897results about How to "Ease of evaluation" patented technology

The invention relates to the technical field of artificial intelligence, and particularly provides a KASP primer intelligent typing evaluation method and system TAL-SRX based on multi-model fusion. The method comprises the following steps: firstly, optimizing the performance of a basic model and a meta model by using RS, PSO and BO algorithms, constructing a Stacking integrated learning framework to integrate the advantages of an XGBoost learner, and enhancing the stability of the model by adopting five-fold cross validation; then, combining ANN and LSTM to construct a hybrid neural network, capturing a nonlinear relationship and extracting complex features, and meanwhile, introducing a Transform algorithm to capture a global dependency relationship in a high-dimensional feature space; and finally, fusing the two machine learning algorithms through a soft voting integration strategy, and outputting a KASP mark typing effect score. According to the method, the KASP marker with an excellent typing effect can be accurately screened, relatively high robustness and good and reliable evaluation performance are achieved, and technical support is provided for work such as molecular marker-assisted breeding.

The invention discloses a learning plan recommendation method and system, and an education robot, and relates to the technical field of education robots, demand identification is carried out by using a trained personalized learning demand identification model, and personalized learning demands are output and identified; matching learning resources and learning plans according to personalized learning requirements of the target students, outputting an optimal learning path, collecting feedback data for comprehensive evaluation, generating a state coefficient according to evaluation data, and if the state coefficient does not exceed expectation, optimizing the personalized optimized learning path of the target students by using a multi-target optimization algorithm, so as to improve the learning efficiency of the target students. Classifying the target students based on a pre-trained density clustering algorithm; and after the abnormal learning behavior is identified, starting a multi-level alarm mechanism, and outputting a corresponding response strategy by an emergency response strategy library according to the abnormal learning behavior. And the current learning path is adjusted and optimized, so that the learning path of the student is more matched with the current learning resource and learning plan.

The invention relates to the technical field of public security monitoring, in particular to a public security situation prediction method and system based on multi-modal data fusion. The method comprises the following steps: performing continuous video monitoring and non-inductive discovery on a public place to obtain non-inductive discovery data of the public place; performing multi-modal data feature extraction on the public place non-inductive discovery data to obtain security threat multi-modal feature data; performing situation assessment on the security threat multi-modal feature data to obtain security threat event level data, and performing event level division on the security threat multi-modal feature data according to the security threat event level data to obtain security threat risk event data; and performing automatic response control according to the security threat risk event data to obtain response control early warning data, and performing multi-view evaluation optimization according to the response control early warning data to obtain response control early warning optimization data. According to the invention, the potential security situation can be predicted, and the level of public security management is improved.

The invention relates to the technical field of power grid monitoring, in particular to a power grid abnormal state monitoring and early warning system which comprises a historical reference establishment module, an operation deviation analysis module, an abnormal threshold adjustment module and an abnormal early warning module. According to the invention, through real-time data acquisition and historical data analysis, the monitoring and early warning efficiency of power grid operation is improved, and through collection of power grid operation records, extraction of data of differential time periods, evaluation and analysis of the average load and volatility of the data, and construction of a historical power grid operation reference, the operation trend and potential abnormity can be captured more accurately. The method reduces the uncertainty of judgment depending on experience, can recognize a deviation mode and an adjusted load detection threshold value by comparing current data with a historical reference, enables the abnormal early warning to be more accurate, avoids excessive false alarms and missed alarms, optimizes the evaluation of early warning levels and the implementation of early warning notification, guarantees the safety and stability of a power grid, and improves the safety and reliability of the power grid. And the economic efficiency and the reliability of power grid operation are also improved.

The invention relates to the technical field of equipment monitoring, in particular to a machine room maintenance monitoring system which comprises a temperature threshold adjusting module, a temperature anomaly evaluation module, a performance anomaly recognition module and a maintenance sequence adjusting module. According to the method, the temperature threshold value monitored by the equipment in the machine room is dynamically adjusted based on the environment and the service life of the equipment, the evaluation process of the temperature anomaly is optimized, the temperature anomaly can be recognized in real time at the initial stage of temperature deviation, the response time is shortened, and damage caused by overheating of the equipment is reduced; the performance loss proportion and the performance decline rate are calculated, performance abnormity is evaluated according to the performance loss proportion and the performance decline rate, a maintenance team can make a quick response, problem diffusion and loss are reduced, the maintenance priority is adjusted according to the importance and the fault severity of equipment, rationalization and systematization of maintenance work are achieved, major faults are effectively avoided, and the service life of the maintenance team is prolonged. And meanwhile, the overall stability and safety of the machine room equipment are improved.

The invention relates to the technical field of remote sensing, and discloses a remote sensing ecological index assessment method based on deep learning, which comprises the following steps: acquiring Landsat remote sensing image data and land utilization data of a research area, and extracting remote sensing ecological index factors of the land utilization data and the preprocessed Landsat remote sensing image data; normalizing the remote sensing ecological index factors and performing dimension reduction analysis to obtain dimension-reduced remote sensing ecological index factors, inputting the dimension-reduced remote sensing ecological index factors into a time sequence convolutional network model, and generating a comprehensive feature vector through causal convolution, extended convolution and global average pooling operations; performing linear combination on the comprehensive feature vector and the principal component analysis feature value, and constructing a time sequence convolutional neural network remote sensing ecological index model for evaluating the ecological environment quality condition of the research area; according to the method, the accuracy, the reliability and the practicability of remote sensing ecological index evaluation are improved by adding the remote sensing ecological landmark factors reflecting human disturbance and water environment change and constructing the time sequence convolutional neural network remote sensing ecological index model.

The invention relates to a pipe network operation risk monitoring method and system for an urban drainage system, and the method comprises the steps: obtaining a pipe network distribution drawing of the urban drainage system, carrying out the key site analysis, carrying out the pipe network topology through a key site analysis result and the corresponding pipe network distribution, and obtaining a pipe network topological relation of the urban drainage system, acquiring real-time monitoring data of each key site, analyzing the operation change trend of all key sites of the same drainage event in combination with a pipe network topological relation to obtain a site operation association relation of the pipe network, acquiring historical risk parameters of each key site, and calculating the risk parameters of each key site; and performing data training on the historical risk parameters according to the site operation association relationship, constructing a pipe network risk monitoring model of the urban drainage system, obtaining external environment change parameters of the urban drainage system, and inputting the external environment change parameters into the pipe network risk monitoring model for risk analysis to obtain pipe network operation risk monitoring data. The method and the device have the effect of improving the accuracy of pipe network risk monitoring.

The invention provides a data quality monitoring and repairing system, and belongs to the technical field of data repairing, and the system comprises an identification module which obtains data of a plurality of data sources, and automatically identifies a plurality of data quality problems existing in the data; the first determination module is used for determining a root cause and an influence path of a data quality problem based on a causal relationship reasoning technology; the evaluation module is used for evaluating the severity and the influence range of each data quality problem; the selection module is used for distributing priority for each data quality problem on the basis of a predefined rule according to the severity and the influence range, determining a processing sequence and selecting an optimal repair strategy for automatic repair; and the automatic repairing module is used for continuously monitoring the repaired data in real time, detecting a new data quality problem by utilizing a stream processing technology and a real-time analysis algorithm, and automatically repairing the data. The problems that data is inaccurate due to low identification efficiency and automatic and real-time processing of the data quality problem cannot be guaranteed are solved.

The embodiment of the invention discloses a method for evaluating the ablation cracking degree of an environment-friendly gas high-voltage circuit breaker contact material. The method comprises the steps that thermal parameters and mechanical parameters of the high-voltage circuit breaker contact material are collected; establishing a three-dimensional geometric model of the high-voltage circuit breaker contact according to the thermal parameters and the mechanical parameters; analyzing the high-voltage circuit breaker contact according to an extended finite element method, and obtaining simulation results of temperature and stress distribution of the high-voltage circuit breaker contact in the working process; according to the simulation result, the ablation degree and the cracking degree of the high-voltage circuit breaker contact material are determined, and a corresponding evaluation result is obtained. Through accurate three-dimensional modeling, data processing and evaluation result obtaining, evaluation and optimization of the high-voltage circuit breaker contact material are facilitated, and a reference is provided for design and application of the high-voltage circuit breaker contact material under the large-current open-end condition.

The invention relates to the technical field of radio orientation, in particular to a measuring method capable of tracking and recording motion trail information of a vibrating rod, which comprises the following steps: arranging at least four UWB (Ultra Wideband) base stations, fixing UWB transmitters at the joint of the vibrating end of the vibrating rod and a rubber tube, and fixing IMU (Inertial Measurement Unit) sensors at the vibrating end of the vibrating rod; according to the data output by the UWB transmitter and the IMU sensor, the position and posture of the vibrating rod in the detected target mold are calculated through a positioning algorithm, and the motion trail of the vibrating rod is further calculated; and according to the vibrating time, the vibrating area and the vibrating area overlapping degree of the vibrating rod, the vibrating state of the concrete is analyzed. The UWB technology and the IMU technology are fused, so that high-precision positioning of the vibrating rod can still be kept under the ground where wireless signals such as GPS signals cannot be transmitted, and the vibrating quality can be conveniently detected and evaluated.

The invention discloses a water quality detection system and detection method for groundwater resource assessment, and relates to the technical field of environmental monitoring, the method comprises the following steps: obtaining multi-source data of groundwater in a target area through a sensor network, and combining environmental parameters to carry out data fusion to generate a comprehensive result; performing water quality scoring based on the fused data, constructing a water quality spatial distribution diagram, and predicting a water quality change trend; and determining a pollution source according to trend data tracing, formulating a pollution control scheme in combination with multi-source data, and generating a treatment suggestion. The technical problem of insufficient water quality detection precision in existing groundwater resource evaluation is solved, and the technical effect of performing high-precision detection and comprehensive evaluation on groundwater quality data is achieved.

The invention provides a self-adaptive correction pharmaceutical process quality detection method and robot, and belongs to the technical field of medicine quality detection. The method comprises the following steps: acquiring hyperspectral image historical data of a drug, and carrying out data preprocessing; key spectral features in historical data of the hyperspectral image are extracted, and a multivariate correction model is constructed; collecting hyperspectral image real-time data of drugs, and forming and dynamically updating a data set; when the dynamic capacity of the data set reaches a set threshold value, the data set is used for real-time training of a multivariate correction model, and adaptive correction and real-time automatic updating of the multivariate correction model are carried out by dynamically adjusting parameters. The robot comprises a data acquisition module and a quality detection module which are connected with each other; the data acquisition module is integrated with hyperspectral imaging equipment; and the quality detection module carries a multivariate correction model. Through continuous learning, a multivariate correction model is automatically updated in real time to adapt to dynamic changes in the pharmaceutical process, and the production efficiency and the quality control precision are improved.

The invention relates to the technical field of sleep monitoring, in particular to a sleep monitoring method and system, computer equipment and a storage medium, and the sleep monitoring method comprises the following steps: monitoring the pressure change of the head and neck of a user through a miniature differential pressure sensor, collecting pressure data in real time, recording a respiratory cycle according to pressure fluctuation, and determining the sleep state of the user; and identifying high and low pressure peak values, and generating head pressure monitoring data. According to the invention, by monitoring the pressure change of the head and the neck in real time, the capturing capability of tiny breathing changes is improved, refined breathing cycle recording and high-precision pressure peak value recognition are realized, potential breathing abnormity can be found at the initial stage, timely response is ensured, and the breathing frequency and amplitude are comprehensively analyzed; the breathing modes are accurately classified, so that the recognition of the abnormal breathing state is quicker, more accurate and more detailed dynamic monitoring and data analysis are realized, the evaluation of the sleep depth and stage is enhanced, and the comprehensive evaluation of the sleep quality is more scientific.

The invention relates to the technical field of valve data prediction, in particular to a water supply valve control method and system based on fault monitoring, and the method comprises the steps: collecting the pressure and caliber data of a valve, and obtaining a pressure sequence and a valve caliber sequence; respectively predicting the pressure sequence and the valve caliber sequence at historical moments based on an HTFE algorithm to obtain a predicted value of the pressure and a predicted value of the valve caliber at the current moment, and correcting the sum of differences between the predicted value of the pressure and the predicted value of the valve caliber and respective actual values by using the reliability of the predicted value at the current moment; and obtaining the fault degree at the current moment. According to the method, the reliability of the prediction value is calculated by considering the relevancy and the change stability between the two dimension data and combining the relevancy and the stability, so that the evaluation on the reliability of the prediction result is increased, and the reliability of the prediction result is ensured.

The invention relates to the technical field of steel reinforced concrete vibration testing, and discloses a method and system for testing the vibration performance of a steel reinforced concrete frame structure, and the state of the steel reinforced concrete frame structure is judged through vibration amplitude and frequency changes under the conditions of different excitation modes, measurement point positions, sampling frequencies, collection durations and the like. By accurately measuring and analyzing the change of the damping ratio, potential damage or performance change of the structure can be found in time, and a powerful basis is provided for safety assessment and maintenance of the structure; the signal is decomposed and analyzed based on Hilbert-Huang transform, the internal composition and the change rule of the structure vibration signal can be analyzed more finely, and the method has advantages for identifying weak vibration characteristics and early damage signs of the structure in a complex stress state, so that the test structure can be evaluated in time in the early damage stage, and the test accuracy is improved. And therefore, coping treatment can be conveniently carried out.

The invention relates to the technical field of engineering design, in particular to a truss-web type steel-concrete composite beam deflection calculation method and system.The method comprises the steps that a mechanical simplified model of a truss-web type steel-concrete composite beam is built, calculation assumption is determined, and a calculation scheme is obtained; determining the acting mode of the external load on the composite beam; basic calculation parameters of the combined beam mechanical simplified model are determined; according to the action mode, calculating a rigidity reduction coefficient of the simply supported beam caused by shear deformation of a truss web member, a slippage effect of a steel-concrete connection interface and local deformation of a node under the action of different loads; calculating the equivalent flexural rigidity of the truss-web type composite beam; and calculating the deflection of the truss-web type steel-concrete composite beam under the action of different loads. By means of the method, the problem that an existing calculation method is poor in pertinence and applicability is effectively solved, so that the stress condition of the truss web type steel-concrete composite beam structure is correctly reflected; according to the method, multiple influences caused by discontinuity of the truss web members, including shear deformation of the truss web members, a slippage effect of a steel-concrete connection interface and additional deflection caused by local deformation of the nodes, are considered, and the midspan deflection of the truss-web type steel-concrete combined simply supported beam can be conveniently evaluated.

The invention discloses a multi-agent interpretable text psychological health level assessment method and system, and the method comprises the steps: training an original agent model according to a multi-field data set and a psychological dialogue assessment data set, and obtaining a dialogue reference model and an assessment reference model; according to the two reference models, respectively processing the user question data and the multi-round dialogue data to obtain a multi-dimensional dialogue feedback data set and an evaluation feedback data set; according to the multi-dimensional dialogue feedback data set and the evaluation feedback data set, training the two reference models to obtain a dialogue effect evaluation model and an evaluation effect evaluation model; according to the two evaluation models and the psychological dialogue evaluation data set, training the two reference models to obtain a target dialogue model and a target evaluation model; according to the target dialogue model and the target evaluation model, mental health level evaluation is completed. According to the invention, through the multi-dimensional feedback mechanism training model and the cooperative work of multiple models, the interpretability and confidence of the psychological health assessment system are improved.

The invention relates to a radiation control method and device applied to a mobile C-shaped arm, equipment and a medium, and the method comprises the steps: obtaining the radiation data of each direction of a to-be-detected region corresponding to a skeleton instrument, and determining the radiation intensity of the to-be-detected region according to the radiation data of each direction; under the condition that the radiation intensity exceeds a preset threshold value, according to a difference value between the radiation intensity and the preset threshold value, obtaining an adjustment parameter through a trained neural network PID model; and adjusting operation parameters of the skeleton instrument according to the adjustment parameters. The C-shaped arm skeleton instrument has the effect of improving the anti-radiation performance and safety of the C-shaped arm skeleton instrument.

The invention discloses a peripheral-central-peripheral evaluation system based on electroencephalogram-myoelectricity dual-channel bidirectional mixed feedback, and the system is characterized in that a muscle electrical stimulation module is used for providing a current with a specific frequency, and carrying out the electrical stimulation of muscles; the transcranial direct current stimulation module is used for providing current with specific frequency and performing electrical stimulation on the cerebral cortex; the myoelectricity acquisition module and the electroencephalogram acquisition module respectively acquire electric signal data of tested muscles and cerebral cortex and transmit the electric signal data to the upper computer feedback module, and the upper computer feedback module is used for controlling the muscle electrical stimulation module and the transcranial direct current stimulation module to generate electrical stimulation and acquiring data acquired by the myoelectricity acquisition module and the electroencephalogram acquisition module; and the data is evaluated after being processed, and an evaluation result is sent to the terminal display module for real-time display. According to the scheme, in the evaluation process, through a bidirectional mixed feedback mechanism, accurate positioning of cerebral cortex and muscle group stimulation targets is achieved, the electrical stimulation and evaluation effect can be optimized, quantitative evaluation is conducted on the result through a bidirectional evaluation model, scientific quantitative indexes are provided, the influence of individual differences and state changes on the evaluation result is reduced, and the evaluation accuracy is improved. And the reliability and stability of evaluation are improved.

A plant illumination method, a plant illumination system, a storage medium and a program product relate to the field of plant illumination, and the method comprises the following steps: determining a target illumination duration, a target spectrum and a target illumination intensity required by a target plant according to a growth stage of the target plant in a target planting area; according to the sunlight illumination intensity of the target planting area, the accumulated duration of the target plant under the effective sunlight illumination is determined; if the accumulated duration is lower than the target illumination duration, determining whether the growth stage is in a preset key growth stage or not; if yes, the photovoltaic panel is set to be in a tracking mode; if not, the photovoltaic panel is set to be in a conventional mode; and when the light supplementing duration exceeds the difference value between the target illumination duration and the accumulated duration, controlling the photovoltaic panel to stop light supplementing. By implementing the method, the illumination efficiency and effect of plant growth are improved.

The invention relates to the technical field of data analysis and evaluation, in particular to a color fastness evaluation method and system, and the method comprises the following steps: setting test conditions to simulate external factors encountered in the actual use or processing process of textiles; under the simulated test condition, treating the textile sample for a preset time; acquiring a multispectral image of the textile by adopting a multispectral imaging technology, and capturing color change of the textile; developing a self-adaptive image enhancement algorithm, and performing preprocessing; comparing spectral information of the multi-spectral image of the textile before and after processing based on chromatic aberration evaluation analysis so as to quantify the color fastness; analyzing the multispectral image after image enhancement by using a deep learning technology, extracting color attenuation features, and comprehensively evaluating a color fastness score; according to the color fastness evaluation method, basic color difference information is considered, advanced features such as textures, edges and local color modes of the textile are further included, and the accuracy and reliability of color fastness evaluation are remarkably improved.

The invention relates to the technical field of additive manufacturing digital twinning, solves the technical problems of large calculation amount and long rendering time of a traditional 3D printing simulation method, and particularly relates to a 3D simulation system for quickly realizing additive manufacturing, which comprises the steps of loading and processing a model file, loading and processing a slice file and loading 3D model section data. According to the method, the cutting plane information of a target renderer or a material is dynamically set, the display and progress of a 3D printing model are controlled and managed, user interaction is controlled through a scroll bar, and through the operation, parts in the additive manufacturing printing process can be rapidly achieved on a Unity platform. According to the method, a vivid 3D printing simulation effect is quickly generated by utilizing a simplified geometric model, material attributes and an efficient rendering algorithm, and the whole printing process is simulated, so that a user can obtain visual preview of a printing object in a short time and can watch the whole printing process, and printing parameters can be better evaluated and adjusted.

The invention provides a method for evaluating the ultimate bearing capacity of a seawater pipeline with a double-ball defect, which comprises the following steps of: S1, expressing an ultimate bearing capacity model of a pipeline with an axial double-ball defect as a product of a distance influence factor # imgabs0 # and an ultimate bearing capacity model # imgabs1 # of a pipeline with a single-ball defect; s2, the corrosion depth ratio N, the corrosion diameter # imgabs2 and the center distance ld of the axial double-sphere defect of the seawater pipeline with the double-sphere defect to be detected are collected; and S3, inputting the data acquired in the step S2 into the ultimate bearing capacity model of the pipeline with the axial double-spherical defect to obtain the ultimate bearing capacity. On the basis of a single-point defect seawater pipeline ultimate bearing capacity model, the influence of the double-point center distance is considered, a double-point defect seawater pipeline ultimate bearing capacity evaluation model is established, and the actual engineering value is achieved for accurately evaluating the actual corrosion defect ultimate bearing capacity of the seawater pipeline.

The invention relates to the technical field of data analysis, in particular to an innovation achievement evaluation system based on big data, and the system comprises a data capturing module which screens an original data set of an enterprise scientific and technological innovation project, extracts time series data, and distinguishes the data into cost data, an innovation project milestone date and a quality index; the cost change rate, the project progress deviation and the quality change index of the multiple innovation projects are calculated, the average value and the standard deviation of multiple indexes are counted, and preliminary project deviation analysis is established; according to the method, time sequence data of science and technology projects are accurately analyzed, cost, progress and quality indexes are monitored and analyzed, time sequence decomposition and independent analysis of seasonal and trend deviation are carried out, the recognition precision of periodic changes and long-term development trends of the projects is improved, key data changes can be accurately captured, and the recognition accuracy is improved. Accurate positioning and prediction of key change points in scientific and technological progress are promoted, and evaluation and optimization effects on research and development achievements are greatly enhanced.

The invention discloses a pipeline defect detection quantitative evaluation method based on CCTV and deep learning, and the method comprises the steps: constructing an improved pipeline crack recognition model, and making a data set for training the pipeline crack recognition model; acquiring an image from the pipeline video, inputting the acquired image into a pre-trained pipeline crack identification model, and identifying and segmenting an image containing a crack; removing redundant pixels in the image, reducing the width of the crack to be one pixel width, and obtaining coordinates of a pipeline center O (xc, yc) in the crack image and a plurality of specific pixels Ai (xi, yi) on the crack by using a circle detection method; the length, the shape and the width of the pipeline crack are obtained through a CDQ quantification method; according to the calculated crack length and width data, evaluation is carried out according to a crack grade judgment principle, and corresponding repair processing is carried out; according to the method, an improved segmentation model and an innovative crack quantification method are combined, basic crack information can be represented, and the severity of the drainage pipe is preliminarily evaluated in the drainage pipe which is poor in environment and complex.

The invention relates to the technical field of medical image analysis, in particular to a training method and system for a cataract recognition model and a storage medium. The method comprises the following steps: acquiring cataract eye image samples and non-cataract eye image samples in different stages; lesion region segmentation and image synchronous simulation enhancement are carried out on the cataract eye image samples and the non-cataract eye image samples in different stages to obtain cataract lesion enhanced image samples and non-cataract enhanced image samples; extracting local detail features of the lesion and extracting structural features of a mask pair, and constructing a multi-task convolutional neural network learning framework to train a cataract recognition model and optimize focus weighted loss; and obtaining a cataract related key feature set and carrying out online incremental learning updating to generate a cataract recognition online incremental updating model. According to the invention, the accuracy and robustness of cataract recognition model training can be improved.

The invention discloses a hybrid energy storage power optimization method for wind-solar coupling prediction and a control system, relates to the field of wind-solar coupling prediction, and solves the problem that most of hybrid energy storage of an existing distributed energy system is static optimization calculation in power distribution research. In order to solve the problems that in the prior art, a wind-light prediction technology is not used for carrying out expected power estimation on an energy storage system, and description of power change characteristics of hybrid energy storage under actual output is lacked, the invention provides the following scheme that the method comprises the following steps: step 1, constructing a multi-scene model for wind-light coupling output prediction; step 2, optimizing and calculating target power of hybrid energy storage under predicted output by PSO (Particle Swarm Optimization); and step 3, a hybrid energy storage power distribution and optimization method. According to the method and the control system, the complexity of processing the constraint conditions in the optimization process can be reduced, and the power characteristics of the whole system and the energy storage units in the hybrid energy storage can be evaluated more simply, conveniently and accurately.

The invention relates to a detection kit and a detection method for EBV latent infection of epithelial cells and B cells in a thymus and salivary gland tissue microenvironment, and the detection method comprises the following steps: dyeing thymus and salivary gland tissues by coupling quantum dots and a specific antibody; the conjugate of the quantum dots and the specific antibody is combined with a specific antigen of a tissue sample to realize marking of EBV latent infection epithelial cells and B cells by the quantum dots. According to the invention, the existing multicolor immunofluorescence technology is improved, the specific antibody and the quantum dot are coupled and purified and then are directly combined with the antigen on the tissue slice to realize direct marking of the quantum dot coupled antibody on the tissue sample, so that EBV latent infection cells in the thymus and salivary gland tissue microenvironment are positioned and quantified; the virus infection condition of the patient is evaluated from the pathogenesis level, the disadvantage that the previous evaluation means are limited is overcome, and the method has good sensitivity and accuracy.

The invention provides an anti-cancer drug collaborative prediction method based on knowledge graph embedding, GNN and Bi-LSTM, and belongs to the technical field of anti-cancer drug collaborative prediction. In order to solve the problem of poor collaborative prediction effect of anti-cancer drugs, the technical scheme adopted by the invention comprises the following steps: respectively obtaining corresponding relations of drug, protein, cell line and tissue interaction from a database; constructing a knowledge graph according to the corresponding relationship, and grouping all the drug combinations based on the collaborative scores of different drug combinations; based on a drug combination grouping result and the constructed knowledge graph, extracting a neighbor entity set interacting with the drug or the cell line, constructing an attention layer of the knowledge graph, and iteratively updating the initial entity embedding representation of each entity in the neighbor entity set; taking an entity neighborhood structure generated by iterative updating as input, extracting a high-order structure and a semantic relationship through a graph neural network, and learning medicine and neighborhood representation thereof; the method is applied to collaborative prediction of anti-cancer drugs.

The embodiment of the invention provides a model evaluation method and device, a storage medium and a program product. In the model evaluation method, after a reply result output by a target model for a target question is obtained, a question type corresponding to the target question can be obtained, and a target evaluation model matched with the question type in a plurality of evaluation models is called to evaluate the reply result. In the embodiment, in the multiple evaluation models, different evaluation models correspond to different question types, and different evaluation models have different parameter quantities. Based on this, a scheme for cooperatively evaluating the reply results output by the target model for different types of questions based on the evaluation models with different parameter quantities is realized, so that the evaluation model with low parameter magnitude can be conveniently adopted to evaluate the reply results of some questions with low complexity, the resource consumption is effectively reduced, and the evaluation efficiency is improved. And an evaluation model with relatively high parameter quantity can be adopted to evaluate reply results of some questions with relatively high complexity, so that the evaluation accuracy is improved.