213 results about "Time complexity" patented technology

The invention discloses a multi-user short-term power load prediction method for an improved Transform model, and relates to the technical field of novel power systems, and the method comprises the steps: S1, setting an internal decomposition block based on a moving average and Fourier transform thought, integrating the internal decomposition block into a frame, and decomposing an input data sequence into trend, season, holiday and festival variables and residual variables; s2, an Nystrom method is used for approximating a softmax matrix in a standard self-attention mechanism, and the softmax matrix is used for coping with challenges of secondary time complexity and memory use in Transform; s3, performing deep feature mining on the seasonal part after time sequence decomposition by using convolution operation; and S4, probability prediction based on Monte Carlo random inactivation. And designing an internal decomposition module of the depth model, and extracting an internal complex time sequence trend of a hidden state in the model by the module, so that the model has a progressive decomposition capability of a complex time sequence. Different from a traditional load prediction method for exploring time characteristics of a single user, a space-time attention mechanism is improved, and accurate multi-user short-term load prediction is provided.

The invention relates to the technical field of short-term load prediction in a power system, and discloses a short-term load prediction system based on a coder-decoder architecture, and the system is characterized in that a coder is used for extracting local features of a power load mode; and the decoder is used for converting the local features of the power load mode into predicted power load values and outputting the predicted power load values. The encoder is realized by a multi-scale expansion causal convolutional network MSDCC, and the decoder is realized by a bidirectional long short-term memory network BiLSTM. The prediction system construction method comprises the following steps: extracting related data from a historical database, preprocessing and analyzing the data, and constructing a predictor matrix; an MSDCC encoder is constructed; a BiLSTM decoder is constructed; and combining the encoder-decoder architecture to construct a short-term load prediction model. According to the method, the size of the feature map is effectively limited, model parameters are reduced, overfitting is avoided, calculation requirements are controlled, non-linear features are efficiently captured, meanwhile, time keeping complexity is low, and therefore the method has the advantages of being high in prediction efficiency and high in prediction precision.

The invention discloses a large language model output quality guarantee method, device and system based on chain type deterministic reasoning verification, and belongs to the field of artificial intelligence safety, deterministic calculation and reasoning verification. The method comprises the following steps: classifying each reasoning step generated by LLM into a deterministic step (bT: deductive reasoning, verified facts and mathematical proof) or a non-deterministic step (bF: inductive reasoning, unverified references and creative guess), and accumulatively constructing a reasoning chain; chain certainty verification chain (chain) = foldr (AND, true, chain) and time complexity O (n) are carried out in each k steps; if the bF is returned, a chain pollution elimination theorem is applied to prove that once the chain contains non-deterministic steps, the whole chain loses deterministic guarantee; and a deterministic boundary report is generated, and a non-deterministic initial position is marked. The device comprises an inference classifier, a reference verifier, a mathematical checker and a chain verification engine, and all the modules can be independently realized (formalized verification / database query / rule engine / mixed mode).

The invention relates to the technical field of data processing, and discloses a space-time K function graph generation method and system based on distribution and aggregation, a terminal and a storage medium, the method comprises the following steps: obtaining a position data set of a target area, constructing a first matrix, performing data point access on the position data set to obtain an access result, and performing range query according to the access result to obtain a space-time K function graph; obtaining a first range query set; constructing a second matrix, and performing element allocation on the second matrix according to the first range query set to obtain a target second matrix; constructing a third matrix, performing numerical calculation on all elements in the third matrix to obtain a target calculation result, and performing optimization processing on the first matrix according to the target calculation result to obtain a target first matrix; and generating a space-time K function graph according to the target first matrix and the space-time K function minimum value and the space-time K function maximum value of the randomly generated data set. According to the method, the time complexity of generating the space-time K function diagram is reduced while the precise solution and the same space complexity are kept.

The invention discloses a multi-source time sequence test data time alignment method and a storage medium, and the method comprises the steps: obtaining the data of each product in a test, and calculating the maximum data collection frequency and time sequence data reference time axis of all products; performing preliminary time alignment on each product; static products and dynamic products are distinguished, and different types of products are subjected to time alignment according to the time sequence data reference time axis in different modes. According to the method, the unique reference time axis is calculated by calculating the maximum data acquisition frequency, the time complexity is O (1) when time matching is carried out on data at each moment, and time sequence data time preliminary alignment work can be completed at a higher speed when the data size is too large. According to the invention, for the time sequence data which is subsequently added into time alignment calculation, the current time sequence data which is subjected to time alignment cannot be influenced. Meanwhile, data preprocessing work is carried out according to the data features of different data sources, so that the time alignment work of the multi-source time sequence data is more complete.

This application provides a point cloud decoding performed by an electronic device. The method includes: obtaining a current point to be decoded from a point cloud code stream, the current point comprising N types of attribute information, and N being an integer greater than 1; determining M attribute groups of the N types of attribute information of the current point, M being a positive integer less than N; and perform uniform decoding for each attribute information in a same attribute group of the current point using a first operation, the first operation comprising at least one of a neighbor point determination and an attribute predicted value calculation. According to this application, the point cloud decoding method avoids the repeated searching, reduces time complexity during attribute decoding and improves attribute decoding efficiency.

The invention relates to an engineering optimization method and system based on a whale optimization algorithm, and a medium, and belongs to the technical field of intelligent optimization algorithms. Comprising the following steps: S1, initializing parameters and populations: setting a population scale, a maximum number of iterations, dimensions and upper and lower bounds of variables, generating an initial population, and initializing a global optimal solution and a current optimal solution; s2, executing an individual variation strategy: recording a global optimal individual position and a current iteration optimal individual position; s3, executing a unified search strategy; s4, executing a group communication strategy; and S5, judging whether the current number of iterations reaches the maximum number of iterations, if so, outputting a globally optimal solution, otherwise, adding 1 to the current number of iterations and returning to the step S2 to continue iteration. The algorithm provided by the invention aims to improve the global search capability and convergence precision of the algorithm and maintain the time complexity equivalent to that of the original WOA at the same time.

The invention discloses an improved two-dimensional ordered statistical constant false alarm rate detector implementation method based on a CPU. The implementation steps are as follows: inputting a two-dimensional data matrix; performing initialization setting; for each detection unit, judging whether the sample value of the detection unit is the maximum value in a nine-block grid taking the detection unit as the center or not before detection; if yes, judging whether the sample value of the detection unit is greater than k arbitrary values of all reference unit sample values multiplied by a threshold factor or not through parallel processing; if yes, the detection unit is judged as a target, and the adjacent undetected unit does not need to be detected subsequently. According to the method, the number of detection units is greatly reduced, the processing time is shortened, meanwhile, in the detection process, a CPU does not need to open up an extra memory space to store all reference units to be sorted, data sorting work which is complex in traditional two-dimensional ordered statistics constant false alarm rate and long in consumed time is ingeniously avoided, the time complexity and the space complexity of an algorithm are reduced, and the detection efficiency is improved. And the real-time performance of target detection is ensured.

The application discloses a kind of combination packet classifier and method based on hash and TCAM, adopt the combination of imperfect hash classification module and TCAM classification module, the complement of imperfect hash classification and TCAM classification on input key value space is realized;Compared with the traditional TCAM packet classification solution based on, the application reduces hardware resource utilization, system power consumption;Compared with the packet classification solution based on perfect hash, the application reduces the design cost of hash function and the time complexity of generation process under the premise of guaranteeing the function of classifier.The application can consider low space resource consumption, low power consumption and low delay in the classification stage, saves design cost in the classifier generation stage, and reduces the time complexity of generation process;Meanwhile, the application can better balance different types of hardware resources on reconfigurable hardware platforms such as FPGA, and provides a new way of thinking for the design of classifier in network tasks such as firewall packet filtering and routing table lookup.

An implicit video recommendation algorithm, by calculating the proportion of the time length of the video watched by the user to the total time length of the video, then the data dispersion situation data is counted, the data is taken as the parameter, the data table of the user and the video is established, then the cosine similarity is calculated, after sorting, the video with the highest similarity is recommended, according to the accuracy, recall rate of the training result, the proportion of data division is dynamically adjusted, so as to achieve the purpose of optimizing the final recommendation result. By deeply mining the relationship between users and videos, this can solve the problem of long tail commodity recommendation and data sparsity, and the time complexity of this algorithm is low, and the operation speed is fast.

The invention discloses a time domain transformation optimization method in a vector network analyzer, relates to the technical field of signal processing, and provides an optimization algorithm for establishing a mathematical relationship between a variable time axis and IFFT (Inverse Fast Fourier Transform) based on IDFT (Inverse Fast Fourier Transform) aiming at the defects of a traditional time domain transformation algorithm in the vector network analyzer. The final result is completely consistent with the variable time axis IDFT, the time complexity is reduced to be equivalent to that of the IFFT, the calculation speed is increased by thousands of times compared with that of the IDFT when the number of data points reaches tens of thousands, and the requirement of the vector network analyzer for observing the time domain fine response can be efficiently and flexibly met.

An exemplary sliding window scheduling method and system are disclosed. The exemplary sliding window scheduling method and system can schedule multiple packets in a given scheduling frame with a sliding window scheduling frame. The scheduling operation can be performed using bitmap operators and can achieve a lowest time complexity of O(1) per matching computation and per port using distributed parallelization hardware. The exemplary sliding window scheduling method and system can be performed in the context of a queue-proportional scheduler (QPS) as well as iSLIP. In alternative embodiments, the SW-QPS operation can be performed in a batching window rather than in a sliding window.

The purpose of this invention is to address the issue of accuracy in processing spinning time-series data streams, which are characterized by high noise levels and distinct time-domain features generated during the operation of spinning workshops, by employing a fuzzy hierarchical clustering method that integrates time-domain characteristics. This method reduces the impact of noise during the classification process and considers both time-domain features and noise effects. The technical solution of this invention is to provide a fuzzy hierarchical clustering analysis method for spinning time-series data that integrates time-domain characteristics. This invention proposes a fuzzy hierarchical clustering analysis method for spinning time-series data that integrates time-domain features. It iterates between the DTS feature matrix and the MTS feature matrix, considering the time-domain characteristics and noise effects in the spinning time-series data. Without increasing time complexity, it incorporates the time-frequency characteristics and noise effects in the spinning time-series data during the iteration process. Compared with the latest methods, this invention can more accurately process newly generated time-series data in spinning manufacturing.

The invention provides a dynamic network capacity calculation method and device based on topology change characteristics, and belongs to the technical field of communication. With the maximum hop count and the maximum spanning time slice number as constraints, augmented path search is carried out on the residual network constructed based on the time extension graph in a breadth-first search mode, that is, the hop count of the search path and the spanning time slice number thereof are concerned all the time; according to the method, the search paths exceeding the set maximum hop count and the set maximum spanning time slice number are pruned, that is, some overlong search paths are removed, so that the range needing to be searched is reduced, the time complexity of searching the augmented path is reduced, and when the augmented path is searched, the search is preferentially performed in the same time slice, so that the search efficiency is improved. And it is ensured that the target augmented path with the minimum number of spanning time slices can be searched preferentially.

The invention discloses a time series data frequent pattern mining method, and belongs to the technical field of CPU utilization rate data processing. According to the invention, the problems of high calculation complexity and high power consumption of the existing method are solved. According to the method, the Boolean mask vector M is introduced, elements unfavorable for the effectiveness of the lower bound are selectively shielded, and the tightness and the filtering capacity of the lower bound are further improved. Firstly, LBDIAG with high calculation efficiency is used for carrying out preliminary filtering, then a traditional tight lower bound is used for carrying out a fine cascade filtering strategy, the number of candidate pairs entering DTW distance accurate calculation can be reduced, and the time complexity of calculation is reduced through combination of a sliding window technology and inner product operation. And lower bound calculation tasks of diagonals are allocated to different threads by utilizing the parallel calculation capability of the GPU, so that the processing speed of large-scale data is remarkably improved, meanwhile, a memory access mode is optimized, the global memory access frequency is reduced, and the power consumption in the calculation process is reduced. The method can be applied to CPU utilization rate data processing.

The invention discloses a scheduling table generation method and system for a time-triggered service, and mainly solves the problem of high time solving complexity in a large-scale scene in the prior art. According to the implementation scheme, a TTE scheduling model and a service conflict graph are constructed, and a maximum connected subgraph is obtained; independently dividing each sub-graph into sets, dividing services into link-conflict-free groups in sequence in each set, and distributing the groups to free time slots in a time axis; judging whether a group of which the time slot is not successfully distributed exists in the current maximum connected sub-graph; if so, trying to distribute again through backtracking operation; if not, obtaining the sending moment of the service at the source end according to the time slot where the group of the service is located and the intra-group time slot offset of the service; and calculating the sending time and the receiving time of each switch and the receiving time of the destination end according to the sending time of the source end, and generating a conflict-free scheduling table. According to the method, the time complexity of scheduling solution is reduced, and the method can be flexibly applied to scheduling planning in various scenes such as unicast service, multicast service, redundant service and mixed service.

Disclosed in the present invention is a real-time recursive CRC method. The method comprises: sequentially extracting from a real-time bit stream a1a2a3… binary sequences S1S2…Si… at non-fixed lengths; bisecting into left and right parts each binary sequence segment Si, so as to obtain sub-sequences, and repeatedly bisecting into left and right parts the sub-sequences until the lengths of finally obtained sub-sequences are less than a preset value; and calculating a CRC check value and check matrix of each final sub-sequence, and performing level-by-level upward merging calculation on left and right sub-sequences, so as to obtain a final check value of the entire real-time bit stream. Further disclosed in the present invention are a corresponding CRC apparatus and a computer storage medium and a program for implementing the CRC method. In the present invention, a method combining iteration with recursion is used to calculate, in real time, a check value of a variable-length binary sequence stream, thereby achieving logarithmic calculation time complexity and meeting CRC operation requirements of a real-time ultra-high-speed data stream.

The invention provides a continuous learning data space relation extraction method and system based on parameter regularization and a dynamic memory, belongs to the technical field of knowledge maps, and aims to provide a continuous learning data space relation extraction method and system based on the parameter regularization and the dynamic memory by using a BERT pre-training language model as an encoder, model parameter regularization, dynamic memory construction and memory sample data enhancement technologies. Embedding space representation of samples in old tasks is maintained, so that old relation knowledge is reserved, a memory selection strategy is optimized, and limited memory space is efficiently utilized; the problems that old sample embedding space deviation caused by encoder parameter changes and a typical sample selection strategy are not reasonable are solved, the diversity of a sample memory set is also considered, and the prediction accuracy of the model is improved under the condition that the time complexity is hardly increased. The method is suitable for a continuous learning relation extraction technology in the field of knowledge graph data space construction.

The present application describes a multi-key fully homomorphic encryption (MK-FHE) scheme that enables secure and efficient multi-party computation by integrating learning with errors (LWE), ring learning with errors (RLWE), and NTRU-based encryption primitives. The scheme supports dynamic key management, parallelizable bootstrapping, and low-overhead homomorphic operations. Key innovations include a hybrid product mechanism for merging ciphertexts across cryptographic structures, a single-key blind rotation algorithm optimized for Fourier domain operations, and a noise-refreshing procedure that bounds error growth during homomorphic evaluations. This scheme achieves quasi-linear time complexity relative to the number of participating parties, making it suitable for resource-constrained environments such as federated learning and secure cloud-based AI inference.

The embodiments of this specification disclose a method, device, and medium for deduplicating a three-dimensional point set, which relate to the technical field of data processing and are used to solve the problem of high time complexity and low efficiency of existing deduplication methods. The method includes: matching a corresponding numerical comparison tolerance based on the required three-dimensional model accuracy of the three-dimensional point set to be deduplicated and the current computing resources; determining the number of one-dimensional space partitions corresponding to the three-dimensional point set to be deduplicated according to the numerical comparison tolerance, so as to partition the space corresponding to the three-dimensional point set to be deduplicated based on the number of one-dimensional space partitions and obtain corresponding spatial cells; preprocessing the current point in the three-dimensional point set to be deduplicated to obtain a first index value corresponding to the coordinates of the current point in the three-dimensional point set to be deduplicated; traversing the current cell and neighbor cells of the current point in the three-dimensional point set to be deduplicated to calculate a second index value corresponding to the coordinates of the current point represented by each cell; and deduplicating the three-dimensional point set to be deduplicated based on the first index value and the second index value.

The invention provides a fair clustering method and device for large-scale data, equipment and a medium, and the method comprises the steps: constructing a to-be-classified data set after carrying out the normalization processing of the obtained to-be-clustered large-scale data, solving the to-be-classified data set through a greedy algorithm, obtaining an initial solution, and carrying out the clustering of the to-be-clustered large-scale data. And then sampling the data set, combining the sampled data set with the initial solution to form a weighted instance of the original data, executing a linear programming algorithm to obtain an intermediate solution, and iterating the intermediate solution through a local search algorithm meeting a fair condition to obtain a local optimal solution. The solution of the constant approximation ratio can be obtained within the time complexity in linear relation with the data scale, that is, the algorithm operation time is in direct proportion relation with the data scale, and compared with a traditional method, the solution efficiency and the solution quality are remarkably improved.

The present application relates to the field of computers, and in particular to the field of cloud storage technology, and discloses a data storage method, apparatus, device, and storage medium for improving data storage efficiency and avoiding file fragmentation, wherein the method comprises: receiving a data write request, and after determining that the data to be written is to be written into a target cache area, determining a first cache interval corresponding to the data to be written, and if it is determined that there is an overlapping area between the first cache interval and at least one second cache interval, merging to obtain a merged cache area, and then saving the data to be written based on the merged cache area. In this way, when storing files, not only can the cache interval be accurately and quickly located according to the offset, but also data to be written of any length can be processed, thereby reducing the time complexity of the operation and improving data storage efficiency.

The invention discloses a video coding method, device and equipment and a computer readable storage medium, and the method comprises the steps: selecting an optimal prediction mode and a suboptimal prediction mode from four prediction modes: a vertical prediction mode, a horizontal prediction mode, a diagonal left lower prediction mode and a diagonal right lower prediction mode; selecting a first candidate prediction mode from a vertical rightward prediction mode, a horizontal downward prediction mode, a vertical leftward prediction mode, a horizontal upward prediction mode and a direct current prediction mode according to the optimal prediction mode and the suboptimal prediction mode; and selecting a second candidate prediction mode from the upper block prediction mode, the left block prediction mode and the most probable prediction mode. And performing video coding by using a target prediction mode determined from the optimal prediction mode, the first candidate prediction mode and the second candidate prediction mode. According to the method, the calculation time complexity is reduced, the video coding efficiency is improved, and the real-time performance of video coding is ensured.

The embodiment of the present application discloses a kind of artificial address text similarity processing method and system filled in.This embodiment of the present application removes the regular character appearing in address in a reasonable way, so as to remove the error similarity contribution caused by address regular character.Using lower time complexity, the common substring of all more than two lengths in two addresses is calculated.A designed increasing function can map the length sequence of common substring into [0,1] space, and the function reasonably describes the characteristics of human-filled address data.A kind of address clustering algorithm capable of identifying not similar but actually identical address is designed.

The embodiment of the application provides a kind of data storage method, device, equipment and storage medium, related to computer technical field, the method comprises: for the transaction processing event of to-be-processed event, corresponding parameter query event is constructed, and transaction information in to-be-processed event is saved in transaction time sequence library with parameter query event.For the parameter processing event of to-be-processed event, corresponding parameter maintenance event is constructed, and parameter maintenance event and exchange rate parameter are associated and saved in parameter time sequence library, so it is not necessary to record corresponding exchange rate information for each transaction, avoid data storage redundancy, improve the utilization of storage resources and the processing performance of system.Parameter time sequence library and transaction time sequence library store and manage data using the same time axis, so that the retrieval time complexity is lower.The storage mechanism of event binding data is used to store the time sequence data of transaction and the time sequence data of exchange rate parameter, so that the recorded time sequence data directly reflects the business processing process.

This invention discloses a method for determining conflicts between track segments, relating to the field of rail transit technology. When the first track segment and the second track segment are in opposite directions, if the starting point of the first track segment can be found along the direction of this segment along the same segment, and the search path length is greater than 0 and less than the sum of the lengths of the two track segments; or, if the first track segment and the second track segment are in opposite directions, if the ending point of the first track segment can be found along the opposite direction of this segment along the same segment, and the search path length is greater than 0 and less than the sum of the lengths of the two track segments, then the two track segments are in conflict. This invention accurately determines whether two track segments conflict. This invention only requires one search to determine whether there is a conflict, greatly reducing the judgment time and the time complexity of the algorithm; this algorithm is particularly advantageous in the rail transit industry, where timeliness is crucial.

The present invention discloses a method, device and electronic device for scanning duplicate codes, the method comprising: obtaining all classes loaded when the target application package is running; based on the functions of each class, classifying classes representing the same function into the same group; performing code comparison between classes in each group, screening classes with duplicate codes for duplicate class marking. The technical solution provided by the present invention, combined with the project engineering characteristics, divides the classes into dimensions based on functions, narrows the comparison range, and only compares within the group. Classes in different groups must not be duplicates of each other. The optimized time complexity is O(n)=(n2‑nk) / 2k, where k is the number of groups and n is the number of all classes. The time complexity is about one kth of the conventional solution, which greatly improves the scanning speed of duplicate codes.

The invention relates to the technical field of wireless communication, in particular to a low-complexity signal detection method of an MIMO-OTFS system, which comprises the following steps: setting a grouping symbol number according to a channel condition, and calculating a maximum iteration number; in each iteration: calculating a filtered received signal through the channel matrix and the received signal; constructing a matrix psi, processing a structure matrix psi S of the matrix psi by adopting an RCM algorithm to obtain a permutation matrix, and calculating a reordering matrix according to the permutation matrix; performing Cholesky decomposition on the reordering matrix to obtain a banded matrix; calculating the SINR of each symbol which is not detected at present; all symbols which are not detected at present are arranged in a descending order according to the SINR size, the first S symbols in the sorting result are taken to be detected one by one and output, and the channel matrix and the received signal are updated; judging whether the maximum iteration number is reached or not, and if not, continuing the next iteration; according to the method, the time complexity is reduced while the bit error rate performance is ensured through an iterative strategy of reliability symbol detection and interference elimination.

In an example, a particular type of deep learning model is used in the global model of the GDMix model: a Factorization Machine. A Factorization Machine combines a Support Vector Machine (SVM) and Matrix Factorizations. It has the advantage of modeling data with huge sparsity well, while maintaining a linear time complexity. A modification may be further made to the Factorization Machine by introducing L2 norm reduction. This acts to divide calculations made by the Factorization Machine into a portion that can be precomputed and a portion that cannot be precomputed. The portion that can be precomputed is then precomputed in an offline manner. As such, when the model is operated in an online manner, the Factorization Machine only needs to compute the portion that cannot be precomputed, reducing the number of operations that need to performed at runtime and greatly improving processing speed over prior machine learned models.

The invention discloses a data restoration method, a distributed storage system, equipment, a medium and a product, and relates to the technical field of data storage, to-be-stored data is obtained and is subjected to block processing to obtain at least one data block, then the stored data of the at least one data block is subjected to mapping processing to obtain a multivariate polynomial coefficient, and the multivariate polynomial coefficient is subjected to data restoration; and completing the work of storing the storage data to the storage node while constructing the minimum storage regeneration code. Data symbols (storage data) are mapped into coefficients of a multivariate polynomial instead of traditional matrix elements, and repair processing in the subsequent data repair process is facilitated. And reconstructing a multivariate polynomial from the polynomial coefficient of the help node by using a corresponding multivariate polynomial interpolation mode to complete restoration. And by utilizing the low complexity characteristic of a polynomial interpolation mode, the high calculation load of matrix inversion is avoided, the time complexity of the repair process is reduced to # imgabs0 #, and the repair efficiency is remarkably improved.