Understanding predictive analytics and Decision Strategy Manager

The following terms are often associated with the configuration, implementation and use of the predictive analytics and Decision Strategy Manager.

Predictive Analytics Director

Predictive Analytics Director (PAD) is a desktop application used to develop and create predictive models. Predictive models exported in PAD can be used to create and define predictive model rules, which can then be used directly in flows, or combined with other components in a strategy. PAD develops the means to differentiate between cases on the basis of likely future behavior.

Powerful and reliable predictive models deliver the key insights that enable opportunities and risks to be evaluated, constituting the foundation of personalized strategies. PAD reveals the relationships in your data, the critical information, and the interactions that drive customer behavior in an intelligent data mining process that knows what needs to be done. Your role is to define the objectives, and judge the results.

Adaptive Analytics Engine

The main process of the Adaptive Decision Manager. The engine is responsible for storing sufficient adaptive statistics analyzing them, and producing individual scoring models that are used in PRPC. These statistics keep the relevant values for adaptive models defined in decision strategies. From these statistics, the adaptive analytics engine creates scoring models that are published to the adaptive data store. PRPC retrieves the scoring models from the database, and uses them to calculate the prediction.

Adaptive Data Store

The database scoring adaptive statistics and adaptive models.

Adaptive Model

Adaptive models are ADM scoring models that output predictions calculated and adapted in real time as responses are captured after executing a strategy. Models in ADM are configured through adaptive model rules. Adaptive model rules define the settings that influence the behavior of the adaptive models in ADM. Adaptive models are created by executing strategies with adaptive model components. When adding the adaptive model component in the strategy, you configure the proposition the adaptive model is going to model, and the interpretation of the outputs. Adaptive models belong to the self-learning aspect of Decision Management, and typically used in the absence of historical records to make predictions.

Adaptive Statistics

The persistent information resulting from running a strategy containing adaptive models.

Behavioral Profile

A behavioral profile represents a univariate model. The probabilities of a positive outcome for each interval/category are score bands and can be used to predict in the same way as those of any other model.

Case

A case can be any person, company, or event that exhibits some defined behavior. For example, good and bad outcomes.

Coefficient

A weight that is used for each predictor in the logistic regression formula. The coefficient is an indication of the importance of a predictor. Negative coefficients imply the presence of predictors with very similar behavioral profile. If present, they can lead to over fitting and unreliable models. Consider reanalyzing the predictor grouping to ensure predictors with highly correlated behavior are placed in the same predictor group.

Coefficient of Concordance

The Coefficient of Concordance (CoC) is a non-parametric coefficient sensitive to the complete range of score bands irrespective of their distribution.The CoC measures how well the scores generated by the model separate positive from negative outcome cases using the statistic known as coefficient of concordance. CoC can vary between 50% (a random distribution of positive and negative cases by score band) and 100% (a perfect separation). The minimum is 50% because the scores are simply used in reverse if a set of scores orders negative cases before positive cases. Its virtue as a measure is that it encourages models to be predictive across the score range. If the desired operational circumstances (volume or quality of business) are unknown, CoC generates powerful and generalized models.

Data Source

Data about customers, and their previous behavior. This data can be used for modeling and strategy design. A source should contain one record per customer with the same structure for each record. Ideally, data should be present for all fields, and customers, but some missing data can be tolerated.

Decision

The result of running a strategy in the interaction context. Several decisions can be involved in a single interaction.

Interaction

Some contact with the customer in real time, or offline.

Interval

Typically, the values of numeric predictors are grouped in intervals. Each interval provides a useful building block for understanding behavior.

Model Attributes

Model attributes include various descriptions and settings created during model development. These can be made available to the decision making system at decision time.

Next Best Action

The Next Best Action (NBA) strategy allows applications to take the best decision in a multidimensional context (retention, recruitment, risk, recommendation, etc.).

Outcome

The field representing the behavior to be predicted.

OXL

Omega XML Language. The XML file format of predictive models as published using Predictive Analytics Director.

Population

The group of cases with known behavior, which is consistent with the group of cases whose behavior is to be predicted. In predictive analytics, it is from the population that samples are extracted for modeling and validation.

Prediction

The outcome to be predicted, which is specific to a form of behavior at a given point in time.

Predictive Model

An algorithm that delivers predicted behavior and values for one or more segments segments given the input of the required data about a case. Predictive models are developed in Predictive Analytics Director.

Predictive Performance

Some measure of the scores or segments generated by models. Performance can be measured in terms of predictive power, value, or rate achieved under selected conditions.

Predictive Power

For scoring models, the predictive power is the measure of the ability of a model to separate cases with a positive outcome from those with a negative outcome. These models use behavior defined in terms of two opposite types of outcome, either a symbol indicating which type of behavior, or the probability of being one of the types.

Predictors

Predictors are properties considered to have a predictive relationship with the outcome. Predictors contain information available about the cases whose values may potentially show some association with the behavior you are trying to predict. Examples include:

• Demographic
  For example, age, gender, and marital status.
• Geo-demographic
  For example, home address, and employment address.
• Financial
  For example, income, and expenditure.
• Activity or transaction information
  For example, the amount of loan taken out of the price of the product.

Propensity

The probability of positive behavior or membership.

Proposition

A product offer. By product is meant tangible product offers (a handset, or a subscription), or less tangible ones (benefits, compensations, or services).

Proposition Bundling

Proposition bundling is a method of combining and presenting a number of propositions as a coherent and justifiable set in terms of cross-product eligibility, propensity, and likelihood of interest linked to the call reason. The proposition set is provided in a bundle, such as the cheapest proposition is offered at a reduced price or for free, a discount is given on all propositions, and there are additional free propositions.

Sample

A sub-set of historical data extracted by applying a selection and/or sampling method on the data source. To be meaningful and reliable, it is essential that sufficient records are used and that the distribution of values and patterns of behavior are representative of those in the population.

Score

The value calculated by a model. Score intervals are aggregated under a score band. 

Score Band

A score band is a set of score intervals.

Scorecard

The scorecard decision rule calculates segments by combining a number of properties. The resulting segmentation is translated in a score.

Scoring Model

The value calculated by the model, known as the score, places a case on a numerical scale. High scores are associated with good performance and low scores are associated with bad performance. Typically, the range of scores is broken in intervals of increasing likelihood of one of the two types of behavior (positive, or negative), based on the behavior of the cases in the development sample that fall into each interval.

Segment

A group of customers defined by predicted behavior, score, and characteristics. Segments are implemented through segmentation components in a strategy, and they drive the decision flow by placing a customer in a given segment for which actions/results are defined.

Strategy

The reasoning built up by a set of components that allow you to define the business strategy. A strategy provides the decision support to manage the interaction in the context of the decision hierarchy. Each component has a well defined functionality. A strategy can reference other decision rules (scorecards, predictive models, decision tables, decision trees, adaptive models, and strategies), and import data and propositions.

Trend Detection

Trend detection is possible by comparing the performance of multiple models. To make this possible, the models triggered by the same proposition are configured with different performance window sizes to determine the time frame in number of cases over which the performance is calculated. Implementing trend detection requires a combination of strategy design patterns, and using compatible adaptive model rules with different memory settings.

Value Field

Typically, a value field is data that is known after the time of prediction (for example, the cost of recruitment, purchased value, subsequent probability of up sell or cross sell). Frequently, value fields are frequently associated with one of the outcome values: people who respond are likely to have a cost of sale and a purchased amount, people who do not respond are not.

Wide of Scheme

When other samples are analyzed, values may be detected that were not present in the development sample. From these values, Wide of Scheme cases are formed, and the corresponding values reported separately.

These landing pages support your development and ongoing management of decisioning capabilities.

• Adaptive Models — Provides the facilities to track the performance of adaptive models, train models, and manage the ADM adaptive models in your application.
• Batch — Supports offline strategy executions.
• Strategies — Provides the facilities to define and manage the your application's decision hierarchy.
• Visual Business Director — Allows you to perform historical analysis. VBD enables business users to visualize the business based on different views (proposition, channel, customer, etc.), and examine the success levels down to any level. The Visual Business Director landing page provides the facilities to visualize decision results, and monitor business metrics with a 3D graphical view of the different dimensions and measurements, such as accept and conversion rate, average price, volume, number accepted, and number of processed responses
• Services — Allows you to define the connections to the servers running Interaction Services, Adaptive Decision Manager, and Visual Business Director. Additionally, it allows you to control the update frequency of ADM models.

DSM includes five rule types:

• Adaptive Model rules
• Strategy rules
• Interaction rules
• Predictive Model rules
• Scorecard rules