296results about How to "Predictable" patented technology

The invention discloses a traffic flow prediction method based on a genetic algorithm optimized LSTM neural network. The traffic flow prediction method based on the genetic algorithm optimized LSTM neural network comprises the steps of: S1, acquiring traffic flow data, performing data normalization pre-processing, and dividing the traffic flow data into a training data set and a test data set; S2,predicting various parameters of a model by adopting the genetic algorithm optimized LSTM neural network; S3, inputting genetic algorithm optimized parameters and the training data set, and performing iterative optimization of an LSTM neural network prediction model; and S4, predicting the test data set by using the trained LSTM neural network model, and evaluating the model error. According to the traffic flow prediction method based on the genetic algorithm optimized LSTM neural network in the invention, by utilization of the rapid optimization feature of the genetic algorithm and the LSTMneural network on parameter combination, the relatively high prediction precision can be obtained; furthermore, the method has good applicability on data samples in different intervals; the calculation amount is reduced through the model; and the prediction performance is better.

The invention discloses a time sequence classification early warning method for a storage device. The method comprises the steps of collecting storage device parameters in real time; cleaning data; performing ARIMA time sequence analysis; and performing logistic regression analysis and early warning mechanism output. Under the background of a big data environment, time sequence prediction analysisis performed by adopting an ARIMA model according to historical data and hard disk SMART information obtained by statistics; the correlation between a SMART eigenvalue and a fault rate of the storagedevice is analyzed; and an eigenvalue more suitable for a Logistic model is selected out to perform classification prediction. A machine learning method is adopted for predicting the fault rate of the storage device, so that the problems of classification singleness and low early warning intensity in final prediction of the storage device are solved, the defects of hysteresis, low accuracy, pooractual early warning effect and difficult application to the big data environment for a disk early warning mechanism in the prior art are overcome, the occurrence probability of each early warning intensity can be predicted, and an effective solution is provided for real-time operation maintenance and monitoring in a data center environment.

The present invention discloses an integrated similarity measurement and bi-directional random walk based pharmaceutical relocation method. When calculating pharmaceutical similarity and disease similarity, other than taking advantage of pharmaceutical characteristics information and disease characteristics information respectively, an integrated similarity measurement method further takes full account of effects on similarity measurement due to pharmaceutical-disease related information in a current data set, so that a calculated similarity value can better reflect similarity between pharmaceuticals and similarity between diseases. On this basis, a pharmaceutical-disease heterogeneous network is built, and based on the heterogeneous network, a bi-directional random walk algorithm is taken to predict a candidate disease for all pharmaceuticals. The method disclosed by the present invention is simple and effective, and compared with other methods, tests on multiple data sets prove that the method disclosed by the present invention has better prediction performance in pharmaceutical relocation.

The invention provides an online and intelligent optical cable monitoring and fault positioning system based on a GIS platform. The online and intelligent optical cable monitoring and fault positioning system comprises a monitoring unit and a positioning unit, wherein the monitoring unit comprises an intelligent diagnosis mainframe, a diagnosis unit and a PC terminal display unit, the intelligent diagnosis mainframe comprises a central processing unit and an optical time domain reflectometer module, an optical path switching module and an optical power monitoring module connected with the central processing unit, the diagnosis unit comprises a power module and a light source device, an optical coupling module and an optical protection module connected with the power module, and the positioning unit comprises a fault locator arranged in an optical cable. Information is processed by calling optical cable information in time, and meanwhile the system can flexibly utilize a geographic information system (GIS) technology, adopts rack-type integration, and is high in integration level, good in stability and high in expandability. When breakdown occurs, the system can clearly display fault GPS points and fault types, maintainers can perform accurate positioning through a GIS map, and accordingly the maintenance efficiency is improved.

The invention discloses a travel time prediction method for optimizing an LSTM neural network through a particle swarm optimization algorithm, and the method comprises the following steps: S1, collecting travel time data, performing data normalization, and dividing the data into a training set and a test set proportionally; S2, optimizing each parameter of an LSTM neural network prediction model by using the particle swarm optimization algorithm; S3, inputting the parameters, optimized through the particle swarm optimization algorithm, and the training set, and performing the iterative optimization of the LSTM neural network prediction model; S4, predicting the test set through the trained LSTM neural network model, and evaluating a model error. The method is quick in optimization. Compared with a random forest, SVM and KNN in the traditional prediction algorithm, the method of the invention has the least mean square error and square error for the data prediction, and the model reducesthe calculation burden, so the method shows better prediction performance.

The invention relates to a small straight-line section interpolation method of a numerical control system on the basis of multicycle optimal corner, comprising the following steps of: on the basis of processing precision and limitation of maximum processing speed, determining interpolation parameters of an optimal corner with multicycle trasition of each corner on a processing path according to geometric parameters at the corner on the processing path, the maximum acceleration of each driving shaft of a machine tool and an optimized target; adjusting the interpolation parameters of the optimal corner of each corner and leading the processing speed of two ends of each small straight-line section to meet reachability requirement; and then carrying out interpolation of straight line sections and corners to all the small straight-line sections on the processing path, connecting interpolate point sequences of the straight line sections with interpolate point sequences of corners, sequentially outputting the interpolate point sequences in real time, and driving a numerical control machine tool to conduct actual processing. The method can effectively improve the whole processing speed, has fast calculating speed, meets the requirement of real time processing, has stable and reliable working performance and strong practicability and can be applicable to various three-shaft numerical control machine tool with different types.

A method for selecting wire of single phase earthing in neutral point noneffective earthing system includes making instant short circuit between neutral point and ground at position near two ends voltage polarity from positive to negative over zero by short circuiter set between neutral point and ground when single phase earthing is occurred in neutral point noneffective earthing system and state is stabilized in order to generate a short circuit pulse current being used to judge earthing line through short circuit detector. The device for realizing said method is also disclosed.

The invention discloses a drug repositioning method based on multi-information fusion and a random walk model. According to the method, disease-target-drug heterogeneous network is constructed through integrating existing disease data, drug data, target data, disease-drug associated data, disease-gene associated data and drug-target associated data; the basic random walk model is extended to the constructed heterogeneous network; and candidate therapeutic drugs are recommended for diseases through effectively utilizing global network information. The method disclosed by the invention is simple and effective; and compared with other methods and proved by tests on a standard data set, the method has good prediction performance in the aspect of drug repositioning.

The invention discloses a chilled fresh meat quality non-destructive testing method based on a hyperspectral imaging technology. The method includes the steps of: 1) collecting and preparing a large number of chilled fresh meat samples, and collecting hyperspectral data of the chilled fresh meat samples by means of a hyperspectral imaging system; 2) according to national standard demand, testing physical and chemical reference values of to-be-tested quality indices of the chilled fresh meat samples by means of physical and chemical methods; and 3) with the acquired hyperspectral data and the physical and chemical reference values of the chilled fresh meat samples, establishing a hyperspectral prediction module of the to-be-tested quality indices of the chilled fresh meat by means of data processing methods such as machine learning, chemometrics and the like. When the method is applied to detection of content of TVB-N in beef, abnormal samples of chilled fresh beef are rejected through a MCS method, and then chilled fresh beef sample sets are divided through a CG algorithm; a CARS algorithm is employed to establish a PLSR model by optimally selecting a characteristic wave band of the content of TVB-N in the beef. The model has excellent prediction function.

The invention discloses a method and a system for automatically selecting an application proxy server. The method and the system are used for selecting the optimal proxy server at last by comprehensively considering factors such as server load, resource requirements of applications, current distribution of applications and exclusiveness between applications. The method and the system are characterized in that in the load estimation aspect, the load balance of multi-dimension resources is taken into account so that the load condition of each server can be estimated more accurately, in the aspect of estimating the resource requirements of applications, historical data are used for estimating the resource requirements of applications, and certain foreseeability and relatively high accuracy are achieved, in the application distribution aspect, all applications at present are counted accurately and the optimal running positions of the applications are entirely controlled from a server cluster, and in the application exclusiveness aspect, the exclusive applications possibly competing with each other in resources are analyzed according to the resource utilization situations of all the applications, and therefore, the exclusive applications are prevented from running in the same server and the resource requirements of the applications are met to the utmost extent.

The invention provides a queuing-number obtaining method and a system provided with appointment function. The method includes the steps: an absence service appointment is applied to an appointment managing system through internet, telephone or messages to obtain a service number, a printed code and appointed service time; the service number is taken out from a number-generating machine with the printed code within the appointment time and the calling of a pager is waited. The system includes a background service manager, the number-generating machine, the pager and the appointment managing system. The invention can avoid the defects in the prior art that the service number only can be obtained by reaching the service hall and the waiting in a queue always takes so much time as the invention leads the queuing-number obtaining system to be provided with the appointed service function in virtue of the obtaining of appointed service number and the learning of appointed service time, so as to lead people to expect the service queuing state. The queuing-number obtaining method and system can save time for clients and enhance the service quality.

The invention provides an epidemic disease forecasting method, system and equipment, and relates to the technical field of epidemic disease forecasting. The method comprises the steps that the epidemic disease relevant data are acquired, wherein the relevant data include the internet data, the weather data and the biological information data acquired by the intelligent terminal, and the internet data include the search data and the social network data; the data cleaning is performed on the relevant data; an integrated forecasting model is established according to the relevant data after data cleaning; and forecasting is performed according to the integrated forecasting model and the forecasting result is outputted. According to the epidemic disease forecasting method, system and equipment,the data sources are wider, the forecasting effect is better, the forecasting accuracy is higher, the cost is lower and the effectiveness is higher and the possible epidemic disease outbreak can be forecasted earlier.

The invention provides a collaborative anti-tumor multi-drug combination effect prediction method based on a deep learning algorithm and pharmacogenomics. The collaborative anti-tumor multi-drug combination effect prediction method comprises the following steps: (1) mining and preprocessing large-scale pharmacogenomics data; (2) effectively integrating different feature information and constructing a modeling sample; (3) constructing a collaborative anti-tumor multi-drug combination prediction model based on large-scale sample data and a deep learning algorithm; and (4) performing parameter optimization and performance improvement of the model. According to the method, an artificial intelligence deep learning algorithm and pharmacogenomics are effectively combined, the limitation that a traditional collaborative drug combination prediction method can only be used for predicting the synergistic effect between every two drugs is overcome, and the specific collaborative anti-tumor multi-drug combination can be screened out for different tumor cells through the gene level; therefore, theoretical basis and technical support are provided for solving the problem of tumor drug resistance,and more effective treatment schemes are further provided for clinical tumor treatment.

The invention discloses a method with a strong discrimination power, which is used for calculating and recognizing active ingredient of a natural product on the basis of compound characteristics. The method is characterized in that whether a molecule has biological activity or not can be calculated and recognized in an extractive molecule of the natural product with a known structure by utilizing descriptors of compound characteristics. The method comprises the following steps: step A, a training set and a predicting set of natural product extractive molecules used for model building are structured; step B, molecular structure files of the training set and the predicting set are collected; step C, the descriptors of the compound characteristics can be figured out by utilizing the molecular structure files; step D, classification modeling can be performed on the training set by using a machine learning software according to the descriptors of the compound characteristics; and step E, the molecule of the predicting set has biological activity or not can be identified by the machine learning software with the combination of step D. The method of the invention has a good application prospect in high flux virtual screening and the study of synergistic effect of natural products.

The invention discloses a shopping mall building air conditioner cooling load prediction method based on GBDT, a storage medium and equipment, and the method comprises: collecting cooling load data, and carrying out the normalization processing to serve as the cooling load energy consumption prediction; establishing a load prediction model based on a gradient lifting decision tree algorithm; inputting the preprocessed data into a prediction model for training, selecting a grid search-cross validation mode, and optimizing the three hyper-parameters with the maximum influence on the performanceof the GBDT model; establishing a final cold load prediction model by completing parameter optimization of the prediction model, and obtaining a predicted cold load curve according to the parameters and the structure of the prediction model; and evaluating the prediction performance of the prediction model, adopting the prediction error for evaluation, enabling the deviation between the true valueand the prediction value to form the prediction error, and completing mall building air conditioner cooling load prediction. The method has good prediction precision, universality and applicability,and is especially suitable for large public buildings with periodically changing cold loads.

The invention discloses a method for predicting an N6-methyladenosine modification site in RNA based on stacking integration, belonging to the field of systems biology. The method comprises the following steps: extracting RNA sequence features of three species, namely saccharomyces cerevisiae, homo sapiens and arabidopsis thaliana through six feature extraction methods, and conducting feature fusion to obtain an initial feature space of an original data set; performing dimensionality reduction on the initial feature space by using an elastic network, eliminating redundant and noise features, and reserving important features related to model classification so as to obtain an optimal feature set; inputting optimal feature subsets and corresponding category labels into stacking integration for model training, and evaluating the prediction performance of a model in combination with evaluation indexes to obtain a prediction model; and inputting a to-be-predicted RNA sequence in a test set into the prediction model, predicting the m6A site and outputting the m6A site. The prediction accuracy of the model on the test set reaches 92.30% and 87.06% respectively, and the model has good development potential in the aspect of cross-species prediction and is expected to become a useful tool for identifying the m6A site.

The invention relates to the technical field of photovoltaic power generation, and in particular to a radiation attenuation-considered photovoltaic power short-term prediction method. The method comprises the following steps of: carrying out training by adoption of an indirect prediction method so as to obtain a sunny day surface radiation prediction model; obtaining an attenuation coefficient ofhistory daily surface radiation, establishing a surface radiation attenuation coefficient prediction model according to the attenuation coefficient, and establishing a cloud coverage coefficient prediction model; carrying out training by taking history real surface radiation, temperature and humidity as inputs of a meteorological factor and taking photovoltaic power as an output, so as to obtain aphotovoltaic power prediction model; and predicting photovoltaic power generation power by utilizing the photovoltaic power prediction model by taking a predicted value of the surface radiation as input of a surface radiation value and taking meteorological data of weather forecast as input of the meteorological factor. According to the method, the influence degree of cloud on the prediction precision can be reduced, the process of image analysis is saved and the algorithm is simpler and more efficient.

The invention discloses a knowledge tracking method and device based on a multi-head attention mechanism long-term and short-term memory network and a storage medium. A knowledge tracking model basedon a multi-head attention mechanism long-term and short-term memory network is constructed and used for knowledge tracking, and the model has better prediction performance; wherein the multi-head attention mechanism can capture more dependency relationships, including long-distance dependency relationships, among the input sequence data, so that the internal structure of the input sequence data can be obtained; in the aspect of calculation, attention calculation is carried out in parallel, calculation at the previous moment is not depended on, and the calculation speed is higher; the input sequence data is processed in parallel by using the long-short-term memory network, the information of the input sequence data can be obtained, a multi-head attention mechanism is combined with the long-short-term memory network, better prediction can be provided, and intelligent tutoring, personalized homework arrangement, learning plan generation, evaluation report generation and the like can be performed by using knowledge tracking. The method is widely applied to the field of knowledge tracking.

The invention relates to an applications-oriented control device of speed-regulating clutch and a control method thereof. The device comprises a sensor unit and an electronic speed-regulating controller electrically connected with the sensor unit. The electronic speed-regulating controller comprises a signal collecting module electrically connected with the output end of the sensor unit; a mode switching signal collecting module connected with a plurality of running modes; a CPU control module electrically connected with the output ends of the signal collecting module and the mode switching signal collecting module; a ratio servo valve base point regulating circuit electrically connected to the output end of the CPU control module; and a V-I converting and current amplifying circuit electrically connected with the output end accumulating the output end of the CPU control module and the output signal of the ratio servo valve base point regulating circuit. The output end of the V-I converting and current amplifying circuit drives a ratio servo valve and controls the actions of a clutch. By using a PID control policy, the invention has the advantages of adaptation to dead zone characteristics of different ratio servo valves and wide application range.

The invention discloses an SDN flow prediction method based on RBF neural network, belonging to the technical field of wireless communication network. The SDN traffic prediction algorithm based on theRBF neural network provided by the invention has excellent non-linear characteristics, is particularly suitable for processing highly non-linear system, and is trained by studying historical data records, so that a properly trained neural network has the ability of inducing all data. Secondly, the RBF neural network has a flexible and effective way of learning. Compared with other neural networks, the structure of RBF neural network is simpler and the learning speed is faster. Therefore, the RBF neural network can predict the complex and changeable network traffic more accurately. The invention utilizes POX and Mininet to simulate the proposed algorithm, and the simulation results show that the proposed algorithm can accurately predict the SDN traffic change trend, and has better prediction performance and lower prediction error.

The invention belongs to the technical field of reservoir parameter prediction, and particularly relates to a porosity prediction method based on a multilayer long-term and short-term memory neural network model. A neural network model based on multi-layer long short-term memory comprises an input layer configured to represent original logging parameters of porosity; the hidden layer is formed bysuperposing a plurality of long short-term memory (LSTM) models; and the output layer is used for outputting a predicted value of the output porosity through the full connection layer from the last hidden layer. According to the invention, the LSTM is superposed; a plurality of LSTM models are used for framework prediction; the output of the front layer is used as the input of the LSTM model of the rear layer; by using the deep LSTM model, long-term information can be memorized continuously, time information can be screened more strictly, prediction can be performed well at the inflection point of porosity change, and particularly, a high-precision target parameter prediction value can be obtained under the conditions of few well positions and few well logging parameter dimensions.

The invention discloses an XGBoost prediction method of an intelligent parameter optimization module. The XGBoost prediction method belongs to the technical field of model parameter optimization and machine learning prediction, and comprises the steps of selecting a parameter group used by an XGBoost model, constructing an XGBoost model parameter group optimization module based on a genetic method, selecting a data sample set, taking 90% of samples in the data sample set as a training set, learning, training and verifying a proposed XGBoost prediction model with a genetic optimization parameter module by adopting the sample set, and comparing results. The XGBoost prediction method has the advantages that the problem that an appropriate parameter group cannot be found by a prediction modelof a data set of a large sample is solved, and a better prediction result can be obtained compared with an XGBoost model for adjusting parameters based on experience. The classification prediction result for a liver disease data set shows the effectiveness of the XGBoost prediction method.

The invention provides an EWT-PDBN combination-based short-term wind power prediction method. The method comprises the following steps of A, collecting numerical weather forecast data and historical wind power data of a wind power plant; b, performing preprocessing and normalization processing of all the acquired data; c, decomposing the normalized historical average wind power data by using an empirical wavelet transform signal decomposition technology; d, performing correlation screening of the decomposed different intrinsic mode component function sub-sequences, respectively taking the screened group sub-sequences and other data subjected to normalization processing as input data, and inputting the input data into a particle swarm optimization deep belief network model for prediction toobtain group prediction data; and E, superposing a group of prediction data to reconstruct a group of data, and then performing reverse normalization processing of the group of data to obtain the result as the final wind power prediction result. The method is advantaged in that through EWT-PDBN combined prediction, the wind power prediction result with high precision and small error is obtained.

The invention discloses a device for generating power and desalinating seawater by using tidal energy, which is characterized in that tidal energy seawater and a tidal energy sea dam are fixedly connected with a U-shaped louver-type water inlet valve dam, the U-shaped louver-type water inlet valve dam and a U-shaped louver-type water drain valve dam form a closed buoyancy pool, a piston vessel is arranged in the center of the closed buoyancy pool and is movably connected with the closed buoyancy pool, the piston vessel is movably connected with a pool cylinder sleeve through a piston and a piston ring, the bottom of the closed buoyancy pool with the pool cylinder sleeve inside is fixedly connected with the pool cylinder sleeve through a pool cylinder sleeve water inlet valve and a pool cylinder sleeve water inlet and drain passage, the pool cylinder sleeve is fixedly connected with a high-pressure water conservation reservoir through a high-pressure water inlet valve and a high-pressure water inlet passage, the high-pressure water conservation reservoir is fixedly connected with a power generator set or a seawater desalinating device through a water diversion pipe, and the U-shaped louver-type water drain valve dam with a louver-type valve inside is connected with the sea. The device can convert tidal energy into power for desalinating seawater through buoyancy, gravity, pressure and hydraulic power, not only has a simple structure and is safe and reliable, but also is economical and practical, has long-term benefits and realizes sustainable development of renewable energy.