Abstract: Systems and methods are provided for producing a mining model accuracy display that depicts the model's accuracy at predicting a state for a multiple-state variable. The model predicts a state and provides an associated probability for each case. Points are graphed such that one coordinate of the data point corresponds to a number N of cases and the other coordinate corresponds to the number of correct predictions made in the top N cases by probability.

Abstract: Disclosed is a technique for obtaining an estimate and variance of each variable based on a constraint manifold. Particles (or samples) are sampled in order to filter and fuse ambiguous data or information on at least one state variable of a system using the particles. The sampling is carried out in consideration of an influence which non-linearity of the constraint manifold of a system model, an observation model or another system model exerts on a probability distribution of the state variable. With this construction, it is possible to reduce decrease of fusion and filtering performance, decrease a Gaussian approximation error, and detect mismatched information.

Abstract: A facial expression recognition system that uses a face detection apparatus realizing efficient learning and high-speed detection processing based on ensemble learning when detecting an area representing a detection target and that is robust against shifts of face position included in images and capable of highly accurate expression recognition, and a learning method for the system, are provided. When learning data to be used by the face detection apparatus by Adaboost, processing to select high-performance weak hypotheses from all weak hypotheses, then generate new weak hypotheses from these high-performance weak hypotheses on the basis of statistical characteristics, and select one weak hypothesis having the highest discrimination performance from these weak hypotheses, is repeated to sequentially generate a weak hypothesis, and a final hypothesis is thus acquired.

Abstract: A computer implemented method and system for assessing the risk of an anthropogenic event are disclosed. An anthropogenic event is defined and recursively decomposed into causal conditions with ascertainable probabilities. The causal conditions and probabilities are transformed into a comprehensive probability of the event. The risk related to the occurrence of the event, such as a monetary loss, is computed based on its derived probability.

Abstract: A method for monitoring a system includes receiving a set of training data. A Gaussian mixture model is defined to model a probability distribution for a particular sensor of the system from among a plurality of sensors of the system based on the received training data. The Gaussian mixture model includes a sum of k mixture components, where k is a positive integer. Sensor data is received from the plurality of sensors of the system. An expectation-maximization technique is performed to estimate an expected value for the particular sensor based on the defined Gaussian mixture model and the received sensor data from the plurality of sensors.

Abstract: Representative embodiments are directed to systems and methods for analyzing the selection of measurements of events associated with a communications network. In one embodiment, the selection of measurements includes creating a representation of a network. After the creation of the network representation, multiple sets of contemplated measurements associated with the network representation are created. A metric is calculated that is related to the respective diagnostic effectiveness for each set of contemplated measurements when the respective set of contemplated measurements is provided to a network diagnostic algorithm.

Abstract: A technique for designing and testing drive-test plan for gathering location-dependent RF data is disclosed. In accordance with some embodiments of the present invention, one candidate drive-test plan is chosen for implementation over a second based on an economic cost-benefit analysis of both plans. This is in marked contrast to, for example, a selection of drive-test plans, or the design of a drive-test plan, based on a calibration-cost analysis, in which the data estimated to be the most effective to calibrate a radio-frequency tool is sought for a given cost or the least cost. Although a data-estimated-to-be-most-effective-to-calibrate-a-radio-frequency-tool vs. cost analysis is a species of cost-benefit analyses in general, it is not an economic cost-benefit analysis because a data-estimated-to-be-most-effective-to-calibrate-a-radio-frequency-tool vs. cost analysis has deficiencies that an economic cost-benefit analysis does not.

Abstract: Embodiments of the present invention provides a system that optimizes a regression model which predicts a signal as a function of a set of available signals. These embodiments use a genetic technique to optimize the regression model, which involves using a portion of the sample signals used to generate each parent regression model from a pair of best-fit parent regression models to generate a child regression model. In addition, in embodiments of the present invention, the system introduces “mutations” to the set of sample signals used to create the child regression model in an attempt to create more robust child regression models during the optimization process.

Abstract: A method, apparatus and software are disclosed for simulating a running median value for a set of input data values, and for simulating a selected percentile value other than the median. A current estimated value is compared to an input data value. If the input data value is greater than the current estimated value, the current estimated value is increased in accordance with a first adjustment function. If the input data value is less than the current estimated value, the current estimated value is decreased in accordance with the second adjustment function. Different adjustment functions are disclosed for increasing or decreasing the current estimated value. The magnitude of each increase or decreases provided by an adjustment function is independent of the magnitude of the difference between the current input data value and the current estimated value.

Abstract: A system and method is disclosed for assessing a probability of failure of operation of a semiconductor wafer. The method includes inputting risk factor data into a memory and inputting a plurality of wafers into a semiconductor fabrication manufacturing process. A subset of wafers is selected to obtain a sample population and at least one region of each wafer of the sample population is inspected. Circuit design data associated with each wafer of the sample population is obtained and one or more defects that present an increased risk to the operation of a particular wafer are identified. The identification is a function of the risk factor data, the inspecting step and the circuit design data. A probability of semiconductor wafer failure is calculated.

Abstract: Data slices of historical time series are leveraged to facilitate in more accurately predicting like data slices of future time series. Different predictive models are employed to detect outliers in different data slices to enhance the accuracy of the predictions. The data slices can be temporal and/or non-temporal attributes of a data set represented by the historical time series. In this manner, for example, a historical time series for a network location can be sliced temporally into one hour time periods as a function of a day, a week, a month, a year, etc. Outliers detected in these data slices can then be mitigated utilizing the predictive time series model by replacing the outlier with the expected value. The mitigated historical time series can then be employed in a predictive model to predict future web traffic for the network location (and advertising revenue values) with a substantial increase in accuracy.

Abstract: The subject invention provides for systems and methods that facilitate optimizing one or mores sets of training data by utilizing an Exponential distribution as the prior on one or more parameters in connection with a maximum entropy (maxent) model to mitigate overfitting. Maxent is also known as logistic regression. More specifically, the systems and methods can facilitate optimizing probabilities that are assigned to the training data for later use in machine learning processes, for example. In practice, training data can be assigned their respective weights and then a probability distribution can be assigned to those weights.

Abstract: A computer implemented method and apparatus for determining the source of at least one observed event occurring among a plurality of types of components is disclosed. The method, operable on a computer system, utilizes a data structure produced for determining the source of a problem by mapping a plurality of observed events and a plurality of known causing events, wherein the mapping represents a correlation between the observed events and the causing events, determines a distance measure based on mapped correlation values and determines at least one causing event based on a distance measure within a known tolerance above a minimum distance measure.

Abstract: Degree of outlier of one input data is calculated by an amount of change in a learned probability density from that before learning as a result of taking in of the input data. This is because data largely differing in a tendency from a so far learned probability density function can be considered to have a high degree of outlier. More specifically, a function of a distance between probability densities before and after data input is calculated as a degree of outlier. Accordingly, a probability density estimation device appropriately estimates a probability distribution of generation of unfair data while sequentially reading a large volume of data and a score calculation device calculates and outputs a degree of outlier of each data based on the estimated probability distribution.

Abstract: An expert system platform supporting multiple and customizable taxonomies is useable in a number of applications. In one embodiment, the expert system platform includes a plurality of libraries and a plurality of study tables with each being related to a corresponding library. Each of the plurality of libraries corresponds to one of the number of applications and has a question table having a plurality of questions and an answer table having a plurality of answers, each of the plurality of answers is related to one of the plurality of questions, wherein a user of the expert system qualifies for one or more studies or clinical trials based on responses provided by the user. In another embodiment, the expert platform system is useable to identify an individual's potential exposure to chemical, biological, and/or nuclear contamination in a rapidly changing environment.

Abstract: The present invention is directed to a method for information analysis used to discriminate between biased and unbiased fits. A goodness-of-fit parameter becomes a quantity of interest for analysis of multiple measurements and does so by replacing a likelihood function with a probability of the likelihood function W(L|?). The probability of the likelihood function is derived from a new posterior probability P(?|L) and the goodness-of-fit parameter G, wherein a is a set of fitting parameters, L is the likelihood function and W(L|?) is equal to the product of P(?|L) and G. Furthermore, if substantial prior information is available on ?, then the probability of the fitting parameters P(?) can is used to aid in the determination of W(L|?).

Abstract: A method to select features for maximum entropy modeling in which the gains for all candidate features are determined during an initialization stage and gains for only top-ranked features are determined during each feature selection stage. The candidate features are ranked in an ordered list based on the determined gains, a top-ranked feature in the ordered list with a highest gain is selected, and the model is adjusted using the selected top-ranked feature.

Abstract: A method and a processing device are provided for forecasting whether a record of data for an individual is associated with an outcome. Whether an n-gram associated with one of a group of predictors matches an n-gram included in the record of data is determined. When the n-gram included with the one of the group of predictors matches the n-gram in the record of data, a prediction is made as to whether the record of data is associated with the outcome based on a value assigned to the one of the group of predictors.

Abstract: In one embodiment, the present invention includes a method for receiving a set of inspection files each corresponding to an inspection performed on a wafer of a set of wafers, automatically analyzing the set of inspection files to select at least one inspection file corresponding to a predetermined rule set, and identifying the wafer(s) associated with the selected inspection file(s). Other embodiments are described and claimed.

Abstract: The present invention provides a system for analyzing a time series signal by a method of Chaos Theory and calculating a chaos theoretical exponent value.

Abstract: A method and system for processing commonality of semiconductor devices. The method includes providing a first plurality of semiconductor devices, providing a second plurality of semiconductor devices, obtaining a first plurality of measured values corresponding to a characteristic associated with the first plurality of semiconductor devices, obtaining a second plurality of measured values corresponding to the characteristic associated with the second plurality of semiconductor devices, performing a first statistical analysis for the first plurality of measured values, determining a first statistical distribution, performing a second statistical analysis for the second plurality of measured values, and determining a second statistical distribution. Moreover, the method includes processing information associated with the first statistical distribution and the second statistical distribution, and determining an indicator.

Abstract: A method for monitoring instantaneous behavior of a tire in a rolling condition includes acquiring and storing at least one reference curve representing an acceleration profile of at least one specified point of the tire as a function of its position during at least one portion of a revolution of the tire; continuously acquiring signals of acceleration of the at least one point; deriving from the signals of acceleration at least one cyclic curve of acceleration of the at least one point; comparing the at least one cyclic curve with the at least one reference curve; and emitting a signal depending on the comparison that indicates the instantaneous behavior of the tire. The at least one reference curve represents the acceleration profile in at least two directions, including two or more of a centripetal direction, a tangential direction, and a lateral direction. A related system and tire are also disclosed.

Abstract: A method of assessing a pregnant woman's risk of having a fetus with a fetal abnormality by determining the fetus' BPD/OFD ratio and using the BPD/OFD ratio in conjunction with one or more other screening markers for the fetal abnormality. Also provided is a method of determining whether a pregnant woman is screen-positive or screen-negative by comparing the BPD/OFD ratio of the pregnant woman's fetus to a risk cut-off level.

Abstract: A time series is analyzed by multiple functions simultaneously to identify an anomaly for a data point in the series. Data point values are predicted by the multiple functions. An anomaly occurs when an actual data point in the series differs significantly from the data point's predicted value as generated by the functions. If enough statistical models detect an anomaly has occurred for a data point, an anomaly event is generated. The set of functions can include different types of functions, the same function type configured with different constants, or a combination of these.

Abstract: A system and method of accurately predicting and reporting a value of a property based on a weighted average of values predicted by at least two prediction models. The system and method include the steps of accessing predicted values, determined by the prediction models, for the property; determining property-specific proportional prediction error distribution information for each predicted value determined by each prediction model; assigning a weight to the predicted value determined by each prediction model by using the property-specific proportional prediction error distribution information; and generating a property-specific weighted average value based on combination of the weight and the predicted value determined by each prediction model and reporting the property-specific weighted average value to minimize prediction error during prediction of the property value.

Abstract: The present invention provides a method and apparatus for detecting step and impulse disturbances. The method includes determining a pattern based on a plurality of probabilities associated with a corresponding plurality of wafer processing parameters and determining a type of a disturbance based upon the pattern.

Abstract: A number of methods to support process quality and maintenance during control of an industrial process such as welding are provided. The methods provide, among other things: automatic process limit programming based on runtime data; user-initiated process limit programming based on upcoming data; correlate equipment deterioration based on capability measurement; correlate tip dressers/formers to new; detect electrical deterioration; integrate process data with programmed data for a visual aid; and quantify process variation in welding tools (pareto of stddev of the c-factor).

Abstract: A method for evaluating an alarm system for continuous or semi-continuous processes that gives an indication of the actual loading on the operator. A transformations is applied to a data set consisting of alarms and associated times, resulting in the data being much more amenable to common techniques of statistical quality control. The technique requires only a history of alarm events, and the transformation has a strong intuitive appeal.

Abstract: Methods and apparatus for monitoring grid-based computing resources are disclosed. Example embodiments provide for the determination and display of a variety of metrics. Data related to processor utilization by a plurality of applications is accumulated, aggregated, and normalized. At least one metric related to the use of the grid by the plurality of applications is assayed or otherwise determined from the aggregated, normalized data. An archive database for archival storage of at least some of the data can also be provided. In some embodiments, a Web server displays metrics via a Web page. In addition, provision can be made to verify the service level agreement (SLA) compliance of the various resources on the grid.

Abstract: A method of evaluating a water sample for the presence or possible future presence of nitrification comprises obtaining data values of a number of parameters, processing the data values to determine correlation coefficients, to identify any linear dependencies, to standardize the scales, evaluating the data values over a plurality of proliferation time periods and neuron numbers, calculating MSEs and R2's from the evaluations, and estimating a valid likelihood of nitrification of the water sample. A method of evaluating a water sample for the presence or possible future presence of nitrification, comprises obtaining data values of a number of parameters, statistically pre-processing the data values and supplying the pre-processed data values to a neural network. Apparatus, media and processors which are used in performing such methods.

Abstract: A method, and associated storage medium containing software and a system, includes extracting a time domain impulse response from parameters that characterize a communication channel, generating a probability distribution function (PDF) of an output voltage based on the impulse response; and computing a relationship between bit error rate and voltage margin based on the final probability distribution function. Generating the PDF of the output voltage may comprise one or more of the following acts: quantizing the impulse response into a plurality of quantized levels, assigning taps to the quantized levels and determining a number of taps assigned to each quantized level, determining allowable voltage levels for each quantized level, and determining a probability of occurrence of each allowable voltage level, determining a PDF for each voltage level; and convolving all of the PDFs for the various voltage levels to obtain the PDF of the output voltage.

Abstract: A method of monitoring an entity within a process plant wherein the entity includes a plurality of lower level entities, includes acquiring a plurality of use indices, acquiring a plurality of weighting values, and creating an aggregate use index from a combination of the lower level use indices and weighting values, wherein the aggregate use index represents status information regarding the entity. Each use index pertains to status information regarding one of the plurality of lower level entities, and each weighting value pertains to the importance of a lower level entity among the plurality of lower level entities.

Abstract: A method for the determination of a physical property parameter of a seat passenger includes getting a reading of at least two parameters from an occupancy sensor; plotting a probability vector which shows, for each value of the physical property, the probability to cause the readings of the at least two parameters; and correlating the physical property parameter to the range of values of the probability vector with the highest probabilities.

Abstract: A histogram difference method and system for power/performance measurement and management has low data storage requirements while supporting multiple monitoring applications having different update rates. Histogram data for power usage and/or performance mode is collected at a predetermined rate and the histogram data is read out at periodic intervals by the monitoring applications. The monitoring applications subtract the histogram data from previously read histogram data set to determine a interval difference histogram. The minimum and maximum values for the interval are the lowest-valued and highest-valued bin in the interval difference histogram that have a count greater than zero. The average value for the interval is the mean of the interval difference histogram. A conservative bound of the maximum and minimum values for a system can be determined by adding the values of the maximum and minimum values determined for each subsystem in the system.

Abstract: A method and apparatus are provided for correlating events in a system. Problems and other events can, e.g., be detected in a system that generates symptoms or observable events. A computer-accessible codebook is provided that includes a mapping between each of a plurality of groups of possible symptoms and one of a plurality of likely exceptional events (e.g., problems) in the system. The system is monitored and one or more known symptoms generated by the system are detected. A mismatch measure is determined between each of the plurality of groups of possible symptoms in the mapping and the one or more known symptoms using a computer, while disregarding symptoms in the groups of possible symptoms not determined to be known. One or more of the plurality of likely problems is selected corresponding to one of the plurality of groups having the smallest mismatch measure.

Abstract: A method and a system for determining a remaining time before failure of a motor system are provided. The method and the system determine the remaining time before failure based on historical motor data and operational parameter data associated with the motor system.

Abstract: In one embodiment, a pressure transmitter is provided which diagnoses the condition of a primary element and/or an impulse line which connects to a pressure sensor. A difference circuit coupled to the pressure sensor has a difference output which represents the sensed pressure minus a moving average. A calculate circuit receives the difference output and calculates a trained output of historical data obtained during an initial training time. The calculate circuit also calculates a monitor output of current data obtained during monitoring or normal operation of the transmitter. A diagnostic circuit receives the trained output and the monitor output and generates a diagnostic output indicating a current condition.

Abstract: A system and method for detecting erosion in turbine engine blades is provided. The blade erosion detection system includes a sensor data processor and a cluster analysis mechanism. The sensor data processor receives engine sensor data, including exhaust gas temperature (EGT) data, and augments the sensor data to determine sensor data residual values and the rate of change of the sensor data residual values. The augmented sensor data is passed to the cluster analysis mechanism. The cluster analysis mechanism analyzes the augmented sensor data to determine the likelihood that compressor blade erosion has occurred. Specifically, the cluster analysis mechanism performs a 2-tuple cluster feature analysis using Gaussian density functions that provide approximations of normal and eroded blades in a turbine engine. The 2-tuple cluster feature analysis thus provides the probability that the sensor data indicates erosion has occurred in the turbine engine.

Abstract: Computer-implemented systems and methods for performing forecast scoring based upon time series data that is indicative of a data generation activity occurring over a period of time. One or more data stores store historical data and future data. The historical data includes historical exogenous inputs. The future data includes future exogenous inputs. A dynamic statistical model receives the historical data and future data that are stored in the one or more data stores. The dynamic statistical model is to provide forecast scoring based upon the historical exogenous inputs and the future exogenous inputs. An optimizer having an objective function iteratively varies the future data, so that forecasts associated with the varying future data values are generated by the dynamic statistical model for evaluation by the objective function.

Abstract: The process comprises acquisition of experimental data from measurements of physical quantities by means of at least one sensor associated to the object. The process comprises a first reliable estimation of a first range of values with a first method. The process comprises at least one additional estimation of an additional range of values with a different method. The methods each present a predetermined reliability. The ranges of values are successively observed in order of decreasing reliability of the corresponding methods. Each additional range of values is compared with the range of values corresponding to the previous method according to said order. The additional range is chosen as result when the additional range is comprised in the range corresponding to one of the previous methods.

Abstract: A system that determines the performance of an integrated circuit (IC). During operation, the system receives probability distributions for parameters for the IC. Next, the system generates samples of the IC, wherein generating a given sample involves using the probability distribution to assign values to the parameters for components within the IC. The system then calculates output performance metrics for the samples based on the assigned values of the parameters, and uses the calculated output performance metrics to generate a distribution of output performance metrics for the samples.

Abstract: The invention provides methods and apparatus, including computer program products, for correcting outlier values in a series of values representing a predetermined numerical parameter over time. For predetermined time interval with a beginning (b) and end (e) point of the time series, an ex-post forecast time series (s_ep) is calculated using the historical time series (s_h) and a predetermined model function (f). Lower and upper tolerance limit lines are defined using (s_ep) and on a quality function (qf) over the time interval. Beginning at (b), (s_h) is corrected by replacing the first value outside of the tolerance lines by a predetermined value inside the tolerance lines. The (s_ep) is recalculated using (s_h) and (f). The tolerance lines are redefined using the recalculated (s_ep) and (qf) that takes only values where the last outlier value has been replaced. These steps are repeated until all (s_h) outlier values have been replaced.

Abstract: A distribution goodness-of-fit test device for testing whether measured data matches an estimated probability distribution has a counting section determination unit, a counting unit and a goodness-of-fit test unit. The counting section determination unit determines according to the number of the measured data, widths of counting sections for counting the measured data. The counting unit counts the numbers of data in the respective determined counting sections. Also, the goodness-of-fit test unit performs a goodness-of-fit test based on the numbers of data in the respective counting sections.

Abstract: Measured values obtained in a measuring machine to be estimated are provided to estimate errors in the measured values. Based on the estimated errors in the measured values, a covariance matrix or correlation matrix of measured values is derived. The covariance matrix or correlation matrix is then subjected to eigenvalue decomposition to derive eigenvalues and eigenvectors. A normal random number with an expected value of 0 and a variance equal to an eigenvalue corresponding to the eigenvalue is generated as a coefficient of coupling for each eigenvector, and all eigenvectors are linearly coupled to generate pseudo-measured values of the measuring machine. The generated pseudo-measured values are subjected to statistic processing to estimate uncertainty of the measuring machine.

Abstract: Some embodiments of the invention provide a system to monitor a patient. In some aspects, a signal representing a value of a physiological parameter is received from a monitoring device. The received value is a value that caused the monitoring device to trigger an alarm associated with the psychological parameter. A notification is determined based on the received signal and on a second signal that represents a value of a second physiological parameter. The notification is then presented to an operator.

Abstract: A method for optimizing an industrial process data is disclosed. The method includes collecting data from a plurality of sensor elements, wherein each sensor element collects data from a portion of the industrial process and verifying the data collected. The method further includes analyzing the data collected for efficiency and generating at least one recommendation for optimizing the industrial process. The method further includes presenting the at least one recommendation generated to an administrator of the industrial process.

Abstract: A sub-system is provided to a discrete event simulator (DES) to expedite simulation execution by first detecting a non-quiescent steady-state condition in the simulated system, and when the steady-state condition is detected, the simulator determines a state, and subsequently simulates the system at a skip-ahead time using this determined state, or a predicted state based on the determined state. Convergence analysis is used to determine whether the system is at, or approaching, a steady-state condition. This convergence skip-ahead process achieves faster analysis by avoiding the computation that would conventionally be required to simulate the system behavior during the time interval that is skipped.

Abstract: A target value processing unit, includes: an input section to which a target value signal showing a target value of a control process is inputted; a target value shaping unit shaping the target value signal inputted to the input section, into a signal form which is proper for a control treatment of a regulator implementing the control process; and an output section outputting to the regulator a shaped target value signal which is shaped by the target value shaping unit. The target value processing unit realizes the high-level control process without improving the regulator.

Abstract: The subject invention provides for systems and methods that facilitate optimizing one or mores sets of training data by utilizing an Exponential distribution as the prior on one or more parameters in connection with a maximum entropy (maxent) model to mitigate overfitting. Maxent is also known as logistic regression. More specifically, the systems and methods can facilitate optimizing probabilities that are assigned to the training data for later use in machine learning processes, for example. In practice, training data can be assigned their respective weights and then a probability distribution can be assigned to those weights.

Abstract: Summarization of video content including sports.