8 results about "Survival data" patented technology

The invention belongs to the technical field of biological medicine, and discloses a portal cholangiocarcinoma postoperative risk prediction model based on a gamma-glutamyltransferase (GGT) dynamic change track and a construction method thereof. The method comprises the following steps: (1) collecting clinical characteristic data and survival characteristic data (including survival time and survival state) of a patient suffering from porta hepatis cholangiocarcinoma subjected to radical resection; 2) on the basis of GGT detection results at different time points in a perioperative period, dividing patients into different GGT dynamic change track types by using a latent category hybrid model; and 3) taking the clinical characteristic data and the GGT dynamic change track type as independent variables, taking survival data as response variables, screening prediction factors through LASSO regression, and constructing a multi-factor Cox regression model, namely a postoperative risk prediction model. According to the method, the dynamic evolution rule of the GGT is introduced, so that the limitation of single-time static detection is made up, the quantitative evaluation of the survival risk is realized, and a scientific basis is provided for individualized treatment and follow-up visit management.

The application discloses a kind of based on artificial intelligence's pharmaceutical clinical test safety evaluation method, comprising the following steps: collecting heterogeneous clinical survival data with deletion characteristics and time series drift characteristics, and constructing original survival data set;Based on propensity score model and survival outcome model, implement deletion correction, obtain individual non-compliance risk score;Through stratified weighting rule and two-side order-preserving prediction mechanism, calibrate and update individual risk prediction interval;Calculate individual risk upper and lower bound probability, define three-domain risk threshold rule, generate real-time risk decision result through three-domain decision mapping and dynamically evaluate overall safety trend.The application improves the robustness and interpretability of clinical safety risk decision.

The invention relates to the technical field of feature screening, in particular to a multi-label data feature screening method and system based on an immune genetic algorithm. Each label population is constructed by single-label prior features, and recognition of complex information between features-features, features-labels and labels is realized through multi-population parallel operation, immunization, elite retention, concentration inhibition and global feature migration strategies. Therefore, the optimal feature subset screened out by each label has a good prediction effect on the corresponding label result, and the accuracy of multi-label medical disease prediction is improved. By constructing classification and survival multi-label models and calculating chromosome fitness values, the method can be simultaneously suitable for feature screening of medical multi-label classification and survival data.

The invention provides a model hypothesis-independent survival analysis marker discovery method and device, and the method comprises the steps: collecting the follow-up survival data of a plurality of patients, and building a survival data set, samples of the data set comprise multi-dimensional features of biomarkers reflecting prognosis conditions of patients, actual observation time and labels indicating whether interested events are observed or not; constructing a model-free survival analysis marker selection optimization model, wherein the optimization model comprises a nucleation survival outcome dependency module for quantifying a dependency relationship between each feature and a survival outcome and a nucleation minimum redundancy module for eliminating redundant features; and solving the optimization model by utilizing the survival data set to obtain a screening result of the biomarker for predicting the prognosis condition of the patient. According to the method, the biomarker combination with high predictability and low redundancy can be effectively screened out under the condition that a preset model form is not needed, so that the discrimination capability and interpretability of the biomarker combination are improved.

The invention discloses an information processing method, device and equipment for individual prognosis gene recognition and a medium, relates to the field of artificial intelligence and biomedicine crossing, is applied to a computer device, and comprises the following steps: screening out undetermined genes belonging to a shared causal gene set from a cross-section gene sequencing sample of a current object; the shared causal genes in the set are genes obtained by performing survival distribution simulation and independence test on cross-section genes and survival data of historical objects to eliminate hybrid associated genes; inputting the actual expression quantity of the to-be-determined gene into the target gene expression prediction model to output the expected expression quantity of the to-be-determined gene, and determining the difference between the actual expression quantity and the expected expression quantity of the to-be-determined gene as the residual error of the to-be-determined gene; and comparing the total gene expression residual distribution with the residual of the to-be-determined genes so as to identify the individual prognosis genes of the current object exceeding a preset group normal fluctuation range from the to-be-determined genes. And completing accurate individual prognosis gene identification suitable for each group on the basis of the cross-section gene data.

The application discloses application of a group of biomarkers in preparation of products for survival prognosis prediction of non-small cell lung cancer patients, wherein the biomarkers contain PD-L1, CTLA-4, CD226, Nectin-2 and / or CD276. And based on immunohistochemical semi-quantification of the above biomarkers in non-small cell lung cancer patients, combined with survival data of non-small cell lung cancer patients, LASSO-Cox regression analysis is conducted to construct a risk score formula, which can effectively distinguish non-small cell lung cancer patients into high-risk patients and low-risk patients, and the high-risk patients distinguished are non-small cell lung cancer patients who can benefit from chemotherapy combined with immunotherapy.

The present application relates to the technical field of diabetes complication health management, in particular to a diabetes complication management method and system based on multi-modal data fusion and reinforcement learning, the method acquires continuous glucose monitoring high-frequency data, low-frequency discrete clinical indicators and survival data; uses a sliding window and a time sequence alignment algorithm to perform feature extraction and interpolation, and constructs a time-varying covariant state vector; uses a Cox proportional risk model containing an L1 regularization term to solve the optimal regression coefficient and calculate the instantaneous risk ratio, and then constructs a dynamic nomogram model to map the disease-free survival probability; the state vector, survival probability and risk ratio are input into the policy network of a reinforcement learning intelligent agent as the environment state, and the optimal intervention strategy is output in the preset action space; finally, the risk improvement indicators of the last moment and the current moment are used as reward signals to update the network parameters in a closed loop. The present application solves the problem of heterogeneous data fusion and realizes safe and personalized dynamic closed-loop health management.

This application discloses a data processing method, apparatus, device, storage medium, and program product. The method includes: performing self-supervised contrastive representation learning based on multiple samples to obtain a first encoder; the multiple samples include unlabeled image data of multiple users, and the first encoder is an updated base encoder; performing supervised contrastive representation learning based on the first encoder and supervisory information to obtain a second encoder; the supervisory information includes survival data of the multiple users, the survival data including temporal event data of censored and uncensored samples, and the second encoder is the updated first encoder; determining a prognostic assessment model based on the second encoder; the prognostic assessment model is used to predict the prognostic risk of users; performing prognostic analysis on users based on the prognostic assessment model, and outputting the prognostic analysis results.