Abstract: A trend of attributes associated with plural market participants is determined. A representation of the trend is computed, and market models for the market participants are built according to the representation of the trend.

Abstract: To analyze a time series of data that exhibits seasonal effects, the time series is processed to obtain a representation in the frequency domain. According to the representation, plural cycle lengths are identified as representing different seasonal effects of the data in the time series, where a first of the plural cycle lengths is greater than a second of the plural cycle lengths.

Abstract: To detect a change in data produced by a system, predicted data values for plural time points are computed. Actual data values for the plural time points are received, and residual values are derived from differences between the predicted data values and actual data values. Based on the computed residual values, a time point at which the change in data occurred is determined.

Abstract: User input regarding a target system on which a software application is to be deployed is received. A benchmark system from plural candidate benchmark systems is matched to the target system. An estimated performance of the software application on the target system or an estimated utilization of resources of the target system by the software application is computed based on information relating to the matched benchmark system.

Abstract: There is provided a system and method for estimating a parameter that represents data describing a physical system. An exemplary method comprises randomizing data representative of a population of items for which the parameter is known. The method may additionally comprise generating data representative of a pseudo population of items using a known perturbation, the data representative of the pseudo population of items being included with the data representative of the population of items for which the parameter is known to form a revised population and selecting a bootstrap sample of a minimum sample size of the revised population. A sensitivity study is performed on the parameters of the items comprising the bootstrap sample to determine a level of change of a predicted parameter value relative to a parameter value of the sample.

Abstract: Aggregation of data values in a data set is computed to produce aggregate values. The aggregate values are partitioned into plural aggregate value sets. Fitting with respect to the plural aggregate value sets is performed. Based on the fitting, at least one time point corresponding to occurrence of a change in the data values is determined.

Abstract: Systems and methods are disclosed for cumulative attribute forecasting using a probability density function of a current-to-future value ratio. Some method embodiments may comprise determining an unconditional probability density function for a ratio of a cumulative attribute's current value to an associated future value for the cumulative attribute, and determining an unconditional PDF for a future value of the cumulative attribute. The unconditional probability functions are combined with a current value of the cumulative attribute in a manner that yields a conditional probability density function for the future value. The conditional probability density function may then be used to determine forecasts such as a most likely future value, a mean value, a median value, a confidence interval for the future value, and a probability of achieving a future value within some specified range.

Abstract: A time series of data values representing occurrences of events at plural time points is received. Durations between successive events are computed, and a burst of activity based on the computed durations is detected. It is determined that a change has occurred in response to detecting the burst of activity.

Abstract: To perform forecasting, a first data collection having data values at first intervals is received, and a first forecasting model is built based on the first data collection. A second forecasting model is built based on a second data collection having intervals aggregated from intervals of the first data collection, wherein the second forecasting model is at a different aggregation level than the first forecasting model. At least one metric is computed by performing at least one test based on at least one of the first and second data collections to indicate which of the first and second forecasting models has a better forecast quality.

Abstract: To forecast data, an initial collection of data having a first length is received. In response to determining that the first length of the initial collection of data is insufficient for performing forecasting using a forecasting algorithm, an order of the initial collection of data is reversed to provide a reversed collection of data. Forecasting is applied on the reversed collection of data to estimate additional data values to combine with the initial collection of data to provide a second collection of data having a second length greater than the first length. The forecasting algorithm is applied on the second collection of data.

Abstract: To detect data change, data values are separated into plural sets. Predefined values representative of a predetermined effect are calculated for respective plural sets. Adjusted data values are calculated by removing impact of calculated predefined values from the data values in the respective plural sets. The data change is detected based on the adjusted data values.

Abstract: Embodiments include methods, apparatus, and systems for multi-period-ahead forecasting. One embodiment is a method that applies a first forecasting algorithm to static historical data to generate a first forecast into a future time period and applies a second forecasting algorithm to dynamic data obtained for a current time period to generate a second forecast. The first and second forecasts are combined to generate forecasts for future time periods.

Abstract: Disclosed herein are systems and methods for profile-based forecasting with dynamic profile selection. Some method embodiments may comprise determining a reference set of profiles from a source set of profiles, and using the reference set of profiles to generate a forecast. The reference set determination comprises at least comparing a current, partial profile to each profile in the source set to obtain a similarity measurement for each profile in the source set.

Abstract: To perform data quality assurance, data values from a data source at discrete time points up to time point t are received. At least one estimated value is computed based on at least some of the received data values, and the received data values and estimated data values are applied to an algorithm. A data quality determination of the data value for time point t is performed based on the algorithm.

Abstract: To analyze a time series of data that exhibits seasonal effects, the time series is processed to obtain a representation in the frequency domain. According to the representation, plural cycle lengths are identified as representing different seasonal effects of the data in the time series, where a first of the plural cycle lengths is greater than a second of the plural cycle lengths.

Abstract: To perform forecasting, a first data collection having data values at first intervals is received, and a first forecasting model is built based on the first data collection. A second forecasting model is built based on a second data collection having intervals aggregated from intervals of the first data collection, wherein the second forecasting model is at a different aggregation level than the first forecasting model. At least one metric is computed by performing at least one test based on at least one of the first and second data collections to indicate which of the first and second forecasting models has a better forecast quality.

Abstract: A trend of attributes associated with plural market participants is determined. A representation of the trend is computed, and market models for the market participants are built according to the representation of the trend.

Abstract: According to information regarding an enterprise offering, characteristics of a life cycle of the enterprise offering are determined. A model including functions representing segments of the life cycle is produced based on the determined characteristics.

Abstract: To forecast data, an initial collection of data having a first length is received. In response to determining that the first length of the initial collection of data is insufficient for performing forecasting using a forecasting algorithm, an order of the initial collection of data is reversed to provide a reversed collection of data. Forecasting is applied on the reversed collection of data to estimate additional data values to combine with the initial collection of data to provide a second collection of data having a second length greater than the first length. The forecasting algorithm is applied on the second collection of data.

Abstract: Systems and methods are disclosed for identifying recurrent patterns. In one embodiment, a method comprises: estimating a power spectral density of a time series; determining a duration-level associated with a peak in the power spectral density; and aggregating the time series at the duration-level to obtain a recurrent pattern.