65results about How to "Accurate and Effective Prediction" patented technology

The invention belongs to the field of traffic engineering, and discloses a multi-factor short-term traffic flow prediction method based on neural network LSTM. The multi-factor short-term traffic flowprediction method based on the neural network LSTM comprises the following steps: step 1, obtaining traffic flow data of a period of time, and preprocessing the traffic flow data to obtain short-termtraffic flow data; step 2, screening the short-term traffic flow data according to weather records and holiday records, and dividing data sets; step 3, performing data cleaning, data reconstruction,and normalization; and step 4, establishing an LSTM neural network model, selecting the data set according to the weather conditions and holiday conditions of the date to be predicted, using the selected data set to train the LSTM neural network model and adjust the LSTM parameters, and obtaining the traffic flow of the date to be predicted based on the established LSTM neural network model. The invention provides a more detailed idea, excludes influences of other factors on the traffic flow, such as weather factors and holiday factors, and relatively improves the prediction accuracy, so thatthe traffic flow prediction of a certain period in the future is more accurate and effective.

Embodiments of the invention provide a transport capacity configuration method and device, and relate to the technical field of computer applications. The method comprises the following steps of: determining at least one influence factor influencing on-duty number of delivery staffs; determining a predicted numerical value of the at least one influence factor on the basis of historical statistical data of any service area; predicting a delivery staff configuration number of the service area by utilizing a transport capacity prediction model on the basis of the predicted numerical value, wherein the transport capacity prediction model is obtained through carrying out training on the basis of the on-duty number of the delivery staffs and a historical numerical value of the at least one influence factor in the historical statistical data; and configuring the service area according to the delivery staff configuration number. According to the technical scheme provided by the embodiments, the transport capacity configuration accuracy is improved.

The invention provides an airport passenger volume prediction method based on typical day identification, which comprises the following steps: collecting airport flight history day operation data setP, extracting historical day time feature vector and historical day flight feature vector for each historical day operation data; The similarity of feature dataset is quantified and the optimal weighting factor is obtained. A typical day identification model is established to predict the passenger volume in different time periods. The method for predicting airport passenger volume based on typicalday identification provided by the invention has the following advantages: the invention realizes accurate and effective prediction of airport passenger volume in different time periods, thereby effectively improving airport operating pressure early warning capability.

The invention relates to an output power prediction method for photovoltaic generating devices; the method comprises: predicting by using a physical model prediction, and a neural network prediction of environmental factors and a rolling a neural network prediction; setting the physical model prediction as P0, the neural network prediction of the environmental factors as P1, the rolling a neural network prediction as P2, and final power prediction output value as p; if P1 is more than P0 and P2 is more than P0, P is equal to P0; if the P1 is more than P0 that is more than P2, P is equal to P2; if the P2 is more than P0 that is more than P1, P is equal to P1; if P1 is less than P0 and P2 is less than P0, P is equal to (P1+P2)/2. According to the output power prediction method for photovoltaic generating devices provided by the invention, relatively complex mathematic relation is changed to be simple and reasonable mathematic relation, thus being beneficial to applying to actual engineering, improving predicting speed, reducing predicting difficulty, and being capable of significantly improving accuracy of prediction value of photovoltaic module output power, controlling error range and eliminating outlier.

The invention discloses a hydroelectric generation prediction method based on an extreme learning machine. The method comprises the following steps: obtaining parameter data information and preprocessing data from a hydroelectric generation system; dividing the data into two mutually exclusive parts, performing data training on one part, and performing data testing on the other part; acquiring training data, and establishing a model by adopting the training data; obtaining an optimal model by adopting methods of cross validation, grid search and model evaluation and model training; adopting the trained optimal ELM model to predict test data, obtaining and outputting a prediction result, wherein ELM is an extreme learning machine model. Through the method, higher learning speed and better generalization ability are displayed; the hydroelectric generation is more accurately and effectively predicted, the cost is reduced, and the learning rate is improved.

The invention discloses an electric power quotation method and terminal equipment. The method comprises steps of acquiring a sample data set, wherein sample data in the sample data set comprises electric power transaction prices in a preset historical time period and influence factors corresponding to the electric power transaction prices, and the influence factors are used for representing an electric power supply and demand relation between electric power production cost and an electric power market; obtaining a label corresponding to the sample data, wherein the label is used for representing an electric power expected quotation corresponding to the sample data; and based on the sample data in the sample data set and the corresponding label, training a quotation model to predict an electric power expected quotation at a target moment. Therefore, the quotation model is trained on the basis of the sample data in the historical time period and the corresponding power expected quotation, so that the power expected quotation at the target moment can be more effectively and accurately predicted by the quotation model when the sample data at the target moment is input into the trainedquotation model.

The invention discloses a coal mine water inflow prediction method and system based on an LSTM algorithm, and belongs to the technical field of artificial intelligence and coal mining water disaster prevention and control, and the method comprises the following steps: 1) analyzing and screening coal mine water inflow related factors to construct a coal mine water inflow characteristic matrix; 2) performing data processing on the coal mine water gushing characteristic matrix, wherein the data processing comprises characteristic matrix variable correlation screening and characteristic matrix variable normalization processing; 3) constructing a coal mine water inflow prediction model based on an LSTM algorithm, constructing an LSTM network structure, and training the prediction model, the LSTM network structure having a memory unit, an input gate, a forgetting gate and an output gate; 4) carrying out model prediction evaluation and model usage. According to the method, the problem that inan existing coal mining water disaster prevention and control technology, the water inflow in the coal mining sink cannot be accurately predicted is solved, the water inflow in the coal mining sink can be accurately and effectively predicted, and safe coal mining is guaranteed.

The invention provides a pension insurance payment income prediction method. The method comprises the steps of obtaining population prediction data of a prediction point; according to the population prediction data of the prediction point, a pension insurance coverage rate, a pension insurance payment rate, a labor participation rate and an unemployment rate, calculating insured payment population prediction data of the prediction point; according to a current per-capita wage, a wage increase rate, an inflation rate, a payment wage rate and the insured payment population prediction data of the prediction point, calculating payment wage data of the prediction point; and according to the payment wage data of the prediction point, calculating pension insurance payment income data. The invention furthermore provides a pension insurance payment income prediction apparatus. According to the pension insurance payment income prediction method and apparatus, the pension insurance payment income is accurately and effectively predicted by accurately calculating the population prediction data of the prediction point, the insured payment population prediction data of the prediction point and the payment wage data of the prediction point.

The invention discloses an ozone concentration prediction model coupled with a landscape pattern, which is characterized in that an ozone concentration multiple regression model module is constructed by taking an index value of a sampled urban landscape pattern as basic data, and the input end of the multiple regression model module is connected with the output end of a collinearity analysis unit; the input end of the collinearity analysis unit is connected with the output end of the data processing unit, the data processing unit is in bidirectional connection with the data verification unit, and the output end of the collinearity analysis unit is connected with the input end of the correlation analysis unit. The invention relates to the technical field of ozone concentration prediction. According to the ozone concentration prediction model coupled with the landscape pattern, ozone concentration values in spatial distribution are obtained through an interpolation technology, different buffer area ranges are delimited, ozone values in the ranges are measured, correlation analysis and multiple regression modeling are carried out on the ozone values based on different buffer areas. Therefore, a proper buffer area range is determined, the correlation between the urban landscape index and the ozone concentration is explored, and the ozone concentration prediction based on the landscape pattern is realized.

The invention provides a code generation method and device, electronic equipment and a readable storage medium. The method comprises the steps of obtaining a webpage screenshot of a target front-end webpage and a webpage source code corresponding to the webpage screenshot; extracting a first feature of the webpage screenshot based on the convolutional neural network visual model with the attention mechanism, and extracting a second feature of the webpage source code based on the language processing network model with the attention mechanism; fusing the first feature and the second feature to obtain a fused feature; and generating a code of the target front-end webpage based on the fusion feature and the decoding model. The attention mechanism is added into the image processing module and the natural language processing module, and the webpage screenshot and the characteristics of the corresponding webpage source code are extracted and fused based on the attention mechanism, so that accurate and effective prediction of codes can be realized only through key calculation on a part of vocabularies with strong correlation and an image local region; therefore, the operation amount can be effectively reduced, and the operation efficiency and accuracy are effectively improved.

The embodiment of the invention provides a credit risk index prediction method and device. The method comprises the steps that: credit risk indexes and corresponding to-be-selected bank macroscopic indexes are determined; data quality analysis and correlation analysis are performed on the credit risk indexes and the to-be-selected bank macroscopic indexes, and target macroscopic indexes in the to-be-selected bank macroscopic indexes are determined according to analysis results of the data quality analysis and the correlation analysis; and a linear regression prediction model is constructed according to the credit risk indexes and the target macroscopic indexes, and credit risk index prediction data are obtained according to the linear regression prediction model. According to the method and device of the invention, credit risk indexes in the financial risk control field can be effectively and accurately predicted.

The invention relates to a fault determination method and system for a variable pitch system of a wind turbine generator. The fault determination method comprises the following steps of obtaining a variable pitch system dynamical model; according to the variable pitch system dynamical model, obtaining data after the fan variable pitch system breaks down; training a convolutional neural network according to the data after the fault occurs, and determining a trained convolutional neural network; obtaining a current pitch angle, a current pitch angle measurement value, a current main transmission shaft torque, a current actual rotating speed of a generator rotor and a current rotating speed measurement value of the generator; determining a currently predicted pitch angle according to the current pitch angle measurement value, the current main transmission shaft torque, the current actual rotating speed of the generator rotor, the current rotating speed measurement value of the generator and the trained convolutional neural network; and determining a fault diagnosis result of the fan variable pitch system according to the currently predicted pitch angle and the current pitch angle. According to the fault determination method and system for the variable pitch system of the wind turbine generator, the fault can be accurately and effectively detected and determined, and meanwhile, the change of the fault amplitude is accurately and effectively predicted.

The invention relates to a nearest neighbor regression-based time sequence underwater acoustic channel quality prediction method, which comprises the following steps of: initializing: receiving an initial data packet by an underwater sensor network node to obtain an identifier, residual energy consumption and a signal-to-noise ratio of a neighbor node, and establishing a channel quality matrix comprising the identifier, residual energy consumption and the signal-to-noise ratio of the neighbor node; in the active packet sending step, the node entering the active packet sending state adopts a time sequence underwater acoustic channel quality evaluation algorithm based on nearest neighbor regression to obtain a neighbor channel quality evaluation value of the node, confirms a forwarding nodeof the next hop according to the evaluation value, adds an identifier of the forwarding node into a data packet, and broadcasts the evaluation data packet; and a passive receiving step: after receiving the data packet, the node in the passive receiving state updates the channel quality matrix and judges whether the node itself is the forwarding node or not through the comparison identifier.

The invention discloses a goaf subsidence region sinkage monitoring system and method, belonging to the field of environment treatment of mining areas. The method comprises the following steps: making a monitoring hole from the ground to the goaf floor, arranging a steel base retaining cylinder on the bottom of the hole, and anchoring to wall rock through an annular bulge under the action of structural adhesive or cement mortar; respectively and mutually nesting a constant-speed sinkage layer retaining cylinder, a low-speed sinkage layer retaining cylinder and a slight-speed sinkage layer retaining cylinder with the steel base retaining cylinder layer by layer through a seal ring and balls, and anchoring with the affiliated layer through the annular bulge; and by using a first fixed end, a first measuring tape, a second fixed end, a second measuring tape, a third fixed end and a third measuring tape, reading the sinkage values of the layers by the aid of a ground pulley and a counterweight-weight composite structure, thereby implementing the monitoring on the goaf subsidence region. The system and method can implement real-time continuous small-scale monitor on sinkage development of the goaf subsidence region, and provides powerful support for environment treatment of mining areas.

The invention discloses an insect pest early warning system based on the Internet of Things. The insect pest early warning system comprises a data acquisition subsystem, a data storage subsystem, a data analysis subsystem and an insect pest early warning subsystem, the data acquisition subsystem comprises an environment information acquisition module and an image acquisition module; the data storage subsystem comprises an insect pest basic database, a crop characteristic database and an environment information database; and the data analysis subsystem is used for processing the data acquired by the data acquisition subsystem through the crop phenological period prediction model and the insect pest effective accumulated temperature prediction model in combination with the insect pest basic database and the prediction characteristics of each model, and outputting the processed data to the insect pest early warning subsystem. According to the method, the occurrence date of each insect state of different insect pests is predicted by combining the characteristics of the crop phenological period prediction model and the insect pest effective accumulated temperature prediction model, the prediction is accurate and effective, the pre-prevention work of crops is facilitated, the operation is simple, and the method has a relatively high application value.

The invention relates to a method and kit for detecting an SNP (Single Nucleotide Polymorphism) site rs10081191 of a susceptible gene region 7q36.3 related to a disease risk of NIHL (Noise Induced Hearing Loss), and application. The method is Sequenom typing. The invention further relates to the kit for detecting the susceptible site and application of the kit. In addition, the SNP site rs10081191of the region 7q36.3 is associated with the disease risk of the NIHL, and the invention discloses the SNP and application thereof. The SNP is located in a human genome region 7q36.3, the 7q36.3 is determined to be a new susceptible gene region of the NIHL, and the SNP can be effectively used for screening susceptible groups of the noise induced hearing loss diseases and determining susceptible individuals of the noise induced hearing loss diseases.

The invention discloses an age prediction method based on hybrid double models. The age prediction method comprises the following steps: constructing a first deep neural network; initializing the first deep neural network; obtaining a soft label of a face age picture; constructing a first target function, performing loss calculation on the output of the first deep neural network and the soft labelby adopting the first target function, and updating the first deep neural network; constructing a second deep neural network; initializing the constructed second deep neural network; obtaining a label distribution label of the face age picture; constructing a second target function, performing loss calculation on the output of the second deep neural network and the label distribution label by adopting the second target function, and updating the second deep neural network; and using the trained first deep neural network and second deep neural network to perform result extraction on a to-be-tested photo, and predicting the age. According to the method, two networks adopting the new inventive method are integrated, so that the accuracy of age prediction can be effectively improved, and errors can be effectively reduced.

The invention belongs to the technical field of data processing, and relates to an infectious disease early warning method and device, a storage medium and electronic equipment. The method comprises the following steps: acquiring a sample data set, wherein the sample data set comprises close connection information of a target crowd and definite diagnosis information of the target crowd; performing quantification processing on the close connection information to obtain feature data; dividing a sample data set according to a cross validation method, and training a model according to feature data and definite diagnosis information corresponding to the divided sample data set to obtain an initialized model; verifying the feature data in the initialization model to obtain a risk assessment model; and obtaining the close connection information of the to-be-detected crowd, performing risk assessment on the close connection information of the to-be-detected crowd according to the risk assessment model to obtain the probability of close connection and definite diagnosis of the to-be-detected crowd, and performing early warning on infectious diseases according to the probability of close connection and definite diagnosis. According to the invention, data guarantee is provided for obtaining a risk assessment model with good interpretability, and an automatic and intelligent infectious disease early warning mode is provided.

The invention discloses an ocean oil spill emulsification prediction method and device and a storage medium. The method comprises the following steps: obtaining a remote sensing image set of a marinetest field which is provided with oil spill samples and comprises a plurality of oil spill sample putting points in advance; obtaining a corresponding relation table of the hyperspectral sensitive wave band and the emulsification degree based on the remote sensing image set,; obtaining an oil spill emulsification change initial model according to the hyperspectral sensitive wave band and emulsification degree corresponding relation table, the remote sensing image set, the influence factor change data and the oil spill emulsification occurrence time; correcting the initial oil spill emulsification change model to obain an oil spill emulsification change model. The oil spill emulsification degree change and the oil spill distribution change of any oil spill sea area are predicted by applying the oil spill emulsification change model, and the continuous emulsification change process in the oil spill is fully considered, so that the development process of oil spill emulsification is accurately and effectively predicted.

The invention belongs to the technical field of logistics combat readiness material consumption prediction, and discloses a combat readiness material consumption prediction model construction method, which comprises the steps of S1, dividing the whole combat service into a plurality of stages according to the nonlinear and uncertain characteristics of combat readiness material demands; s2, according to the actual situation of each stage, the types of reserve materials needing to be consumed in each stage are determined; s3, scientifically measuring and calculating the war readiness material consumption quantity based on the basic data by using the material consumption data obtained by deducing the war chess for ten times, and analyzing and concluding a war readiness material consumption rule to obtain a material consumption calculation model; s4, a tactical classification and summarization method is applied, and the total consumption amount of the reserve materials obtained through superposition is counted; the method solves the problem that a logistics support consumption prediction model in the prior art is poor in accuracy and effectiveness, and is suitable for the construction of a logistics combat readiness material consumption prediction model.