Prediction Type Definition

[This topic is pre-release documentation and is subject to change.]

A Prediction represents the outcome of a machine learning model analysis on a data set.

Prediction Properties

Property names appended with asterisks* are required.

Property JSON type Description
predictionName* string Name of the prediction
displayName dictionary Localized display names for the link
description string Localized description of the link
tenantId string Name of the owning hub (readonly)
involvedInteractionTypes string[ ] Interaction types involved in the prediction
involvedKpiTypes string[ ] KPI types involved in the prediction
involvedRelationships string[ ] Relationships involved in the prediction
negativeOutcomeExpression* string Negative outcome expression
positiveOutcomeExpression* string Positive outcome expression
primaryProfileType* string Primary profile type
provisioningState string ["Provisioning" | "Succeeded" | "Expiring" | "Deleting" | "HumanIntervention" | "Failed" ]
scopeExpression* string Scope expression
autoAnalyze* boolean True if analysis is automatically performed; otherwise false
scoreLabel* string Score label
mappings* {string score, string grade, string reason}[ ] Link mapping definitions for the prediction
grades {string gradeName, integer minScoreThreshold, integer maxScoreThreshold}[ ] Grades defined for this prediction
systemGeneratedEntities {string generatedInteractionTypes[ ], string generatedLinks[ ], generatedKpis} System generated entities