112results about How to "Improve the evaluation effect" patented technology

The invention discloses a method for controlling aeration in a biochemical reaction tank in real time. The method comprises the following steps of: arranging a monitor at a water inlet end of the reaction tank, in the tank, at a water outlet end and at a sludge reflux port of a sedimentation tank; transmitting water quality and water quantity data acquired by the monitor to a central control system through a PLC (Programmable Logic Controller); calculating oxygen demand, a dissolved oxygen target value and sludge reflux quantity of the biochemical reaction tank by a real-time biochemical reaction model module in the central control system according to the water quality and water quantity data; converting the oxygen demand of the biochemical reaction tank into air supply by an intelligent conversion module; calculating air valve opening and fan pressure data of the reaction tank by an air supply pipe network aerodynamic model module according to the air supply; transmitting the sludge reflux quantity, the air valve opening and the fan pressure data to an executing mechanism by the central control system; and controlling opening and closing of a valve, a fan and a reflux pump by theexecuting mechanism. By applying the method, real-time on-line control over the aeration of the biochemical reaction tank is realized.

The invention relates to a user-service awareness assessment method and device. The method includes: acquiring IP (internet protocol) messages of a control plane and a user plane on the mobile internet; integrating the IP messages with identical quintuple information into a service flow; integrating http streaming media service flows in the service flow to each http streaming media service operation record under single-user single-service granularity according to a redirector server IP address, user IP addresses, a GTP (GPRS tunneling protocol) tunnel ID (identifier) and time parameter; according to playing related parameters and download related parameters of the http streaming media service flows in the http streaming media operation records under the single-user single-service granularity, calculating a KQI (key quality indicator) which is enabled to indicate true service awareness of each single user. Therefore, assessment precision and assessment effect are improved.

The invention discloses a credit assessment method. The method comprises the steps that the historical credit data are acquired and an AI risk control engine is constructed according to the historicalcredit data; when a loan request is received, the user pre-loan data of the borrowing user corresponding to the loan request are acquired and the user pre-loan data are inputted to a preset expert engine and the AI risk control engine to perform credit assessment; and the expert assessment result of the preset expert engine and the AI assessment result of the AI risk control engine are acquired and the corresponding decision and suggestion are generated based on the expert assessment result and the AI assessment result. The invention also discloses a credit assessment device and equipment anda computer readable storage medium. The credit decision and suggestion are acquired by using the double-engine assessment mode so that the adverse influence of the limitation of human thinking on engine construction and the credit assessment process can be greatly reduced, the assessment capability of the risk control model can be enhanced and the rate of bad loans can be effectively reduced.

The invention provides a scene classification method and device for automatic driving vehicle road testing. The method comprises the following steps: determining one or more pre-created road scenes and scene composition elements of the road scenes, and dividing the scene composition elements into a first composition element and a second composition element; calculating to obtain a first score according to the information amount contained in the first component element and a preset score corresponding to each scene component element; in a real scene test or a simulated scene test of the vehicle, obtaining test information generated by each scene component element and each scene component element in the second component elements, and inputting the test information into a predetermined modelstructure to obtain a second score; and calculating to obtain a third score according to the first score and the second score so as to realize classification of the road scene based on the third score. According to the invention, the complexity of dividing different types of road scenes is realized, and the test efficiency and the evaluation effect of the automatic driving vehicle are improved.

The invention discloses a main shaft bearing state evaluation method and device based on a deep fusion network, and the method specifically comprises the steps: converting a vibration signal, a current signal and a temperature signal of a main shaft bearing under different working conditions into a frequency spectrum signal through fast Fourier transform, taking a low-frequency component of the frequency spectrum signal, and processing the low-frequency component to obtain a training set and a test set; constructing an Inception branch network based on an Inception module and a convolutional neural network, constructing a central layer fusion network based on features extracted by each Inception branch network, and forming a deep fusion network by the branch networks and the central layerfusion network to realize state evaluation of the variable working condition spindle bearing. By fully mining the coupling characteristics of various sensing information, the defect that a single signal characteristic is difficult to comprehensively characterize the running state of the main shaft bearing is overcome, and the accuracy of main shaft bearing state evaluation is improved; coupling information of all sensing data is fully extracted from the perspective of multi-source information fusion, and the robustness to noise and different working conditions is enhanced.

The invention discloses a method for detecting the quality of oranges in different maturity on the basis of a taste-smell fingerprint spectrum. The method comprises the following steps of washing the oranges, peeling the oranges, juicing the oranges, and obtaining the orange juice; carrying out an electronic nose experiment: placing the filtered orange juice in a closed vessel, standing, detecting an overhead air by adopting an electronic nose, extracting characteristic data from the smell fingerprint spectrum, and adopting the extracted characteristic data as an original data of the electronic nose part; carrying out the electronic tongue experiment: placing the filtered orange juice in a special beaker for an electronic tongue, enabling an electronic tongue sensor array to contact sample liquid, generating a sensor response signal, extracting characteristic data from the taste fingerprint spectrum, and adopting the extracted characteristic data as the original data of the electronic tongue part; processing response values from different systems in three different preprocessing methods, so that the generalization capability of a model can be improved. By adopting the method, the high-efficient data fusion of the taste and smell fingerprint spectrum can be realized, the quality of the orange in different maturity can be directly, conveniently, objectively and rapidly evaluated and predicted, and the popularization and application value is relatively high.

The invention relates to a search engine user information demand satisfaction evaluation method capable of integrating multiple views and semi-supervised learning. The method is divided into the following six stages: preprocessing data, training a subview satisfaction model, distributing a dummy tag for unlabeled data, training a user satisfaction model based on the multiple views and the semi-supervised learning, and carrying out evaluation. Through a semi-supervised learning method, a small quantity of labeled data and a great quantity of unlabeled data are used for improving the performance of an evaluation model, and a multi-view learning method is imported to overcome the problem that a traditional single-view based semi-supervised learning method is always caught in local optimum. The search engine user information demand satisfaction evaluation method has the beneficial effects: (1) under the condition of the small quantity of labeled data, the search engine user information demand satisfaction can be effectively evaluated; (2) the small quantity of labeled data and the great quantity of unlabeled data can be used for improving the evaluation performance of the user satisfaction model; and (3) a search process of the user can be independently described from angles of behaviors and time, and the model can be prevented from being caught into the local optimum through mutual learning.

A coal seam blasting pressure relief effect assessment method includes steps of arrangement of microseismic sensors, calculation of blast hole microseismic event energy, determination of judgment criteria and the like. The coal seam blasting pressure relief effect assessment method is high in operability and precision and reliable in result, and the problem of difficulty in quantitative assessmentof blasting effects is solved.

The invention discloses a cellular automaton-based arid region natural oasis space dynamic simulation method. The cellular automaton-based ecological water delivery driven arid region natural oasis space dynamic simulation method includes:, based on normalized vegetation indexes, underground water burial depth observation data and other multi-source information obtained through remote sensing images, simulating the total oasis area by adopting an underground water bearing capacity function and an oasis area dynamic function; constructing a cellular NDVI, a neighborhood vegetation rate, a neighborhood flooding rate and a cellular state change probability function according to the NDVI sequence; calculating the coating conversion probability at the t + 1 moment by adopting the cellular NDVI,the neighborhood vegetation rate and the neighborhood flooding rate at the t moment; and calculating a coating conversion score of each cell by adopting a weighted average method, sorting the coatingconversion scores from large to small, setting a coating conversion dynamic threshold in combination with the oasis area simulation result, finally judging how to change the state of the cell in combination with the coating conversion static threshold, obtaining an oasis space dynamic simulation result, and reflecting a model simulation effect through precision evaluation.

The invention provides a method for predicting the waving form of thin cold-rolled strip steel in the width direction. The method is implemented on the basis that the residual stress value and the coordinate values of the thin cold-rolled strip steel are obtained and comprises the following steps that (1), a left auxiliary vector and a right auxiliary vector are calculated; (2), the residual stress positive maximum value vector including the endpoint value and the coordinate vector are calculated; (3), the residual stress negative minimum value vector including the endpoint value and the coordinate vector are calculated; (4), the width vector is calculated; (5), the amplitude vector is calculated; and (6), a function for the waving form in the width direction is determined. By the adoption of the method, a waving area and the waving form in the width direction of the thin cold-rolled strip steel can be predicted, and the capacity of evaluating the strip steel quality during cold rolling production is improved.

An embodiment of the invention relates to the technical field of wireless communication, in particular to a simulation method and a simulation device, which are used for reflecting time-variation characteristic of signal channels on the premise of efficiency improvement and are capable of reflecting processing gain of transmission modes and algorithms such as detection algorithm. The simulation method includes: by means of dynamic pre-simulation, determining corresponding relation between a dynamic index value and a static index value which correspond to simulated scenes; according to the determined corresponding relation between the dynamic index value and the static index value, determining the dynamic index value, corresponding to the static index value, obtained by means of static simulation. By the simulation method and the simulation device, both simulation efficiency and performance of systems can be considered. Simulation efficiency is improved greatly since the simulation process is simplified and loss of the simulation performance is low. Assessment performance of dynamic simulation can be acquired according to the static simulation efficiency, so that the assessment performance is improved on the basis of guaranteeing performance reliability.

The invention discloses a dairy cow feed intake evaluation method for optimizing a BP neural network based on a genetic algorithm. The method comprises the following steps: 1, determining a topological structure of the BP neural network; 2, preprocessing sample data; 3, initializing a population; 4, calculating a population fitness value; 5, selecting, crossing and mutating; 6, judging whether evolution is completed or not; 7, constructing a model; 8, verifying the model. According to the method, a feed intake evaluation model is established by adopting a method of optimizing a BP neural network weight threshold by adopting a genetic algorithm, and scientific basis and theoretical guidance are provided for accurately evaluating the change of the feed intake of dairy cows and reasonably controlling the change rule of the feed intake of the dairy cows. According to the method, the defect that the BP neural network may fall into local optimum is overcome, the convergence rate of the modelis improved, and a relatively high evaluation effect is achieved.

The invention discloses an LTE (long-term evolution) communication system speech quality evaluation method. The method includes the following steps: initializing E-Model model parameters; extracting relevant parameters; extracting speech segments; calculating speech encoding and decoding loss; calculating delay loss; calculating jitter loss; calculating MOS (mean opinion scores) and transmission grade factors R of speech evaluation, and converting the transmission grade factors R into corresponding MOS. An RTP (real-time transport protocol) and an RTCP (real-time transport control protocol) of an application layer are introduced into a speech quality evaluation scheme of an LTE communication system, and speech quality loss caused by delay and jitter in the LTE communication system can be evaluated effectively; the RTCP is introduced into an estimation method of LTE communication system network packet delay, and evaluation performance of the speech quality loss caused by delay can be improved effectively; the RTP is introduced into an estimation method of LTE communication system network packet jitter, evaluation performance of the speech quality loss caused by jitter can be improved effectively, and dispatching control and wireless resources allocation can be performed adaptively.

A variable selection method based on partial least square aims at resolving the problems of poor prediction capability and incapability to improve prediction performance of the existing variable selection method. The variable selection method based on the partial least square includes preprocessing variables to be analyzed to obtain a standardized variable set and calculating regression vector according to the standardized variable set. Elements of the regression vector indicates contribution rate of corresponding input variables to output, and the order of the contribution rate is reduced gradually. The method further includes ordering the input variables according to the order of the contribution rate of the regression vector to obtain an input variable order Xior, conducting a correlation test on all the input variables in the input variable sequence Xior and conducting selection calculation on the input variables to obtain regression coefficient. The variable selection method is applied to a process control system in the fields of steel, boiler, chemical industry, pharmacy and the like.

The invention discloses a sleep disorder comprehensive testing and training method and a system based on a virtual reality situation. The method comprises the following steps: S1, detecting the sleepquality of a target user to judge whether the target user is a sleep disorder user with sleep disturbance; if so, executing step S2; 2, carrying out targeted symptom analysis on that sleep disorder user according to the sleep quality detection result, so as to obtain the cause of the sleep disorder for the sleep disorder user; S3. constructing a first virtual scene for the cause of sleep disorderuser disorder, and providing the sleep disorder user with guided breathing training adapted to the cause of sleep disorder in the first virtual scene; S4. after completing the guided breathing training in step S3, performing post-training sleep quality evaluation on the sleep disorder user, and when the evaluation is unqualified, returning to the step S3. The method and system can locate differentsymptoms of different users, has stronger pertinence in training, and enable each user to understand its own problems.

The invention provides a credibility evaluation method for a complex simulation system based on task orientation. The credibility evaluation method comprises the following steps of carrying out pre-analysis on the complex simulation system; for a simulation task to be executed by a complex simulation system, segmenting the simulation task into a plurality of sub-tasks according to different stagesof an execution process, constructing an information network corresponding to each sub-task, and superposing the information networks according to a time sequence to form a system cooperative network; converting and mapping the system cooperative network into a credibility index network of the complex simulation system, and calculating a credibility conduction coefficient between nodes accordingto the comprehensive credibility of cooperative operation of every two adjacent nodes and the credibility of a single node; and forming a credibility conduction chain according to the credibility conduction coefficient in the index network and the credibility of a single node and a time sequence, and deducing to obtain the final credibility. The invention provides an implementation method for credibility evaluation of a complex simulation system, and provides technical support for evaluation and optimization improvement of the complex simulation system in the fields of aviation, spaceflight, military affairs, traffic and the like.

The present disclosure provides for systems and method for producing hybrid seed. In various embodiments, the disclosure provides a system for the high-throughput, nondestructive sampling of seeds. In another embodiment, a high-throughput, nondestructive method for producing hybrid seeds comprises removing a sample from a plurality of seeds in the population while preserving the germination viability of the seed and analyzing the sample for the presence or absence of one or more genetic markers indicative of a male-sterile gene.

The invention discloses a method for identifying a rigidity weak link of an elastic supporting beam by utilizing a curvature angle index, belongs to the field of mechanical vibration, and solves a keyproblem of identifying the rigidity weak link of a combined machine tool structure. A machine tool structure which can be equivalent to an elastic supporting beam structure is adopted. The method comprises the steps of taking the finite element model of the elastic support beam structure as a numerical example, acquiring dynamic response data of each node of the numerical model by utilizing a state space method, establishing a relationship between data fitted by a least square method and actually measured data, and identifying a rigidity weak link by adopting a curvature angle new index. Modal parameters under all orders of modals are obtained through a hammering method, relative data such as displacement are obtained through a modal state space method, and inaccuracy caused by a displacement relative measurement method and interference of noise on the data obtaining process in the experiment test are eliminated through system reconstruction. According to the method, the problem of inaccurate joint surface modeling in finite element analysis is solved, and the problems of algorithm convergence and blindness in an Update method can also be solved.

The invention discloses a coherent filter updating device considering scene classification during target tracking. The device comprises a target blocking module, a parameter acquisition module, a target scene judgment module and an adaptive learning updating module. In a practical target tracking process, many realistic conditions will impose influences, e.g., a target deforms or is sheltered, particularly long-time big-area sheltering happens; and due to the fact that existing coherent filters have limited understanding on the scenes, the performance of a tracker will be substantially lowered. In order to improve tracking robustness, the adaptive learning updating device is arranged, a coherent filter is subjected to adaptive updating according to different scenes of the target, and the influences of long-time big-area sheltering, appearance changes of the target and other different scenes on tracking performance are solved.

The invention belongs to the technical field of tourist sports behavior monitoring, and discloses a big data analysis based tourist sports behavior real-time information monitoring method. The methodincludes: through a physiological data acquisition module, acquiring data of heart rate, blood pressure and body temperature of tourists by electronic physiological detection equipment; through a positioning module, positioning location of the tourists by positioners; through an image acquisition module, acquiring image data of tourist amount in tourist areas by cameras; through a travel route planning module, planning routes of the tourists by tourist software; through a big data analysis module, performing big data analysis on the tourists according to big data resources by cloud servers; through a sports estimation module, estimating sports ability of the tourists by estimation software. The defect, namely complete dependence on medical staff's experiences, of the conventional user sports ability estimation methods can be overcome, and estimation effect and accuracy to real sports ability of users can be improved effectively.

The embodiment of the invention provides a data processing method and a device of an intelligent customer service system, and electronic equipment, and relates to the technical field of artificial intelligence. The method comprises the following steps: firstly, acquiring session service data and user behavior data of the intelligent customer service system in a set period; determining an index value of each quality evaluation index in a set period based on the session service data and the user behavior data; and determining the service health degree of the intelligent customer service system according to the index value of the quality evaluation index, the service health degree being used for representing the service quality of the intelligent customer service system in a set period. According to the method, the service health degree is determined through the session service data and the user behavior data of the intelligent customer service system; the service health degree is used for evaluating the service quality of the intelligent customer service system, a test set does not need to be maintained, data adopted for evaluation is richer, comprehensive evaluation on the service quality of the intelligent customer service system can be achieved, and the evaluation effect is improved.

The invention provides a phase-based partial discharge generation device and a phase-based partial discharge generation method with adjustable discharge capacity and density. The phase-based partial discharge generation device comprises a main control unit, a PWM circuit, a filtering circuit, a pulse generation circuit, a phase shaping circuit, a key and a knob switch, wherein the pulse generationcircuit is used for generating a discharge pulse according to an instruction of the main control unit; the PWM circuit is used for generating a pulse voltage according to the instruction of the maincontrol unit, after the pulse voltage is stabilized and smoothed into a DC voltage by means of the filtering circuit, inputting the DC voltage into the pulse generation circuit to perform voltage regulation on the pulse generation circuit, and further outputting stable discharge voltage; and the phase shaping circuit is used for shaping a phase pulse output from the main control unit into a sinusoidal signal to serve as a phase output signal. The phase-based partial discharge generation device and the phase-based partial discharge generation method can realize verification and performance evaluation of a partial discharge monitoring system and equipment thereof, thereby guaranteeing the safe operation of the partial discharge monitoring system, finally guaranteeing the safe operation of cables, and guaranteeing the stability and reliability of power supply.

The invention discloses a video target real-time tracking method and system based on depth feature fusion and adaptive correlation filtering. The system comprises a feature extraction module, a feature fusion module and a correlation filtering module. The feature extraction module uses a lightweight network model to extract multi-layer depth features, and the real-time performance of a target tracking task is ensured. The feature fusion module provides a multi-layer depth feature fusion strategy of typical correlation analysis for the problem that independently extracted multi-layer features are not complete enough in representation of a target. According to the method, the target expression capability and the target and background distinguishing capability are improved, the feature redundancy is reduced, and the calculation amount of subsequent related filters is reduced. The correlation filtering module provides a correlation filter updating strategy based on response value dispersion analysis for solving the tracker drifting problem caused by challenges that the target is deformed, shielded, moved out of the view, and is rotated existing in a target tracking task, filter template updating is carried out in a self-adaptive mode, and the specific problems are relieved.

The invention provides a battery cell modeling method. The battery cell modeling method comprises the following steps: S1, measuring the three-dimensional size of a battery cell; S2, extruding the battery cell along the thickness direction by adopting a first pressure head to obtain an F1-d1 curve of the battery cell; S3, calculating an equivalent sigma-epsilon curve of the battery cell; S4, selecting an equivalent material according to the equivalent sigma-epsilon curve, endowing parameters of the equivalent material to the battery cell model, performing benchmarking correction on a simulation curve F1'-d1' curve when the battery cell model is extruded by a first pressure head model, and establishing an equivalent stiffness finite element model of the battery cell; S5, extruding the battery cell along the thickness direction by adopting a second pressure head to obtain an F2-d2 curve; and S6, on the basis of the step S4, replacing the first pressure head model with a second pressure head model for simulation and benchmarking correction, and establishing a final battery cell equivalent mechanical model. According to the battery cell equivalent mechanical model established by the method, the simulation precision of the battery pack can be improved, the stress of the battery cell is evaluated more accurately by utilizing a simulation means, and the evaluation capability of the CAE technology on the safety of the battery pack is expanded.

The invention discloses a financial risk management system and method for a financing supply chain. The method comprises the following steps: acquiring supply chain information of a preset central enterprise terminal, performing feature construction processing on acquired data, calculating supply chain risk evaluation data, and obtaining a variable threshold value of a rule variable according to the rule variable; and determining an early warning state of the central enterprise terminal according to the risk coefficient and a variable threshold value, responding to the early warning state, and performing financial risk control on the central enterprise terminal lower than the threshold value. The collected enterprise end data has good authenticity, comprehensiveness and real-time performance, coefficient judgment and risk evaluation are facilitated, judgment of the existing state of a central enterprise can be achieved, the financial risk of a financing supply chain is reduced, and development of a central enterprise end and property safety of supply chain investors are facilitated.

The invention provides a method and a system for evaluating influence of low-illumination stress on the quality of greenhouse tomatoes. The method comprises the following steps: collecting air temperature and photosynthetically active radiation information in a sunlight greenhouse under preset conditions; calculating a heat effect, a radiation effect, a temperature-light effect and an accumulated temperature-light effect of the sunlight greenhouse tomatoes in the flowering and fruit period in each day by combining the upper limit temperature and the lower limit temperature required by the growth and development of the sunlight greenhouse tomatoes in the flowering and fruit period; and calculating a difference value between the CK and the accumulated temperature-light effect between the different low-illumination treatment, calculating a difference value between the CK and the tomato nutritional quality index between the different low-illumination treatment, establishing a dynamic relationship between the accumulated temperature-light effect difference value and the nutritional quality index difference value, and simulating the dynamic influence of the low-illumination stress on the tomato nutritional quality index. According to the method, the advantages of the heat radiation product algorithm are used for reference, the photosynthetic active radiation in the original heat radiation product index is replaced by the light effect by using the temperature-light effect algorithm, the evaluation effect is improved, and the influence of the light stress on the tomato nutritional quality index can be evaluated more accurately.

The invention discloses a fire extinguisher equipment condition monitoring device based on the Internet of Things and a working method of the fire extinguisher equipment condition monitoring device. The invention relates to the field of data acquisition and monitoring. The device comprises a central control and processing module, an environmental condition acquisition module, an NB-IoT communication module, a GSM communication module, a fire extinguisher equipment pressure dial image acquisition module, a motion detection module and an RS232 interface module. The environmental condition acquisition module, the NB-IoT communication module, the GSM communication module, the fire extinguisher equipment pressure dial image acquisition module, the motion detection module and the RS232 interfacemodule are connected with the central control and processing module; and the NB-IoT communication module and the GSM communication module are respectively connected with the server through the base station. According to the invention, through acquisition and local processing of data such as environment temperature and humidity and a fire extinguisher equipment pressure dial plate, and uploading of the data to the server by using the NB-IoT module or the GSM module, remote monitoring and management of the fire extinguisher equipment condition are realized.

The invention discloses a multi-population collaborative entropy cascading method for electronic medical record knowledge reduction performance evaluation. The method includes: dividing a large-scale electronic medical record data set into different electronic medical record optimized data subsets by using a Map operation and a Reduce operation in cloud computing; constructing a neighbouring evolution population similarity matrix, and designing a knowledge reduction performance computing method by using an evolution population collaborative entropy; analyzing an evolution population distribution law affecting the electronic medical record knowledge reduction performance, and constructing a cascading evaluation index matrix and performing matrix optimization; and finally evaluating the electronic medical record knowledge reduction performance evaluation precision, and outputting the electronic medical record knowledge reduction performance optimal evaluation precision. The method has good application value in large-scale electronic medical record knowledge reduction qualitative quantitative intelligent analysis and relevant disease auxiliary diagnosis effect evaluation.

The invention provides a department limit analysis system and method with exponential decay as a model. The system comprises an electronic medical record system, a hospital information system, an assessment module and a hospital office system. The assessment module conducts analysis according to data of a medical statistics system, the basic limit and the per capita limit of each department are obtained through an allocation algorithm, and then the basic limits and the per capita limits are fed back to a patient in-hospital charging system and the hospital office system. The department limit analysis system and method improve the informatization level and assessment capacity of hospitals, and can effectively inhibit the problems of excessive medical treatment and department payment uneven distribution.

The invention discloses a compressed domain based mobile phone football video quality evaluation device and method considering scene classification. The evaluation device comprises a coding compressing module, a parameter acquiring module and a quality evaluating module that are connected in sequence. The evaluation method comprises a first step that the coding compressing module codes and compresses an input original video; a second step that the parameter acquiring module performs half decoding on a coded and compressed video so as to extract compressed domain parameters and processes the compressed domain parameters, wherein processed compressed domain parameters are used for video quality evaluation in following steps; and a third step that the quality evaluating module computes quality of the mobile phone football video according to an established video evaluation model. In the process of quality evaluation of the football video, the evaluation accuracy is improved due to consideration of three different football video scenes, the football video scenes are classified by use of the compressed domain parameters and the quality of the football videos of different scenes is evaluated, so that feasibility and real-time performance of the evaluation are improved effectively.