC model threshold to be determined, and where the L1 regularized loss function is constructed based on a plurality of photoresist coefficients to be determined. The method further includes determining values of a plurality of photoresist coefficients and an OPC model threshold based on a loss function of the LASSO linear regression to establish an OPC model.

Disclosed in the present invention is an image classification method based on collaborative optimization algorithms and a feature selection mechanism, which effectively improves the image feature extraction quality and the image classification accuracy. The technical solution comprises: step S1, preprocessing a collected image; step S2, constructing an image feature extraction model on the basis of a genetic programming algorithm, performing image feature extraction, using the concept of individual information optimization to assign position and velocity information to each individual in the algorithm, and updating the position and velocity information of each individual to adjust a selection operator; step S3, constructing a feature selection model to perform selection on the extracted features; step S4, by using the selected features as inputs, training an SVM classifier; and step S5, using the trained SVM classifier to classify express delivery images.

The invention discloses a large-scale satellite measurement and control scheduling method and system based on a self-organizing learning algorithm, and the method comprises the steps: building a large-scale satellite measurement and control scheduling model, and enabling the optimization target of the large-scale satellite measurement and control scheduling model to maximize the task income; generating a measurement and control scheduling scheme by using a regular algorithm; and when the termination condition is not satisfied, calculating the income, updating the state of the next moment, calculating the strategy probability, generating an action selection operator according to the strategy probability, generating a new measurement and control scheduling scheme by using the selection operator, repeating the process until the termination condition is satisfied, and outputting the final scheme. According to the method, a large-scale satellite measurement and control scheduling integer programming model is constructed, genetic algorithm parameters are trained through a deep reinforcement learning algorithm, and a satellite measurement and control scheduling scheme is generated by using the optimal genetic algorithm parameters, so that the diversity of populations can be kept, falling into a local optimal solution is avoided, and the robustness and toughness of search are improved.

The present invention belongs to the field of big data, and specifically relates to a method, a system, and a device for selecting a problem-solving layer of a multi-modal data model. This method is used to optimize the problem-solving layer in the partitioned order product space. It includes a preselection stage and an optimization stage. The specific steps are as follows: S1: Represent the problem-solving layer as chromosome individuals to generate a quasi-initial population. S2: Define a fitness function. S3: Iteratively update the quasi-initial population and save the optimal individual in each round of iterative update. S4: Improve the selection operator, crossover operator, and mutation operator in the classical genetic algorithm to obtain a new adaptive genetic algorithm. S5: Combine the randomly generated chromosomes and the sub-optimal population as the initial population in the optimization stage. S6: Iteratively update the initial population and select the optimal problem-solving layer after the iteration ends. The present invention solves the problems such as difficult convergence and easy entrapment in local optima when the existing classical genetic algorithm is used to handle this problem.

The present disclosure provides a method, an electronic device, a storage medium and a program product for establishing an OPC model. The method comprises: iteratively determining, based on a least absolute shrinkage and selection operator (LASSO) linear regression algorithm, an intermediate model threshold value and a corresponding physical effect coefficient value of the OPC model in each iteration; in response to the iteration satisfying a predetermined condition, determining a final model threshold value based on at least a comparison of a last model threshold value corresponding to a last iteration among the determined intermediate model threshold values and a predetermined threshold value; and establishing the OPC model based on the final model threshold value and the physical effect coefficient value corresponding to the final model threshold value.

The invention relates to the technical field of homomorphic encryption and distributed computing, and discloses a parallel collaborative privacy computing method under the condition of non-uniform computing power of an edge network, which is applied to a star network topology structure consisting of a main node and K edge nodes, and aims to solve the problems of minimum absolute shrinkage and operator selection. And solving and constructing a minimum absolute shrinkage and selection operator problem based on an alternating direction multiplier method by adopting the alternating direction multiplier method, segmenting the compression matrix to obtain a plurality of sub-problems, solving each sub-problem by utilizing edge nodes, and obtaining a minimum absolute shrinkage and selection operator problem based on the minimum absolute shrinkage and selection operator problem. The solution process is as follows: a main node generates encryption and decryption parameters, Euler function values corresponding to the encryption and decryption parameters, encrypted plaintext parameters and quantized plaintext parameters and shares the parameters to an edge node, homomorphic iterative calculation is performed, and three times of communication with the main node is performed, so that parallel collaborative privacy calculation is realized; according to the method, the calculation overhead is reduced while the encryption and decryption precision is not influenced, and meanwhile, the overall calculation speed is increased.

The application discloses a deep optical neural network training method and system based on a hybrid mutation strategy genetic algorithm, and the method comprises the following steps: S1, sequentially stacking a linear operation layer based on MZIs, a nonlinear activation layer based on EOA and a Dropmask based on a mask to build an N-layer deep DONN; S2, preprocessing a data set with different characteristic categories to conform to the data input size of the DONN; S3, uniformly initializing the DONN population, combining the MSE and the Accuracy between the real value and the predicted value as the fitness evaluation function of the individual; S4, taking the exponential ranking selection ERS and the uniform crossover UC as the selection operator and the crossover operator in the training process, adopting a hybrid mutation strategy, and distributing three operators, namely, the single-point mutation SM, the uniform mutation UM and the Gaussian mutation GM, to different individuals for mutation according to a dynamic game probability; and S5, adopting a double-elite reservation strategy, reserving two individuals with the optimal MSE and Accuracy performance to the next generation, and through iterative evolution, until a termination condition is met, and a DONN individual with the globally optimal network parameter is obtained.

The application discloses a drilling accident prediction method and device based on network optimization, equipment and medium, through multiple populations, each population evolves along different directions, the evolution in the population and the existing work remain consistent, the exchange between the populations is introduced, and the solving efficiency is improved. The application proposes a drilling accident prediction method based on BP neural network, relying on the BP neural network, the intelligent level of accident early warning is significantly enhanced; a multi-population based method is proposed to search for the optimal parameters of the BP neural network, the global search ability is enhanced, and the accuracy of the early warning is improved; the selection operator, the exchange operator and the crossover operator of the multi-population are proposed, the shortcomings of the existing genetic algorithm are improved, and the method can be widely applied to the technical field of data processing.

A distributed stochastic nonsmooth optimization method based on smoothing and momentum techniques is provided. The distributed stochastic nonsmooth optimization method solves nonsmooth composite optimization problems with constraints and stochastic factors in a distributed manner and achieves a higher convergence rate, lower computational complexity, and lower storage overhead. The method includes: using a distributed algorithm based on smoothing and momentum techniques to process all the agents in a loop, and providing initial states, step sizes, recommended ranges of smoothing parameters, and so on of the algorithm; and specifying performance metrics of the algorithm, and depicting clustering comparison results according to the metrics. The method is applicable to composite optimization problems with nonsmooth terms, including but not limited to clustering problems, least absolute shrinkage and selection operator (LASSO) regression in machine learning and compressed sensing problems in sensor networks, and is applicable to large-scale distributed nonsmooth optimization scenarios with high-dimensional complex constraints.

The invention belongs to the field of unbalanced data classification, and relates to a genetic programming text data oversampling method based on a multi-layer selection double-criterion evaluation mechanism. According to the method, firstly, an existing unbalanced data set is divided, the number of samples needing to be generated is calculated, and a target majority class and a target minority class are obtained; secondly, an evolution process of genetic programming needs to be designed, and each sub-population independently evolves a target sample by using the designed evolution process; a fitness function adopts a dual-criterion evaluation function based on distance and angle, and a selection operator adopts a selection algorithm which is improved based on a tournament selection method and is suitable for dual-criterion evaluation; the objective of sub-population evolution is to obtain an optimal synthetic instance. Then, selecting an optimal individual from each evolved sub-population as a synthetic instance, and finally, combining the selected individual with an original data set to obtain a complete balanced data set.

Disclosed is an estimation method for joint causal effects of multiple exposures based on high-dimensional independent variables, including the following steps: reducing a dimension by using a modified adaptive least absolute shrinkage and selection operator (LASSO); calculating balance weights by using a nonparametric multiple treatments covariate balancing generalized propensity score (npmtCBGPS) method, and determining an optimal value of a tuning parameter by taking a minimum multiple treatment dual-weighted coefficient (mtDWC) as a criterion; and estimating joint causal effects of multiple continuous exposure factors on an outcome variable by using an inverse probability weighting (IPW) method. According to the present invention, in a framework of a GOAL method, a multiple treatments GOAL (mtGOAL) method by combining the npmtCBGPS method with the adaptive LASSO, and a method capable of estimating joint causal effects of multiple continuous exposure factors on an outcome variable in the presence of high-dimensional covariates are proposed.

The invention provides a fireman physical fatigue identification method based on LASSO multi-feature weighting and a long and short term memory model, and belongs to the technical field of physiological signal processing and fatigue detection. The method comprises the steps that electrocardiosignals of firemen are collected for time domain analysis, frequency domain analysis and nonlinear analysis, and physical fatigue influence variables of the firemen are obtained; performing Wilcoxon symbol rank test and pairing t test on the physical fatigue influence variables of the firefighter to screen out preliminary input features; selecting key features based on the preliminary input features through a minimum absolute shrinkage and selection operator; and inputting the key features into a long-short-term memory model to output a fireman physical fatigue identification result. Sensitive HRV indexes are screened out by using a statistical test method, redundant and irrelevant features are effectively removed by combining an LASSO feature selection method, and a most representative key feature set is reserved. The data dimension is reduced, the calculation efficiency of the model is improved, and the over-fitting risk caused by feature redundancy is also avoided.

The invention discloses a polarization compass orientation method based on real-time discriminant analysis of interpretable course information, which comprises the steps of course data modeling, real-time monitoring, discriminant analysis and noise suppression, modeling by adopting a sparse local Fisher discriminant analysis method, enabling different types of course data states to be more separated according to an inter-class dispersion matrix of a weight coefficient matrix, and improving the orientation accuracy. Similar and same course data states are more clustered by using an intra-class matrix; a minimum absolute contraction selection operator optimization model is introduced, so that a course data discrimination model is simpler and more interpretable; course process data are analyzed through monitoring indexes, so that evolution and instability of dynamic output data of the polarization compass can be monitored in real time, and fine description of the data process can be realized; noise of discriminant analysis is effectively suppressed while real course signals in dynamic course data output by the polarization compass are reserved to the maximum extent. The method solves the problem that the orientation precision of the polarization compass is affected due to weak machine learning generalization ability.

The invention relates to the technical field of artificial intelligence, and provides an integrated agent-assisted genetic programming image classification method, which comprises the following steps: S1, inputting an image training set; s2, randomly generating a genetic programming population according to a program structure, a function set and a terminator set defined in the base method; s3, in the new generation, evaluating each individual by using a real fitness function; s4, extracting features from the evaluated individuals according to tree features, father nodes and class labels, constructing a proxy training set and training a proxy model; and S5, in each cycle of the evolutionary learning process, through elite operator, selection operator, crossover operator and mutation operator operations, selecting individuals with good fitness (i.e., high classification accuracy) to generate a new generation of offspring. According to the method, a basic agent model is constructed by using different types of features, and a dynamic weighting strategy is developed to allocate different weights for the agents, so that compatibility of all genetic programming-based image classification methods is realized, and the evolutionary learning process of the methods is remarkably accelerated.