12657 results about "Prediction methods" patented technology

An automatic playback overshoot correction system predicts the position in the program material where the user expects to be when the user stops the fast forward or reverse progression of the program material. The invention determines the position where the program material was stopped. The media controller transitions to the new mode that the user selected, starting at the stopped position with an overshoot correction factor added or subtracted from it. The invention adapts to the user by remembering how much the user corrects after he stops the fast forward or reverse mode. Correction factors are calculated using the user's corrections and adjusting the correction factors if the user continues to make corrections. The invention also uses a prediction method to correctly place the user within the program upon transition out of either mode and determines if the speed of the fast forward or reverse modes and then automatically subtracts or adds, respectively, a time multiple to the frame where the transition was detected and positions the user at the correct frame. The time multiple is fine tuned if the user is consistently correcting after the fast forward or rewind mode stops. Another method initially tests the user's reaction time using a test video and asks the user to press the fast forward or reverse button on his control device during the test video and then asks the user to position the video to the place that he expected the system to have been. This time span is then used whenever the user uses the fast forward or reverse modes and is adjusted with a multiple for each speed. A final method allows the user to simply set a sensitivity setting that the system will use as a correction factor and a multiple is subtracted or added to the release frame whenever the user uses the fast forward or reverse modes, respectively.

The invention discloses an attention CNNs and CCR-based text sentiment analysis method and belongs to the field of natural language processing. The method comprises the following steps of 1, training a semantic word vector and a sentiment word vector by utilizing original text data and performing dictionary word vector establishment by utilizing a collected sentiment dictionary; 2, capturing context semantics of words by utilizing a long-short-term memory (LSTM) network to eliminate ambiguity; 3, extracting local features of a text in combination with convolution kernels with different filtering lengths by utilizing a convolutional neural network; 4, extracting global features by utilizing three different attention mechanisms; 5, performing artificial feature extraction on the original text data; 6, training a multimodal uniform regression target function by utilizing the local features, the global features and artificial features; and 7, performing sentiment polarity prediction by utilizing a multimodal uniform regression prediction method. Compared with a method adopting a single word vector, a method only extracting the local features of the text, or the like, the text sentiment analysis method can further improve the sentiment classification precision.

A method and apparatus is provided for decoding an encoded baseline video stream and an enhancement stream. The baseline video stream is decoded, upscaled and enhanced by applying adaptive filters specified by the enhancement stream. Baseline upscaled images are then coded to motion compensate enhanced high resolution images using previously decoded enhanced images, thus recycling these enhanced images. The enhancement stream provides the best predictor method for the decoder to combine blocks from previous enhanced images and upscaled images to produce a motion compensated enhanced image. Likewise, forward and backward motion compensated images are blended according to feature classification and filter extraction methods provided by the enhancement stream to produce a bidirectionally predicted frame. Lastly, the decoder applies residual data from the enhancement stream to produce a completed enhanced image.

A video compression method, and more particularly, a prediction method for efficiently eliminating redundancy within a video frame, and a video compression method and an apparatus using the prediction method are provided. There is provided a method for encoding video based on a multi-layer structure, including performing intra-prediction on a current intra-block using images of neighboring intra-blocks of the current intra-block to obtain a prediction residual, performing prediction on the current intra-block using an image of a lower layer region corresponding to the current intra-block to obtain a prediction residual, selecting one of the two prediction residuals that offers higher coding efficiency, and encoding the selected prediction residual.

The invention discloses a motion vector forecasting method in a resolution quantizing structure, wherein, forecasting motion vectors of macro blocks in enhancement layers are acquired by utilization of relativity of motion vectors of a time domain, a space domain and an interlayer domain. The realization process is that: on the time domain, candidate motion vectors of the time domain are acquired from motion vectors of macro blocks with the same position with the prior frame; on the space domain, candidate motion vectors of the space domain are acquired from motion vectors of adjacent blocks; on the interlayer domain, candidate motion vectors of the interlayer domain are acquired from corresponding motion vectors with low space domain resolution hierarchy; candidate motion vectors with minimum motion estimation cost are selected to be forecasting motion vectors by selection among the candidate motion vectors of the time domain, the space domain and the interlayer domain. The invention has the advantages of capability of acquiring more accurate forecasting motion vectors, capability of compressing code rate of motion vectors, and obtaining of performance gain.

A prediction method, merged method, has been introduced in the enhancement layer of a multiple layer video coding. This merged method was designed to efficiently handle the prediction of the non moving parts in coding of the enhancement layer VOP or frame. All the information for this merged mode prediction is obtained from the base layer, and no additional side information is transmitted. This prediction mode when used together with combination of the existing forward mode, backward mode, and interpolated mode, can improve the coding efficiency for enhancement layer video coding, especially in low bit rate coding. The method can be used in most multiple layer video coding schemes, especially in spatial scalability video coding.

A driving behavior prediction apparatus includes, for accurately predicting a driving behavior, a position calculation unit for a subject vehicle position calculation, a route setting unit for setting a navigation route, a distance calculation unit for calculating a distance to a nearest object point, a parameter storage unit for storing a template weighting factor that reflects a driving operation tendency of a driver, a driving behavior prediction unit for predicting a driver's behavior based on vehicle information and the template weighting factor, a driving behavior recognition unit for recognizing driver's behavior at the object point, and a driving behavior learning unit for updating the template weighting factor so as to study the driving operation tendency of the driver in a case that a prediction result by the driving behavior prediction unit agrees with a recognition result by the driving behavior recognition unit.

The invention relates to a bus arrival time prediction technique based on bus GPS data, which comprises sending out from the existing bus GPS data, making full use of auxiliary information such as pulling in, door opening and the like of a bus GPS apparatus, adopting a space dividing method to divide zones for bus lines, accounting the history arrival time of partitioned areas to each station in front according to the actual arrival time of a bus, and leading the bus arrival time to continuously do self-adapting update as the accumulation of history data, and meanwhile, predicating the bus arrival time by means of road traffic status which is calculated by GPS information of a taxi or road traffic status calculated by other systems, and improving the predicating accuracy.

The present invention combines an inter-prediction method using an AMVP mode and an inter-prediction method using a merge mode so as to use the same candidates. The method for decoding video data proposed by the present invention comprises receiving mode information on an inter-prediction method of a current block, determining, on the basis of the received mode information, whether the inter-prediction method to be applied to the current block is an AMVP mode or a merge mode, and selecting a candidate to derive motion information of the current block.

A video coding method using a multi-layer structure, and more particularly, a method and apparatus for facilitating a search for an intra-prediction mode in an upper layer using an intra-prediction mode in a lower layer while efficiently and compressively encoding the searched intra-prediction mode are provided. The intra-prediction method includes searching for an optimum prediction mode of a current block among a predetermined number of intra-prediction modes, and obtaining a directional difference between the searched optimum prediction mode and an optimum prediction mode of a lower layer block corresponding to the current block.

An object path prediction method, apparatus, and program and an automatic operation system that can secure safety even in situations that can actually occur are provided. For this purpose, a computer having a storage unit that stores the position of an object and the internal state including the speed of the object reads the position and internal state of the object from the storage unit, generates trajectories in a space-time consisting of time and space from changes of the positions that can be taken by the object with the passage of time based on the read position and internal state of the object, and predicts probabilistically paths of the object by using the generated trajectories.

The invention discloses a power grid security situation predicting method based on an improved deep learning model and belongs to the technical field of power system safety. The power grid security situation predicting method includes: performing power grid security situation evaluation through power grid data collection and preprocessing; aiming at the characteristic that indicator data of power grid security situation evaluation are high in relevance and dimension, providing an improved self-coding network method to lower dimension of the indicator data, and utilizing a data sample after dimension reduction and a power grid security situation value corresponding to a next time monitoring point; adopting an improved deep belief network to build a deep learning situation predicting model with multi-input and multi-output for power grid security situation prediction. By the power grid security situation predicting method, speed and accuracy of power grid security situation prediction can be improved effectively.

The invention belongs to the technical field of network security and particularly relates to an attack-oriented network security situation prediction method, device and system. The method comprises the following steps: detecting and collecting alarm data and network environment operation and maintenance information in a network countermeasure environment, obtaining an element set required by network security situation prediction, wherein the element set comprises three types of information of an attacker, a defense party and a network environment; evaluating the attacker capability and the level of the defense party, establishing a dynamic Bayesian attack graph, and calculating an attack phase number and an attack state occurrence probability vector; and combining a vulnerability scoring standard and network asset information, and performing time-space dimension quantification on the network security situation value. According to the method, dynamic association of the situation elements of the defense party, the attacker, the environment information and the like is achieved, the actual environment of the network is better conformed to, the future situation and the attack occurrencetime can be accurately predicted, higher prediction efficiency is achieved, and storage scale and timeliness of network security situation awareness are optimized, so as to provide more effective guidance for network protection.

The invention belongs to the petroleum exploration field, and concretely relates to a compact sandstone reservoir complex netted fracture prediction method. The method comprises the steps of: building a geological structure model and a fracture growth model; testing magnitudes and directions of ancient and modern crustal stresses; completing a rock mechanic parameter experiment; testing rock mechanic parameters and fracture stress sensitivities; developing a fracture rock multistage composite rupture criterion; performing a rock deformation physical test to obtain a peak value intensity; building a relation model between single axle state stress-strain and fracture bulk density; building a relation model between triaxial state stress-strain and fracture bulk density and occurrence; building a relation model between single axle state stress-strain and fracture bulk density; calculating and stimulating fracture parameters under modern conditions; and verifying the reliability of a fracture quantitative prediction result. The method can accurately obtain compact sandstone reservoir complex netted fracture parameters, and perform quantitative characterization, is suitable for quantitative prediction of any fracture mainly with a brittle reservoir, and reduces exploitation risks and costs.

The present invention represents a procedure for compensating the heave movement of offshore cranes. The dynamic model of the compensation actuator (hydraulically operated winch) and the load hanging on a rope are derived. Based on this model, a path-tracking control unit is developed. To compensate the movement of the ship/watercraft caused by waves, the heave movement is defined as a time-varying disturbance and is analyzed with respect to uncoupling conditions. With a model expansion, these conditions are satisfied, and an inversion-based uncoupling control law is formulated. To stabilize the system, an observer is used for reconstructing the unknown state by means of a force measurement. Furthermore, the compensation efficiency can be improved by predicting the heave movement. There is proposed a prediction method in which no ship/watercraft models or properties are required. The simulation and measurement results validate the heave compensation method.

The embodiment of the invention discloses an image prediction method and a relevant device. The image prediction method comprises the following steps: determining the motion vector predictors of K pixel samples in a current image block, wherein K is an integer greater than 1, the K pixel samples comprise a first vertex angle pixel sample of the current image block, the motion vector predictor of the first vertex angle pixel sample is obtained on the basis of the motion vector of a preset first spatial domain adjacent image block of the current image block, and the first spatial domain adjacent image block is adjacent to a first vertex angle pixel sample spatial domain; and predicting the pixel value of the current image block on the basis of a non-translational motion model and the motion vector predictors of the K pixel samples. According to the scheme in the embodiment of the invention, the computation complexity of image prediction based on the non-translational motion model is lowered.

The invention discloses a prediction method for a life cycle of a battery. The prediction method comprises the following steps: drawing a curve of the number of discharge and capacity fading rate; determining a function model of the capacity fading rate Qloss of the battery and the number of discharge N of the battery; substituting data of a certain impact factor impacting the life of the battery into the function model, wherein Qloss is 50-80 percent of the nominal capacity of the battery; calculating to obtain the expected cycle period N of the battery. According to the prediction method disclosed by the invention, as long as little cyclic test is performed on the battery, the residual life of the battery can be judged, and therefore, the test time is greatly shortened and the test cost is greatly reduced; the objective condition of capacity fading during the actual use process of the battery can be truly and comprehensively described, the life and the residual life of the battery can be obtained, the performance of the battery can be helped to be determined, and the life of the battery is prolonged. As the established model is universal, the prediction method disclosed by the invention is suitable for establishing life fading models of various storage batteries.

A data-driven lithium ion battery cycle life prediction method based on an AR (Autoregressive) model and an RPF (Regularized Particle Filtering) algorithm relates to a lithium ion battery cycle life prediction method and belongs to the technical field of data prediction. The invention solves the problems in the existing lithium ion battery cycle life prediction method that the model-based prediction method is complicated in modeling, and parameters are difficult to identify. The data-driven lithium ion battery cycle life prediction method combines time sequence analysis with particle filter method and comprises the following steps: the AR model is firstly utilized to realize the multi-step prediction on battery performance degradation process time sequence data; and then, aiming at the problem of uncertainty expression of the cycle life prediction result, the regularized particle filtering method is introduced, and a lithium ion battery cycle life prediction method framework is proposed. The method proposed by the invention can be used for effectively predicating the cycle life of a lithium ion battery and realizes the output of probability density distribution of the predication result, has good computational efficiency and uncertainty expression ability.

The embodiment of the invention discloses a self-adaptable prediction method, comprising the steps as follows: block division is carried out on micro-blocks required to be predicted in frame by a plurality of block division methods; a plurality of prediction modes are used to predict each sub-block in each block division result, thus gaining the sub-block residue of the current sub-block; according to the type of the sub-block, a plurality of conversion types are selected to convert the residues of each sub-block, thus gaining the ratio of bit rate and distortion of the current sub-block; the micro-block is predicted according to the minimum ratio of the bit and the distortion out of all bit rate/distortion ratios, the prediction mode and the conversion method. The embodiment of the invention discloses a self-adaptable prediction device which comprises a block division module, a prediction module, a conversion module, a selection module and a frame prediction module; the method and the device can smartly carry out the coding in frames or between frames of the micro-blocks, adapt for the local vein statistical characteristics of the micro-block better and improve the coding efficiency.

A method for predicting motion vectors to improve compressibility in an image compression codec which processes videos, and an image encoding/decoding apparatus and method using the same. A method for predicting a motion vector used during differential encoding of a motion vector for image encoding, the method including generating a motion vector list with candidate motion vectors for adjacent blocks of a target block, a predictive motion vector of which is to be obtained; calculating each distance between motion vectors included in the motion vector list; and determining a predictive motion vector for the target block by removing motion vectors in order of large distances between the motion vectors.

The invention discloses a knowledge graph-based enterprise risk prediction method and system. The method comprises the steps of 1) constructing entities of an enterprise knowledge graph, establishingrelationships between the entities, and extracting knowledge from set data sources to form the enterprise knowledge graph; 2) capturing news information of enterprises and related entities, processingthe captured news information based on the enterprise knowledge graph, marking relationships between the related entities and events, and storing marked information in a structured knowledge base toform an enterprise news information knowledge graph; and 3) performing enterprise risk prediction based on the enterprise knowledge graph and the enterprise news information knowledge graph. The blankof an existing knowledge graph in the field of enterprise risk analysis is filled; and the constructed enterprise knowledge graph and enterprise news information knowledge graph have relatively highpracticality.

The invention discloses a method for predicting an hourly cooling load of a central air-conditioner in an office building on line based on indoor temperature and humidity parameters. The method for predicting the cooling load comprises the following steps of: performing time sequence prediction on outdoor meteorological parameters and air-conditioner operation input parameters, establishing an Online support vector regression (SVR) dynamic prediction model of the air-conditioner cooling load by using the data, predicting 24-hour air-conditioner cooling load in the current day in advance, and performing compensation by using a residual sequence of the actual value and the predication value of the 24-hour air-conditioner load in the previous day. The predication data of the air-conditioner cooling load prediction model established by the method is high in reliability; and the method can be applied to occasions for prediction of the hourly cooling load of the central air-conditioner in the office building in a single building or a large range, energy-saving control of a central air-conditioner system, energy consumption prediction of the air-conditioner, power peak clipping in areas and the like.

The invention discloses a power load forecasting method based on a long short term memory (LSTM) neuron network. The power load forecasting method comprises inputting power load data and a region feature factor at a historic moment through an input unit; carrying out training and modeling on the power load data and the region feature factor at the historic moment by means of an LSTM network in order to generate a deep neural network load forecasting model which is a single-layer multi-task deep neural network model or a double-layer multi-task deep neural network model used for power supply load forecasting; forecasting the power load in an area needing to be forecasted by means of the deep neural network load forecasting model, and generating a forecasting result of the power load in the area; and outputting the forecasting result of the power load in the area through an output unit. According to the invention, a multi-task learning power load forecasting model is constructed based on the LSTM network in the deep learning field, power consumption loads in multiple areas can be forecasted accurately, and the forecasting effect is improved.

The present invention discloses a deep learning method-based short-term traffic flow prediction method. The influence of the traffic flow rate change of the neighbor points of a prediction point, the time characteristic of the prediction point and the influence of the periodic characteristic of the prediction point on the traffic flow rate of the prediction point are considered simultaneously. According to the deep learning method-based short-term traffic flow prediction method of the invention, a convolutional neural network and a long and short-term memory (LSTM) recurrent neural network are combined to construct a Conv-LSTM deep neural network model; a two-way LSTM model is used to analyze the traffic flow historical data of the point and extract the periodic characteristic of the point; and a traffic flow trend and a periodic characteristic which are obtained through analysis are fused, so that the prediction of traffic flow can be realized. With the method of the invention adopted, the defect of the incapability of an existing method to make full use of time and space characteristics can be eliminated, the time and space characteristics of the traffic flow are fully extracted, and the periodic characteristic of the data of the traffic flow is fused with the time and space characteristics, and therefore, the accuracy of short-term traffic flow prediction results can be improved.

An in-frame predication method for video encode in order to improve video encode quality is disclosed. The original video stream taken by camera is used as input, which is input to computer by video acquisition card, and then processed by computer and the JVT video encode technique. An operation rule for calculating the DC predication mode by use of the samples of the decoded pixel in adjacent blocks is defined. Multiple predication modes can be recombined and sorted.

The invention relates to the technical field of artificial intelligence, in particular to training and prediction methods and devices of a federal transfer learning model. The training method comprises the following steps: a first terminal acquires an encrypted migration model issued by a parameter server in an ith training period, wherein the encrypted migration model is generated by the parameter server according to K encrypted sharing models uploaded by K terminals of K participants in an (i-1)th training period; and the first terminal updates an encryption sharing model in a first local neural network model of the first terminal according to the issued encryption migration model in the ith training period, and trains and updates the first local neural network model according to the first data. Therefore, in the federal transfer learning training process, the privacy of the data of each participant is ensured, and the training efficiency of the model is effectively improved, and theuniqueness of each terminal model is improved while the generalization ability of the model is ensured.

A computer implemented system and methods for forecasting the future values of a time series with multiple seasonal cycles are described. The forecasting method of the present invention include 1) two multi-seasonal exponential smoothing methods, 2) the seasonal sub-series method splitting a time series into multiple sub-series, and 3) the aggregate-to-period forecasting method.

The embodiment of the invention provides a disease prediction method, device, a medium and electronic equipment. The disease prediction method comprises steps: activity trajectories of multiple usersare acquired; if a patient diagnosed with an infectious disease appears in the multiple users, based on the activity trajectories of the multiple users in a predetermined time period, the position relationship between other users in the multiple users and the patient in each time point is determined; according to the position relationship between other users and the patient in each time point, thecontact degrees between other users and the patient are calculated; and target users with the contact degrees with the patient higher than a predetermined threshold are predicted as a high-risk diseased population. In the technical scheme provided by the embodiment of the invention, persons closely contacted with the infection source of the infectious disease (that is, the patient) are determinedthrough the position information of the user, persons at risk of infection can be further accurately determined, corresponding measures are taken timely, and the health control level of the user is improved.

A method and device for coding a digital image using intra-mode block prediction, wherein a list of prediction modes for each combination of prediction modes of the neighboring blocks is obtained. The modes assigned to each combination of prediction modes may be divided into two groups. The first group includes n (where n is smaller than the overall number of available modes) most probable prediction modes and the second group includes the remaining modes. The modes in the first group are ordered according to their probability. This order may be specified as a list of modes ordered from most probable to the least probable mode. The modes belonging to the second group may be ordered in some predetermined manner, which may be specified depending on the information already available to the decoder.

The invention discloses a CNN and LSTM-based rolling bearing residual service life prediction method, and relates to the field of rolling bearing life prediction. The method aims to solve the problemthat residual service life (RUL) prediction of a rolling bearing is difficult in two modes of performance degradation gradual change faults and sudden faults. The method comprises the following stepsof: firstly, carrying out FFT (Fast Fourier Transform) on an original vibration signal of the rolling bearing, then carrying out normalization processing on a frequency domain amplitude signal obtained by preprocessing, and taking the frequency domain amplitude signal as the input of a CNN (Convolutional Neural Network); The CNN is used for automatically extracting data local abstract informationto mine deep features, and the problem that a traditional feature extraction method depends too much on expert experience is avoided. the deep features are input into an LSTM network, a trend quantitative health index is constructed, and a failure threshold value is determined at the same time; And finally, smoothing processing is carried out by using a moving average method, eliminating local oscillation, and a future failure moment is predicted by using polynomial curve fitting to realize rolling bearing RUL prediction. And the prediction result can be well close to the real life value.