19 results about "Regression modelling" patented technology

The invention discloses a power grid transformer technical renovation and overhaul time domain interval evaluation method. The method comprises the steps of 1, establishing an evaluation basic data set through multi-source data acquisition and data preprocessing; 2, screening key influence variables, quantifying variable influence weights by adopting regression modeling, and realizing accurate quantification of the health state of the transformer in combination with a TOPSIS algorithm; 3, forming a dynamic fault rate model; 4, constructing a system dynamics-Monte Carlo cooperation model, a fault tree-Bayesian network dynamic and static diagnosis model and a long and short term memory network time sequence prediction model; 5, setting an optimization target and a constraint condition, and outputting an optimal time point and a reasonable time domain interval of technical renovation and overhaul of the transformer by adopting a multi-target optimization solution method; and 6, model evaluation precision is tested back and verified through historical data, and key parameters of the model are subjected to feedback adjustment by adopting a particle swarm optimization algorithm. The method effectively solves the problems that a traditional method is low in evaluation precision, single in target and poor in dynamic adaptability, and is convenient to popularize and use.

The invention relates to a dam safety state intelligent prediction method and system based on regression modeling, and the method comprises the steps: carrying out the self-adaptive wavelet packet decomposition of structural data collected by a sensor, so as to extract a time sequence feature, and carrying out the differential coding of an environment variable, so as to generate a covariable feature; the method comprises the following steps: mapping a text event into vector features through a knowledge graph, inputting the features into a mixed primary function library formed by a physical primary function and a data primary function, and constructing a sparse regression model through a dynamic gating mechanism; a physical constraint term based on a stiffness matrix and an external load matching error is introduced into the model, a causal graph is dynamically updated through an incremental PC algorithm, and a causal regularization term is added into a loss function to improve the adaptability of the model. The method is suitable for dam structure health monitoring, intelligent early warning, long-term trend analysis and other application scenes, and has good real-time performance, interpretability and small sample generalization ability.

The invention discloses a regional resident load meteorological sensitivity analysis method, device and equipment based on fused meteorological indexes, and a medium, and relates to the field of power load characteristics, and the method comprises the steps: constructing a multi-factor fused meteorological index model of a single meteorological point, determining a multi-meteorological point fusion meteorological index of the target area according to the resident load proportion corresponding to the target area and a multi-factor fusion meteorological index model, and then constructing a target data set according to the multi-meteorological point fusion meteorological index and the resident load data; and segmenting the target data set to perform piecewise nonlinear regression analysis on the obtained target first data set and the target second data set, and determining the current residential load meteorological sensitivity of the target area through a target piecewise regression analysis function obtained through analysis and the target data set. Therefore, nonlinear segmented regression modeling and sensitivity analysis of the resident load and the fused meteorological indexes in the target area can be realized, and support is provided for refined analysis and management of the resident load.

The invention discloses an agricultural electricity consumption prediction method and system based on solar term segmentation, and the method comprises the following steps: segmenting solar terms; on the basis of correlation analysis of a cause mechanism of agricultural production electric quantity and historical data, the following seven types of main characteristic variables are selected to construct a prediction model, regression modeling and parameter optimization, namely, a set of multiple linear regression model and a solar section model calling and prediction process are independently established for a data set of each solar section. According to the method, the whole year is divided into a plurality of solar sections based on solar change, the segmentation model is established in combination with the power consumption mode of each stage, three types of characteristic variables of weather, economy and power are fused, modeling is performed by using the multiple linear regression model, and the method has a clear application boundary, strong logic interpretation capability and good small sample adaptability; the method can be widely deployed in agricultural main production areas, rural power grid dispatching centers, power transaction agent platforms and other scenes, and the electric quantity prediction precision and agent power purchase market benefits are improved.

A system and method for Medical Claims Risk Score (MCRS) algorithmic underwriting includes a predictive machine learning model configured to generate underwriting decisions on electronic applications. MCRS underwriting applies word embedding modeling, such as GloVe (global vectors), to transform high dimensional MC records into single-code word vectors. These single-code word vectors are employed in regression modeling, and may include summarized embedding coordinates aggregated at the applicant level. Regression modeling uses medical claim codes data and underwriting decision data stored for historical underwriting applicants to train a random forest model to predict relative mortality risk for underwriting applicants. A risk rating may be derived from the underwriting decision data based upon standard quantitative risk ratings of a plurality of risk classes. Other inputs to the random forest model may include cohort level applicant profile data, such as applicant issue age and sex.

The invention provides an LDL-C concentration estimation method based on an interpretability enhancement machine, and belongs to the technical field of medical detection and artificial intelligence. The LDL-C concentration estimation method comprises the following steps: selecting a subject with a complete blood fat detection record from a preset database to obtain initial data; performing data cleaning on the initial data, and longitudinally merging the blood fat measurement data under the two visits to obtain a data set; performing regression modeling on the EBM based on the training set and the test set to obtain an EBM model; calculating a contribution value of each input ratio variable by constructing a square view of a lookup table based on an EBM model, adding all the contribution values, and calculating a final LDL-C predicted value through a connection function g; and performing comprehensive evaluation and optimization on the EBM model by adopting a plurality of statistical indexes, and integrating the EBM model into various medical systems. Through high-quality data, an optimized data processing strategy, an accurate EBM prediction model, an interpretable prediction result and efficient system integration, accurate estimation and wide application of LDL-C are realized, and the method is superior to a traditional formula calculation method.

This application discloses a method, system, device, and medium for production log analysis based on the Industrial Internet of Things (IIoT), relating to the technical field of IIoT. The method includes: extracting an environmental parameter set, a standard setting parameter set, and an actual processing parameter set corresponding to the target production line from the production logs; performing regression modeling based on the environmental parameter set, the standard setting parameter set, and the actual processing parameter set to obtain an initial SVM model; using the standard setting parameter set, the environmental parameter set, and the actual processing parameter set as model training data, and training and testing the initial SVM model based on the model training data to obtain an actual processing parameter prediction model; acquiring the environmental parameters of the area where the target production line is located and the processing parameters of the products, setting standard setting parameter constraints, calling the actual processing parameter prediction model, and selecting a set of standard setting parameters based on an annealing algorithm. This application has the effect of improving production efficiency.

The application discloses a two-stage large language model energy consumption analysis method and system for reasoning tasks, and belongs to the technical field of data mining.The application effectively solves the modeling problem of key dynamic characteristics by adding a response token number interval prediction stage, so that the energy consumption model can indirectly use output length information, and innovatively constructs and fuses enhanced features such as model total parameter count and theoretical floating point operation number, which have clear physical meanings.The system integrates multi-dimensional heterogeneous information such as prompt words, models and hardware, and realizes fine description of energy consumption differences of different reasoning tasks.The proposed two-stage framework of classification prediction and regression modeling has clear logic and is consistent with the energy consumption characteristics of the LLM reasoning process.The first stage focuses on predicting features strongly related to output, and the second stage performs comprehensive energy consumption regression, so that the model structure is reasonable and has strong interpretability.

The embodiment of the invention relates to the technical field of data processing, in particular to a performance testing method and device, equipment and a storage medium, and aims to effectively perform performance testing on a regression modeling process. The method comprises the following steps: creating a model training task through a task creation interface; determining task information corresponding to the model training task from a data source file; determining a model corresponding to the model training task according to the task information; the model training task is executed in a multi-thread concurrent mode within a preset time period; and determining a performance index of the model training system according to the execution information of the model training task.

The invention belongs to the technical field of traffic geographic information analysis, and discloses an urban crowd activity and space structure dynamic deduction method and system based on single traffic flow, a terminal and a storage medium. Identifying a spatial distribution mode through a flow similarity measurement method in combination with a hierarchical clustering algorithm; carrying out quantity aggregation on the identified spatial distribution mode according to grid units, and constructing a spatial grade distribution model; combining the urban built environment features with spatial grade distribution in the spatial grade distribution model, and constructing a comprehensive data set suitable for machine learning regression modeling; and based on the comprehensive data set, in combination with a machine learning model interpretation method of a game theory, quantitatively analyzing a nonlinear influence mechanism of urban built environment characteristics on different spatial distribution modes, and deducing internal relevance between urban crowd activities and spatial structures. The dynamic evolution process of the urban space structure is comprehensively realized.

The invention discloses a typical tree species growth prediction method, system and device based on multi-stage multi-factor regression and a medium, and relates to the technical field of tree species growth prediction.The method comprises the steps that multi-source heterogeneous data are collected and preprocessed; dividing the independent variables into different types of influence factors, and performing statistical test on each factor to obtain a preliminary candidate variable set; calculating a variance expansion factor of the candidate variables, and when the expansion factor exceeds a threshold value, reducing the correlation among the preliminary candidate variables by adopting a collaborative path method to obtain a final variable; performing regression modeling through a three-stage modeling method based on the final variable to generate a regression model, and establishing a regression sub-model for each partition; and based on the obtaining mode of the final variable, extracting a judgment rule of the tree species and outputting the judgment rule in a structured format. According to the method, high-precision prediction of the growth under multi-factor driving can be realized, and the problems of multi-collinearity, unstable variable selection, insufficient nonlinear structure expression and the like in a traditional regression model are solved.

The invention discloses a regression modeling method and device based on a multi-binning scheme, equipment, a medium and a product, and relates to the technical field of financial risk control, and the regression modeling method based on the multi-binning scheme comprises the steps: obtaining each original feature variable in financial customer data, determining at least two binning schemes respectively corresponding to the original characteristic variables and a WOE coding rule of each binning scheme; converting each original characteristic variable based on each WOE coding rule to obtain a plurality of WOE variables corresponding to each original characteristic variable; screening the plurality of WOE variables corresponding to each original characteristic variable to obtain a target WOE variable corresponding to each original characteristic variable; and performing regression modeling according to the target WOE variable of each original feature variable to obtain a target risk control model, each target binning scheme and weight information. According to the method, the constraint of a traditional single binning scheme is broken through, and the risk expression ability of the original feature variables is improved.

A system and method for regression modeling and mapping an interior volume of a fluid containment vessel and interpolating data from multi-point sensor arrays within the fluid containment vessel to detect conditions across the interior volume of the fluid containment vessel. The interpolated data may then be used to control operating equipment associated with the fluid containment vessel to modify the conditions within the fluid containment vessel.

The invention discloses an inference task-oriented two-stage large language model energy consumption analysis method and system, belongs to the technical field of data mining, and effectively solves the modeling problem of key dynamic characteristics by adding a response token number interval prediction stage, so that an energy consumption model can indirectly utilize output length information, and the analysis efficiency is improved. The method innovatively constructs and fuses model total parameter counting, theoretical floating point operation times and other enhanced features with clear physical significance, the system integrates cue words, models, hardware and other multi-dimensional heterogeneous information, fine description of energy consumption differences of different reasoning tasks is realized, and a proposed classification prediction and regression modeling two-stage framework has the advantages of high efficiency and high reliability. And the logic is clear and accords with the energy consumption characteristics of the LLM reasoning process. The first stage focuses on predicting and outputting strongly correlated features, the second stage carries out comprehensive energy consumption regression, and the model is reasonable in structure and strong in interpretation.

The present invention provides a blood pressure measurement system and method based on the dynamic characteristics of oscillation waves and regression modeling, which relates to the field of wearable medical health monitoring technology. With signal processing methods as the main body, the time domain, frequency domain, time-frequency and statistical methods are used to extract the key features of the oscillation waves, and then a lightweight deep learning model is used to fine-tune the traditional features. Finally, a mapping relationship between the characteristic band changes and the blood pressure values is established through machine learning. It can not only predict the complete oscillation wave waveform through the characteristic changes of part of the oscillation wave to achieve low-pressure comfortable measurement, but also realize personalized blood pressure measurement by identifying the mutation points of the oscillation wave characteristics. The present invention not only ensures the interpretability and stability of the traditional method, but also uses lightweight deep learning fine-tuning to improve the feature recognition accuracy and individual adaptability, and can realize edge computing, thereby reducing measurement time, improving user comfort, measurement accuracy and system practicality.

The invention relates to the technical field of construction data acquisition, and particularly discloses a construction data acquisition method for construction full-period automatic management, which comprises the following steps: step S1: based on historical cable laying area data, dividing different laying scene categories, establishing a mathematical relationship between the scene size and the cable laying shape through clustering analysis and regression modeling, and providing a reference model for prediction; s2, after the initial path of the cable is planned, electronic tags are arranged according to a preset interval and are divided into sub-paths, sub-regions are divided by combining the tag distance and the length of the sub-paths, and the cable bending shape of each sub-region is predicted by using a basic relational expression; and S3, summarizing the prediction results of the sub-regions to calculate the total length of the cable, comparing the total length with the actually measured distance of the electronic tag, dynamically correcting the bending shape parameters if the deviation exceeds a threshold value, and finally outputting the total length of the high-precision cable to ensure the accuracy of construction material estimation and cost control.

The invention discloses an ironmaking process anomaly detection method based on a time kernel stationary generalized learning system. The method comprises the following steps: expression of nonlinear kernel generalized features, optimization target construction and solution based on the time kernel stationary width learning system, real-time modeling and anomaly detection strategy, and independent incremental learning mechanism. Firstly, a nonlinear kernel generalized representation extraction strategy is established, then a time matching mechanism between input and output of a model is explored through a time alignment parameter, and the parameter can be explained under a potential variable relation. In the integration stage, a Kullback-Leibler divergence objective function is established, so that a stationary relation in time sequence data is conveniently captured, and regression errors are combined. Then, a double-loop parameter optimization algorithm and an independent incremental learning mechanism are provided, and when additional data are collected, the independent incremental learning mechanism is utilized to maintain the mutual independence of an original model and an updating model, so that the long-term updatable regression modeling and monitoring capability is maintained.

The invention discloses a production log analysis method, system and device based on the industrial Internet of Things and a medium, and relates to the technical field of the industrial Internet of Things, and the method comprises the steps: extracting an environment parameter set, a standard setting parameter set and an actual processing parameter set corresponding to a target production line according to a production log; performing regression modeling according to the environment parameter set, the standard setting parameter set and the actual processing parameter set to obtain an initial SVM model; the standard setting parameter set, the environment parameter set and the actual machining parameter set serve as model training data, the initial SVM model is trained and tested according to the model training data, and an actual machining parameter prediction model is obtained; environment parameters of the area where the target production line is located and machining parameters of the product are obtained, standard setting parameter limitation is set, an actual machining parameter prediction model is called, and a set of standard setting parameters are screened out based on an annealing algorithm. The method has the effect of improving the production level.

The invention provides a new energy electricity price prediction system, which comprises a data acquisition module for acquiring historical electricity market data, new energy power generation data, meteorological data, load and economic data and policy parameters, a data preprocessing module for performing missing value filling and abnormal value elimination on the acquired data, lagging factors, seasonal indexes and policy influence factors are generated; the model integration module comprises an LSTM-Transform hybrid model, a reinforcement learning model and a time sequence prediction model; the transaction simulation module executes medium and long term contract segmentation regression modeling and spot market marginal clearing calculation, the segmentation regression modeling divides a price gradient interval based on a contract electric quantity proportion, and the marginal clearing calculation is combined with a unit clearing curve and a tie line transmission constraint; and the decision optimization module outputs weighted comprehensive electricity prices and energy storage scheduling strategies under different contract proportions. The system realizes quantitative fusion of policy parameters, meteorological factors and market data, and breaks through the external variable integration bottleneck of a traditional time sequence prediction model.