55 results about "Method screening" patented technology

The invention discloses an uncertainty and representativeness combined active learning method and system for medical image data annotation, and relates to the technical field of medical image data. The method comprises the following steps of: training a variational auto-encoder infoVAE on an image pool; m samples are randomly extracted from the image pool and labeled, an initial labeled image set is constructed, and a segmentation model is trained on the labeled image set; t rounds of active learning circulation are carried out, and in each round of t active learning circulation, the following steps are specifically executed: screening candidate samples based on a representative method; screening a final sample based on an uncertainty method; updating the labeled data set and the unlabeled data set; retraining the segmentation model on the updated labeled image set, and optimizing model parameters; and obtaining final model parameters. According to the method, by designing a sample selection strategy integrating uncertainty and representativeness, the global performance of the model and the difficult sample segmentation capability are improved.

Described herein are methods such as multi-omic methods for assessing a disease such as cancer. The multi-omic methods may integrate proteomic, transcriptomic, genomic, lipidomic, or metabolomic data. The method screening diseases or disease states. Also described herein are methods for screening for diseases or disease states from biological samples. The methods may include assessing whether a nodule, mass, or cyst is cancerous.

The invention discloses an ICU patient abnormity identification method based on video monitoring, and relates to the technical field of video monitoring, and the method comprises the steps: collecting historical data, building a standard set and a habit feature set according to the collected historical data, and generating a prediction set based on the standard set and the habit feature set; collecting real-time data, comparing the real-time data with the standard set, the habit feature set and the prediction set to obtain an abnormal coefficient of the data, calculating a comprehensive danger coefficient according to the abnormal coefficient of the data, and triggering early warning based on the comprehensive danger coefficient; a plurality of disease types of a patient are collected in real time, one disease type is screened out through a two-stage recognition method, the danger degree of the patient is quantified through calculation of a comprehensive danger coefficient, visual decision support is provided for medical staff, and the judgment accuracy is improved through comparative analysis of a standard set, a habit feature set and a prediction set.

The invention discloses a photovoltaic micro-grid grid-connected synchronous control method based on Beidou time service, and relates to the technical field of power system synchronous control, and the method specifically comprises the steps: building a unified time reference of a whole system, calibrating node time through Beidou time service, setting a multi-mode data collection method for multi-mode type data, and screening effective data. A deviation range and a grid-connected state label are set; influences of voltage, frequency and phase deviation on grid-connected instability are quantified through mutual information entropy; a deviation contribution degree coefficient and a risk correction factor are calculated; a core feature is extracted; the method comprises the following steps: constructing Beidou time sequence fingerprints, calculating the similarity between the current Beidou time sequence fingerprints and a fingerprint sample library in real time, screening high-similarity samples to extract strategy elements, selecting an optimal control combination through multi-objective optimization, and adjusting output through information interaction between nodes to realize collaborative optimization.

The invention discloses a screening method of mango pulp processing mode markers and an identification method of mango pulp processing modes, and relates to the technical field of analytical chemistry. According to the method, non-targeted metabonomics is combined with quasi-targeted metabonomics and targeted quantitative analysis, so that the characteristic markers capable of accurately distinguishing mango pulp in different processing modes are systematically screened and verified for the first time. The feature marker screened by the method can be used for constructing a discrimination model of the mango pulp processing mode, high-precision and high-reliability processing mode discrimination is realized based on the established multi-model fusion discrimination method, the discrimination efficiency and accuracy of the mango pulp processing mode are greatly improved, and the method is suitable for popularization and application. And a new technical means and theoretical basis are provided for product quality control and market supervision.

Described herein are methods such as multi-omic methods for assessing a disease such as cancer. The multi-omic methods may integrate proteomic, transcriptomic, genomic, lipidomic, or metabolomic data. The method screening diseases or disease states. Also described herein are methods for screening for diseases or disease states from biological samples. The methods may include assessing whether a nodule, mass, or cyst is cancerous.

The invention relates to the field of high-temperature alloy forging, in particular to a high-temperature alloy turbine disc forge piece key process feature recognition method and a related device. The method comprises the following steps: firstly, constructing a multi-dimensional data set fusing raw material parameters, equipment process parameters and quality detection indexes, and carrying out feature pre-screening in a Filter stage by adopting the combination of a Pearson correlation coefficient and minimum redundancy and maximum correlation to obtain candidate feature subsets; on the basis, a Bayesian optimization strategy taking a random forest as an agent model is introduced, and iterative search is performed on a feature combination space through an expected improved acquisition function to obtain an optimal key process feature set. According to the method provided by the invention, the calculation efficiency is remarkably improved through screening of a hybrid key process feature recognition method, and meanwhile, the product performance consistency is remarkably improved by predicting the key process features and regulating and controlling the process parameters in the production process.

The invention discloses a deep learning electroencephalogram signal identification method and management system. According to the method, original electroencephalogram signal data is converted into a form which can be effectively processed by a machine learning model through feature extraction and conversion. The beneficial effects of the step are mainly embodied in the following aspects: firstly, by converting text data into numerical feature vectors, the model can quantify information in the electroencephalogram signals, which is the basis of a machine learning algorithm for pattern recognition; secondly, the most useful features for the classification task are selected, so that the complexity of the model can be reduced, the calculation efficiency is improved, and the risk of over-fitting is reduced; features are screened through a statistical method or a machine learning method, it can be ensured that the model pays attention to signal attributes most related to electroencephalogram signal classification, and the situation that the learning process of the model is dominated due to the fact that certain feature factor value ranges are large is prevented.

The invention develops a sarcopenia prediction system for an adult group. The sarcopenia prediction method comprises the following steps: determining a crowd range and a standard included in the prediction system; determining a sarcopenia diagnosis standard and collecting outcomes in the target population; selecting a tree-based machine learning model according to the data features; screening the features by using a Boruta algorithm to construct a first set of screening model with high recall rate; screening features by using a pre-SHAP analysis method to construct a second set of auxiliary models for balancing the recall rate and the accuracy rate; optimal models of the system are determined by comparing model performance parameters, and finally prediction is achieved by combining the models. Features used by the system only depend on basic demographic features and physical activity questionnaires without clinical indexes, meanwhile, high performance is achieved in a model parallel connection mode, and the system is suitable for early screening detection of sarcopenia.

The invention discloses a plasma biomarker combination for auxiliary diagnosis of Alzheimer's disease and a screening method thereof. The screening method provided by the invention has the characteristics of high detection and analysis speed, high reproducibility and low sample consumption; the method is used for screening out a plasma biomarker combination consisting of A [beta] 1-42 / A [beta] 1-40, NFL, pTau181 and pTau217. The plasma biomarker combination realizes the efficient diagnosis performance of the Alzheimer's disease, the sensitivity is 85.5%, and the specificity is 98.1%. The method has an important application value.

This disclosure provides a model training method, a standard output selection method, a selection method, and an apparatus. It relates to the fields of Internet of Things (IoT) and artificial intelligence (AI). The model training method includes: acquiring target hard drive attribute information, wherein the target hard drive satisfies a target selection threshold, the target selection threshold including a threshold defined for at least one hard drive attribute information of the target hard drive; acquiring a first hard drive attribute expectation value defined for the target hard drive by a first service, wherein the first service has first service type information; inputting the target hard drive attribute information, the first service type information, and the first hard drive attribute expectation value into a prediction model to be trained to obtain a prediction selection threshold, wherein the prediction selection threshold is used when the target hard drive is applied to the first service; and training the prediction model using the target selection threshold and the prediction selection threshold.

The invention relates to the technical field of microbial screening, in particular to a method for high-throughput screening of flavone-transformed bacteria and a screening system thereof. The method comprises the following steps: firstly, inoculating a to-be-detected strain into a basal culture medium containing a flavone substrate for fermentation, then scanning fermentation broth at two different time points by using a full-wavelength microplate reader through a quartz microplate, calculating an absorbance difference value at at least one preset wavelength, and finally, calculating the absorbance difference value at least one preset wavelength. The strains with the absorbance difference larger than a preset threshold value are screened as primary screening positive strains. Compared with the traditional technology, the method has the advantages that the screening efficiency is improved by hundreds of times, the sample consumption is reduced by 75%, the industrial bottleneck that multi-structure flavonoids need to be repeatedly developed is broken through, and an innovative tool is provided for microbial resource development.

The invention discloses a sensitizing drug for resisting breast cancer by radiotherapy and a screening method and application thereof, and belongs to the field of biomedicine.The method comprises the steps that an in-vitro radiotherapy resistance model is constructed, a clinical segmentation radiotherapy dose increasing mode is simulated, stable acquired radiotherapy resistance is induced to be generated, and a carrier is provided for drug screening; natural plant small molecule compounds are used for treatment, and through concentration and time gradient dual control, the reversion ability of the natural plant small molecule compounds to radiotherapy resistance phenotypes is investigated; a multi-dimensional joint detection system is established, multi-level comprehensive evaluation is performed from cell survival and the like, the limitation of a single index is broken through, and the screening accuracy and reliability are improved; dNA damage repair is dynamically monitored through immunofluorescence, comet electrophoresis and the like, and the drug action mechanism is revealed. The medicine screened by the method is suitable for primary and acquired radiotherapy resistance patients, can effectively improve the radiotherapy curative effect, provides a new medicine source and scheme for clinical problems, and opens up the application of the compound in preparation of radiosensitizers or adjuvant therapy medicines.

The present invention discloses a high-throughput computational screening method for catalytic separation dual-function MOFs, belonging to the field of computational chemistry and nanocomposite catalytic materials. The screening method is to accurately and efficiently screen out promising MOFs materials with catalytic separation dual functions from the MOFs candidate pool through high-throughput calculations of multi-level theories step by step. The advantages of the present invention are: 1) the use of high-throughput parallel calculations effectively shortens the data acquisition time; 2) the use of high-throughput step-by-step calculations avoids the waste of resources caused by traditional experimentally intensive "trial and error" experiments, reduces research and development costs, and shortens the research and development cycle; 3) the screening method provided by the present invention takes into account both catalytic activity and the timely separation of by-products, and has good practical value.

This method includes a step of performing a whole genome analysis of the bacteria contained in the sample; and a step of comparing the base sequence of the obtained whole genome with the base sequence of SEQ ID NO: 1 or 2, and determining whether the bacteria in the sample are pan-resistant K. pneumoniae based on the degree of match.

The invention provides application of a biomarker in preparation of a product for predicting damp-heat syndrome of chronic atrophic gastritis, and the biomarker comprises TM7x bacteria, and further comprises at least one of kiwifruit bacteria, actinomycetes, twin coccus and phenylacetic acid. Compared with traditional gastroscopy, the tongue coating biomarker obtained by screening through a bioinformatics method provides a non-invasive diagnosis method, so that the compliance of a patient is improved. According to the method, the medical cost is reduced, the diagnosis feasibility and accuracy are improved, and the method not only can be used for diagnosing the chronic atrophic gastritis damp-heat syndrome, but also can be popularized to early screening and prevention.

Described herein are methods such as multi-omics methods for assessing a disease. The multi-omics methods may integrate intronic and proteomic, transcriptomic, genomic, lipidomic, or metabolomic data. The method screening diseases or disease states. Also described herein are methods for screening for diseases or disease states from biological samples. Also described herein are multi-omics databases and methods of using them.

The invention provides a medical clinical data processing method and system, computer equipment and a medium, and belongs to the field of medical data processing.The medical clinical data processing method includes the steps that demographic statistics features, lifestyle features, biochemical index features, physical examination result features and disease history features are extracted from a CHARLS data set, and an original feature set is constructed; randomly disrupting feature values of the original feature set, and constructing corresponding shadow features; importance scores are calculated, and significance features are determined according to the original feature set with the importance scores larger than the maximum value of the importance scores of the shadow features; and training an XGBoost model by taking the saliency characteristics as input and taking the risk probability of the ADL disability as output to obtain a prediction model capable of predicting the risk probability of the ADL disability. The saliency features screened by the method inevitably have non-linear strong correlation with the health result, the model obtained through training can assist doctors in health state diagnosis, and the problem that the extracted features are low in effectiveness when an existing model faces data related to sarcopenia is avoided.

The invention relates to the technical field of bioinformatics, and relates in particular to a translation efficiency prediction method, a screening method, a model training method, and an electronic device. The translation efficiency prediction method is implemented on the basis of a translation efficiency prediction model, and the translation efficiency prediction model comprises an embedding module, an encoding module, and a prediction module. The method comprises: acquiring sequence data of a messenger ribonucleic acid (RNA); extracting features of the sequence data via the embedding module, so as to obtain a feature representation of the messenger RNA; inputting the feature representation into the encoding module and performing convolution processing on the input feature representation via the encoding module, encoding a convolution result on the basis of an attention mechanism, and outputting an updated feature representation on the basis of an encoding result; and performing translation efficiency prediction on the updated feature representation via the prediction module, so as to obtain a translation efficiency prediction result.

The application provides methods for screening associated objects and recommending same style products. The method for screening associated objects includes: obtaining a first feature vector set corresponding to a first object and at least one second feature vector set corresponding to at least one second object; determining an object similarity between the first object and the second object based on multiple vector similarities between the feature vectors in the first feature vector set and the feature vectors in the second feature vector set, wherein the multiple vector similarities include vector similarity between the image feature vectors and the text feature vectors; and screening the at least one second object based on the object similarity to obtain an associated object of the first object. The solution proposed by this application improves the accuracy of determining object similarity, thereby allowing for more precise screening of the associated objects for the first object.

The application discloses a generator set artificial intelligence system machine learning sample screening method and device, obtains generator set operation data, adopts a preset cleaning method for one or more times of data cleaning on the generator set operation data, obtains a candidate sample set and an invalid sample set, if multiple times of data cleaning are adopted, different cleaning methods are adopted for each time of data cleaning, sorts the candidate sample set by adopting a preset sorting method, obtains a sorted sample set, extracts the sorted sample set according to a preset extraction method, obtains a learning sample subset, repeats the previous step for multiple times, adopts different sorting methods and the same or different extraction methods each time, extracts different learning sample subsets, and combines all the learning sample subsets into a learning sample set. The method is used for screening the generator artificial intelligence system machine learning sample, has low requirements on the professional quality of an operator, has high automation degree in a processing process, and has good extensive distribution of the screened learning sample.

To provide a dispersibility evaluation method, a screening method and a dispersibility prediction method of a dispersion system having three components of a powder, a dispersant and a dispersion medium as constituent components, and to provide a system capable of proposing a combination of the constituent components having high dispersibility.SOLUTION: One constituent component is specified from a powder, a dispersion medium, and a dispersant which are constituent components of a dispersion system, and then, with respect to a dispersant-treated powder obtained by surface-treating a powder with a dispersant in advance and an untreated powder for a plurality of constituent components, a mean particle diameter in a dispersion obtained by mixing each of the powders with the dispersion medium is measured, the suitability of each constituent component is determined according to a parameter based on the mean particle diameter, and a Hansen sphere forming a boundary between the two is obtained. The dispersibility is evaluated based on the distance between the center position coordinate of the Hansen sphere and the position coordinate of the Hansen solubility parameter of the specific component to be evaluated.SELECTED DRAWING: Figure 2

The invention discloses a transportation method and system for garment processing based on Internet of Things control, and relates to the technical field of Internet of Things control. Transportation environment data is collected and preprocessed; the environment position data, the AGV position data and the transportation target data are defined as nodes, node path edges are defined according to the distance between the nodes, and a three-dimensional communication topological graph is generated; determining a transportation path of the AGV from the initial position to the transportation target position according to the path edge weight in the three-dimensional communication topological graph, and generating an AGV transportation instruction; an AGV transportation instruction is executed, electromagnetic interference and physical obstacles of a transportation path are monitored, and bimodal obstacle avoidance is carried out on the electromagnetic interference and the physical obstacles; and after the bimodal obstacle avoidance is completed, measuring the three-dimensional space deviation between the actual arrival position and the target position of the AGV, and outputting a clothing adjustment measure. According to the method, the transportation path is screened by constructing the three-dimensional communication topological graph and the DFS method, and accurate transportation of the garment materials in the interference environment is achieved.

The invention provides a high-throughput screening method for screening an antibody adapted to TR-FRET application development. Conventional antibody screening methods include ELISA, SPR and the like, the methods are long in time consumption, high in sample demand and tedious in operation, and the screened antibodies are not necessarily suitable for TR-FRET application and development. The method is an antibody screening method developed and optimized based on the TR-FRET technology, and has the characteristics of simple steps, low sample demand, easy realization of high-throughput automatic detection and the like; and the technical methods are completely consistent, so that the screened antibody is completely suitable for TR-FRET application and development, and the affinity and matching degree are better. In addition, the antibody screened by the method is also suitable for ELISA and FACS detection application and development. The method can be used for screening research and development tool antibodies, diagnosis antibodies, antibody drugs, nano antibodies and the like, so that the sample consumption and the manpower demand are greatly reduced while the screening efficiency is improved and the project progress is accelerated.

A multilevel sensitivity analysis method for nuclear power plant main steam system model parameters based on Morris and GLUE comprises the steps that a main steam system model is operated, and parameter-observed quantity data pairs are generated through the model; screening out main parameters which have obvious influence on the output of the main steam system model through a Morri method; generating a plurality of groups of parameters in a parameter prior distribution range by adopting LHS, and performing GLUE method calculation; the convergence of the Morri method and the GLUE method is analyzed by calculating the Pearson's correlation coefficient, the KL divergence and the divergence difference delta KL; the method comprises the following steps: determining modeling parameters having significant influence on the output of a main steam system model, finding most sensitive key parameters by using a DBSCAN clustering method, evaluating the efficiency improvement performance of the main steam system model, and verifying the efficiency improvement effect during multilevel sensitivity analysis by comparing original calculation dimensions and calculating required time. According to the method, sensitivity analysis is performed on the model parameters of the main steam system of the nuclear power plant, the search space of parameter tuning is optimized, and the model optimization efficiency is improved.

The application belongs to the field of pharmaceutical analysis, and particularly relates to a method for establishing a traditional Chinese medicine fingerprint spectrum of Gastrodia and Uncaria Decoction. The method is realized through the following steps: firstly, preparing a reference solution and a sample solution, and then detecting by using high performance liquid chromatography to establish the fingerprint spectrum. The method provided by the application can effectively represent the change of the component quantity among samples by reasonably selecting the reference component and using the multi-wavelength method according to the maximum absorption wavelength of the component, so that the quality of the sample can be more accurately judged. The application further provides a preparation method screening and optimization method of Gastrodia and Uncaria Decoction, the prepared method has good repeatability and stability, and can more comprehensively monitor the quality of Gastrodia and Uncaria Decoction, and the method provides a valuable reference for the quality evaluation and standardization of traditional Chinese medicine decoction.

The invention relates to the field of resume data analysis systems, in particular to a talent feature extraction system and method based on resume data analysis, and aims to solve the problem that screening logic of the talent feature extraction system and method based on resume data analysis excessively depends on educational background indexes. Educational background weight setting is fixed and cannot be flexibly adjusted according to differentiated requirements of different posts for educational background; the problems that deep mining and quantitative evaluation of non-educational core features are lacked, and educational difference is difficult to make up through dimensions such as skill proficiency, work achievements and growth potential are solved; by setting a multi-mode educational background screening function and a characteristic weight self-defined adjusting mechanism, the most important rigid screening logic of an existing system educational background is broken through, candidates which do not reach the target educational background but have the capability reaching the standard are not directly screened out in a flexible mode, the situation that high-quality talents are missed due to limitation of the fixed educational background is effectively avoided, and the reliability of the system is improved. And the inclusiveness of talent selection is improved.

The invention relates to the technical field of pharmaceutical chemistry and bioinformatics, in particular to a drug molecular structure three-dimensional modeling and optimization processing method, which comprises the following steps: constructing a three-dimensional virtual space containing a pathogen and a drug agent; setting a connection logic based on distance and energy, so that the drug agents are self-assembled to form molecular clusters when meeting conditions; dynamic simulation is operated, and molecular clusters are driven to form space surrounding on the surface of the pathogen; calculating the surface coverage rate of the cluster to the pathogen key sites and generating a fitness score; and carrying out updating such as hydrophobic modification and the like on the drug molecule topological structure based on the score by utilizing an evolutionary algorithm, and iteratively generating an optimal structure. According to the method, a multi-agent cooperation mechanism is introduced into a calculation model, and drug self-assembly and virus space encirclement are incorporated into screening indexes, so that the method can discover a novel drug structure for physically blocking virus infection by forming a large-volume cluster, and the problem of screening leakage in a traditional method is effectively solved.

The present invention provides a method and system for pruning federated learning model layers based on differential privacy, wherein the method includes: measuring the correlation between the local model and the global model in the federated learning model based on the gradient sign and gradient size in the local model and the global model, screening the relevant layers in the local model with correlation higher than a preset correlation, and pruning the relevant layers with correlation lower than the preset correlation; using a Top‑K-based gradient sparsification method to screen the parameters with gradient values greater than the preset gradient value in the relevant layers with higher than the preset correlation for double pruning; considering the historical parameter information of the server-side global model in the previous round when aggregating the global model on the server side; using a preset local differential privacy mechanism to perturb the parameters of the local model and inject noise into the local model. This application reduces the uploaded model parameters and ensures that a small amount of noise is injected into the model, thereby achieving higher communication efficiency while ensuring that the model has good privacy and accuracy.

Described herein are methods such as multi-omic methods for assessing a disease such as cancer. The multi-omic methods may integrate proteomic, transcriptomic, genomic, lipidomic, or metabolomic data. The method screening diseases or disease states. Also described herein are methods for screening for diseases or disease states from biological samples. The methods may include assessing whether a nodule, mass, or cyst is cancerous.