36 results about "Linear prediction model" patented technology

The invention provides a Prony algorithm-based cable segment wave velocity acquisition method, device and system, and a medium, and the method comprises the steps: testing a to-be-tested cable, and obtaining a cable signal transfer function; on the basis of the cable signal transfer function, in combination with cable joint distribution, constructing a cable signal approximation function based on a Prony method; constructing a linear prediction model of the cable signal based on the cable signal approximation function; performing denoising processing on the linear prediction model based on a singular value decomposition method to obtain a denoised linear prediction model; solving the denoised linear prediction model to obtain an attenuation coefficient; and calculating the segmented wave velocity of the cable based on the attenuation coefficient. According to the method, the Prony estimation method is combined with the singular value decomposition noise reduction algorithm, high-precision extraction of attenuation constants and segmented wave velocity decoupling are achieved, and then the electrical distance positioning precision is improved.

This application relates to a real-time power load forecasting method and system based on a hybrid model. The method includes: acquiring current cycle load data and combining it with historical load data to form a load sequence, and acquiring corresponding weather data; performing timestamp alignment, anomaly processing, and normalization on the load sequence and weather data to obtain a standardized input sequence; determining the order parameters of the linear forecasting model and the set of hyperparameters to be optimized, consisting of the network structure and training parameters of the nonlinear forecasting model, and optimizing them through a combined optimization algorithm to update the training configuration of the two models; outputting the first and second forecast sequences for the next cycle in the current cycle, respectively, and using the second forecast sequence as a trend term to compensate the first forecast sequence to obtain a fused forecast sequence; acquiring the fused forecast sequence for the current cycle from the previous cycle, constructing a residual sequence with the current cycle load data and determining the deviation term, calibrating the current cycle fused forecast sequence online, and outputting the calibrated load forecast result.

This invention discloses a method for calculating needle-free injection depth based on energy analysis, comprising the following steps: Step 1, obtaining the jet process parameters during a single needle-free injection, and calculating the initial peak jet power and the total jet energy of this injection; Step 2, calculating the initial injection depth based on the initial peak jet power using a preset initial depth linear model; Step 3, calculating the diffusion depth increment based on the total jet energy using a preset diffusion depth linear model; Step 4, calculating the predicted maximum injection depth based on the sum of the initial injection depth and the diffusion depth increment. This invention decomposes the injection process into two stages: initial impact and subsequent diffusion, and establishes linear prediction models for each stage based on the jet energy parameters, thereby achieving reliable and accurate calculation and prediction of the injection depth for large-volume needle-free injections.

The invention discloses a circuit breaker actuation time prediction method and system, a storage medium and electronic equipment. The method comprises the following steps: constructing a multi-parameter nonlinear prediction model of a nonlinear relationship between multiple characteristics of a circuit breaker and actuation time; determining a key feature based on a rate of change of each of the plurality of features of the circuit breaker; constructing a grid on a plane based on the key features, and performing prediction by using a multi-parameter nonlinear prediction model based on feature data corresponding to each point on the grid so as to determine an action time prediction value corresponding to each point on the grid; calculating a partial derivative based on the action time predictor to determine a gradient corresponding to each point; and obtaining target key feature data, and based on the target key feature data, the motion time prediction value corresponding to each point on the grid and the gradient corresponding to each point, performing motion time prediction by using a two-point cubic Hermite interpolation algorithm, and obtaining a motion time prediction value corresponding to the target key feature data.

The invention provides a model prediction control method for a paper pulp kappa number in a batch cooking process, and belongs to the technical field of pulping and papermaking, the method comprises the following steps: firstly, collecting multi-dimensional measurable state variables related to delignification in the cooking process to form an original state vector; then, utilizing a Koopman operator and extended dynamic mode decomposition to map the original state vector to a high-dimensional linear space to obtain a dimension raising state vector; a Koopman global linear prediction model is established on the basis of historical data offline training, and a free liquid temperature setting sequence in a future control time domain is optimized in a rolling mode through a model prediction controller; the first control quantity of the temperature sequence is applied to the cooking process in real time after each time of optimization, and rolling optimization is repeatedly executed in combination with feedback correction in the next sampling period, so that the problem of steady-state deviation of an end-point kappa value in batch cooking application is solved, and the adaptability to model mismatch and the robustness of a control system are improved.

This invention relates to the field of intelligent manufacturing and industrial data analysis technology, specifically to a method, system, and storage medium for identifying production bottlenecks based on multi-dimensional data fusion. The method includes: acquiring production data and constructing a multi-dimensional data vector; updating the weights of a linear prediction model based on the multi-dimensional data vector; performing multi-dimensional data fusion using the updated weights to obtain a comprehensive index value; obtaining a production bottleneck severity score based on the comprehensive index value; determining the production bottleneck status based on the production bottleneck severity score and the comprehensive index value; and identifying factors influencing the production bottleneck based on weight changes. This invention dynamically adjusts the weights using a gradient descent method to score the bottleneck, and dynamically updates the judgment threshold using an exponentially weighted moving average. Finally, it performs multi-dimensional cause analysis by comprehensively considering the weight contribution, change trend, and data anomaly, achieving adaptive and accurate location and cause identification of production bottlenecks.

The application discloses a method for predicting total intestinal digestible calcium and phosphorus content of growing pigs based on a linear model, and relates to a method for estimating total intestinal digestible calcium and phosphorus content, comprising the following steps: S1, arranging collected data sets through a preprocessing module to obtain an initial data set containing input features and target variables; the target variables include total intestinal digestible calcium content ATTD_Ca and total intestinal digestible phosphorus content ATTD_P measured through a digestion test; S2, performing feature analysis and screening on the initial data set through a feature engineering module to obtain a data sample constructed by a feature subset adapted to the target variables; S3, respectively constructing linear prediction models related to the prediction of each target variable, and training, optimizing and verifying the linear prediction models by using the data sample; and S4, importing feature data of a new feed formula to be predicted, selecting the data sample through S1-S2, and obtaining a prediction value of the target variable through the linear prediction model.

The application discloses a kind of solubility carbon black electron transfer capacity quantitative prediction method, more than three temperature is selected and is prepared carbon black using biomass, is configured into mixed solution by adding water, mixed solution is cultured by shaker, solution is centrifuged, supernatant is filtered, and solubility carbon black solution is obtained;The concentration of solubility carbon black solution is determined, and the cyclic voltammetry curve of carbon black solution is determined using electrochemical workstation and three electrode system, the peak potential corresponding to each scanning speed in the curve, peak current are obtained, according to scanning rate, peak potential, peak current, solubility carbon black solution concentration, electron transfer rate constant k 0 It is calculated by using Bulter-Volmer equation, linear prediction model is constructed by the logarithmic value of electron transfer rate constant k 0 And pyrolysis temperature, the electron transfer capacity of carbon black at different temperatures is predicted using the model;The method is simple and easy to operate, and the electron transfer capacity of solubility carbon black of known raw material type can be directly predicted.

The invention relates to the technical field of intelligent meteorological early warning, and discloses an intelligent monitoring system based on a rainfall sensor, and the system comprises a sensing collection module which collects signals of the rainfall sensor and synchronizes environmental parameters; the data processing and modeling module is used for receiving the original data, performing preprocessing and feature extraction, and constructing a prediction model fusing rainfall and environmental factors; the model learning and self-adaptive updating module is used for optimizing model parameters according to actual rainfall and prediction errors to realize dynamic updating and continuous optimization; the output and communication module is used for judging the rainfall level and sending prediction and alarm information to a remote terminal through multi-channel communication; and the system control and energy supply module regulates the operation of the module and guarantees the stable and efficient operation of the system in combination with a multi-source power supply and scheduling strategy. According to the method, the moving average preprocessing, the multi-feature linear prediction model, the multi-channel communication and the multi-source energy supply technology are fused, and the prediction precision and the environment adaptive capacity of the rainfall monitoring system are improved.

The invention relates to a method for the adaptive observation of a system at a time t, the system comprising:-a set of inputs comprising the control setpoint ut and modeled in the form of vectors: Ut; -a set of output quantities, the set of output quantities being modeled in the form of vectors: Yt; and-a set of quantities representing the state of the system, the set of quantities modeled in the form of vectors: xt; the method is implemented in a computing unit and comprises: a first step E1 of programming a set of linear differential equations relating to the input, the system state and the output; a step E2 of determining a slow-varying parameter; and-an adaptive observation step E3 comprising numerical calculation of the quantity xt of the system state at the moment t by solving the set of linear differential equations.

The invention provides a battery thermal management system control method based on multi-mode switching and predictive optimization, and the method comprises the following steps: building a discrete time state space model based on the physical structure of a battery module; performing first-order Taylor expansion at a nominal working point for a nonlinear coupling term in the discrete time state space model so as to obtain a linear prediction model adapted to online optimization; designing a model prediction controller based on the linear prediction model; a dual-threshold trigger logic based on an instantaneous temperature tracking error and a battery output power demand is defined, so that a battery thermal management mode is adaptively switched. According to the method, the prediction controller based on the linear dynamic state space model is established, and a mode switching mechanism based on temperature error and battery power dual-threshold triggering is introduced, so that the model prediction controller can realize accurate tracking and quick response of the battery temperature, the health state of the battery is improved, and the service life of the battery is prolonged.

The application discloses a system control method and device based on model predictive control and reinforcement learning, comprising: acquiring a training data subset of a controlled system and a total state vector at a current moment, the historical state vector at the current moment comprising a control input vector and a state output vector before the current moment, and a plurality of training data subsets being obtained by unsupervised clustering of training data based on historical state vectors at historical moments; determining an MPC intelligent agent of a local working condition corresponding to a training data subset with the highest similarity according to the similarity between the historical state vector at the current moment and the historical state vector at the historical moment; and generating a control input vector at the current moment according to parameters of the MPC intelligent agent and the total state vector and sending the control input vector to the controlled system, wherein the parameters of the MPC intelligent agent are obtained by training a linear prediction model and reinforcement learning based on the training data subset corresponding to the local working condition. According to the application, the complexity of system control can be reduced, and the interpretability and safety can be improved.

The invention relates to the technical field of agricultural product quality monitoring and risk early warning, and discloses an agricultural product quality risk early warning system and method based on digital twinning, and the method comprises the steps: constructing a metabolic network response digital twinning model under a multi-stress condition, bidirectional data mapping of the physical agricultural product and the digital model is realized; constructing a metabolic toughness potential energy landscape model, and representing the stability and critical transformation characteristics of an agricultural product quality system; extracting an early warning signal based on a critical moderation theory; analyzing an interaction effect among various environmental stress factors, and constructing a toughness map of an agricultural product quality system; integrating analysis results to form an early warning decision and providing an intervention suggestion; according to the method, the limitation of a traditional linear prediction model is broken through, potential risks can be identified in the quality appearance stability stage, early warning several days ahead of time is achieved, and powerful technical support is provided for agricultural product cold-chain logistics and warehouse management.

A braking pressure control method for new energy vehicles based on the Koopman operator includes: establishing the dynamic equilibrium equations and nonlinear model of the braking system; constructing an observation function set, identifying the Koopman operator matrix using an extended dynamic mode decomposition algorithm, constructing a global linear prediction model, and correcting the Koopman operator matrix; constructing an augmented state equation in the discrete-time domain, designing a linear extended state observer based on an improved Kalman filter algorithm, and outputting a lumped disturbance estimate; injecting the lumped disturbance estimate into the prediction equation of the linear model predictive control, and obtaining the optimal control sequence by solving the objective function of a quadratic programming problem; converting the first control increment of the optimal control sequence into a motor drive signal, which is applied to the master cylinder motor of the electro-hydraulic braking system for physical pressure build-up. This invention can solve the problems of low hydraulic pressure control accuracy, poor anti-interference ability, and high computational load in existing technologies.

The invention discloses a superconducting optical fiber preparation method and a high-sensitivity landslide monitoring method, which are applied to the technical field of landslide monitoring. The monitoring method comprises the following steps: constructing a three-dimensional distributed sensing network based on a biological superconducting sensing optical fiber in combination with a multi-source sensor; emitting a detection light signal, and collecting a returned backscattering light signal; demodulating the collected optical signals to obtain strain, temperature or vibration parameters along optical fiber distribution points; in combination with multi-source data about the landslide mass and a landslide geomechanical model, the stable state of the landslide mass is evaluated through a threshold criterion or a machine learning algorithm, a prediction result is obtained by using a linear prediction model, and safety operation is executed based on the prediction result. According to the invention, the sensing of the landslide disaster from the millimeter level to the micro-nano level is realized, the sensitivity is high, the anti-interference capability is strong, and a brand new technical means is provided for the early precise early warning of geological disasters.

The invention discloses a method and a system for constructing an investment cost estimation model of a light-storage integrated station. The method comprises the following steps of: acquiring engineering quantity, equipment composition, energy storage scale, sending-out line parameters and a construction period financial structure in research data of a photovoltaic project; respectively quantifying the data into a direct current side installed capacity, an energy storage configuration proportion, a transmission project line length, a power transmission line voltage grade, a financing interest rate and a construction period; the total investment is divided into basic cost, energy storage increment cost, capitalization interest cost and outgoing line cost, and the four costs are composed of direct current side installed capacity, energy storage configuration proportion, outgoing engineering line length, power transmission line voltage grade, financing interest rate and construction period; forming an overall nonlinear prediction model based on the four costs; and early-stage investment estimation, scheme comparison and selection and power station economy evaluation of the photovoltaic project are carried out based on the overall nonlinear prediction model. According to the invention, accurate prediction of new energy projects of different scales is realized through the overall nonlinear prediction model.

The present application relates to the field of image processing and power monitoring technology, and particularly relates to a substation multi-target identification and tracking method based on image processing technology, comprising the following steps: S1: real-time acquisition of substation monitoring video stream, output of filtered image; S2: extraction of moving target contour, generation of target contour set; S3: calculation of spatial gradient amplitude and electromagnetic interference quantization parameter, formation of feature vector; S4: input of feature vector into classifier, output of device target label or biological target label; S5: establishment of linear prediction model, output of predicted coordinate sequence; S6: when the predicted coordinate sequence intersects with the safety area of the device target, an intrusion alarm signal is generated. The present application, through the combination of target classification identification and trajectory prediction, realizes the accurate identification and early warning of biological targets in the substation under the electromagnetic interference environment, and improves the response capability and safety control level of the system to dynamic intrusion behavior.

This application provides a method, system, medium, and device for real-time prediction of marine engine operating parameters based on the DLinear algorithm. This application replaces explicit sequence decomposition with feature decoupling, simplifying multivariate prediction into multiple univariate prediction problems through feature decomposition, and uses linear mapping instead of complex time series modeling. This method simplifies the Dlinear algorithm, making it suitable for time series prediction of parameters with weak correlations. For parameters with stronger correlations, the original Dlinear algorithm structure can be used. This application collects parameter data from marine engines across different load ranges, establishing a feature decoupling prediction model and a trend / seasonal bilinear prediction model. This avoids the impact of full-load testing on engine lifespan, solves the problem of insufficient effective data samples for modeling, and can quickly capture the nonlinear characteristics of multivariate co-evolution.

The invention discloses a method and a device for identifying rare earth ore soil pollution by using a leaf spectrum, a data acquisition device is adopted to obtain reflection spectrum data of vegetation leaves, the reflection spectrum data is input into a computer, and a processed spectrum is obtained in a data preprocessing module of the computer; meanwhile, a surface soil sample is collected, and the heavy metal content in the surface soil sample is obtained; the processed spectrum is decomposed into a plurality of intrinsic mode function components and a residual term in a multivariate empirical mode decomposition module, then correlation analysis is carried out on the plurality of intrinsic mode function components and the heavy metal content in a data processing cabinet, intrinsic mode function component wave bands with weak correlation are removed, and the heavy metal content is obtained; and finally, constructing a non-linear prediction model by adopting a random forest regression algorithm, and outputting a predicted soil heavy metal content value in the non-linear prediction model. According to the method, the rare earth ore soil heavy metal pollution is accurately identified, and the precision higher than that of a current popular method can be obtained.

The invention relates to the technical field of power equipment detection and state evaluation, in particular to a time domain alternating current equivalent direct current resistance test and analysis method based on big data. According to the method, the equivalent direct current resistance is calculated through two independent paths of a linear prediction model and time domain signal iterative optimization, the time sequence characteristics of the equivalent resistance and the working condition correlation characteristics are fused, the high-dimensional characteristic matrix is constructed, the data internal law and the working condition influence mechanism are fully mined, and rich characteristic support is provided for the model; the geometric cascade forest model is good at basic feature classification, the gradient boosting tree and the isolated forest fusion model give consideration to health state evaluation and anomaly detection, the two complements each other to form dual research and judgment, the comprehensiveness and accuracy of state recognition are improved, and the accuracy of state recognition is improved through dual prediction result comparison and anomaly feedback list output. And data or model problems can be found in time, a direction is provided for subsequent optimization, and the stability and credibility of an analysis result are ensured.

The invention discloses a high-dynamic motion control method and system for a four-wheel-foot robot. The method comprises the steps of acquiring a robot system state in real time through a sensor; then, dynamically updating a linear prediction model and compensating a nonlinear error by adopting a self-adaptive time-varying model prediction control algorithm fused with affine compensation, and solving an expected foot end force on line; then, based on closed-chain Jacobian mapping, the foot end force is mapped into a hub driving torque and a leg joint torque containing a structural stiffening torque at the same time, so that leg shear deformation caused by high-dynamic motion is cooperatively inhibited; and finally, outputting the calculated torque instruction to each execution motor. According to the method, through innovation of an algorithm level, on the premise that hardware is not changed, the problem that the control precision, the system stability and the algorithm real-time performance of the wheel-foot robot in a high-dynamic scene are difficult to consider at the same time is effectively solved, and the motion performance of the robot under extreme working conditions such as rapid acceleration and large-slope climbing is remarkably improved.

The application discloses a mixed tank level control system and method based on a Kupman network, and belongs to the technical field of tank level control. The method specifically comprises the following steps: collecting key process data of a mixed tank, and constructing an original state vector of tank level control; inputting the original state vector into a pre-trained Kupman network; mapping a nonlinear dynamic process of the mixed tank level to a high-dimensional observable space; establishing a linear prediction model; combining a tank level balance to generate a tank level prediction trajectory; based on the tank level prediction trajectory, using a predictive compound PID strategy to solve a comprehensive feeding amount adjustment amount, and issuing the comprehensive feeding amount adjustment amount to a feeding execution unit to maintain the tank level within a target range; after executing the comprehensive feeding amount adjustment amount, calculating a prediction deviation, and iteratively correcting the Kupman network and the tank level balance. The application improves the tank level control accuracy and continuity, reduces the risk of full tank or lack of material, and enhances the self-adaptive ability to working condition switching and sudden disturbance.

The invention discloses a method for predicting low-temperature performance of aged asphalt, and belongs to the technical field of performance evaluation of road materials. According to the method, firstly, an asphalt sample is subjected to simulated aging treatment, then improved column chromatography is adopted for separating and measuring the content of saturates, aromatics, colloids and asphaltene of aged asphalt, a chromatographic column is of an activated aluminum oxide and silica gel double-layer adsorption structure, gradient elution is combined, and the separation efficiency and reproducibility are remarkably improved. The method comprises the following steps: firstly, measuring a creep rate m value of aged asphalt, analyzing the correlation degree between the m value and components of the aged asphalt by using grey correlation, and establishing a multiple linear prediction model based on ridge regression by taking the contents of four components as independent variables and the m value as a dependent variable; the model effectively overcomes the multicollinearity problem among components, component analysis data are input into the model for unknown samples, the low-temperature performance of the samples can be predicted, and the research and development efficiency and the engineering quality control level are greatly improved.

This invention discloses a method for optimizing phase change sweating cooling structures, comprising the following steps: S1, generating co-simulation data based on physical field dynamic mapping; S2, constructing an SA-ResBPNN phase change nonlinear prediction model; and S3, performing NSGA-II multi-objective optimization based on the above model. A Fluent-Matlab co-simulation platform based on physical field dynamic mapping is constructed: it enables rapid reconstruction of porous medium properties under varying structural parameters, avoiding repeated geometric modeling and mesh generation, and automating the entire process from sample generation, coupled solution, to result extraction. A self-attention residual neural network (SA-ResBPNN) for phase change nonlinearity is proposed: it adaptively identifies and amplifies the feature weights of key segments; based on the cross-layer transmission of main physical features, it drives the deep network to focus on fitting high-order nonlinear mutations caused by phase change, thereby effectively improving the prediction accuracy of cooling performance.

The application provides a model predictive control method for pulp kappa number in batch cooking process, and belongs to the technical field of pulping and papermaking. The method comprises the following steps: first, collecting multi-dimensional measurable state variables related to delignification in the cooking process to form an original state vector; then, using a Koopman operator and extended dynamic mode decomposition, the original state vector is mapped to a high-dimensional linear space to obtain an upgraded state vector; based on historical data offline training, a Koopman global linear prediction model is established, and a free liquid temperature setting sequence in a future control time domain is optimized by a model predictive controller; after each optimization, the first control amount of the temperature sequence is applied to the cooking process in real time, and the rolling optimization is repeatedly executed in combination with feedback correction in the next sampling period. The method overcomes the end-point kappa number steady-state deviation problem existing in batch cooking applications, and improves the adaptability to model mismatch and the robustness of the control system.

The invention belongs to the technical field of fish and soft-shelled turtle polyculture control, and discloses a fish and soft-shelled turtle polyculture stocking parameter determination method and device. The determination method comprises the following steps: firstly, obtaining a culture target label of fish and turtle polyculture and corresponding historical stocking parameters; screening stocking parameter information corresponding to each breeding target label, and constructing a linear prediction model and a nonlinear prediction model corresponding to each breeding target label; further optimizing to obtain an optimized prediction model; and when a breeding target instruction input by a user is received, analyzing the breeding target instruction to determine a corresponding breeding optimization target and optimization prediction model so as to output optimized stocking parameters corresponding to the breeding target instruction. According to the determination method, key stocking parameters of economic benefits and ecological effects of fish and turtle polyculture can be accurately determined, great significance is achieved for reducing the culture cost and improving the culture benefits, and meanwhile optimization of the internal structure of a culture system, reduction of carbon emission and formation of a healthy and low-consumption ecological culture mode are facilitated.

The application provides a kind of multi-feature subjective probability type prediction modeling and aggregation method and system, comprising: obtaining expert probability type judgment result, and establishing the linear prediction model of expert probability type judgment result based on lens model;Generate prediction result through the linear prediction model, determine the error of prediction result;According to the prediction result containing prediction error, extract potential information source, and decompose problem variable characteristics and expert behavior characteristics;Estimate the weight given by expert to different information source based on the decomposed expert behavior characteristics, cluster expert weight, form multiple cluster groups;From each cluster group, select the expert with the closest estimated weight distance from the cluster center as the representative of each cluster group, and aggregate the probability type judgment through multiple heuristic methods, the application solves the problem that existing multi-element expert probability type judgment prediction is difficult to accurately aggregate.

The invention provides a sow perinatal insulin resistance state prediction method and device based on machine deep learning, and belongs to the crossing field of a computer technology and a biological monitoring technology. The method comprises the following steps: acquiring a training data set containing easily measured sign parameters of a plurality of samples and corresponding insulin resistance indexes; based on the data set, training a machine learning prediction model capable of learning and establishing a nonlinear mapping relationship between the easily measured sign parameters and the insulin resistance indexes; and obtaining easy-to-measure physical sign parameters of an individual to be measured, and inputting the easy-to-measure physical sign parameters into the trained model to obtain a predicted insulin resistance index. The device comprises a memory and a processor and is used for executing the method. According to the method, the machine learning model capable of capturing the complex nonlinear relation is utilized, the problems that in the prior art, intrusive detection is high in cost and low in efficiency, and a linear prediction model is insufficient in accuracy are solved, and rapid, low-cost and high-precision prediction of the key physiological indexes is achieved.

The invention provides a geopolymer underwater non-dispersible concrete mix proportion optimization method and a geopolymer underwater non-dispersible concrete mix proportion optimization system, and relates to the technical field of concrete mix proportion optimization. A multi-dimensional nonlinear prediction model covering fluidity, dispersion resistance and mechanical properties is constructed based on experimental data, and a coupling relation and a mixing amount optimization rule between material properties are effectively revealed; by combining the particle swarm optimization algorithm, the intelligent optimization design of the concrete mix proportion under the constraint of multiple performance indexes is realized, the construction flowability and durability requirements are ensured, and the compressive strength and the structural stability are improved, so that the scientificity and the accuracy of the design are improved, the test cost is reduced, and the design period is shortened.

A decoding device includes a transformer sets a decoded luminance component of a prediction target block to the same number of samples as that of the chrominance component corresponding to the decoded luminance component of the prediction target block and generates a luminance reference signal. A specificator specifies luminance pixels having minimum and maximum pixel values of the decoded luminance component adjacent to the decoded luminance component of the prediction target block, respectively, outputs luminance pixel values obtained from specified luminance pixels, and outputs chrominance pixel values from pigment pixels corresponding to the luminance pixels. A derivator derives a linear prediction parameter from the two pixel values and a linear prediction model. A chrominance linear predictor obtains chrominance prediction signal applying the linear prediction model based on the linear prediction parameter to the luminance reference signal. The chrominance prediction and residual signals are summed to generate a reconstructed chrominance signal.