40results about How to "Troubleshoot technical issues with low accuracy" patented technology

The invention provides a text label generating method and device. Regarding targeted text, keyword extraction is conducted by the adoption of strategies corresponding to all label types, so that candidate labels of the label types are subjected to cross validation among the different label types after the candidate labels of all the label types of the targeted text are obtained, and finally, according to the validated candidate labels, a targeted label of the targeted text is determined. Because label extraction and cross validation are conducted separately regarding the different label types which include entity words, fragment text and / or topics, the accuracy of the label extraction is improved, and the technical problem is solved that the accuracy of the label extraction in the prior art is not high.

The invention provides an electricity stealing risk prediction method and device based on deep learning. A solution is provided for the problems that in a traditional electricity larceny prediction method, an electricity larceny user cannot be positioned due to the fact that a topological structure in a power distribution network is not known, and the electricity larceny prediction accuracy is lowdue to the fact that a non-efficient identification feature is used in the traditional electricity larceny prediction method. The prediction method comprises the following steps: firstly, predictingthe daily power consumption mode quantity of a user by utilizing daily historical power consumption information of a low-voltage user; secondly, according to analysis of power utilization factors of the users in nearly three days of the day, an analytic hierarchy process is adopted to calculate the abnormal rate of power utilization, and finally, based on K <->, the abnormal rate of power utilization is calculated. And the Means method is used for grading the electricity stealing. Technical effects of improving prediction accuracy and realizing electricity larceny grade division are achieved.

The invention provides a method and device for researching a near-surface three-dimensional velocity field of physical geography. The method comprises the steps of obtaining geological information of a research area according to the areal geology of the research area and research content; establishing a three-dimensional geological element conceptual model according to collected data; determining the preliminary survey position of a near-surface survey and a near-surface preliminary survey scheme, carrying out construction according to the determined preliminary survey position of the near-surface survey and the determined near-surface preliminary survey scheme, and obtaining site construction data; adjusting the three-dimensional geological element conceptual model according to the obtained site construction data until the geological elements in the three-dimensional geological element conceptual model are adaptive to wave velocity; establishing a geology and speed field model with the surficial geology coordinated with wave velocity layers on the basis of the three-dimensional geological element conceptual model with the geological elements adaptive to the wave velocity. According to the method and device for researching the near-surface three-dimensional velocity field of physical geography, the technical problem that in the prior art, a near-surface velocity model is not accurate is solved, and the effect of effectively improving the accuracy of the near-surface velocity model is achieved.

The invention discloses an adjacent shot interference suppressing method and device for independent simultaneous sweeping seismic data of seismic vibrators. The adjacent shot interference suppressing method includes that separating non-common shot point gather seismic data from ISS seismic data of the seismic vibrators; carrying out vector median filtering on the separated non-common shot point gather seismic data to obtain non-common shot point gather seismic data performed with adjacent shot interference suppressing; separating common shot point gather seismic data from the non-common shot point gather seismic data performed with the adjacent shot interference suppressing. The adjacent shot interference suppressing method and device for the independent simultaneous sweeping seismic data of the seismic vibrators solve the technical problem of low seismic data precision of the prior art, realize to effectively eliminate the adjacent shot interference in the ISS seismic data and enable the seismic data precision to be improved.

The invention discloses a game questionnaire data processing method and device. The game questionnaire data processing method comprises the following steps: obtaining the unique user identifier of a user in the game, wherein the unique user identifier is used for uniquely identifying the game role of the user in the game; obtaining the questionnaire data of answers of the user; and matching the unique user identifier of the user with the questionnaire data of the answers of the user, and storing a matching relation of the unique user identifier and the questionnaire data. The game questionnaire data processing method is capable of effectively fusing the role information and the questionnaire data to comprehensively analyze problems, and therefore, the problem of low accuracy of traditional game questionnaire results is effectively solved.

The invention provides a fault diagnosis method and apparatus for a main sintering exhaust fan. The fault diagnosis method comprises: collecting a vibration signal of a main sintering exhaust fan; and carrying out wavelet packet decomposition on the vibration signal, extracting a test sample and acquiring fault information of the main sintering exhaust fan by using an extreme learning machine model. With the wavelet packet analysis method, effective extraction and decomposition can be carried out on high-frequency and low-frequency components; while information integrity is ensured, characteristics of non stability and short duration of the signal can be reflected with deepening of a fault degree; the severe working environment of the main sintering exhaust fan and the harmonic characteristic of the vibration signal can be adapted effectively. Meanwhile, because of the powerful classification capability of the extreme learning machine, adaptability to various fault types of the main sintering exhaust fan is improved. Therefore, with the fault diagnosis method, a technical problem of poor accuracy of fault diagnosis in the prior art is solved.

The invention provides an information matching method and a device, the methods include: a match initiating side sends information matching request to a match initaiating recipient, the matching recipient verifies the information in the information matching request, if the verification passes, a first response to the information matching request is constituted, sending to the matching initiator, verifying the information in the first response after the matching initiator receives the first response, if the verification passes, sending a second response to the matching receiver, obtaining a first calculation result according to the first commitment and the second commitment by the matching initiator, obtaining a second calculation result according to the first commitment and the second commitment by the matching receiver; after receiving the second response, the matching receiver compares the first calculation result with the second calculation result. If the comparison result is equal,the matching is successful. If the comparison result is not equal, the matching fails. The invention solves the technical problem that the accuracy of the matching result is not high in the prior art.

The embodiment of the invention provides a method and device for controlling follow of an unmanned aerial vehicle. The control method comprises the steps that basic category information of a basic target object is acquired; whether or not the target object is a standard target object needing to be followed is judged; if yes, the basic target object is marked as the standard target object, and according to a primary navigation signal of the standard target object, the unmanned aerial vehicle is controlled to fly to the position indicated by the navigation signal; an image of a preset range corresponding to the indicated position is acquired; whether or not the target object is contained in the image is judged, and if yes, the current display position of the target object in the image is acquired; and when the current display position is different from the preset display position, control parameters of the unmanned aerial vehicle are adjusted according to relative position information ofthe current display position and the preset display position to enable the display position of the target object to keep at the preset display position. According to the method and device, the technical advantage of precisely following the target object is achieved.

The invention discloses a gait monitoring method based on an animal treadmill, and the method comprises the steps: collecting an animal gait image, and carrying out the preprocessing of the animal gait image, and obtaining a preprocessed image; establishing a gait recognition model, recognizing skeleton points of animals in the preprocessed image through the gait recognition model, mapping the recognized skeleton points to obtain space rays, and matching the space rays; after the matching is completed, establishing a world coordinate system, performing calculation according to the position ofthe skeleton point of each animal in the world coordinate system to obtain a space average coordinate of each animal, and setting the space average coordinate as a tracking point; and recording path information of the tracking points in a world coordinate system in real time through adoption of a tracking method to obtain gait monitoring data. According to the invention, the animal skeleton pointsgiven in the gait image are recognized, the tracking points are constructed through the space rays, the gait of the animal is accurately recognized and recorded, and the accuracy of experimental result data is improved.

The invention discloses an information processing method and device, wherein the method comprises the following steps that the information input by a user is acquired; the information which is input by the user is analyzed to obtain document information; and the document information is sent to a server, wherein the server provides an interface which transfers the document information to handle business for the user. According to the information processing method and device, the technical problem of the prior art that the data accuracy is low when the document information of the user is acquired by the user through manually filling in forms is solved, and the technical effect of improving the data accuracy is achieved.

The invention discloses an object-level sentiment classification method based on a segmented convolutional neural network, and the method comprises the steps: carrying out the feature extraction of ato-be-classified text, and carrying out the context feature division of the to-be-classified text, and obtaining a first text-above feature and a first text-below feature; inputting the data into a preset segmented convolutional neural network; carrying out the convolution operation on the convolution layer to obtain local features of the preceding text and the following text; enabling the segmented pooling layer to perform maximum pooling operation on the local features of the preceding text and the following text to obtain the pooling features of the preceding text and the following text; enabling the feature fusion layer to perform feature fusion on the pooling features of the preceding text and the following text and the attention feature representation extracted by the attention module; enabling a softmax layer to process the fused features; and outputting sentiment classification results. According to the invention, the technical problems that the context feature with finer granularity cannot be obtained, and the emotion classification result accuracy is not high due to the facts that the existing emotion classification method does not consider the influence of the context divided by the object on the classification effect, and the maximum pooling operation is generally adopted in the pooling layer of the convolutional neural network.

The embodiment of the invention provides a method and device for controlling flight of an unmanned aerial vehicle. The control method comprises the steps that according to a primary navigation signalof a target object, the unmanned aerial vehicle is controlled to fly to the position indicated by the navigation signal; an image of a preset range corresponding to the indicated position is acquired;whether or not the target object is contained in the image is judged, and if yes, the current display position of the target object in the image is acquired; and when the current display position isdifferent from the preset display position, control parameters of the unmanned aerial vehicle are adjusted according to relative position information of the current display position and the preset display position to enable the display position of the target object to keep at the preset display position. According to the method and device, the technical advantage of precisely following the targetobject is achieved.

An embodiment of the invention provides a UAV following control method and device. The control method comprises that according to a primary navigation signal of a target object, a UAV is controlled tofly to a position indicated by the navigation signal; an image of a preset range corresponding to the indicated position is obtained; whether the image includes the target object is determined, and if YES, a present display position of the target object in the image is obtained; and when the present display position is different from the preset display position, a control parameter of the UAV isadjusted according to relative position information of the present and preset display positions, so that the display position of the target object is kept in the preset display position. The target object is followed accurately.

The invention provides a branch horizontal well deployment method and device. the method comprises the steps that the drilled vertical well position, the position of a main horizontal section of a branch horizontal well and the position of a branch horizontal section of the branch horizontal well within a work area where a to-be-deployed branch horizontal well is located are obtained; the positions of a primary virtual well and a secondary virtual well are determined; according to the drilled vertical well oil reservoir section, the oil reservoir section of the primary virtual well is determined; according to the drilled vertical well oil reservoir section and / or the primary virtual well oil reservoir section, the oil reservoir section of the secondary virtual well is determined; according to the primary virtual well oil reservoir section, the main horizontal section oil reservoir section is determined; according to the secondary virtual well oil reservoir section, the branch horizontal well oil reservoir section is determined; and according to the main horizontal section oil reservoir section and the branch horizontal well oil reservoir section, the branch horizontal well is deployed. According to the branch horizontal well deployment method and device, the technical problem that accuracy for determining the branch horizontal well oil reservoir section is low due to the fact that existing drilled wells are few is solved.

The invention discloses a preparation method of a cell staining quality control product. According to the preparation method, E-Cadherin immunomagnetic beads and CEA immunomagnetic beads are specifically combined with lung cancer circulating tumor cells, and CD45 immunomagnetic beads are specifically combined with white blood cells; by staining the enriched lung cancer circulating tumor cells and leukocytes, the method can conduct quality control on a staining kit for the lung cancer circulating tumor cells. According to the preparation method of the cell staining quality control product, through specific binding of the antigen and the antibody, the credibility of a quality control result is high.

The invention discloses a user film-watching information prediction method and device, equipment and a storage medium, and the method comprises the steps: obtaining a dimension value corresponding to a preset user film-watching evaluation dimension of each day within a preset number of days, and determining a preset time attenuation weight factor and a preset holiday weight factor corresponding to the preset user film-watching evaluation dimension; according to the obtained dimension values and weight factors, view evaluation indexes corresponding to different time periods in the preset number of days are determined; inputting the film watching evaluation indexes corresponding to different time periods into a pre-trained full-connection neural network to obtain a film watching score corresponding to a preset number of days; and predicting a film watching probability and a loss probability of the user according to the film watching score. According to the method, the user loss probability can be accurately reflected through the user film watching probability, then different user groups are accurately classified based on the user loss probability, and guidance and basis are provided for user retention and activity promotion through a refined and targeted marketing strategy.

The invention discloses a method and device for obtaining bidding product information. The method comprises the steps of obtaining a bidding announcement document based on a document object model tree structure of a preset bidding announcement page; disintegrating the bidding announcement document into multiple statements; according to a preset rule, dividing the statements into a first statement set and a second statement set; obtaining a first training set according to the first statement set, and obtaining a second training set according to the second statement set; training and obtaining a classifier of a Bayesian classification model according to the first training set and the second training set; obtaining target bidding product information from the bidding announcement document to be processed according to the classifier. According to the method and device for obtaining the bidding product information, the technical problem is solved that bidding product information obtained in the prior art is low in accuracy.

The invention relates to a positioning method, device and system, a terminal, a storage medium and a program product. The method is applied to the positioning system, the positioning system comprises a first terminal and a second terminal, the first terminal and the second terminal are used for receiving and transmitting positioning signals at the same time, and the method comprises the following steps: acquiring a first positioning signal received by a target receiving end; and if the first positioning signal is a target positioning signal transmitted by a target transmitting end, performing positioning processing according to transmitting information of the first positioning signal transmitted by the target transmitting end and receiving information of the first positioning signal received by the target receiving end, obtaining a relative position relation between the first terminal and the second terminal; wherein the target receiving end is one of the first terminal and the second terminal, and the target sending end is the other one of the first terminal and the second terminal. By adopting the method, the accuracy of terminal positioning can be improved.

The invention discloses a video cover selection method and device, equipment and a storage medium, and the method comprises the steps: obtaining multiple to-be-processed key frames in a video, and determining the facing angle of a target person in the to-be-processed key frames; identifying a first posture action of the target person, and obtaining a preset picture matched with the first posture action; correcting each action angle of a second posture action of the reference figure in the preset picture according to the facing angle, and determining an action score of the first posture action according to the facing angle and each corrected action angle; and determining a position score of the target person in the to-be-processed key frames, and selecting a video cover of the video from each to-be-processed key frame according to the action score and the position score. According to the invention, the selected video cover is more accurate and better meets user requirements.

The invention provides a target information identification method and device, equipment, a readable storage medium and a product, and relates to NLP in artificial intelligence. According to the specific implementation scheme, an identification request is obtained, and the identification request comprises a privacy policy file to be identified; according to the recognition request, the privacy policy file to be recognized is compared with a preset target information word list, an intersection vocabulary set of the privacy policy file to be recognized and the target information word list is determined, the target information word list comprises standard target information vocabularies meeting preset regulations, and the target information word list comprises the standard target information vocabularies meeting the preset regulations; the non-standard target information vocabularies do not accord with preset regulations; and determining vocabularies in the intersection vocabulary set as target information corresponding to the privacy policy file to be identified. Therefore, the identification operation of the target information which accords with the preset regulation and does not accord with the preset regulation in the privacy policy file to be identified can be quickly realized, and the efficiency and accuracy of target information identification are improved.

The invention discloses a driving state detection method and system, a data processing device and a storage medium. The method comprises the steps of acquiring electrocardiogram data of a driver in a to-be-detected vehicle; when the electrocardiogram data is equal to the target electrocardiogram data, determining health state information corresponding to the target electrocardiogram data; and determining the driving state of the driver according to the health state information. The accuracy of driving state detection of the driver is improved, and life safety of the driver is guaranteed.