131results about How to "Improve the accuracy of evaluation" patented technology

The invention discloses a pronunciation quality evaluation method of language study machine in the computer subsidiary language study and phonetic technique domain, which is characterized by the following: extracting exercise phonetic feature; exercising standard pronunciation model; forming standard pronunciation network; detecting phonetic end; extracting evaluation phonetic feature; searching optimum path; calculating the mark of pronunciation quality. The method displays objective and stable evaluation, which constitutes imbedded English study system and mutual human-machine education and oral English self-detection.

The invention discloses a method for partial reference evaluation of wireless videos based on space-time domain feature extraction, which relates to a method for evaluating video quality. On the basis of performing profound understanding and detailed analysis on the existing video objective quality evaluation model, by combining visual characteristics of human eyes, the invention provides the method for the partial reference evaluation of the wireless videos based on the space-time domain feature extraction to improve the conventional SSIM model. The method takes the fluency of the time domain and the structural similarity and the definition of the space domain of the videos as main evaluation indexes, and under the condition of ensuring the evaluation accuracy, the method extracts characteristic parameters of ST domains (the space domain and the time domain), establishes a new evaluation model, reduces the reference data needed by the evaluation and lowers the computational complexity to ensure that the evaluation model is suitable for evaluating the quality of wireless transmission videos in real time.

The invention discloses a spoken language evaluation method based on deep learning and a spoken language evaluation system. The method provided by the invention is characterized in that voice segment intonation degree can be evaluated by adopting the deep learning algorithm, and the evaluation of the intonation accuracy of the tested voice can be acquired; the voice emotion degree can be evaluated by adopting the deep learning algorithm, and the evaluation of the emotion accuracy of the tested voice can be acquired; and the overall evaluation of the pronunciation quality of the whole sentence can be carried out by adopting the deep learning algorithm. By establishing the deep belief network mode, the DBN(deep belief network) model can be used for the spoken English test, and the evaluation of the pronunciation of the spoken English can be more comprehensive and more accurate, and at the same time, the deep learning algorithm has the higher evaluation accuracy by comparing with the emotion evaluation of the shallow model.

The embodiment of the invention provides a method and a device for obtaining transverse relaxation time distribution of a target storage layer. The method comprises the steps of setting corresponding experiment temperature and pressure according to temperature and pressure information of the target storage layer, and obtaining nuclear magnetic resonance intrinsic transverse relaxation time distribution of free state formation water of the target storage layer, nuclear magnetic resonance intrinsic transverse relaxation time distribution of crude oil of the target storage layer and nuclear magnetic resonance intrinsic transverse relaxation time distribution of formation water of the target storage layer; and calculating a hydrogen-containing index of the crude oil, a hydrogen-containing index of formation water, a diffusion coefficient of the crude oil and a diffusion coefficient of the formation water, then replacing intrinsic relaxation and diffusion relaxation comprehensive contribution of the crude oil by intrinsic relaxation and diffusion relaxation comprehensive contribution of the free state formation water, acquiring a rotating echo signal of the target storage layer under a complete water containing state, and performing inversion on the rotating echo signal to obtain the nuclear magnetic resonance transverse relaxation time distribution of the target storage layer under the complete water containing state.

The invention discloses a construction enterprise credit evaluation method and system based on big data technology, which is used to provide an enterprise credit evaluation mechanism applicable to the construction industry. In this method, the multi-dimensional credit characteristics of multiple construction enterprises are obtained through big data technology; the fuzzy consistent judgment matrix is ​​established according to the multi-dimensional credit characteristics of construction enterprises; the initial credit evaluation index model is determined according to the fuzzy consistent judgment matrix. The number of indicators and the weight coefficients corresponding to each indicator; use the initial credit evaluation indicator model to output the initial credit diversity of each construction enterprise; use the initial credit diversity of each construction enterprise as training sample data, and conduct in-depth analysis of the BP neural network model Learning and training, output the final credit evaluation index model when the BP neural network model converges; use the final credit evaluation index model to conduct credit evaluation on the target construction enterprise, and output the credit score of the target construction enterprise.

The invention provides a video wonderful degree evaluation method. The method comprises steps of obtaining the target video; splitting the target video into a plurality of video clips; extracting video feature and audio feature of each video clip; and inputting the video feature and the audio feature corresponding to each video clip into a pre-trained regression neural network model to obtain respective wonderful degree score of each video clip, and performing comprehensive operation on the wonderful degree score of each video clip to obtain the wonderful degree score of the target video. According to the method, after the wonderful degree score of each video clip is determined, the wonderful degree of the target video is comprehensively determined according to the wonderful degree scores of the plurality of video clips, and compared with the prior art of only detecting whether the wonderful segments are included or not, the method comprehensively evaluates the wonderful degree of the whole target video by integrating the wonderful degrees of all video clips, so that the evaluation accuracy is higher.

The invention discloses a human body posture behavior intelligent evaluation method. The method comprises the following steps: S1, extracting human body joint point information; S2, constructing feature vectors; S3, realizing action alignment; S4, analyzing action matching similarity; S5, realizing behavior evaluation. According to the method, the human body part feature vectors are constructed byutilizing the human body joint point information, the standard degree intelligent scoring of the to-be-evaluated human body posture behavior based on the reference human body posture behavior is realized on the formulated behavior similarity scoring system, and the evaluation accuracy is high.

The invention discloses a traffic information evaluation method and a traffic information evaluation device, and relates to a traffic information technology. The traffic information evaluation method in the embodiment of the invention comprises the following steps of: performing vehicle-mounted positioning equipment measurement, and determining the measured data of merged road sections, wherein each merged road section is formed by merging at least two road units which are sequentially connected; and evaluating traffic information data according to the measured data of each merged road section. Each of main roads and secondary main roads has a great number of signal lamps, and the influence of each signal lamp on the measured data is reduced by the merged road sections, so that the influence of the signal lamps on traffic information evaluation is reduced, and traffic information evaluation accuracy is improved.

The invention belongs to the technical field of marine coral biological research, and particularly relates to a coral distribution and health condition evaluation method based on deep clustering analysis. Acquired video images are spliced into a large image through an improved image splicing algorithm, target recognition detection is performed on warm water coral through a deep learning algorithm,then image segmentation is performed on a recognition target through a k-means segmentation algorithm, and finally coral distribution and health conditions are researched. According to the method provided by the invention, the image splicing efficiency is improved; smooth seamless splicing of the coral images is realized; the coral in the image is rapidly detected by using the YOLOV3, the coral coverage rate is improved by calculating the coral area through image segmentation, so that the overall calculation rate of coral distribution and health conditions is greatly improved, and a good effect is achieved for researching the coral reef distribution condition and health conditions.

The invention discloses a system for evaluating urban bus stop service accessibility, and belongs to the field of intelligent transportation. The system comprises a data acquisition unit, a data processing and analysis unit and an accessibility diagnosis unit. Data acquired by the data acquisition unit is transmitted to the data processing and analysis unit, the data processing and analysis unit analyzes the data to obtain bus station service accessibility and transmits the bus station service accessibility to the accessibility diagnosis unit, and the accessibility diagnosis unit acquires thebus station service level of a certain area. According to the invention, the accessibility of urban bus stops can be diagnosed and evaluated, so that the problems of excessive mechanization and insufficient precision of traditional bus accessibility evaluation are solved.

The invention provides an intelligent electric meter state evaluation method based on neural network health degree evaluation, which is a method for establishing and applying a model based on a neuralnetwork, and comprises the steps of A1 a model initial data acquisition method of acquiring network-accessed intelligent electric meter operation related data as initial input data of a neural network model through a power grid marketing business system; A2 screening model input data; A3 model parameter optimization of determining each parameter of the neural network model, and performing parameter training on the neural network model to establish an evaluation model based on the neural network, wherein the evaluation model comprises an electric meter index evaluation model and an electric meter state comprehensive evaluation model; A4 in the application of the state evaluation structure, according to the evaluation result of the evaluation model and the operation cycle of the electric meter participating in evaluation, analyzing the internal relation between the evaluation model and the operation cycle, and predicting the operation sampling inspection time and replacing selection parameters of the electric meter. According to the invention, the electric meter big data accumulated in power grid operation can be utilized to evaluate the on-grid electric meter.

The invention provides a speech level evaluation method and system. The method comprises the steps of receiving speech voice data to be evaluated; conducting voice recognition on the speech voice data to obtain recognition text data; conducting characteristic extraction on the speech voice data and the recognition text data respectively to obtain speed evaluation characteristics. The speed evaluation characteristics comprise an acoustic speech evaluation characteristic and a semantic speech evaluation characteristic. According to the speed evaluation characteristics and a pre-constructed speech level evaluation model, evaluating the speech level of the speech voice data to obtain an evaluation result of the speech level. According to the speech level evaluation method, when the speech level is evaluated, no human participation is involved, therefore the practicability, the objectivity and the accuracy are improved, and the cost is lowered.

The invention discloses a digitized quantitative evaluation method of oilfield waterflooding process. The method comprises the following steps of 1, creating a waterflooding process precision management effectiveness evaluation index system structure; 2, creating each level evaluation element fuzzy judgment matrix; 3, according to the fuzzy judgment matrix, creating a fuzzy consistent matrix of evaluation elements; 4, according to the fuzzy consistent matrix, determining each evaluation index weight by using hierarchy rank-ordering; 5, combining oilfield waterflooding development status and waterflooding management check requirements, making a degree of membership table of different index ranges of each evaluation factor; checking the degree of membership according to the value of the eachevaluation factor, and obtaining quantitative results according to the weight. The digitized quantitative evaluation method improves the efficiency and timeliness of analysis evaluation; the digitized quantitative evaluation method takes full advantage of oilfield informatization construction achievements, ensures the timeliness and authenticity of data, has high accuracy, overcomes problems of difficulty in index data acquisition and the like, and provides basis for the quantitative management check of the waterflooding process.

The invention discloses a supercritical fluid well drilling and completion simulation test device comprising a main circulation loop formed by connecting a storage tank, a booster pump, a heater, a well drilling and completion simulation test platform, a separator and a cooler in turn, wherein the well drilling and completion simulation test platform comprises a shaft the bottom of which is provided with a simulated core; a drilling tool is arranged above the simulated core, and is provided with a hollow drilling tool channel for the supercritical fluid to flow; an annular cavity is formed between the drilling tool and the shaft; and a shaft outlet communicated with the inner chamber of the shaft is arranged on the side wall at the upper part of the shaft. By the supercritical fluid well drilling and completion simulation test device in the invention, the supercritical fluid well drilling and completion test process can be simulated more objectively and truly, and the shaft parameter distribution and the action condition on the stratum in the process of stimulating supercritical fluid well drilling and completion can be reflected intuitively. The supercritical fluid well drilling and completion simulation test device can be used for multiple functional expanding tests besides supercritical fluid well drilling, thus the device has wide applications.

The invention provides a system and a method for evaluating the damage on reservoir by cement slurry. The system comprises a cement slurry pollution device, a stratum environment simulation device, arock core fixation device and a control device, wherein the control device is connected to the cement slurry pollution device, the rock core fixation device is respectively connected to the cement slurry pollution device and the stratum environment simulation device, the control device is used for setting the working parameters for the cement slurry pollution device, the rock core fixation deviceis used for fixing rock so as to make one end surface of the rock contact the test solution in the cement slurry pollution device and make the other end surface of the rock core contact the stratum water in the stratum environment simulation device, and the pressure difference between the two end surfaces of the rock core is the pressure difference between the wellbore and the stratum during cementing. According to the present invention, with the evaluation system, the accuracy of the evaluation of the damage on the reservoir by the cement slurry is improved.

The invention provides a fuel cell engine performance evaluation method, which adopts a fuzzy analytical hierarchy process to comprehensively evaluate each index of the performance of a fuel cell engine. According to the fuel cell engine performance evaluation method, the performance of the engine is evaluated through using at least two levels of indexes, the defects that an existing fuel cell engine system is one-sidedness and cannot be quantified are overcome, the problem that consistency detection of an analytic hierarchy process judgment matrix is complex and tedious is solved, the performance of the fuel cell engine is quantified, and evaluation is accurate and comprehensive.

The invention provides a transmission line security reliability evaluation method, which comprises the following steps: S1, acquiring the transmission line satisfies N-1, N-2 standard proportions, calculate N-1, N- Pass Rate Score = N-1, N-2 Pass rate* 100; 2, respectively calculate that proportion of the heavy-load line, the proportion of the light-load line, the maximum load rate and the load unbalance degree, and marking the rationality of the power flow distribution according to the difference between the average value of the four indexes of the heavy-load line proportion, the proportion of the light-load line, the maximum load rate and the load unbalance degree and the four indexes; 3, calculating that abundance of the short-circuit current to obtain a rationality score of the short-circuit current; S4: Scoring the variable capacitance load ratio according to the difference between the variable capacitance load ratio and the standard interval; 5, scoring that reliability of electric power use according to the load reduction probability and the expected power supply shortage quantity in a year; S6: Integrating Step S1-S3 socres. The invention realizes the comprehensive evaluation through the combination of safety reliability and electric power reliability.

The present invention provides a method and a device for determining a shear-wave attenuation factor in multi-pole array acoustic logging. The method comprises the following steps: acquiring multi-pole array acoustic logging data of a target stratum; according to the multi-pole array acoustic logging data of the target stratum, determining a corresponding flexural-wave attenuation factor; and determining the shear-wave attenuation factor at preset frequency by applying the flexural-wave attenuation factor, and evaluating the target stratum according to the shear-wave attenuation factor. By theadoption of the method provided by the invention, the shear-wave attenuation factor in multi-pole array acoustic logging can be effectively and accurately acquired, and then the evaluating accuracy of a target reservoir as well as the accuracy and the reliability of a reservoir exploration-development scheme can be effectively improved.

The invention discloses a device for evaluating the performance of an SCR catalyst for synergistically removing nitric oxide, benzene and methylbenzene. The device comprises a first gas distribution system, an NH3 gas distribution system, a gas mixing system, a temperature control system, a reaction system and an analysis system which are respectively in electric control connection with an instrument control system, wherein the first gas distribution system, the gas mixing system, the reaction system and the analysis system are sequentially arranged and connected from upstream to downstream along a gas path, the NH3 gas distribution system is connected with the reaction system, and the temperature control system is respectively connected with the gas mixing system and the reaction system.According to the invention, the denitration performance and the volatile organic compound removal performance of the SCR catalyst can be simultaneously detected, the safety is high, the temperature window is wide, the flow rate is adjustable, the proportion selectivity of each pollutant component is high, the application range is wide, the reliability is good, and the performance of the SCR catalyst for synergistically removing nitric oxide, benzene and toluene can be detected on line.

The invention discloses a search quality evaluation method and device and electronic equipment and a search engine bad case mining method and device and electronic equipment. The search quality evaluation method comprises the steps that quality characteristics, exerting an influence on search quality, of a to-be-evaluated search result are acquired; and the search quality of the to-be-evaluated search result is acquired through a search result classifier generated in advance according to the quality characteristics. By the adoption of the method, the search result classifier can be updated according to a newly added quality evaluation result to improve model accuracy of the search result classifier, and therefore the effect of improving evaluation accuracy of the search quality is achieved.

The invention relates to an evaluation method of refining device parts prone to corrosion. The method includes: determining a main failure mode and a corresponding inspection plan by performing RBI risk analysis on equipment and pipelines; using HAZOP to perform formal, systematic and strict inspection on a newly-built or existing process device and project nature so as to evaluate the hazard possibility of a single device and possible influence on a whole set of device; using an LEC analyzing method to perform semi-quantitative safety evaluation on the hazard sources in an operation environment with potential hazard; analyzing and correlating the steps above to determine the parts prone to corrosion and corrosion processes so as to perform corrosion evaluation analysis and carry out preventive monitoring; comprehensively using the HAZOP, LEC and RBI analyzing methods to perform corrosion evaluation analysis on parts without evident corrosion problems so as to carry out preventive monitoring. The method is highly scientific, easy to use, high in evaluation accuracy and capable of effectively carrying out preventive monitoring.