3517 results about "Linear regression" patented technology

A method of calibrating glucose monitor data includes collecting the glucose monitor data over a period of time at predetermined intervals. It also includes obtaining at least two reference glucose values from a reference source that temporally correspond with the glucose monitor data obtained at the predetermined intervals. Also included is calculating the calibration characteristics using the reference glucose values and corresponding glucose monitor data to regress the obtained glucose monitor data. And, calibrating the obtained glucose monitor data using the calibration characteristics is included. In preferred embodiments, the reference source is a blood glucose meter, and the at least two reference glucose values are obtained from blood tests. In additional embodiments, calculation of the calibration characteristics includes linear regression and, in particular embodiments, least squares linear regression. Alternatively, calculation of the calibration characteristics includes non-linear regression. Data integrity may be verified and the data may be filtered.

A method of calibrating glucose monitor data includes collecting the glucose monitor data over a period of time at predetermined intervals, obtaining reference glucose values from a reference source that temporally correspond with the glucose monitor data obtained at the predetermined intervals, calculating the calibration characteristics using the reference glucose values and corresponding glucose monitor data to regress the obtained glucose monitor data, and calibrating the obtained glucose monitor data using the calibration characteristics. In additional embodiments, calculation of the calibration characteristics includes linear regression and, in particular embodiments, least squares linear regression. Alternatively, calculation of the calibration characteristics includes non-linear regression. Data integrity may be verified and the data may be filtered. Further, calibration techniques may be modified during a fast rate of change in the patient's blood glucose level to increase sensor accuracy.

A device and method for measuring the concentration of analytes in the blood of a portion of tissue. The device includes a sensor module, a monitor, and a processor (separate from or integral with the sensor module). The sensor module includes a radiation source for emitting radiation to the tissue; a collimator and narrow band filter for processing the radiation after it has transmitted through or been reflected by the tissue; and one or more sensors for sensing the transmitted or reflected radiation. The one or more sensors send a signal to the processor which algorithmically converts the radiation using linear regression or orthogonal functions to determine the concentration of one or more blood analytes. The device self-calibrates to eliminate error caused by variables such as skin character. The sensor module is integrated to reduce size and weight such that it is inobtrusive, and the monitor is compact for transport

Methods and apparatus are provided for determining a diabetic patient's carbohydrate to insulin ratio (CIR), carbohydrate to blood glucose ratio (CGR), and insulin sensitivity factor (ISF) using the patient's record of blood glucose readings, carbohydrate consumption and insulin doses. The method provides the sensitivity factors that best account for the patient's observed blood glucose changes by linear regression of appropriately transformed variables. An apparatus that can collect and store the blood glucose readings, insulin dosages, and carbohydrate intake data and process these data according to this invention can generate statistically characterized sensitivity factors to advise the diabetic patient on optimal bolus insulin dosages.

Crop yield may be assessed and predicted using a piecewise linear regression method with break point and various weather and agricultural parameters, such as NDVI, surface parameters (soil moisture and surface temperature) and rainfall data. These parameters may help aid in estimating and predicting crop conditions. The overall crop production environment can include inherent sources of heterogeneity and their nonlinear behavior. A non-linear multivariate optimization method may be used to derive an empirical crop yield prediction equation. Quasi-Newton method may be used in optimization for minimizing inconsistencies and errors in yield prediction. Minimization of least square loss function through iterative convergence of pre-defined empirical equation can be based on piecewise linear regression method with break point. This non-linear method can achieve acceptable lower residual values with predicted values very close to the observed values. The present invention can be modified and tailored for different crops worldwide.

In a method of recalibrating continuous glucose monitoring data from a user, operable on a digital processor, an indication from the user that the user has taken a meal is received (806). A self-monitored of blood glucose levels from the user (810) at two separate times during a day corresponding to when the user has taken a meal. A glucose signal is received from a continuous glucose monitoring sensor (818) at times corresponding to the two separate times that the user has taken a meal. Two reconstructed blood glucose values based on the glucose signal from the continuous monitoring sensor at times when the at least two self-monitored of blood glucose levels are received from the user. A linear regression is performed (822) using y=ax+b, wherein x corresponds to the two reconstructed blood glucose values and y corresponds to the two self-monitored of blood glucose levels thereby generating an estimation of a and b. A recalibration signal, including the estimation of a and b, is transmitted to the continuous glucose monitoring sensor (824) based on the linear regression.

A method for calibrating non-invasive breathing monitors with sensors placed on the rib-cage and abdomen of a subject includes determining an initial scaling factor and an optimal multiplicative factor for the readings of the rib-cage and abdomen sensors using one of a least squares, linear regression, or multi-linear regression techniques. A current scaling factor is determined on a periodic basis using qualitative device calibration techniques. The current scaling factor is used to monitor breathing and diagnose obstructive apneas. Furthermore, the optimal multiplicative and the current scaling factor are used to determine the current tidal volume.

A system and method for calibrating a sensor of a characteristic monitoring system in real time utilizes a self-calibration module for periodic determination of, and compensation for, the IR drop across unwanted resistances in a cell. A current-interrupt switch is used to open the self-calibration module circuit and either measure the IR drop using a high-frequency (MHz) ADC module, or estimate it through linear regression of acquired samples of the voltage across the sensor's working and reference electrodes (Vmeasured) over time. The IR drop is then subtracted from the closed-circuit value of Vmeasured to calculate the overpotential that exists in the cell (Vimportant). Vimportant may be further optimized by subtracting the value of the open-circuit voltage (Voc) across the sensor's working and reference electrodes. The values of Vmeasured and Vimportant are then controlled by respective first and second control units to compensate for the IR drop.

Disclosed is a method of measuring a parameter of a litho-graphic process, and associated inspection apparatus. The method comprises measuring at least two target structures on a substrate using a plurality of different illumination conditions, the target structures having deliberate overlay biases; to obtain for each target structure an asymmetry measurement representing an overall asymmetry that includes contributions due to (i) the deliberate overlay biases, (ii) an overlay error during forming of the target structure and (iii) any feature asymmetry. A regression analysis is performed on the asymmetry measurement data by fitting a linear regression model to a planar representation of asymmetry measurements for one target structure against asymmetry measurements for another target structure, the linear regression model not necessarily being fitted through an origin of the planar representation. The overlay error can then be determined from a gradient described by the linear regression model.

Methods and systems are disclosed for learning a regression decision graph model using a Bayesian model selection approach. In a disclosed aspect, the model structure and / or model parameters can be learned using a greedy search algorithm applied to grow the model so long as the model improves. This approach enables construction of a decision graph having a model structure that includes a plurality of leaves, at least one of which includes a non-trivial linear regression. The resulting model thus can be employed for forecasting, such as for time series data, which can include single or multi-step forecasting.

A method of performing a formation rate analysis from pressure and formation flow rate data. Pressure and flow rate data are measured as fluid is withdrawn from a formation. Variable system volume is accounted for. The pressure and flow rate data are correlated using a multiple linear regression technique. Time derivative terms related to pressure and flow rate are smoothed using a summation technique, thereby providing better correlations than using the time derivatives directly. Formation parameters comprising formation permeability, formation pressure, and fluid compressibility may be determined from the correlation.

New and efficient protocols are provided for privacy-preserving machine learning training (e.g., for linear regression, logistic regression and neural network using the stochastic gradient descent method). A protocols can use the two-server model, where data owners distribute their private data among two non-colluding servers, which train various models on the joint data using secure two-party computation (2PC). New techniques support secure arithmetic operations on shared decimal numbers, and propose MPC-friendly alternatives to non-linear functions, such as sigmoid and softmax.

A device and method for measuring the concentration of analytes in the blood of a portion of tissue. The device includes a sensor module, a monitor, and a processor (separate from or integral with the sensor module). The sensor module includes a radiation source for emitting radiation to the tissue; a collimator and narrow band filter for processing the radiation after it has transmitted through or been reflected by the tissue; and one or more sensors for sensing the transmitted or reflected radiation. The one or more sensors send a signal to the processor which algorithmically converts the radiation using linear regression or orthogonal functions to determine the concentration of one or more blood analytes. The device self-calibrates to eliminate error caused by variables such as skin character. The sensor module is integrated to reduce size and weight such that it is inobtrusive, and the monitor is compact for transport.

A compilation time estimator provides a quantified estimate of the optimizer compilation time for a given query optimizer. The estimator automates the optimizer to choose the right level of optimization in commercial database systems. The estimator reuses an optimizer's join enumerator to obtain actual number of joins, but bypasses plan generation to save estimation overhead, and maintains a small number of interesting physical properties to estimate the number of plans by using a linear regression model. The estimator uses the number of generated plans to estimate query compilation time.

The invention discloses a sensor-fault diagnosing method based on the online prediction of a least-squares support-vector machine. In the method, a least-squares support-vector machine online-predicting model is established, and then the measured data of a sensor is acquired on line and used as an input sample of the least-squares support-vector machine online-predicting model to realize that the output value of the sensor at the next moment is predicted in real time as the predicting model is trained on line. Whether sensor faults occur or not is detected by comparing residual errors generated by the predicting value and the actual output value of the sensor. When the faults occur, the unary linear regression for a residual-error sequence is carried out by a least-squares method to realize the identification of the deviation and drift faults of the sensor, and furthermore, measures can be more effectively taken to carry out real-time compensation for the output of the sensor. Through the sensor-fault diagnosing method, the online fault diagnosis of the sensor can be rapidly and accurately realized, and the sensor-fault diagnosing method is particularly applicable to diagnosing the deviation faults and the drift faults of the sensor.

The invention discloses a convolutional neural network structure-based traffic flow prediction method. The method comprises the following steps of 1) establishing a traffic flow data set and preprocessing the data set: establishing the traffic flow data set according to obtained traffic flow data, preprocessing the data set, constructing a data set sample matrix, and dividing the data set into a training set and a test set; 2) establishing a single-layer conventional convolutional neural network, removing a pooling layer, constructing a feature extraction network of a road traffic flow matrix,adding a sigmoid nonlinear regression layer to a full connection layer, and constructing a road traffic flow nonlinear regression prediction network; and 3) training the convolutional neural networkand realizing real-time prediction of short-term traffic flow: defining a model objective function, taking the training set as an input of a convolutional neural network model, solving an optimal parameter of the model to finish model training, and performing real-time traffic flow prediction on the test set by utilizing the trained convolutional neural network model. The short-term prediction accuracy of the traffic flow is improved.

Machine executable method of analyzing growth curve data to identify the transition from a baseline phase into a growth phase. Applications of the method include analysis of results from time-dependent monitoring of amplicon synthesis in a nucleic acid amplification reaction to quantify a starting amount of a nucleic acid template in a test sample. The method advantageously simplifies the quantitation by circumventing the need to establish thresholds used for calculating initiation of the growth phase, to calculate derivatives, or to perform linear regression analysis.

A rate control system is disclosed for video coding applications. The rate controller assigns a quantization parameter for video data in a picture in response to complexity indicators indicative of spatial complexity, motion complexity and / or bits per pel of the picture. A virtual buffer based quantizer parameter is proposed based on a virtual buffer fullness analysis and a target rate estimate, which is derived from the complexity indicators. A second quantizer parameter is proposed from a linear regression analysis of quantizer parameters used to code previously coded pictures of similar type (e.g., I pictures, P pictures or B pictures). A coding policy decision unit defines a final quantizer parameter from a comparison of the two proposed quantizer parameters.

A robust method for determining the boundaries of cells and the associated reliability of the RF coverage within these boundaries is presented. The invention accurately determines the average range from the base station to the cell edge from RF signal strength measurements with a linear regression approach. The accuracy of this estimate is quantified both as a range uncertainty (e.g. + / -100 meters) and as a cell coverage reliability (i.e. area / edge) through 1) simulation, 2) analysis of real data, and 3) theoretical analysis. It is shown that if the estimate of the cell radius meets the desired accuracy, then the corresponding estimates of coverage reliability (both area and edge) are more than sufficiently accurate. It is recommended that radio survey analyses incorporate this test as part of the coverage validation process.

A system and method for measuring or quantifying the probability of default of a borrower. Credit factors from companies that banks have extended loans to are inputted and collected into a processor. The method employs a process utilizing an optimization function and a standard multivariate nonlinear regression to process client information and to provide an output value whose value is indicative of the likelihood or risk of default by a particular borrower.

The aesthetic quality of a picture is automatically inferred using visual content as a machine learning problem using, for example, a peer-rated, on-line photo sharing Website as data source. Certain visual features of images are extracted based on the intuition that they can discriminate between aesthetically pleasing and displeasing images. A one-dimensional support vector machine is used to identify features that have noticeable correlation with the community-based aesthetics ratings. Automated classifiers are constructed using the support vector machines and classification trees, with a simple feature selection heuristic being applied to eliminate irrelevant features. Linear regression on polynomial terms of the features is also applied to infer numerical aesthetics ratings.

The invention discloses a network device performance monitoring method and a network management system. The network device performance monitoring method includes: the network management system acquires key performance indicator (KPI) data of a network device periodically; the acquired KPI data are summarized according to set time windows to obtain KPI data summarizing values of the time windows; a linear regression algorithm model is utilized, KPI data of future N time windows including the current time window are predicted according to KPI data summarizing values of N time windows before the current window; whether the KPI data of the further N time windows exceed a set alarm threshold value is judged; and a alarm is given out during yes judgment. The network device performance monitoring method and the network management system can achieve early warning according to KPI historical data.

The invention relates to a battery energy storage system peak clipping and valley filling real-time control method based on load prediction and belongs to the field of power system automatic control. The control method provided by the invention comprises the following steps of: firstly searching similar history daily load data, carrying out expanding short-term load prediction by adopting a linear regression analysis method, building a battery energy storage system peak clipping and valley filling real-time optimization model, solving the battery energy storage system peak clipping and valley filling real-time optimization model by adopting a dynamic programming algorithm, and obtaining the output power of a battery energy storage system at each moment. The control method provided by the invention comprises battery charging and discharging frequency constraint and discharge depth constraint in the real-time optimization model, is used for researching relation between battery life and the charging and discharging frequency and the relation between the battery life and the discharge depth and is beneficial to prolonging the battery life. Minimum load variance is taken as a target function, the peak-to-valley of a load curve can be reduced, the load curve is smoother while constraint conditions are met, and the peak clipping and valley filling application requirement can be met. Local part of the load curve can be smoother by adopting load smoothness constraint.

A time-resolved, fluorescence spectrometer makes use of a RadiaLight® optical switch and no dispersive optical elements (DOE) like gratings. The structure is unique in its compactness and simplicity of operation. In one embodiment, the spectrometer makes use of only one photo-detector and an efficient linear regression algorithm. The structure offers a time resolution, for multiple species measurements, of less than 1 s. The structure can also be used to perform fluorescence correlation spectroscopy and fluorescence cross-correlation spectroscopy.

A method and system for identifying users for advertising. Users, attributes, and first web sites provided by ISPs accessed by the users are identified. First data including content of the first web sites and user time spent thereon are received from ISPs and analyzed to determine first attribute values indicative of user interest. Second data received from ISPs include content of second web sites and user time spent thereon. Second attribute values, derived from questionnaires completed by the users, indicate interest in each attribute by each user. Third attribute values are determined by combining the first and second attribute values. The third attribute values are processed to identify users to which a product or service may be advertised. The identified users are communicated to a service provider or product provider. The first, second, or third attribute values may have missing values, which are determined by correlation and linear regression.

The invention discloses a short-term and medium- and long-term electric power load prediction method based on machine learning model. Firstly, preprocessing is conducted on data, including smootheningabnormal data and filling missing data. Factors of affecting load changes will be analyzed, including historical data, time periodicity, and weather variable characteristics. Domestication will be conducted on all input variables for accelerating learning speed and raising prediction precision. The invention is advantageous in that linear regression is compared, and the performance of the vectorregression and gradient lifting regression in the short-term and medium- and long-term electric power load prediction is supported; with the prolongation of the prediction time, the performance of thegradient lifting regression model is better that that of the other two models; the AdaBoost algorithm which uses the gradient lifting tree as a basic classifier is brought forward, and load prediction is conducted, and the precision of electric power load prediction can be effectively raised.

A system and method for calibrating a sensor of a characteristic monitoring system in real time utilizes a self-calibration module for periodic determination of, and compensation for, the IR drop across unwanted resistances in a cell. A current-interrupt switch is used to open the self-calibration module circuit and either measure the IR drop using a high-frequency (MHz) ADC module, or estimate it through linear regression of acquired samples of the voltage across the sensor's working and reference electrodes (Vmeasured) over time. The IR drop is then subtracted from the closed-circuit value of Vmeasured to calculate the overpotential that exists in the cell (Vimportant). Vimportant may be further optimized by subtracting the value of the open-circuit voltage (Voc) across the sensor's working and reference electrodes. The values of Vmeasured and Vimportant are then controlled by respective first and second control units to compensate for the IR drop.

A fully automatic, computationally efficient segmentation method of video employing sequential clustering of sparse image features. Both edge and corner features of a video scene are employed to capture an outline of foreground objects and the feature clustering is built on motion models which work on any type of object and moving / static camera in which two motion layers are assumed due to camera and / or foreground and the depth difference between the foreground and background. Sequential linear regression is applied to the sequences and the instantaneous replacements of image features in order to compute affine motion parameters for foreground and background layers and consider temporal smoothness simultaneously. The Foreground layer is then extracted based upon sparse feature clustering which is time efficient and refined incrementally using Kalman filtering.