Fair Isaac Corporation

Gerald Fahner

Joseph P Milana

Christopher Allan Ralph

Scott Michael Zoldi

Liang Meng

Brad Vancho

Zhenyu Yan

Shafi Ur Rahman

Amit Kiran Sowani

Computer-implemented systems, methods and products for modeling sensitivities to potential disruptions by observing performances of entities in a first sub-population and a second sub-population using a machine learning model comprising a set of predictors and a binary indicator variable associated with a first entity subjected to a first event associated with the first sub-population, the machine learning model trained to predict an expected performance for the first entity based on at least one of a known attribute associated with the first entity in relation to the first event and a value of the binary indicator variable associated with the first event.

Supervised machine learning-based modeling of sensitivities to potential disruptions

A method, a system, and a computer program product for generating a refined synthetic data from one or more sources of data. One or more source data are received from one or more data sources. One or more encoded source data are generated from the one or more source data. A synthetic data is generated by decoding one or more encoded source data. One or more variables in the synthetic data are selected and one or more predetermined identifiability values and one or more predetermined anonymity values are associated with them. The generated synthetic data including the selected variables is decoded using associated one or more predetermined identifiability values and one or more predetermined anonymity values. The decoded synthetic data is outputted.

Enhanced data privacy through structure-preserving autoencoder with latent space augmentation

A method, a system, and a computer program product for generating an interpretable set of features. One or more search parameters and one or more constraints on one or more search parameters for searching data received from one or more data sources are defined. The data received from one or more data sources is searched using the defined search parameters and constraints. One or more first features are extracted from the searched data. The first features are associated with one or more predictive score values. The searching is repeated in response to receiving a feedback data responsive to the extracted first features. One or more second features resulting from the repeated searching are generated.

Interpretable feature discovery with grammar-based bayesian optimization

A sensitivity index model for predicting the sensitivity of an entity to a potential future disruption can be trained using a process that includes dividing a population of entities for which data attributes are available into matched pairs in a first sub-population and a second sup-population based on matching propensity scores for the entities using supervised machine learning, modeling outcomes for the two sub-populations, using the resultant models to calculate expected performances of the entities under differing conditions, and generating the sensitivity index model using supervised learning techniques based on quantification of differences between the calculated expected performances for the entities.

Systems and methods for generating concise explanations of scored observations that strike good, and computationally efficient, trade-offs between rank-ordering performance and explainability of scored observations are disclosed. The systems and methods described herein for explaining scored observations are based on a framework of partial dependence functions (PDFs), multi-layered neural networks (MNNs), and Latent Explanations Neural Network Scoring (LENNS).

Fast automatic explanation of scored observations

Optimal strategies for providing offers to a plurality of customers are generated. A plurality of categorical attributes (for example, gender and residential status) and ordinal attributes (for example, risk score and credit line utilization) can be determined. Values of one of more categorical attributes can be changed as per a transition probability table. Some probabilities can be varied to determine a first tradeoff, based on which a first updated strategy can be generate Further, noise can be added to one or more ordinal attributes. Standard deviation of a noise distribution associated with the noise can be varied so as to determine a second tradeoff, based on which a second updated strategy can be generated. The second updated strategy can be an update of the first updated strategy. Offers can be provided to the plurality of customers in accordance with the second updated strategy.

Generating optimal strategy for providing offers

Profiles characterizing each of a plurality of consumers are received. Thereafter, each profile is associated with one of a plurality of customer segments (e.g., matched pairs, etc.). Thereafter, a coupon effectiveness index is determined for each of the plurality of consumers for an offering based on the associated customer segment. The coupon effectiveness indices model characterizes causal effects estimates determined using historical data of purchases of individuals having varying coupon treatments for the offering. Subsequently, provision of at least a portion of the determined coupon effectiveness indices is initiated. Related apparatus, systems, techniques and articles are also described.

Coupon effectiveness indices

A method and system for estimating potential future outcomes resulting from decision alternatives is presented to enable lenders to make lending related decisions. The estimation is based on a propensity score variable that encompasses an effect of multiple covariates associated with one or more individuals for whom the estimation is being performed. For consistency with empirical testing, the estimation approach assumes conditions of unconfoundedness and localized common support. According to the unconfoundedness assumption, for a given variable, the potential outcomes are conditionally independent of the decision alternatives. According to the localized common support assumption, an overlap is ensured between individual accounts that are categorized together as potentially having the same future outcome. The outcomes and an effect (e.g. comparison) of the outcomes may be displayed graphically.

Causal modeling for estimating outcomes associated with decision alternatives

A computer-implemented simulator models the entire analytic value chain so that data generation, model fitting and strategy optimization are an integral part of the simulation. Data collection efforts, data mining algorithms, predictive modeling technologies and strategy development methodologies define the analytic value chain of a business operation: data&#x2192;models&#x2192;strategies&#x2192;profit. Inputs to the simulator include consumer data and potential actions to be taken regarding a consumer or account. The invention maps what is known about a consumer or an account and the potential actions that the business can take on that consumer or account to potential future financial performance. After iteratively performing simulations using varying inputs, modeling the effect of the innovation on a profit model, the simulator outputs a prediction of the commercial value of an analytic innovation.

Growing community of inventors

Austin, TX, United States of America

Gerald Fahner