13results about How to "Predictable" patented technology

This invention belongs to the field of tourist flow prediction and provides a tourist flow prediction method based on BeiDou positioning and structured gated recurrent units. The technical solution involves constructing an adaptive adjacency matrix based on adaptive graph convolution and a random initialization method; randomly initializing the temporal and spatial embedding vectors of the tourist flow data sequence and superimposing the original tourist flow data sequence using a broadcast mechanism to obtain a multi-layered spatiotemporally embedded tourist flow data sequence; combining the adaptive adjacency matrix and the multi-layered spatiotemporally embedded tourist flow data sequence, and extracting the temporal features of tourist flow by embedding a graph convolutional neural network into a gated recurrent unit; based on the temporal features of tourist flow and the constructed structured GRUs network, tourist flow prediction data at different time granularities are obtained, and the feature matrix can be dynamically learned to achieve better prediction performance.

This invention discloses a method for optimizing process parameters in coaxial powder-feed laser cladding, belonging to the field of laser cladding technology. The method includes the following steps: First, using laser cladding process parameters as independent variables and cladding layer performance evaluation indicators as target response values, relevant data are collected using an orthogonal experimental design. Then, using range analysis and variance analysis, the range analysis table, variance analysis table, and mean response diagram of the target response values ​​for the process parameters to the cladding layer performance evaluation indicators are calculated. Significantly influential process parameters are selected as optimization variables, and regression prediction models are established for each target response value and the optimization variables. A multi-objective optimization model for laser cladding process parameters is constructed. Finally, a fast non-dominated genetic algorithm is used to solve the Pareto front solution set of the multi-objective optimization series to obtain the optimal combination of laser cladding process parameters. This invention features a simple calculation process and high versatility; the composite material cladding layer obtained using this method has a smooth, continuous, and highly flat surface.

The application discloses a kind of multi-plane fault-tolerant checking method and system of TTE switch, it is related to communication technical field, in the system, GT interface module is used to convert optical fiber signal into the data of pre-set format;MAC layer is used to convert the data of pre-set format into TTE standard format data, obtains TTE network data frame;Input shunt module is used to frame information and special field corresponding to different service frame type according to TTE network data frame analysis, and frame information and special field are divided into same two-way;According to special field contained in one of them, TTE network data frame is distributed to corresponding processing plane and frame processing is carried out, and data frame after frame processing is obtained, another way is further input to checking data storage module.The application uses non-heterogeneous COM / MON, can realize the method of comparing all effective fields of various types of frames, effectively reduces the occupation of resources.

The invention discloses a model preferential method, device and equipment based on feature weight and a storage medium. The method comprises the following steps: acquiring historical predicted power data and historical model input data corresponding to at least two candidate power prediction models, and historical real power data corresponding to the historical predicted power data; on the basis of the historical predicted power data, the historical model input data and the historical real power data, generating a performance feature vector corresponding to each candidate power prediction model; and determining a target power prediction model from at least two candidate power prediction models based on the performance feature vector and the to-be-optimized performance feature weight corresponding to each performance feature. According to the technical scheme, the prediction precision of power prediction is improved.

The application provides a method for constructing a T7RNAP expression unit, which comprises the following steps: inserting a designed T7RNAP expression unit into a test plasmid to express T7RNAP in a host cell, then determining the T7RNAP enzyme activity E expressed by the plasmid, if E < E0, the threshold value of the T7RNAP enzyme activity in the host cell, then the T7RNAP expression unit is successfully constructed, if E ≥ E0, the DNA structure of the designed T7RNAP expression unit is adjusted, and the above steps are repeated until E < E0, then the T7RNAP expression unit is successfully constructed, and finally the plasmid with the T7RNAP enzyme activity lower than E0 is selected. The application also provides a method for adjusting the DNA structure of the designed T7RNAP expression unit in the above steps. The method for constructing the T7RNAP expression unit has the characteristics of predictability, rapidity, customization and the like, thereby solving the problems of the previous single-plasmid T7 expression system, such as incapability of construction, long construction period and easy mutation of the system, and providing a convenient method for the non-model strains to use the T7 expression system.

The invention discloses a 6G edge network user prediction method based on fairness federation. The method comprises the following steps: step 1, initializing a 6G edge network user prediction mechanism based on fairness federation; 2, the edge core network completes local model updating according to a personalized enhancement strategy; 3, the central core network integrates knowledge to construct a global model according to a fairness aggregation strategy; 4, based on local model updating and a global model, training a 6G edge network user prediction mechanism based on fairness federation; and 5, performing 6G user prediction by using the trained 6G edge network user prediction mechanism based on the fairness federation. The method has the characteristics of personalized service enhancement, fair strategy implementation, prediction performance improvement and data privacy protection.

This invention discloses an integrated reaction device for chlorination and iron extraction from steel tailings and tail gas purification, relating to the field of steel tailings treatment technology. The device includes a chlorination reaction mechanism and a tail gas purification mechanism connected thereto. The tail gas purification mechanism sequentially includes a condensation component, a coalescence component, and a collection component. The coalescence component includes a first baffle channel with adjustable channel size and baffle members disposed within the first baffle channel. The baffle members include a first baffle plate and a second baffle plate. The surface of the second baffle plate has a surface structure for enhancing aerosol adhesion and coalescence, and the coalescence intensity of this surface structure changes with the size of the first baffle channel. This invention, through the synergistic effect of staged condensation and the variable baffle coalescence structure, enables metal chloride aerosols to achieve stable coalescence and collection conditions under different tail gas treatment volumes, realizing the integrated operation of the chlorination and iron extraction process from steel tailings and the tail gas purification process.

The invention relates to a lake and reservoir water quality prediction method and system based on multi-source data fusion, and the method comprises the steps: obtaining multi-source and multi-site data in real time, carrying out the data preprocessing, and removing the influence of sudden change data on prediction; performing feature extraction on time sequence data in the multi-source multi-site data through a time sequence attention mechanism, and processing the time sequence data in the multi-source multi-site data to obtain a predicted value of a target index; analyzing multi-site data in the multi-source multi-site data, and obtaining upstream site information in combination with the spatial position of online monitoring equipment and the water flow direction of a reservoir to obtain trend change information of a target index; and inputting the prediction value and the trend change information into a pre-constructed meta-learner at the same time, and outputting a final prediction result of the target index through combined relation learning. Compared with the prior art, the method has the advantages of high model prediction precision and high interpretability.

The application discloses a power system disturbance minimum frequency prediction method based on a physical information embedded layer, divides a system frequency response physical model minimum frequency expression into different parts, constructs physical information embedded layer features, including first feature information, first weight information and first bias information, and further acquires a physical information embedded layer output function; the output function of the physical information embedded layer and system operation feature data are normalized and then combined, serving as an input layer embedded long short-term memory network, and a novel power system minimum frequency prediction model based on the physical information embedded layer is constructed. The method deeply fuses the frequency prediction physical model and the deep learning algorithm, enhances the ability of a data-driven model to mine physical knowledge, and can quickly and accurately predict the system minimum frequency after disturbance.

The application discloses a kind of considering the motion control method of paddle drive unmanned vehicle of actuator dynamics, belong to unmanned vehicle formation control technical field.The existing paddle drive unmanned vehicle is difficult to effectively control in multi-modal switching and high dynamic operation.The geometric relationship between the follower and the follower is established to establish a tracking error model;Construct a finite time preset performance function embedded in overshoot and stable time and other classical control indicators, and perform error transformation through tangent function;Based on the backstepping method, an adaptive control law is designed, a first-order command filter is introduced to suppress calculation expansion, and RBF neural network is used to compensate model uncertainty and external disturbance online.While considering the physical limitations of the actuator, high-precision and strong-robustness motion tracking control of the paddle-driven unmanned vehicle in complex environments is achieved.

The application discloses a kind of high-precision prediction method of friction coefficient based on ceramic coating, it is related to ceramic coating friction coefficient prediction technical field, comprising the following steps: step one: obtaining the experimental data and literature auxiliary data of APS spraying ceramic coating;Step two: complete field alignment, training / test set division, missing value processing and training set enhancement;Step three: construct five-level physical feature chain, convert original variable into descriptor with tribological physical meaning;Step four: descriptor input machine learning model for training and prediction;Step five: robustness verification and explainability analysis are carried out, the method constructs five-level physical feature chain, and further maps original material parameter and working condition parameter into physical descriptor related to contact stiffness, pore weakening, bearing capacity and lubrication response, so that model input no longer stays in surface experimental variable, but can more fully characterize tribological mechanism.

This invention relates to the field of medical data analysis technology, specifically to a method for constructing an early warning model for urosepsis. The method includes: collecting multimodal time-series clinical data of patients; performing time-series alignment and imputation of missing values ​​on the data; fusing features from different modalities and time points based on a dynamic weight allocation mechanism to generate a comprehensive feature representation of the patient; employing a hierarchical cascade classification strategy, first using a lightweight rapid screening model to preliminarily identify high-risk patients, then applying a refined assessment model to conduct a detailed risk assessment of high-risk patients; and finally generating early warning signals and clinical decision recommendations. This invention effectively integrates multimodal time-series data, focuses on key information through dynamic weights and attention mechanisms, and balances screening efficiency and assessment accuracy by combining a two-stage classification strategy. It solves the problems of low sensitivity, neglect of dynamic time-series changes, and difficulty in clinical deployment in existing technologies, enabling early and accurate risk warning.

The application discloses a kind of text-to-image auxiliary machine translation methods based on large model, comprising: obtaining source text, generating initial image using diffusion model;Text semantic graph of source text and visual semantic graph of initial image are constructed, and the diffusion model is optimized based on consistency reward function using reinforcement learning, to obtain the optimized image;Degenerate text is obtained by degeneration processing to source text, and the text features of source text and degenerate text and the image features of optimized image are extracted, and after fusion, input large model to obtain multi-modal semantic representation;Diffusion model and large model are sequentially trained, and translation result is generated according to the large model after training.The application does not need artificial marking of graph-text data, improves the semantic consistency of generated image and source text through reinforcement learning feedback, realizes the deep cooperation of text and image using multi-modal fusion and sequential training, and improves the performance of multi-modal machine translation.