56 results about "PROJECTIONS PREDICTIONS" patented technology

An evidence based cost modeling and predictive analysis system, and an incentives based plan to reduce healthcare costs are disclosed. An analytics system may generate incremental expenditures among overweight and obese individuals, predictive forecasts of future medical costs, and predictive forecast of cost reduction based on financial incentives to recipients. The forecasts may include statistical trends, prevalence of diseases based on body mass index, and medical evidence associated with specific illnesses. A computer based program may process and analyze dependent and independent variables in electronically stored information (for example insurance, health and medical records). A health insurance provider may provide an annual rebate on paid premiums to recipients based on a qualifying annual BMI as an incentive. The recipients may receive the rebates in a qualified health reimbursement account (HRA) managed by the recipients towards future healthcare related expenditures.

An on-demand and real-time evidence based cost modeling and predictive analysis system, and a financial incentives based plan to reduce healthcare costs. An analytics system that includes a data aggregator and regression models generates incremental expenditures among overweight and obese individuals, predictive forecasts of future medical costs, and predictive forecasts of cost reduction based on financial incentives to recipients. The forecasts may include interactions, personalized variables, statistical trends, prevalence of diseases based on body mass index and / or age, and medical evidence associated with specific illnesses. A computer-based program may process and analyze variables in healthcare records. A health insurance provider may provide an annual rebate on paid premiums to recipients based on a qualifying annual BMI as an incentive. The recipients may receive the rebates in a qualified Healthcare Individual Reimbursement Account (HIRA) managed by the recipients towards future healthcare related expenditures.

Various embodiments of the invention provide methods, systems, and computer program products for generating an outbound forecast. In particular embodiments, campaign parameters are defined for at least two outbound call campaigns that include for each campaign a time period over which the campaign is to be carried out, a target value identifying a number of an outbound communication result to occur over the time period, and sources for handling outbound communications. Further, a historical dataset is selected that includes historical data on the corresponding outbound communication result for each campaign based on past outbound communications. In various embodiments, an outbound forecast is generated based on the historical dataset and campaign parameters for the outbound call campaigns. The forecast provides a number of the outbound communication result forecasted to be achieved over the time period for each campaign in an attempt to meet the target value for the corresponding campaign.

The invention discloses a prediction model construction method and device, a prediction method and device, equipment and a medium. The construction method comprises the steps of obtaining a sequence value of a historical power load of a prediction target; based on a preset value range of model parameters, determining the model parameters by utilizing the correlation characteristics of the sequencevalues; and constructing a prediction model based on the model parameters, with the prediction model being used for predicting the power load of the prediction target in the next time period. According to the embodiment of the invention, the obtained historical power load sequence value of the enterprise user is analyzed. , The model parameters are determined according to the model parameters, and then the prediction model is constructed by utilizing the determined model parameters, so that the electricity selling enterprise can accurately predict the medium and long-term power load of the enterprise in the future by utilizing the constructed prediction model, a reliable basis is provided for the electricity selling enterprise in an electricity selling quantity declaration service, and deviation assessment is reduced.

A data analysis and combination primary function nerve network-based wind power prediction method is disclosed and relates to the field of wind power prediction. The method comprises the following steps: collected wind power data is subjected to data analyzing operation via variable mode decomposition operation, sample entropy technologies and phase space reconstruction technologies; four groups of subsequences are obtained; primary function nerve networks are built via orthogonal polynomials, and a combined primary function prediction model containing four groups of primary function nerve networks is built; a state transition algorithm is used for optimizing weight and threshold values of the primary function nerve networks, reconstructed subsequences are used as input for the primary function nerve networks, and the optimized prediction model combination primary function nerve networks are used for predicting wind power. Prediction accuracy of the method is markedly higher than that of a BP network and that of an RBF nerve network.

The invention discloses a sales prediction method and system, and relates to the field of sales prediction, and the method comprises the steps: obtaining a predicted category and corresponding time information; obtaining corresponding feature data according to the predicted category and the corresponding time information; and inputting the feature data into a preset model combination correspondingto the predicted category, and predicting the sales volume of the predicted category in the time corresponding to the time information. According to the invention, the system is directly adopted to replace manpower to predict the sales volume, so that the capability requirement on purchasing personnel is reduced; and a computer system is adopted for prediction, so that the dimensions and the datavolume considered during prediction are greatly increased, and a foundation is laid for accurate prediction. In addition, the sales volume is predicted by using a multi-model combination, and the precision is greatly improved compared with prediction by using only one model. And each category has a corresponding preset model combination, so that the matching degree is better, and the prediction result of each category is further improved.

The invention discloses a prediction model training method and device, a prediction method and device, equipment and a medium. The training method comprises the steps of obtaining a sequence value ofa historical power load and a plurality of influence factors corresponding to the sequence value; extracting power load characteristics and influence factor characteristics from the sequence values and the influence factors; and training a prediction model based on a CatBoost algorithm by using the power load characteristics and the influence factor characteristics, with the prediction model beingused for predicting the power load of the prediction target in the next time period. According to the embodiment of the invention, the power load characteristics and the influence factor characteristics are extracted from the obtained historical sequence values and the corresponding influence factors; and the extracted power load features and influence factor features are trained to obtain a prediction model, so that feature processing in the training process is simple, the prediction model obtained by training is accurate in short-term power load prediction of enterprise users, and a reliable basis is provided for power spot market transaction.

The invention discloses a red tide grade prediction method. The method comprises: prediction algorithm optimization, prediction model construction and prediction result analysis. According to optimization of a prediction algorithm, a C4.5 decision tree classification contribution optimal feature selection method is adopted, the problem that input parameters of the BP neural network are difficult to select is solved, and the problem that the number of nodes of a hidden layer of the BP neural network is difficult to determine is solved by adopting a binary segmentation algorithm. A prediction model construction comprises: constructing a red tide level prediction model by adopting the optimized BP neural network, training the model by using historical case data, and ending the training when the prediction error is within an allowable range or the network training reaches the maximum iteration times. And the prediction result is analyzed, the trained model is used to predict the red tide level, the root-mean-square error of the prediction result is smaller than the prediction result of the traditional BP neural network before optimization, and the prediction precision is higher. The method can provide a new solution for red tide grade prediction.

The invention discloses an intra-frame prediction method and device, an encoder and a storage device. The intra-frame method includes defining the reference lines at a first side, a second side, a third side and a fourth side of a current encoded block, wherein the first side and the second side are adjacent and in an encoding direction of the current encoded block, and the third side and the fourth side are adjacent and in an encoding reverse direction of the current encoded block; obtaining a projection prediction value corresponding to the compensation pixel in the current coding block in each angle mode on a reference line, wherein the projection prediction value comprises a first projection prediction value in an angle mode direction and a second projection prediction value in an angle mode reverse direction; and respectively carrying out weighted average on the first prediction value and the second prediction value of each compensation pixel to obtain an angle mode prediction value of each compensation pixel, the first prediction value being obtained by using the first projection prediction value, and the second prediction value being obtained by using the second projection prediction value. In this way, the spatial redundancy removal effect can be improved.

The invention relates to a sand table manufacturing method based on machine learning, and belongs to the technical field of sand table manufacturing. The problems that a sand table is long in manufacturing cycle, authenticity is lacked during displaying and an adjustment cannot be made in real time in the prior art are solved. The method includes the following steps of establishing a laser holographic projection prediction model and a geomorphic information prediction model to be trained, adjusting projection parameters and geomorphic information of a laser holographic projector in real time through the trained models, conducting sand table projection imaging through the laser holographic projector according to the projection parameter value, and conducting sand table geomorphic manufacturing through a mechanical arm according to the geomorphic information. The sand table manufacturing process is rapid and efficient, the geomorphic state can be truly displayed and simulated through laser holographic projection, the terrain is rapidly piled and adjusted through the mechanical arm, the optimal projection parameters and geomorphic information can be obtained in real time through the models trained through machine learning, the projection state and geomorphology are automatically corrected and adjusted in real time, and the manufactured sand table reaches the optimal projection size and effect.

The invention relates to a regional power load prediction method. The prediction method comprises the following steps: step 1, acquiring power load historical data of a corresponding acquisition area;step 2, establishing an RNN (Recurrent Neural Network) prediction model for power load prediction, and optimizing the RNN prediction model by using a hybrid locust optimization algorithm; and step 3,substituting the historical data into the RNN prediction model to obtain a power load prediction value of the acquisition area. The hybrid locust optimization algorithm is used to optimize the neuralnetwork model to predict the power load. Prediction precision is substantially improved. Problems of large power prediction model training samples and many network adjustable parameters established based on statistical data in a traditional scheme are solved, effective reference is provided for reasonable optimization of use of electric power resources, and electric energy transmission and supplyare balanced.

Disclosed is a machine learning system with date alignment features for improved demand forecasting for products and / or services. The system includes an appliance for more accurately aligning days and weeks between years, including adapting to holidays and special days, in order to ascertain the date in a previous year that most closely aligns with the date in the future for which the forecast is sought. The corresponding day in one or more previous years can then be computed and demand data associated therewith can be retrieved from data storage to be used in forecasting demand on the forecast date. The most closely aligned day from a previous year can be selected such that the aligned day is positioned appropriately within the calendar week and year and the aligned day falls within a week that is positioned appropriately within the calendar month (i.e., first week, last week or middle-month weeks).

Described is a system for learning and predicting key phrases. The system learns based on a dataset of historical forecasting questions, their associated time-series data for a quantity of interest, and associated keyword sets. The system learns the optimal policy of actions to take given the associated keyword sets and the optimal set of keywords which are predictive of the quantity of interest. Given a new forecasting question, the system extracts an initial keyword set from a new forecasting question, which are perturbed to generate an optimal predictive key-phrase set. Key-phrase time-series data are extracted for the optimal predictive key-phrase set, which are used to generate a forecast of future values for a value of interest. The forecast can be used for a variety of purposes, such as advertising online.

The embodiment of the invention provides a network attack prediction method and device, and relates to the technical field of network security. The network attack prediction method comprises the following steps: firstly, obtaining to-be-detected network attack behavior data; preprocessing the network attack behavior data to obtain preprocessed data; constructing an attack event sequence according to the preprocessed data; further performing prediction processing on the attack event sequence through a pre-constructed network attack prediction model to obtain a prediction result; and finally, determining an attacked target address and an attacked probability corresponding to the attacked target address according to the prediction result. Therefore, network attack behaviors can be predicted, the prediction accuracy is high, and the network security can be further maintained in time.

An information processing apparatus (2000) acquires input data (10). The information processing apparatus (2000) extracts a prediction rule (50) used for prediction related to the input data (10) from a usage rule set (60) by using a neural network (30). The usage rule set (60) includes a plurality of candidates for the prediction rule (50) used for prediction related to the input data (10). The prediction rule (50) is information in which condition data (52) representing a basis for prediction and conclusion data (54) representing a prediction related to the input data (10) are associated with each other. The prediction rule (50) used for prediction related to the input data (10) indicates the condition data (52) indicating a condition satisfied by the input data (10). The information processing apparatus (2000) outputs a prediction result (20), based on the conclusion data (54) indicating the extracted prediction rule (50).

The invention discloses a gridding-based power distribution network top-down load prediction information method. The method comprises the following steps: dividing L1 and L2 planning grids, establishing an L2 grid load prediction formula, establishing an L1 grid load prediction formula, establishing a variable measurement economic equation, constructing a scene model, predicting a variable, predicting a predicted value of a future power load, establishing an annual maximum load prediction curve and querying a grid load predicted value. According to the invention, L1 and L2 are used to plan a grid division mode; load prediction is carried out by using L1 and L2 grid regions; each area is predicted separately, the prediction result is clearer, assignment of various variables is introduced ina scene model mode, various factors such as area development data are considered in load prediction, the credibility and accuracy of the prediction result are demonstrated, the load prediction information is more reasonable, and the method has positive significance for future power grid development planning.

Embodiments predict / forecast demand of a product by receiving historical sales data for the product and, using a plurality of different seasonality estimation methods, estimating a plurality of different seasonality estimations for future time periods and determining an approximate error amount for each of the different seasonality estimations. Embodiments determine a weight for each of the plurality of different seasonality estimation methods based on the corresponding approximate error amount and generate an aggregate seasonality model based on the plurality of different seasonality estimations and the weights. Embodiments then determine a demand forecast using the aggregate seasonality model.

The invention discloses a PCA-LSTM bearing residual life prediction method based on multilayer grid search, and the method comprises the steps: firstly extracting a plurality of time-frequency domainfeatures of bearing fault time sequence data, employing PCA to fuse a plurality of feature index quantities, and removing the redundant data of the feature indexes; obtaining required influence faultprincipal component data, namely a group of new comprehensive index time series data, preprocessing the time series data, converting the time series data into equipment degradation degree value data,and inputting the equipment degradation degree value data into a constructed LSTM model to perform fault sequence prediction training; achieving optimal selection of LSTM model parameters with the minimum prediction loss as the target through a multi-layer grid search algorithm, so that an optimal time series data prediction model is obtained, and finally the remaining service life of the bearingis obtained through polynomial curve fitting calculation. Problems of low prediction precision and low prediction speed of bearing life prediction are solved, and the stability and accuracy of bearingresidual life prediction are improved.

The invention discloses an end-to-end multi-target identification, tracking and prediction method, and belongs to the technical field of Internet of Vehicles and intelligent automobiles. The method comprises the following steps: establishing an end-to-end multi-target identification, tracking and prediction model which comprises a target detector, a target tracking module and a trajectory prediction module; the target detection module uses a multi-target detector based on a central point; the target tracking module adopts a graph-based convolutional neural network to track multiple targets; the trajectory prediction module performs motion trajectory prediction on multiple targets based on a graph network, including prediction of a trajectory destination point, information transmission between intelligent agents and generation of a future trajectory; according to the method, end-to-end multi-target identification, tracking and prediction models are taken as a whole, and simultaneous training is carried out by adopting a joint training framework. The three modules are trained at the same time and promote each other, the final trajectory prediction precision is further improved, multi-target trajectory prediction can be better achieved, and the predicted trajectory is more reasonable.

The invention provides a mine water inflow prediction method and system based on coal mine big data, and the method comprises the steps of obtaining coal mine geographic position natural conditions and mining area information, and obtaining coal hydrogeological detection data; determining a main control factor of the mine water inflow according to the acquired information; building a mapping relation between the mine water inflow and the determined main control factors, building a water inflow prediction model, using a TRUST-TECH method to find a plurality of local optimal solutions so as to determine a global optimal solution, and obtaining a working face water inflow prediction result. According to the invention, the main control factors of mine water gushing are extracted, the TRUST-TECH technology is used for establishing the water gushing prediction model, the water gushing damage degree of the working face is predicted, the prediction process is simple, and the prediction result is accurate.

Systems and methods for determining a source of anomaly in a cyber-physical system (CPS) are disclosed. A forecasting tool can obtain a plurality of CPS feature values during an input window and forecast the plurality of CPS feature values for a forecast window. An anomaly identification tool can determine a total forecast error for the plurality of CPS features in the forecast window, identify ananomaly in the cyber-physical system when the total forecast error exceeds a total error threshold, and identify at least one CPS feature as the source of the anomaly.

The invention relates to a semantic analysis method and device based on machine learning, a medium and electronic equipment, and belongs to the technical field of machine learning application, and themethod comprises the steps: converting to-be-processed input information into pre-input information when the to-be-processed input information is received; inputting the pre-input information into apre-trained machine learning model to obtain a prediction semantic template corresponding to the to-be-processed input information; obtaining semantic template constraint information and a predictionsemantic template, inputting the semantic template constraint information and the prediction semantic template into the prediction semantic template constraint model, and outputting a constrained prediction semantic template; converting the input information to be processed into pre-analysis data according to the constrained semantic template; and according to the pre-analysis data, obtaining a semantic analysis result of the to-be-processed input information. On the basis of the preset machine learning model, the predicted semantic template is obtained through analysis according to various input information, so that the semantic analysis accuracy and efficiency are effectively guaranteed.

The invention discloses a medium and long term load prediction method, and the method comprises the steps: finding out an economic factor which has a long term equilibrium relation with electric quantity through a co-integration test in measurement economics based on a Granger causality test and an LSTM; determining economic factors beneficial to electric quantity prediction by using a Granger causality test method so as to reduce the number of input variables of the prediction model; and finally, inputting the economic factor data into the LSTM model for load prediction. A method of combining the Granger causal relationship test and the LSTM time sequence prediction model is introduced into a multivariable system, the medium-and-long-term load prediction model which is not easy to overfit and high in expandability is constructed, and the model is used for medium-and-long-term load prediction. And the predicted medium-and-long-term load has relatively high accuracy.

The invention provides a landslide displacement multilinear prediction method based on an ST-SEEP segmentation method and a space-time ARMA model. The landslide displacement multilinear prediction method comprises the steps of data preprocessing, curve segmentation, spatial weight matrix acquisition, modeling and prediction, and prediction effect evaluation. in data preprocessing step, reading landslide displacement data and coordinate data, and preprocessing the landslide displacement data and the coordinate data; drawing a landslide displacement-time curve in a curve segmentation mode, and providing an ST-SEEP method to conduct segmentation processing on the curve; in spatial weight matrix acquisition step, performing spatial clustering on the monitoring points by adopting a K-means clustering method, and acquiring a spatial weight matrix; modeling and predicting to establish a space-time ARMA model, and predicting a landslide displacement space-time sequence; and the prediction result evaluation adopting an absolute error and a root-mean-square error to evaluate the prediction result. The method has the beneficial effects that quantitative analysis of the spatial relationship of the monitoring points is realized, and the spatial relationship is more effectively utilized; the space-time autoregressive moving average statistical model is introduced into the landslide prediction field, the physical significance of formulas and parameters is clear, the process is clear, and the landslide displacement can be accurately predicted.

The embodiment of the invention discloses a method and device for predicting a destination based on a position trajectory. The method comprises the steps of starting to predict the destination to be reached by a target user according to the real-time position of the target user after the change of the position of the target user is monitored. In the prediction process, different prediction algorithms are selected for prediction according to the number of the target position signaling corresponding to the target user, and when the number of the target position signaling is insufficient, the destination of the target user is predicted by referring to public position signaling, so that inaccurate prediction caused by the insufficient target position signaling is avoided. On the other hand, compared with a method for predicting the destination of the user by adopting a unified prediction method, the method has the advantages that the calculation process of prediction is simplified, the destination to be reached by the user can be predicted according to the real-time position of the user, and the timeliness of prediction is fully considered.

The invention discloses an inorganic arsenic content prediction method and device based on XGBoost and a medium. The method comprises the following steps: acquiring original data; preprocessing the original data to generate modeling data; according to the modeling data, an XGBoost model is trained, and a prediction model is formed; and inputting process condition information and raw material information of a to-be-detected object into the prediction model, and generating a predicted value of the inorganic arsenic content of the to-be-detected object. The XGBoost model is used for predicting the inorganic arsenic content of the to-be-detected object, the prediction precision is high, in addition, the XGBoost model has the unique advantages of short training time and the like, the prediction efficiency of the method can be further improved, the safety of the related to-be-detected object is further improved, and the method can be widely applied to the technical field of data processing.

The invention discloses a turbine backpressure trend prediction method based on a catboost algorithm, and relates to the field of machine learning, and the method comprises the following steps: S1, preparing historical operation data, S2, carrying out the preprocessing of the historical operation data, S3, carrying out the feature engineering dimension raising of the preprocessing data, S4, carrying out the re-sampling and screening of variables having a large correlation degree with the backpressure, S5, determining parameters, carrying out the modeling through catboost, and carrying out the prediction of the backpressure trend of a turbine. The screened historical operation data predicts the back pressure in a certain period of time, and the model trained in the step S6 is used for predicting and replacing with the model trained offline when the error of the current training preparation model and the current prediction model is too large at the same time; the method is indirect and easy to operate, can effectively predict the back pressure of the steam turbine, enables the prediction model and offline training to be parallel, can make up the error of the prediction model in time, and improves the prediction precision.

The embodiment of the invention discloses a load comprehensive prediction method, which comprises the following steps of: analyzing and counting average power load data at the same time point in short-term time and meteorological data corresponding to each time point, and determining meteorological data parameters influencing the change of the power load data; establishing a short-term fractionalBrownian motion model between the power load data and the meteorological data in a short period of time; establishing a medium and long term load prediction model between the power load data and the economic policy data; optimizing the short-term fractional Brownian motion model and the medium-and-long-term load prediction model to obtain a fractional Brownian motion optimization model about the power load within a short medium-and-long term limit; predicting short-term, medium-term and long-term power load data by using a fractional Brownian motion optimization model; according to the scheme,short-term and medium-and-long-term prediction is carried out on the power load, comprehensive prediction on the power load is realized, and the accuracy of prediction data is high.