1289results about How to "Good forecast" patented technology

The present invention is a computer system that analyses the factual and behavioural data of a group of subjects, extracts the common behavioural patterns from the collection, and is capable of forecasting the behaviour of a new subject when his or her factual data are inputted to the system. It does so by first taking a collection of both the factual data and behavioural data from a group of subjects. A clustering engine is employed to compute a set of exemplars that concisely represent the population. Afterwards, the factual data of a subject, and the corresponding behavioural exemplar that he or she belongs, are fed to a learning module so that it can learn the mapping between the subject's factual data and behavioural exemplar. After learning, the system is able to predict the behaviour patterns when factual data of a new subject is presented.

The invention relates to a prediction-based virtual network function scheduling method for 5G network slices, and belongs to the field of mobile communication. The method specifically comprises: establishing a delay-based service function chain queue model for service function chain features having dynamically changing service traffic; establishing a multi-queue cache model, and determining, at different time, priorities requested by the slices and a lowest service rate that should be provided, according to the size of a slice service queue; discretizing the time into a series of consecutive time windows, and establishing a prediction-based traffic perception model by using the queue information in the time windows as a training data set sample; and searching a scheduling method for the best service function chain VNF under the resource constraint that the caching of the slice service queue does not overflow according to the predicted size of each slice service queue and the corresponding lowest service rate. The method realizes on-line mapping of network slices, reduces the overall average scheduling delay of multiple network slices, and improves the performance of network services.

The invention relates to an electric power telecommunication network device fault prediction method based on improved LSTM. The invention provides a data preprocessing and time sequence input construction method for the first time. Compared with the simple recurrent neural network, the LSTM is more likely to learn long-term dependence and can solve the prediction problems related to sequences. Because of the strong association between device alerts, the independence of variables can be ensured through PCA. The target replication strategy is also used for improving the LSTM, which can bring local error information in every step. Compared with a simple target output only in the last step, the strategy can improve the accuracy of the model and reduce the risk of over fitting. In combination with dropout, the invention proposes a prediction model of LSTM, which can achieve better prediction precision by deep learning. At the same time, the LSTM is usd for modeling electric power telecommunication network alarm data for the first time and identifying the timing sequence mode therein.

The invention relates to a photovoltaic array fault diagnosis and early warning method comprising the following steps: combining an Elman nerve network optimized by a non-linear least square method and a decision tree with experience knowledge so as to form a fault diagnosis model; collecting present photovoltaic array operation data and meteorology data, and computing errors when compared with historical normal state data; using the fault diagnosis model to obtain the corresponding fault type and credibility when the error is bigger than a threshold; finally integrally evaluating so as to obtain the final fault type credibility, and selectively carrying out fault early warning according to the credibility values; updating a fault knowledge base according to the field actual measurement conditions. The method combines the LM-Elman nerve network and the decision tree with experience knowledge so as to built the fault diagnosis model, thus improving the history data sensitivity, providing better prediction effect when compared with a BP network, and improving the network convergence speed and training precision; the experience knowledge is supplemented, thus providing stronger robustness; the method can timely detect and diagnose, thus reducing fault incidence rate, and ensuring the photovoltaic power station to stably work.

The invention discloses a computer vision-based dynamic gesture recognition method, and aims at solving the gesture recognition problems under complicated backgrounds. The method is realized through the following steps of: acquiring a gesture data set and carrying out artificial labelling; clustering a labelled image set real frame to obtain a trained prior frame; constructing an end-to-end convolutional neural network which is capable of predicting a target position, a size and a category at the same time; training the network to obtain a weight; loading the weight to the network; inputting agesture image to carrying out recognition; processing an obtained position coordinate and category information via a non-maximum suppression method so as to obtain a final recognition result image; and recording recognition information in real time to obtain a dynamic gesture interpretation result. According to the method, the defect that hand detection and category recognition in gesture recognition are carried out in different steps in the prior art is overcome, the gesture recognition process is greatly simplified, the recognition correctness and speed are improved, the recognition systemrobustness is strengthened, and a dynamic gesture interpretation function is realized.

The invention discloses an acquisition terminal fault prediction method and system based on a Bayesian network optimization algorithm. In view of the potential fault risk in the operation of an acquisition terminal, the operation state of the acquisition terminal is evaluated reasonably, and thus, the fault of the acquisition terminal can be predicted. An acquisition terminal fault prediction model is established by using a Bayesian network algorithm. In view of the fact that the acquisition terminal has many characteristic parameters which are associated complexly, a Bayesian network association diagram constructed by the experts in the power field is simplified by using a maximum principal sub-graph decomposition technology, and then, attribute association oriented mining is carried outon the association diagram through conditional independent test and local score test. Therefore, the Bayesian network algorithm is optimized, the state of the acquisition terminal in operation can beevaluated comprehensively and objectively, and the prediction accuracy of system is improved. The efficiency and feasibility of the method are verified by taking the electricity consumption information acquisition system of the State Grid Chongqing Electric Power Company as an experimental platform.

The invention provides a prediction method of a road traffic flow based on a graph convolution network. The method comprises the following steps of collecting vehicle GPS data in the past period of time; integrating the vehicle GPS data with actual road network information data to obtain a road traffic flow characteristic matrix, and converting a road and a road intersection point in the actual road network through line graph conversion to generate a road adjacency matrix; and based on the road adjacency matrix and the road traffic flow characteristic matrix, acquiring a predicted value of traffic flow data of the road in the next time period by using a spatial characteristic and a time characteristic of traffic flow data of the integrated road of a GCN network and an LSTM network. Through integrally using the spatial characteristic of the road traffic flow data extracted by GCN and the time characteristic of the road traffic flow data extracted by LSTM, and combining a time period characteristic and a weather characteristic, the road traffic flow data is predicted, and a prediction effect is better than the prediction effect acquired through using only time characteristic or the spatial characteristic.

A processor integrated circuit has one or more processor cores and a power management controller in a North-Bridge that generates a first power state recommendation for the one or more processor cores. The North-Bridge also receives a second power state recommendation from a South-Bridge integrated circuit. The North-Bridge determines a final power state for the one or more processor cores based on the first and second power state recommendations.

The invention relates to a low-cycle creep and fatigue life evaluation method under the conditions of high temperature and multiaxial spectrum load. The method comprises the following steps of reading a stress strain history in a multiaxial loading spectrum data block, working out equivalent strain, and finishing a loading history; repeatedly extracting by adopting a relative equivalent strain multi-axis counting method; working out all repeated fatigue damage by adopting a unified multiaxial fatigue damage life prediction model; accumulating the fatigue damage to work out the total fatigue damage; working out equivalent creep stress by utilizing the original loading history; working out creep damage Dc according to the equivalent creep stress and the stress history by combining a creep lasting equation; working out the total damage D caused by a multiaxial load spectrum block at the high temperature; and estimating the multiaxial creep and fatigue life. According to the method, the fatigue damage under the multiaxial stress and the creep damage under the multiaxial stress can be respectively calculated in the whole loading spectrum data block, the fatigue material constant at the room temperature is adopted in the calculation of the fatigue damage, and lasting equation material constant recommended by specification is adopted in the calculation of the creep damage; through experimental verification, the method has a good prediction effect.

Disclosed is a rolling bearing remaining life prediction method based on feature fusion and particle filtering. According to an index calculation process, firstly, original features are extracted from bearing vibration signals, the extracted original features are clustered by the adoption of a relevance clustering method, then, one typical feature is selected from each cluster to form optimal feature sets, and finally the feature sets are fused by the adoption of a weight fusion method into a final recession index. According to a life prediction process, firstly, smoothing and resampling are carried out on the recession index, the time interval is adjusted to be an expected value, state-space model initial parameters are calculated by the adoption of least square fitting, then, model parameters are updated in real time according to new observation data, and finally the remaining life of a bearing can be predicted. According to the rolling bearing remaining life prediction method based on feature fusion and particle filtering, the difference between the life prediction result and a true value is small, and the application effect is good.

The invention discloses an advertisement click-through rate prediction method based on multi-dimensional feature combination logical regression. The method comprises the first step that feature information of a hierarchical structure of the user hierarchy, feature information of a hierarchical structure of the media hierarchy and feature information of a hierarchical structure of the advertisement hierarchy are extracted from the obtained click-through rate data respectively; the second step that multi-dimensional combination is carried out on the feature information of the hierarchical structure of the user hierarchy, the feature information of the hierarchical structure of the media hierarchy and the feature information of the hierarchical structure of the advertisement hierarchy, three-to-three combination is carried out on one-dimensional feature information in the feature information to obtain a three-dimensional feature combination, and a feature vector combined by the three-dimensional feature information is formed to represent a user cluster; the third step that the second step is carried out repeatedly and a learning set of the feature vector combined by the three-dimensional feature information is obtained; the fourth step that the learning set obtained in the third step is used for training and testing a logical regression model, and the logical regression model is used for predicting the advertisement click-through rate.

The invention discloses an intra-frame and inter-frame combined prediction method for P frames or B frames. Whether the intra-frame and inter-frame combined prediction is used or not is adaptively selected through rate-distortion optimization RDO decision; a method for weighting an intra-frame prediction block and an inter-frame prediction block is used in the intra-frame and inter-frame combinedprediction to obtain a final prediction block; the weighting coefficients of the intra-frame prediction block and the inter-frame prediction block are obtained according to prediction distortion statistics of the prediction method; therefore the prediction precision can be improved, and the coding and decoding efficiency of the prediction block are improved; the advantages of intra-frame prediction and inter-frame prediction are fully used, the two methods are selected to combine the optimal prediction parts, the area within the excessive distortion of the intra-frame prediction block and theinter-frame prediction block can be removed to a certain extent to obtain a better prediction effect; and the practicality and the robustness are high.

The invention discloses a number-of-people estimation method based on deep learning semantic image segmentation. The number-of-people estimation method comprises a step 1 of constructing a training sample set including an original picture and a corresponding mask label picture; a step 2 of selecting or constructing a deep network model based on semantic image segmentation; a step 3 of training thesample set and obtaining a human head position prediction network model; and a step 4 of inputting a picture to be detected into the human head position prediction network model to obtain a mask picture, and obtaining the estimated number of people in the picture to be detected and position information of each person in the picture to be detected according to positions of points and the number ofthe points in the mask picture. Compared with an estimation method which is mostly employed in the prior art and based on image partitioning, the number-of-people estimation method is advantaged in that statistical errors brought by image partitioning can be avoided; the estimated number of people in the image area and the position of each pedestrian in the image area can be provided at the sametime.

The invention provides a method for recognizing inter-well connectivity and predicating the oil-water dynamic state. The method comprises the steps that a water drive oil deposit inter-well connectivity evaluation model considering two phases of oil and water is built; a water drive oil deposit oil saturation tracing equation for the two phases of the oil and the water is built; the effective control volume of a single well is calculated through a Koval method; initial values of a connectivity factor, a interference factor and a time constant are given; then, the average oil saturation and liquid production capacity of a jth oil well within a first time step are obtained according to the model and the equation; and the oil saturation of the oil-water front is solved; a minimized objective function is built according to injection and output data of an oil-water well, the objective function is subjected to minimized calculation through a constrained optimization algorithm, and the optimal connectivity factor, the interference factor, the time constant and the average oil saturation are obtained; and the connectivity relationship between injection and production wells is discriminated according to the optimal connectivity factor, and predication of dynamic indicators is conducted according to a model after update.

The invention relates to a virtual network function dynamic migration method based on deep belief network resource demand forecasting, and belongs to the field of mobile communication. The method comprises the following steps: (S1) in view of the dynamic features of SFC business resource demand in a slicing network, establishing a system overhead model of comprehensive migration overhead and bandwidth overhead; (S2) in order to realize spontaneous VNF migration, monitoring the resource utilization condition of virtual network function or link in real time, and discovering the deployed bottom nodes or resource hot spots in the link in time by using an online learning based adaptive DBN forecasting method; (S3) designing a topology awareness based dynamic migration method according to the forecasting result, so as to reduce system overhead; (S4) proposing a tabu search based optimization method to further optimize the migration strategy. The forecasting method provided by the invention not only increase the convergence rate of a training network, but also realizes a perfect forecasting effect; by combining the forecasting method with a migration method, the system overhead and the violation frequency of the service level agreement are effectively reduced, and the performance of network service is improved.

The invention discloses a new short-term wind speed forecasting method of a wind farm, which comprises the following steps: collecting wind speed data of the wind farm, forming the time sequence of the historical wind speed and carrying out normalization treatment; applying the chaos analysis method for analyzing the time sequence of the historical wind speed after the normalization treatment for obtaining phase space reconstruction parameters of a wind power system in the area located by the wind farm, wherein the parameters are delay time and embedding dimension of the time sequence; utilizing the parameters for carrying out treatment on the time sequence of the historical wind speed after the normalization treatment, and obtaining a training sample set required by a support vector regression model for wind speed forecasting; adopting the training sample set for training the support vector regression model; utilizing the support vector regression model after training for carrying out short-term wind speed forecasting on the wind farm, and obtaining the normalized result of the short-term wind speed forecasting of the wind farm; and carrying out anti-normalization treatment on the obtained normalized result of the short-term wind speed forecasting of the wind farm, and obtaining the short-term wind speed forecasting result of the wind farm.

The invention discloses an electric quantity consumption predicting method based on deep learning. According to the invention, a deep learning model is able to train a BP network according to the historical data as of now so as to achieve a better predicting effect. The reasons behind choosing the deep learning for electric quantity predicting are that it has a non-linear adaptive information processing ability unique to the neural network, that it has a strong error tolerance, and that it can be applied to the dynamic analysis for electric quantity consumption and meets a plurality of integrated factors such as time regularity and event suddenness. In the method, first, the intelligently sensed current and voltage data are calculated as electric quantity consumption amount for the training of a neural network so as to predict the electric quantity consumption at a next period. The predicted electric quantity consumption amount and the statistic power using duration are fed back to the user so as to guide him or her to conserve power. The method of the invention is simple and practical in use and can be applied to a smart home system connected via Wifi, and the method can also be used to predict the electric quantity consumption in regional and urban power grids.

The invention discloses an energy-saving routing method of nodes of an energy effective based on a context-aware technology. A CATRP (Context-Aware Technology Routing Protocol) is used as a routing work way of nodes in an internet of things environment, so as to achieve an established energy saving goal. The method disclosed by the invention belongs to the technical field of the internet of things. According to the method disclosed by the invention, the context-aware technology is applied to communicational nodes which work in the internet of things as a core technology for an energy-saving purpose of the invention, and composition modules of the protocol are designed comprehensively from the perspectives of working process, data structure, quantitative algorithm design, and the like. Due to application of the patent, a good foundation for realization of a universal computing service in the whole society in the near future can be laid, right development of the internet of things technology in the future is guided, theoretical basis enrichment and certain contribution for transformation of a wireless sensor network technology at the present stage are made, and the method has important meaning to enhancement of energy saving technology level of the internet of things in the industry, to acceleration of schedule of a national energy-saving routing research subject for nodes of the internet of things, to promotion of domestic demand, and to promotion of development of related industries.

The invention discloses a household appliance operation state non-intrusive detection method based on intelligent electric meter data. The method comprises the following steps: collecting electricityutilization total power data of a user family and predicting consumption power data of an electric appliance; marking the running state of the electric appliance, and performing normalization processing on the acquired power data; building a deep learning network model; training the established LSTM deep learning network to obtain a trained LSTM network model; testing the trained LSTM network model, and checking the network prediction accuracy; collecting power data of any home-entry intelligent electric meter to serve as input of an LSTM network model, and detecting and recognizing the operation states of a plurality of household appliances. The method has a good effect on load operation state identification, compared with a common single prediction network, the training time of the network is greatly shortened, a good prediction effect can also be achieved for different regions by using transfer learning, and the method has very high value in production and life.

The invention relates to a multi-video coding and decoding method and a device. The method mainly comprises the following steps that: firstly, the reference distance between a reference image and a current image is determined according to the time distance and the space distance between the reference image and the current image; a reference image with at least one frame adopted by the current image is chosen and determined according to a reference distance of each candidate reference image and the reference image is adopted to code and decode the current image. Therefore, the invention can realize the optimal predicting effect by adopting the image in a current reference cache under the circumstance that an additional reference cache is not necessary, thereby raising the multi-video coding and decoding efficiency.

The invention relates to a prediction method suitable for a movement track of a non-cooperative spinning object in space. A position coordinate of the object in a camera coordinate system is obtained, and a position coordinate of the object in an inertia coordinate system is solved according to a known camera attitude; and an NAR neural network is constructed, a BPTT algorithm is used to train the neural network, the position value of the object is predicted and output after deviation convergence, and the system robustness and sampling continuity are ensured. Position information is used to calculate an attitude transformation quaternion, the attitude transformation quaternion is calculated via kinematical and kinetic equations according to estimated parameters, and an estimated parameter value of least square regression is used to calculate a prediction result via the equations. The method is applied to an in-orbit service task of the space, the object track can be traced rapidly and accurately, and long-term prediction information of the object track can be obtained after parameter convergence.

The invention discloses a Chinese text sentiment analysis method based on deep learning, and belongs to the technical field of natural language processing. The defects of an unsupervised sentiment analysis method based on English are overcome. The method comprises the following steps: after converting an obtained corpus text into pinyin, pre-training a constructed language model to obtain a pre-trained language model; obtaining a small amount of text data which is in the same field as the corpus text and has emotion categories, converting the text in the text data into pinyin, and training a constructed emotion classification model based on a pre-trained language model to obtain a trained emotion analysis model; and carrying out sentiment classification on the unlabeled text by utilizing the trained sentiment analysis model to obtain a corresponding sentiment category label. The method is used for Chinese text sentiment analysis.

The invention discloses a method for predicting industrial sewage inflow based on an ARIMA model. The method comprises the following steps: analyzing initial time series data to meet a requirement onARIMA model establishment; preprocessing abnormal data by eliminating, filling and the like; removing data noise through moving average filtering; by a unit root testing method, performing ADF testingon the stability of a time series; analyzing and verifying the non-randomness through an autocorrelation coefficient; preliminarily determinating an autoregressive and moving average order of the ARIMA (p, d, q) model, and then performing order determination on the model through combination with an AIC information criterion; optimizing model parameters by a least squares method; finally, testinga residual and evaluating a simulation result to determine a final prediction model. Acquired sewage inflow data are determined, and the obtained prediction model is used for predicting test data, andan output of the model is a prediction result of the sewage inflow. By the method, the model is succinct, the fitting effect of the prediction model is very good, and the precision is high.

The invention belongs to the field of social network modeling and analysis, and particularly relates to a method for predicting the maximum information spreading range on the basis of a random model, and by means of the method, dynamic characteristics of a social network are explored. According to the method, a set of functions capable of describing the network information spreading dynamic characteristics are constructed, a dynamic information spreading model is built according to historical data of social network information spreading, a random model detector is used for predicting the possible maximum information spreading range through a verification and emulation technology, and the node set capable of maximizing the spreading range is found out, wherein information is spread through different node sets. Compared with a traditional spreading range maximization modeling method, the dynamic characteristics of the network can be modeled, so that the initial node set which is predicted out is higher in quality, and the success rate of a network marketing strategy is increased.

The invention belongs to the technical field of urban road traffic flow prediction and analysis and particularly relates to a method for road traffic flow prediction under a suddenly occurred trafficevent. The method comprises the steps that road traffic event alarm information data is preprocessed, and abnormal data is removed and repaired; a place name address space positional information database is established, traffic event position information is obtained and classified, and spatial and temporal distribution characteristics of road traffic flow and spatial and temporal characteristics of road traffic flow under the suddenly occurred event are analyzed; a random forest algorithm, an ARIMA method and a Kalman filtering method are utilized to conduct traffic flow prediction on time series data and space state data under the suddenly occurred event; a weighted least square method is used for conducting fusion processing on prediction results obtained through a time series data prediction method and a spatial series prediction method, and new prediction results are obtained. In the method, three indexes including a comprehensive error percentage absolute value mean value, an error absolute value mean value and a square-error mean value reach an ideal prediction effect.

The invention provides a method for analyzing a data service. The method for analyzing the data service includes the following steps: building a correlation model and a type preference model based on using conditions of a user to the data service, and building a feature matching model; then building a data service relation model by making use of the correlation model and the type preference model, and building a data service integrated analysis model by using the data service relation model and the feature matching model; and finally analyzing the data service by means of the data service integrated analysis model and using the analysis result in data service recommendation. By means of the method for analyzing the data service, the accuracy of the analysis result is improved, the analysis result is used in the data service recommendation, and thus the precision of data service recommendation can be improved.

The invention belongs to the technical field of soil detection, and discloses a soil nutrient prediction and comprehensive evaluation method based on machine learning algorithm. The method includes collecting a soil sample and measuring various soil nutrient indexes and soil moisture; collecting various environmental variable data; conducting spatial interpolation on each soil nutrient index for prediction by combining correlation analysis with random forests to determine the spatial distribution of soil nutrients; comparing the prediction accuracy of a model by calculating an average error, amean absolute error and a root mean square error of a verification point; determining the correlation between the soil nutrient and the soil moisture and between environmental variables and fertilization amount; utilizing a projection seeking model to comprehensively evaluate the soil nutrient and making a spatial distribution map of evaluation grade results. The method attempts to provide a newidea for soil nutrient evaluation from a non-linear perspective by the relationship between nutrient grades and evaluation indexes.

The invention discloses a traffic flow prediction method based on deep neural network integration. The method comprises the following steps: obtaining original traffic flow data, carrying out data preprocessing, constructing sample data, dividing the sample data into a training set and a test set, and enabling the sample data to be one-dimensional time series data consisting of a plurality of traffic flow measurement values; constructing a convolutional neural network prediction model for traffic flow prediction, inputting sample data into the model, training by using a gradient optimization algorithm, and calculating a variance of a prediction error for the trained model on the training set; using the convolutional neural network prediction model as an individual learner according to thevariance of the prediction error, and constructing a convolutional neural network integration model for traffic flow prediction; and predicting the traffic flow data to be tested by using the convolutional neural network integration model. Based on a convolutional neural network model, an integrated learning method is utilized, an improved traffic flow prediction method is provided, and the prediction accuracy is improved.

The invention relates to the technical field of wind power prediction, and discloses a wind power short-term prediction method. The method uses wind speed as an input, adopts a regression model of a least square support vector machine to predict output power of a wind power plant, and parameters of the regression model of the least square support vector machine are optimized by adoption of a chaotic particle swarm algorithm. The wind power short-term prediction method provided by the invention introduces chaotic motion characteristics into an iterative process, uses ergodicity of chaotic motion to improve a global searching capability of the algorithm in a searching process, overcomes the defects that the particle swarm algorithm is easy to fall into a local extreme point and is slow in convergence and low in precision in a later period of evolution, effectively solves the problem of prematurity of the particle swarm algorithm, can ensure global optimum, and achieves a better prediction effect; the method uses the least square support vector machine to predict, avoids the problem of solving quadratic programming, converts the prediction problem to a process of solving a linear equation set, and the solving process is greatly simplified; and the method adopts single wind speed as input data, and thus a prediction model is simpler.

The invention discloses a method for predicting an author cooperation relation in an academic heterogeneous information network. The problem about author cooperation relation prediction is solved by utilizing a heterogeneous information network close to a real world. The method comprises the following steps: acquiring the topological attribute of the network according to different measurement of a metapath in the constructed academic heterogeneous information network, introducing concepts of temporal dynamics, transfer similarity and author attribute to acquire the content information of the network, combining the topological attribute and the content information to acquire a characteristic space based on the metapath and the content information, and finding the optimum weight of each characteristic attribute according to the obtained characteristic attribute set and by a logical regression algorithm to predict the author cooperation relation. According to the method, the potential cooperation relation of scholars can be excavated by utilizing academic big data, the scholars can be helped to perform efficient scientific research cooperation and know the academic circle of the scholars, and in particular a good prediction effect on high-yield scholars and high-frequency cooperation relation is achieved.