41results about How to "Improve comprehensiveness and accuracy" patented technology

The invention discloses a speed limiting information base generation and speed limiting information detection method and system. The speed limiting base generation method includes the steps of collecting original vehicle track point information; conducting data washing for all vehicle track point information in the original vehicle track point information to obtain all vehicle track standardized information, wherein the track standardized information includes the road numbers, the track lengths, the track starting point moments, the track ending point moments and the first average speeds; extracting the first average speeds of all vehicles to form cluster object sets according to the road numbers and preset time periods; calculating vehicle speed limiting information corresponding to roads and the time periods for all the cluster object sets to generate a speed limiting information base. According to the method and the system, the speed limiting information of the whole area of the current road is reflected, the renewing period of the speed limiting information is shortened, and therefore the up-to-date state of the speed limiting information is enhanced, and the speed limiting information of the road can be detected under the condition that the speed limiting information does not completely cover the whole area.

The invention discloses a method for calculating problem similarity with answer relevance ranking evaluation measurement. The method includes the steps that 1, an answer ranking model is learned on a training data set in a question and answer knowledge base; 2, quick information retrieval is used for retrieving related problems of a user problem from the question and answer knowledge base; 3, the same candidate answer set of the user problem and the related problems is ranked according to the learned answer ranking model; 4, question-answer relevance ranking evaluation measurement namely Spearman, Pearson and Kendall relevance coefficients are calculated; 5, the obtained relevance coefficients directly serve as similarity scores of the problems or are added into a similarity characteristic set to train a classifier, and then probability confidence scores output by the classifier serve as the similarity score. The problem similarity is calculated through answer similarity, and the problem that the same problem has wide vocabulary gaps caused by different natural language expressions is solved.

The invention discloses a neural-network-based artifact correction method and apparatus of a CT image. The method comprises: preprocessing of CT data is carried out to generate a database carrying outan input sample of at least one kind of artifact; a neural network for correcting the at least one kind of artifact is constructed; the input sample is inputted into the neural network and the neuralnetwork is trained by a preset objective function to obtain a trained neural network; and a to-be-processed CT image is inputted into the trained neural network to carry out artifact identification and correction, and a corrected CT image and an identified artifact image are outputted. On the basis of deep learning, the CT image including multiple kinds of artifacts is processed by the neural network, so that the quality of the CT image is improved.

The invention discloses a movie recommendation method and system based on a multi-modal knowledge graph, and a terminal and the method comprises the steps: obtaining movie pictures and text information according to single-modal features, extracting entities, attributes and relationships in the movie pictures and the text information, after obtaining the knowledge ontology, performing knowledge fusion and knowledge processing on the knowledge ontology and knowledge graph structured data to obtain a multi-modal knowledge graph; projecting the head entity and the tail entity into a vector space through an embedding layer, mapping the head entity and the tail entity into two embedded vectors, and projecting each relation in the path as an independent embedded vector; utilizing a long short-term memory network model to learn an aggregation path of embedded vector output representation entities and relationships; and projecting the final state of the aggregation path by using a full connection layer, outputting path scores, aggregating all the path scores through average pooling, outputting a prediction probability of watching a movie by the user, and displaying the prediction probability as a recommendation score to the user. Comprehensiveness and accuracy of movie recommendation are improved.

The invention provides a virtual machine hidden process detection method and system. By obtaining three views including a user mode process view, a process link table-based kernel view and a CPU scheduling-based trusted kernel mode process view, cross comparison is performed to realize effective detection of a virtual machine hidden process. By implementing the method and the system, a Hypervisordoes not need to be re-compiled and loaded, own businesses of a cloud platform tenant are not influenced, behaviors of performing direct modification for a kernel object to hide an own process and thelike can be effectively detected, and the comprehensiveness, accuracy and high efficiency of virtual machine hidden process detection are further improved.

The invention relates to a historical physiological data and physical data comparison based status monitoring method and device. The device comprises a wearing mechanism, an alarm module, a physiological status detection module, a physical status detection module, a status information storage module and a comparison module. The method includes the steps: 1) continuously acquiring biological characteristics of a user in a normal status and physical characteristics related to the biological characteristics, and storing the characteristics to form historical data; 2) detecting current biological characteristics and physical characteristics of the user, comparing the biological characteristics and physical characteristics with the historical data obtained in the step 1), and judging whether the biological characteristics, related to the physical characteristics similar to the currently detected physical characteristics and in the historical data, in the historical data in the normal status are similar to the currently detected biological characteristics or not; 3) if yes, storing the currently detected biological characteristics and physical characteristics as components of the historical data, and otherwise, sending an alarm signal.

The embodiment of the invention discloses an oil delivery pump fault identification method, device and system based on multi-source information fusion. The method comprises the following steps that amulti-source heterogeneous risk scene database corresponding to an oil delivery pump is constructed; a target risk scene is screened from the multi-source heterogeneous risk scene database according to the scene importance degree of the risk scenes in the multi-source heterogeneous risk scene database; monitoring parameters corresponding to the target risk scene are acquired, wherein the monitoring parameters at least comprise one or more kinds of a degradation state characterization parameter and a process parameter; and fusion parameters of the oil delivery pump are determined according to the monitoring parameters corresponding to the target risk scene, and the fusion parameters are utilized to perform fault identification on the oil delivery pump. By utilizing the embodiments of the specification, the accuracy of fault identification of the oil delivery pump can be improved.

According to the embodiment, the invention discloses a pressure management method and a pressure management system. The method comprises the following steps: acquiring pressure assessment parameters of a user, wherein the pressure assessment parameters include physiological and psychological measurement data, an objective stress level measurement value, a subjective stress level survey value and stress level impact survey data; in accordance with the physiological and psychological measurement data, the objective stress level measurement value, the subjective stress level survey value and the stress level impact survey data, calculating a pressure value of the user; and generating and outputting a pressure feedback exercise mechanism in accordance with the calculated pressure value, so that the user can conduct self-regulation in accordance with the pressure feedback exercise mechanism and the pressure value of the user is kept at a normal level. With the implementation of the pressure management method and the pressure management system, the impact of such factors as physiological, psychological, objective and subjective factors, personal information, external environment and the like on the pressure value is taken into full consideration, so that the accuracy and comprehensiveness of a pressure assessment result are improved, and a real-time, continuous and quantitative pressure management mechanism is established.

The embodiment of the invention discloses a method and device for obtaining user feedback information. The method comprises: after an application program in a mobile terminal is started, user fingerprint information is collected; according to the collected user fingerprint information, time of using the application program by all users is recorded; and when the time of using the application program by the same user is larger than or equal to a preset threshold value, a dialog box pops up and the user fingerprint information inputted into the dialog box by the user is obtained. According to the invention, the user fingerprint information is collected; the time of using the same application program by all users is obtained; dialog boxes pop up for all users; and the user fingerprint information of all users is obtained. Therefore, the comprehensiveness and accuracy of the user fingerprint information can be improved.

The embodiment of the invention provides a data analysis method and device for a target drug. The method comprises the following steps: acquiring a target spot query request; obtaining drug information corresponding to the target point, indication information corresponding to the drug information and a research and development stage of the indication information according to the target point and apreset relation graph; sorting the drug information and the indication information according to a preset weight value at preset time intervals to obtain a corresponding sorting sequence difference value; and determining clinical indication research and development progress information of the drug corresponding to the target point according to the sorting sequence difference value. By adopting thedata analysis method for the target drugs disclosed by the embodiment of the invention, the clinical indication research and development progress data of different drugs under each target can be updated in real time, the data analysis efficiency is greatly improved, and the use experience of a user is effectively improved.

The invention provides a user text information analysis method. The method includes: processing to-be-analyzed text information; carrying out potential topic mining on the preprocessed to-be-analyzedtext information, and obtaining topic probability distribution of the text; calculating the similarity of the text according to the topic probability distribution, and performing user characteristic value clustering according to the similarity; performing digital marking on the clustered to-be-analyzed text information to obtain to-be-analyzed sample data; and inputting the to-be-analyzed sample data into a pre-established user preference analysis model to obtain a user preference analysis result. According to the scheme, the text similarity between the users is calculated by deeply mining thetext features of the users, and clustering analysis is performed according to the similarity distance, so that the structure of a hidden layer of the deep neural network is simplified, and the learning efficiency of the deep neural network is improved.

The invention discloses credit risk monitoring method, device and equipment, and a storage medium. The credit risk monitoring method comprises the steps: collecting original data from a data source, cleaning the original data according to a preset cleaning strategy, and obtaining cleaning data; performing identity normalization processing on the cleaning data according to user identifiers in the cleaning data to obtain a first user data set with the same user identifier; according to a preset mapping rule, replacing the risk event in the first user data set with a reason code corresponding to the risk event to obtain a second user data set; and generating a credit risk form according to the second user data set. According to the invention, processing such as data cleaning, identity normalization, event cause coding and the like can be carried out on the original data, so that an accurate, ordered and integrated data set is provided for credit risk monitoring, and the accuracy and comprehensiveness of business processing are improved.

The invention relates to a method for predicting complaints caused by power failure. Power failure information, client standing book information and user telephone traffic information in a state gridPMS2.0 system, a marketing system and a 95598 customer service system are collected, and data of each dimension is obtained. The data in the three systems are associated through the user station areanumber and the user account number, so that the telephone traffic data in the power failure period and the telephone traffic data in the non-power-failure period are distinguished; for telephone traffic data in a power failure period, firstly, an entropy method is used for determining the power failure duration, whether the power failure is notified in advance or not, and the weight of features influencing telephone traffic data of the peak power failure time; and then according to the influence weight, the difference of the outage event on the outage traffic data between the transformer areasis shielded, and on the basis, depicting the sensitivity degree of the power consumer according to the daily telephone traffic data and the outage traffic data. According to the method, the prediction range is expanded, and the alarm judgment mode is closer to reality.

The invention provides an automatic scoring system and method based on an intelligent power grid dispatching control system, and the system comprises a plurality of examination systems and at least one judgment system. The judgment system is provided with score points and score bases, and the judgment system and the examination systems are in communication connection and can carry out the mutual remote data retrieval. The actual result of examinee operation generated by the examination system is scored by real-time library comparison, standard configuration file and system parameter comparison, non-standard file comparison, relation library comparison and manual judgment of a referee after the file is uploaded to the referee system. According to the invention, comprehensive evaluation of examinee operation is realized, thereby greatly improving the accuracy and comprehensiveness of scoring, saving a lot of manpower, material resources and time costs, avoiding resource waste, and improving the fairness and fairness of examination.

The invention provides a medical image processing method based on an attention mechanism and related equipment, and the method comprises the steps: obtaining a three-dimensional medical image of a to-be-classified object, inputting the three-dimensional medical image into a feature extraction model for feature extraction, and outputting a three-dimensional feature map of the three-dimensional medical image; wherein the feature extraction model is obtained by training a three-dimensional residual convolutional network based on a double attention mechanism, and the double attention mechanism comprises a channel attention mechanism and a depth attention mechanism, so that the feature extraction model can extract deep semantic feature information of the three-dimensional medical image; the accuracy and the comprehensiveness of the three-dimensional feature map are improved, so that the three-dimensional feature map is subjected to pooling processing, the obtained target feature vector is input into the classification model, and the category information of the three-dimensional medical image can be quickly and accurately obtained.

The invention provides a game plug-in detection method and a storage medium. The method comprises the steps of extracting modifying point data for modifying a game client by analyzing a known plug-insample; extracting feature codes containing modifying points corresponding to each modifying point according to the modifying point date; after starting the game client, matching game client memory data and each feature code one by one. The method can not only detect the known plug-in data and the unknown plug-in data at the same time, expand a detection range, improve the detection accuracy, butalso just detect the client memory data to obtain a detection result, significantly improve the detection efficiency, and have broad application prospects.

The invention provides an interlocking simulation model design method applied to an urban rail and an interlocking simulation model, the method is realized through the interlocking simulation model, the interlocking simulation model comprises an interlocking logic operation module, a station yard equipment information maintenance module, an application information processing module and a network communication module, and the method comprises the following steps: through the interlocking logic operation module, calculating a route selection result according to the route selection instruction and the current equipment state; updating the current equipment state of the station equipment through a station equipment information maintenance module according to the calculation result of the interlocking logic operation module; obtaining first application data through an application information processing module according to the current equipment state of the station equipment in a station equipment information maintenance module; and sending the first application data to the vehicle-mounted equipment and the regional control center through the network communication module. According to the method, the function test of the urban rail train control system is realized through the interlocking simulation model, real interlocking equipment is not needed, and the test cost is reduced.

The invention discloses a data processing system and method based on multiple geographic visualization platforms, and the system comprises a geographic information collection module, a data management center, a geographic visualization module, a call data analysis module, and a call data processing module. Data needing to be called, display data and historical operation behavior data of a user are collected through the geographic information collection module, all the collected data are stored and managed through the data management center, the progress of selecting an adaptive geographic database is analyzed through the geographic visualization module, the adaptation probability is predicted, and the availability degree of the geographic database is analyzed. A city construction database is established on the geographic database with the highest availability degree, a temporary visual platform is established, and the size and contour information of a default displayed picture after data are called out are analyzed by calling a data analysis module; the display picture is automatically adjusted through the calling data processing module, unified calling of user data is achieved to the maximum extent, and resource waste is reduced.

The invention discloses a method and system for identifying a camouflage telephone number, and a server, and relates to the technical field of communication. The method comprises the following steps:in response to a number camouflage query request sent by a terminal, comparing a communication record of a to-be-queried telephone number sent by the terminal and initiated to the terminal with the communication record in an actual bill so as to determine whether the to-be-queried telephone number really initiates the communication and identify whether the to-be-queried telephone number is the camouflage telephone number. By adoption of the method and system, any telephone number that initiates a call or sends a short message to the terminal can be identified accurately, and the accuracy and comprehensiveness of identifying the camouflage telephone number are improved.

The invention relates to an inspection task processing method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring inspection task information; determining to-be-inspected equipment pointed by the inspection task information; querying target inspection information corresponding to the to-be-inspected equipment in an inspection database, wherein the corresponding relation between the equipment and the inspection information is stored in the inspection database; generating a target inspection form according to the target inspection information; and controlling an inspection terminal to inspect the to-be-inspected equipment according to the target inspection form. By adopting the method, the comprehensiveness and accuracy of inspection can be improved.

The invention discloses a bus operation accurate management method and system. The method comprises the steps of operation information collection, arrival passenger flow number calculation, operation time calculation, data updating and calibration, subsequent arrival time calculation, automatic bus scheduling plan adjustment and real-time data updating of on-line vehicles. According to the invention, bus real-time Gps, traffic light waiting, road network congestion degree, platform intelligent monitoring, passenger flow estimation and vehicle intelligent scheduling information are integrated, corresponding combination, analysis and processing are carried out, the bus operation time can be accurately predicted through calculation of the number of passenger flow arriving at the station, and thus accurate and intelligent management of bus operation is realized. Compared with traditional prediction based on a vehicle real-time GPS and a road network congestion degree, the method improves the accuracy and comprehensiveness of a calculation result, facilitates the operation management of buses, and meets the development of intelligent transportation.

The invention provides a data center site selection method and system based on a fuzzy analytic hierarchy process, and the method is characterized in that the method comprises the steps: firstly analyzing related impact factors which need to reach a total target, building a hierarchical relation between the impact factors, taking each impact factor as an evaluation index, and aiming at the same evaluation index, adopting the fuzzy theory to solve priority membership values for the expressions of different schemes, then applying an analytic hierarchy process to solve weights of relative importance degrees of different evaluation indexes of a next layer relative to evaluation indexes of an upper layer, then carrying out priority membership weighted summation on the schemes to be selected, obtaining the evaluation indexes of the upper layer, and so on, finally obtaining a priority membership degree value of the topmost evaluation index and selecting a scheme preferentially.

The invention discloses a conference document information acquisition method. The method comprises the steps of collecting conference video information and conference audio information; performing feature extraction on the conference video information to obtain corresponding lip language features and text features; performing feature extraction on the conference audio information to obtain corresponding audio features; performing feature fusion on the lip language features, the text features and the audio features to obtain fusion features; processing the fusion features to obtain conference document information. According to the conference document information acquisition method, the comprehensiveness and accuracy of conference recording can be improved while the conference recording efficiency is ensured. The invention furthermore discloses a conference document information acquisition device and equipment, and a computer readable storage medium, which all have the above beneficial effects.

The invention provides a vulnerability detection method and device for a web application program and a computer readable storage medium; the method comprises the steps: obtaining request data of a to-be-detected web application program; then propagating the request data through n propagation functions to obtain n pieces of new request data, wherein, the n is a positive integer larger than 1, the output of the propagation functions is the new request data, and the input of the nth propagation function is the output of the (n-1) th propagation function; before the nth propagation function obtains the output of the (n-1) th propagation function, copying the (n-1) th propagation function and the corresponding output from the current thread variable to the global variable, and when the nth propagation function does not obtain the output of the (n-1) th propagation function, obtaining the output of the (n-1) th propagation function from the global variable; obtaining execution parameters; and according to the execution parameters, the request data and the new request data, determining whether the web application program to be tested has vulnerabilities or not. According to the method and the device, the comprehensiveness and the accuracy during vulnerability detection of the web application program to be detected can be improved.

The invention relates to the technical field of intelligent traffic systems, in particular to a V2X-based vehicle obstacle avoidance system, a platform framework, a method and a vehicle. The system comprises a sensing module used for obtaining vehicle condition information, road condition information and obstacle information; the vehicle wireless communication module is used for acquiring corresponding network interaction information; and the information processing module is used for calculating and generating the driving obstacle avoidance information of the own vehicle and distributing the driving obstacle avoidance information to the corresponding function module to execute the obstacle avoidance capability so as to dynamically adjust the driving obstacle avoidance strategy of the own vehicle and assist in adjusting the driving obstacle avoidance strategies of other vehicles. The invention further correspondingly discloses a V2X-based vehicle obstacle avoidance platform framework, a method and a vehicle. The vehicle obstacle avoidance system can assist in realizing cooperative obstacle avoidance of non-network-connected vehicles, intelligent network-connected vehicles and autonomous vehicles, thereby assisting in improving comprehensiveness and accuracy of vehicle cooperative obstacle avoidance.

The invention discloses solar cell satellite-mounted equipment and a cell testing method, and belongs to the technical field of aerospace power supplies, and the solar cell satellite-mounted equipment is characterized by comprising a mounting piece connected with a satellite; the carrying solar cell piece is mounted on the upper surface of the mounting piece; the temperature sensor is mounted on the upper surface of the mounting piece and is used for testing the on-orbit temperature of the solar cell; the illumination sensor is mounted on the mounting piece and is used for monitoring on-orbit illumination angle data of the solar cell piece; the solar cell test system is used for receiving current and voltage data of the temperature sensor, the illumination sensor and the solar cell, and storing and processing the data; the communication module performs data interaction with a satellite; a solar cell piece, a temperature sensor, an illumination sensor and a solar cell test system are carried for data communication. The device has the function of independently and simultaneously acquiring data of multiple sub-batteries, and can acquire current, voltage, battery temperature, illumination angle and test time data on a complete I-V curve in one period.

The invention relates to a data table grading method which can be applied to the technical field of big data or the field of finance. The data table grading method comprises the steps that a data processing language is obtained, and a mapping relation exists between the data processing language and a data table; obtaining a data processing complexity parameter based on the data processing language, the data processing complexity parameter being used for identifying a processing frequency of a data table having a mapping relationship with the data processing language; calculating a data table importance score based on an importance weight factor and the data processing complexity parameter, the importance weight factor being associated with an operation type of a data processing language; and acquiring the importance level of the data table based on the importance score of the data table. The invention further provides a data table grading device, equipment, a medium and a program product.

The invention discloses an electrocardiosignal anomaly detection method, system and device and a storage medium. The electrocardiosignal anomaly detection method comprises the steps of obtaining first data; preprocessing the first data to generate second data; obtaining a preset number of samples from the second data as first samples; inputting the first sample into the grid model for training, and generating an encoder; performing feature processing and dimension reduction on the first data through an encoder to generate a second sample; and according to the second sample, generating an electrocardiosignal detection result through a hidden Markov model. According to the method, the encoder is obtained by training the grid model integrated by the bidirectional circulation gating unit layer and the attention layer, the data processing capacity of the encoder is improved, and the encoder can extract more detail features from original electrocardiosignal data; the hidden Markov model with the Watson hybrid model as the emission density is adopted to recognize the sample, and the accuracy and comprehensiveness of the electrocardiosignal detection result are improved.

The embodiment of the invention discloses a data processing method and device, equipment and a computer readable storage medium. The method comprises the following steps: determining a first data value corresponding to each data participant; the first data value represents the inherent value of the data of the data participant; determining a second data value corresponding to each data participant; the second data value represents an efficiency value generated by the data of the data participant to the collaborative task in the execution process of the collaborative task; based on the first data value and the second data value, determining an excitation value allocated to each data participant, and allocating the determined excitation value to the corresponding data participant; the excitation value represents a feedback reward obtained by the data participant in the process of participating in the collaborative task.

The invention provides a risk prevention method which can be applied to the technical field of big data or the field of finance. The risk prevention method comprises the steps that historical risk records and a risk behavior set are acquired, and the risk behavior set comprises a consumption behavior type set associated with the risk records; extracting a risk behavior association rule based on the historical risk record and the risk behavior set; establishing a risk degree evaluation algorithm based on the risk behavior association rule; calculating a to-be-evaluated record risk degree score based on the risk degree evaluation algorithm to obtain a record risk degree rating; and performing risk prevention processing based on the recorded risk degree rating. The invention further provides a risk prevention device and equipment, a storage medium and a program product.