11 results about "Statistical difference" patented technology

The invention discloses a mass control method and system based on mass spectrum flow data. According to the method, whether signal distribution before different data has good consistency or not can be determined through signal distribution visualization; the MAD can measure whether the distribution of signals in the channel is concentrated or discrete; various statistical index calculations can know whether there is a large gap between the same channel and different data; calculating the proportion of each subgroup in each data through FlowSOM clustering, and analyzing whether the proportions of the same subgroups between different batches have statistical differences or not by combining batch information; and calculating an AOF value of each channel in each piece of data in combination with the subgroup information, and judging whether the marker dyeing effect of the channel is normal or not through the AOF value. According to the method, by comprehensively applying multiple statistical methods, the quality of different batches of CyTOF data can be efficiently and accurately evaluated, potential system errors or batch effects can be identified, and optimization suggestions are provided, so that the reliability of single-cell proteomics research is improved.

The application is an ultrasonic signal disturbance suppression method based on covariance whitening method. The application relates to the technical field of medical ultrasonic sensor signal disturbance suppression, and establishes a second-order statistical model of array elements in the channel by using original RF data collected in multiple probe states, and constructs a whitening transformation or a reference domain covariance alignment transformation in the channel dimension, so that statistical changes related to probe disturbance are normalized and suppressed. The application normalizes and aligns the second-order statistical differences caused by the probe disturbance in the original RF channel space, so as to reduce the sensitivity of subsequent signal analysis, feature extraction, classification model or regression model to non-target acquisition state factors. The corrected data still maintains the original two-dimensional RF matrix form, and when used for traditional beam forming or image reconstruction, diagonal whitening, local strip approximation, scale recovery or double branch processing mode can be adopted to give consideration to statistical stability and array element physical consistency.

The application provides a kind of based on Shearlet domain continuous data prestack consistency processing method, to the data of each block in continuous exploration area respectively carry out surface consistency amplitude compensation and surface consistency deconvolution and other consistency pretreatment, select a processing target model, the application selects a certain amount of good quality data as target model in the block of relatively good quality data in continuous exploration area, the other data in continuous exploration area is corrected as the standard of the target data, estimate the consistency correction factor of Shearlet domain, eliminate the statistical difference between target model data and data to be processed, realize the consistency processing of continuous exploration area. Process prestack data, can eliminate the difference of continuous data in amplitude, frequency and space, has good stability under noise environment.

The invention discloses an accelerated degradation credibility evaluation method based on KL divergence, and relates to the technical field of reliability engineering and life evaluation, and the method comprises the steps: selecting a performance degradation model and an acceleration model which are adaptive according to a product degradation mechanism, constructing an accelerated degradation model, and extrapolating the accelerated degradation model to model prediction distribution of degradation amount under an actual working condition; collecting actual measurement data of the degradation amount of the product under an actual working condition, selecting typical probability distribution as an alternative model, and selecting the alternative model with the maximum log-likelihood function value as actual measurement fitting distribution of the degradation amount; and realizing credibility evaluation of the accelerated degradation test based on the statistical difference between the model prediction distribution and the actual measurement fitting distribution of the KL divergence measurement degradation amount. Therefore, by adopting the accelerated degradation credibility evaluation method based on the KL divergence, the problem that a credibility evaluation index of a traditional accelerated degradation test is insensitive to distribution tail differences is solved, and a quantitative basis is provided for correcting an accelerated degradation model and improving extrapolation accuracy.

The application discloses a mass spectrometry flow data-based quality control method and system. The application can determine whether the signal distribution of different data is consistent by signal distribution visualization; MAD can measure whether the distribution of signals in the channel is concentrated or dispersed; various statistical index calculations can know whether there is a large gap between different data in the same channel; by FlowSOM clustering, the proportion of each subgroup in each data is calculated, and then, whether the proportion of the same subgroup between different batches has a statistical difference is analyzed in combination with batch information; in combination with the subgroup information, the AOF value of each channel in each data is calculated, and whether the marker dyeing effect of the channel is normal is judged through the AOF value. The application can efficiently and accurately evaluate the quality of different batches of CyTOF data, identify potential systematic errors or batch effects, and provide optimization suggestions, so as to improve the reliability of single-cell proteomics research.

The present invention relates to a method, storage medium and device for detecting mutation points of a climate mode index sequence. The observation sequence data is read, the cutoff length and probability level of the climate mode are determined, and the statistically significant differences between different modes are calculated. A loop is started to search for mutation points. If the value of a certain point is significantly different from that of the previous climate mode, the point is considered to be a candidate point for climate mode mutation, otherwise it is moved to the next point. After the candidate point is determined, the mode conversion index of the point is calculated. If the index is negative, it means that the mode mutation fails. If all are positive, it means that the mutation is successful, and the point is officially confirmed as a mode mutation point. The new mode is used as the reference mode, and the above operation is continued until the end of the observation sequence, and all mode mutation points are detected. The present invention solves the problem that the detection performance of the existing method drops sharply at the end of the sequence. The proposed method can detect the climate mode transfer as early as possible and then monitor the change of its amplitude over time. The method is simple to operate, and the user can change the input parameters according to the needs.

A method for bias detection in large language models (LLMs) is disclosed. The method includes receiving a plurality of contrastive questions for a plurality of contexts. A prompt including at least two questions of the plurality of contrastive questions may be received. An LLM may be applied on the prompt to generate a set of reasonings associated with the at least two questions and a set of scores associated with the set of reasonings. Statistical hypothesis testing model may be applied on the set of scores. It may be determined whether the at least two questions are statistically different. A set of biases associated with the LLM may be detected, based on the statistical difference. Rendering of first information including the set of biases may be controlled.

The invention relates to the technical field of data processing, in particular to an economic scene quality evaluation method and device, equipment and a storage medium, and the method comprises the steps: processing an economic scene historical sample based on a preset scene generation model to obtain a scene sample, and obtaining a simulation sample based on the scene sample; obtaining first statistical comparison information based on the economic scene historical sample and the simulation sample, wherein the first statistical comparison information comprises a first probability density comparison result, a first cumulative probability density comparison result and a first quantile comparison result; establishing a first statistic group based on the economic scene historical sample, and establishing a second statistic group based on the simulation sample; and obtaining first statistical magnitude difference information based on the first statistical magnitude group and the second statistical magnitude group. The method is convenient for determining the quality condition of the economic scene output by the economic scene generator.

The invention discloses a brain disease unbalanced data classification method based on feature difference enhancement. The recognition capability of minority class samples is cooperatively improved through contribution degree driven difference enhancement and a dynamic cost sensitive mechanism. The method comprises the following steps: preprocessing rs-fMRI data, and constructing a functional connection matrix; pre-training an auto-encoder according to categories to perform feature reconstruction, and generating a contribution degree matrix by using layer-by-layer correlation propagation; calculating an inter-class statistical difference, and constructing a difference matrix to enhance the feature expression of discriminative connection in the graph neural network; inputting the enhanced features into a graph classification model to obtain a preliminary result; a self-adaptive cost matrix algorithm is adopted, the loss weight is dynamically adjusted by integrating the intra-class dispersion degree and the inter-class separation degree, and the model is guided to focus on minority classes in cooperation with a difference enhancement mechanism. According to the method, a systematic scheme for unbalanced resting state functional magnetic resonance data classification is constructed, and the recognition capability of the model for minority class samples and the overall classification performance are remarkably improved.

The invention discloses a method, equipment and medium for auxiliary identification of acute A-type aortic dissection. The method comprises the following steps: step S1, acquiring detection data of volatile organic compounds VOCs in a serum sample detected based on a gas chromatography-ion mobility spectrometry GC-IMS technology; s2, determining the types of VOCs, carrying out statistical analysis, and screening out characteristic VOCs with statistical differences between the acute A-type aortic dissection group and the control group; s3, on the basis of the screened characteristic VOCs, adopting a machine learning algorithm to construct a classification model for identifying the acute A-type aortic dissection; s4, screening an optimal classification model, and optimizing the optimal classification model to obtain an optimized recognition model; and S5, processing the VOCs data of the serum sample of the to-be-detected object by using the optimized recognition model, and outputting an auxiliary recognition result of the acute A-type aortic dissection.

The present application relates to the technical field of bearing fault diagnosis, and particularly relates to a bearing fault migration diagnosis method based on high-order statistical difference. The method comprises the following steps: collecting vibration signals of a rolling bearing under a source domain working condition as a source domain data set; performing sliding window processing to obtain a plurality of sample segments; a one-dimensional convolutional neural network extracts deep feature representation of the sample segments; a maximum mean r-order difference measure MMRD is calculated in a reproducing kernel Hilbert space; a fault diagnosis model and a feature extractor are trained based on a joint loss function to obtain a fault diagnosis model with domain-invariant characteristics. The present application can improve the feature distribution alignment capability in cross-condition mechanical fault migration diagnosis, thereby improving the fault recognition accuracy and stability of the model under different working conditions.