Abstract: Example systems and methods of developing a learning model are presented. In one example, a sample data set to train a first learning algorithm is accessed. A number of states for each input of the sample data set is determined. A subset of the inputs is selected, and the sample data set is partitioned into a number of partitions equal to a combined number of states of the selected inputs. A second learning algorithm is created for each of the partitions, wherein each second learning algorithm receives the unselected inputs. Each of the second learning algorithms is assigned to a processor and trained using the samples of the partition corresponding to that algorithm. Decision logic is generated to direct each of a plurality of operational data units as input to one of the second learning algorithms based on states of the selected inputs of the operational data unit.

Abstract: Example systems and methods of developing a learning model are presented. In one example, a sample data set to train a first learning algorithm is accessed. A number of states for each input of the sample data set is determined. A subset of the inputs is selected, and the sample data set is partitioned into a number of partitions equal to a combined number of states of the selected inputs. A second learning algorithm is created for each of the partitions, wherein each second learning algorithm receives the unselected inputs. Each of the second learning algorithms is assigned to a processor and trained using the samples of the partition corresponding to that algorithm. Decision logic is generated to direct each of a plurality of operational data units as input to one of the second learning algorithms based on states of the selected inputs of the operational data unit.

Abstract: Example systems and methods of developing a learning model are presented. In one example, a sample data set to train a first learning algorithm is accessed. A number of states for each input of the sample data set is determined. A subset of the inputs is selected, and the sample data set is partitioned into a number of partitions equal to a combined number of states of the selected inputs. A second learning algorithm is created for each of the partitions, wherein each second learning algorithm receives the unselected inputs. Each of the second learning algorithms is assigned to a processor and trained using the samples of the partition corresponding to that algorithm. Decision logic is generated to direct each of a plurality of operational data units as input to one of the second learning algorithms based on states of the selected inputs of the operational data unit.

Abstract: Example systems and methods of developing a learning model are presented. In one example, a sample data set to train a first learning algorithm is accessed. A number of states for each input of the sample data set is determined. A subset of the inputs is selected, and the sample data set is partitioned into a number of partitions equal to a combined number of states of the selected inputs. A second learning algorithm is created for each of the partitions, wherein each second learning algorithm receives the unselected inputs. Each of the second learning algorithms is assigned to a processor and trained using the samples of the partition corresponding to that algorithm. Decision logic is generated to direct each of a plurality of operational data units as input to one of the second learning algorithms based on states of the selected inputs of the operational data unit.

Abstract: Example systems and methods of developing a learning model are presented. In one example, a sample data set to train a first learning algorithm is accessed. A number of states for each input of the sample data set is determined. A subset of the inputs is selected, and the sample data set is partitioned into a number of partitions equal to a combined number of states of the selected inputs. A second learning algorithm is created for each of the partitions, wherein each second learning algorithm receives the unselected inputs. Each of the second learning algorithms is assigned to a processor and trained using the samples of the partition corresponding to that algorithm. Decision logic is generated to direct each of a plurality of operational data units as input to one of the second learning algorithms based on states of the selected inputs of the operational data unit.

Abstract: Example systems and methods of developing a learning model are presented. In one example, a sample data set to train a first learning algorithm is accessed. A number of states for each input of the sample data set is determined A subset of the inputs is selected, and the sample data set is partitioned into a number of partitions equal to a combined number of states of the selected inputs. A second learning algorithm is created for each of the partitions, wherein each second learning algorithm receives the unselected inputs. Each of the second learning algorithms is assigned to a processor and trained using the samples of the partition corresponding to that algorithm. Decision logic is generated to direct each of a plurality of operational data units as input to one of the second learning algorithms based on states of the selected inputs of the operational data unit.

Abstract: A filter selection technique is described for automatically selecting filters and filter parameters to apply to a given input data. The technique first receives input data and accesses a library storing information from previously analyzed data. The technique selects an entry from the library where the entry contains data that is correlated with the input data. The technique then applies a filter to the input data. The filter and filter parameters are determined by the selected entry.

Abstract: A filter selection technique is described for automatically selecting filters and filter parameters to apply to a given input data. The technique first receives input data and accesses a library storing information from previously analyzed data. The technique selects an entry from the library where the entry contains data that is correlated with the input data. The technique then applies a filter to the input data. The filter and filter parameters are determined by the selected entry.

Abstract: A computer-implemented method controls commercial transactions involving a portfolio of financial products by conducting business operations related to commercial transactions between a bank and consumer involving purchase and utilization of the financial products, collecting transactional data related to the financial products, and providing a centralized modeling and optimization tool to predict customer response to changes in an attribute of a financial product under evaluation based on the transactional data and to optimize the variable of the financial product under evaluation. The modeling and optimization tool is configurable to evaluate the financial products in the portfolio under KPIs and business rules selected according to the financial product under evaluation. The optimized variable is transmitted to the bank.

Abstract: Example systems and methods of circular transaction path detection are presented. In one example, a directed graph comprising nodes and directed edges interconnecting the nodes is generated. The directed graph is based on information describing a plurality of parties and a plurality of transactions between the parties. A circular path length of interest is received. Strongly connected components of the directed graph are identified. Within each of the strongly connected components, each circular path having a length equal to the circular path length of interest is discovered. For each discovered circular path, the transactions represented by the directed edges of the path are denoted as related transactions.

Abstract: A system, a computer program product, and a method for order planning and optimization are disclosed. A first data is received, where the first data represents historical shipment data of an item from a distributor to a location. The received first data is processed and a model for at least one shipping pattern of the item from the distributor to the location is determined based on the processed received first data. A forecast for a future shipping demand of the item by the location is generated based on the determined model. At least one shipping pattern of the item from the distributor to the location is optimized based on the generated forecast.

Abstract: Various embodiments herein include at least one of systems, methods, and software for adaptive in-memory customer and customer account classification. Some such embodiments include receiving a rule identifying data attributes that contribute to at least one outcome with regard to at least one product and applying the rule to a dataset replicated from a transactional data environment to an in-memory data environment. Application of the rule results in segmentation of at least one of customers and customer accounts likely to have each of the at least one outcomes, the replicated dataset including customer data. Such embodiments may then output data representative of the segmented at least one of customers and customer accounts likely to have each of the at least one outcomes. The in some embodiments, the rule is applied to define a further rule which may be stored and later utilized to perform further data segmentation.

Abstract: A computing system (100) receives transaction records (130) for loans taken at various interest rates (1904) for a loan segment (902). Performance indicators (1716) indicative of customer behaviors (1702) are computed (1806) using independent demand models (300, 302, 304, 306, and 308). Computing system (100) includes a performance indicator forecaster (112) that determines relationships between the performance indicators (1716) and various prices, or interest rates (1904). These relationships can include profit (1906) and/or volume (1908) relative to the various interest rates (1904). The relationships are utilized to select an interest rate (1912, 2102) for the product segment (902) for implementation by a financial institution.

Abstract: A computer-implemented method for managing operations of a system includes deriving a nonlinear modeling function from a nonlinear response function, defining an allowed range for output values of the nonlinear modeling function, determining a range of a first set of input values of the nonlinear modeling function based on the allowed range of the output values, deriving a nonlinear probability function from the nonlinear response function, receiving the first set of input values, calculating the output values by processing each input value in the first set of input values through the nonlinear modeling function, determining, using the probability function, a relative probability of performing a first future system operation for each input value of the first set of input values and displaying, for each input value in the first set of input values, the corresponding output value and the corresponding probability.

Abstract: Various embodiments herein include at least one of systems, methods, and software for freight market demand modeling and price optimization. Some such embodiments include acquiring historical data regarding hauled loads, bid loads that were not hauled, data representative of at least one of current and expected conditions, and data representing business goals. The acquired data may then be mapped to market segments and a statistical, spot load demand model is generated for each market segment based on a number of factors included in the mapped data including at least a load price factor. A demand and price forecast model may next be generated for each market segment based on the generated model and the data representative of at least one of current and expected conditions. For each market segment, a pricing element may then be determined based on the respective market segment model and forecast in view of the business goals.

Abstract: A computer-implemented method for managing operations of a system includes deriving a nonlinear modeling function from a nonlinear response function, defining an allowed range for output values of the nonlinear modeling function, determining a range of a first set of input values of the nonlinear modeling function based on the allowed range of the output values, deriving a nonlinear probability function from the nonlinear response function, receiving the first set of input values, calculating the output values by processing each input value in the first set of input values through the nonlinear modeling function, determining, using the probability function, a relative probability of performing a first future system operation for each input value of the first set of input values and displaying, for each input value in the first set of input values, the corresponding output value and the corresponding probability.

Abstract: A computer system for modeling a portfolio of products in a financial system to determine the rate of a target product. The products are defined by attribute values, an attribute being any criteria that impacts product rates. Linear associated product rules are used by the computer system to create an optimized scenario of total profit and overall volume of sales for the portfolio. From the optimized scenario a rate for the target product can be determined which maintains a financial institution's strategic and business objectives. The optimizing process includes applying the associated product rules to products actively contributing to key performance indicators. Densification is then used to infer the rate for all other products in the portfolio. Finally, if the starting rate of a product violates an associated product rule, the starting rate is relaxed to avoid the violation.

Abstract: A computer implemented method for determining the reference values of sensitivities and strategies for price optimization demand models from a profit function and current product price. A total profit objective is expressed as the maximization of profit and volume, where a strategy parameter represents the relationship between profit and volume. From the total profit objective, the bounds of the strategy parameter are expressed as conditional inequalities relating the bounds to functions of the unit profit at the current rate and average volume. The strategy parameter is then set to the average of these bounds. The reference elasticity is expressed as a function of the unit profit function and average volume. The resulting reference values can be used in a price optimization system to generate recommended prices that relate to an industry's current pricing scheme.

Abstract: A computer-implemented method controls commercial transactions involving a portfolio of financial products by conducting business operations related to commercial transactions between a bank and consumer involving purchase and utilization of the financial products, collecting transactional data related to the financial products, and providing a centralized modeling and optimization tool to predict customer response to changes in an attribute of a financial product under evaluation based on the transactional data and to optimize the variable of the financial product under evaluation. The modeling and optimization tool is configurable to evaluate the financial products in the portfolio under KPIs and business rules selected according to the financial product under evaluation. The optimized variable is transmitted to the bank.

Abstract: A non-stationary time series model using a likelihood function as a function of input data, base demand parameters, and time dependent parameter. The likelihood function may represent any statistical distribution. The likelihood function uses a prior probability distribution to provide information external to the input data and is used to control the model. In one embodiment the prior is a function of adjacent time periods of the demand profile. The base demand parameters and time dependent parameter are solved using a multi-diagonal band matrix. The solution of base demand parameters and time dependent parameter involves making estimates thereof in an iterative manner until the base demand parameters and time dependent parameter each converge. A non-stationary time series model is provided from an expression using the solution of the base demand parameters and time dependent parameter. The non-stationary time series model provides a demand forecast as a function of time.