> ## Documentation Index > Fetch the complete documentation index at: https://docs.orbit.devotel.io/llms.txt > Use this file to discover all available pages before exploring further. # Train a predictive model > Fit the named predictive model over the tenant's current profiles and return a training report — algorithm, training sample size, evaluation metrics (AUC / log-loss / accuracy for classification, R² / RMSE / MAE for regression), the learned intercept, and the standardised per-feature coefficients (feature importances). The fit is cached briefly per (tenant, model) so an immediately-following score call reuses it. When there are too few resolved-outcome samples to fit, the endpoint returns `422` with a `status: "insufficient_data"` report instead of a trained one. Owner / admin / developer only. ## OpenAPI ````yaml /openapi.yaml post /api/v1/cdp/predictive-models/{model}/train openapi: 3.1.0 info: title: Devotel CPaaS API description: Orbit by Devotel — Communications Platform as a Service API version: 1.0.0 contact: name: Devotel url: https://devotel.io email: support@devotel.io license: name: Proprietary servers: - url: https://api.orbit.devotel.io description: Production security: - Bearer: [] - ApiKey: [] tags: - name: Messages description: >- Send and manage messages across all channels (SMS, WhatsApp, RCS, Email, Viber, etc.) - name: Fax description: >- List and track fax (MMS/T.38) transmissions on Telnyx-backed fax numbers (sending flows through the Messaging API) - name: Agents description: AI agent creation, configuration, and execution - name: Voice description: Voice calls, IVR, conferencing, and SIP trunking - name: OnCall description: On-call rotations and escalation policy planning for incident alerting - name: Webhooks description: Webhook endpoint management and delivery logs - name: Numbers description: Phone number search, provisioning, and configuration - name: Brand Identity description: >- Unified cross-channel brand trust posture (10DLC, toll-free, WhatsApp, RCS, branded calling, number KYC) - name: Contacts description: Contact management, segmentation, and lifecycle tracking - name: Campaigns description: Marketing campaign orchestration and analytics - name: Flows description: Automation flow builder and execution engine - name: Templates description: Message template management and approval workflows - name: Settings description: Organization, channel, and user preference settings - name: Verify description: OTP generation and verification across channels - name: Push description: Push notification delivery via FCM and APNs - name: Integrations description: Third-party service connections and OAuth management - name: CDP description: >- Customer Data Platform — activation surface (CRM object sync, streaming destinations, ad-audience activation) - name: Files description: >- Server-to-server media upload, listing, retrieval, and deletion (signed-URL backed) - name: Commerce description: >- Omnichannel conversational-commerce — persistent cart + checkout state machine, channel-agnostic hosted pay-by-link, native WhatsApp checkout, and AP2-style agent payment mandates. - name: Sync description: >- Real-time shared-state primitive (Twilio Sync parity) — Documents, Maps, Lists, and ephemeral Streams, with change events relayed over the /api/v1/ws/sync WebSocket gateway - name: Risk description: >- Unified cross-channel Trust & Fraud risk scoring — fuses SMS-pumping, URL-reputation, Verify Fraud Guard, and Voice Biometrics signals into one composite verdict queryable before a send or a call. - name: Orby description: >- In-dashboard Orby operator assistant: streamed assistant turns, conversation threads, product knowledge-base search, and the tool-action approval gate. Available to signed-in operators only (dashboard session auth — API-key requests are rejected). paths: /api/v1/cdp/predictive-models/{model}/train: post: tags: - CDP summary: Train a predictive model description: >- Fit the named predictive model over the tenant's current profiles and return a training report — algorithm, training sample size, evaluation metrics (AUC / log-loss / accuracy for classification, R² / RMSE / MAE for regression), the learned intercept, and the standardised per-feature coefficients (feature importances). The fit is cached briefly per (tenant, model) so an immediately-following score call reuses it. When there are too few resolved-outcome samples to fit, the endpoint returns `422` with a `status: "insufficient_data"` report instead of a trained one. Owner / admin / developer only. parameters: - schema: type: string enum: - churn_propensity - conversion_intent - lifetime_value - engagement_fatigue in: path name: model required: true description: Model key from `GET /api/v1/cdp/predictive-models`. responses: '200': description: The model was fitted; a training report is returned. content: application/json: schema: type: object properties: data: $ref: '#/components/schemas/CdpModelTrainReport' meta: $ref: '#/components/schemas/ResponseMeta' '422': description: >- Not enough resolved-outcome samples to fit the model (status `insufficient_data`). components: schemas: CdpModelTrainReport: type: object description: >- A predictive-model training report. When `status` is `trained` it carries the evaluation metrics, learned intercept, and per-feature coefficients; when `status` is `insufficient_data` it carries the reason and the sample count that fell short. required: - status - model - kind properties: status: type: string enum: - trained - insufficient_data model: type: string kind: type: string enum: - classification - regression algorithm: type: string output_unit: type: string feature_names: type: array items: type: string training: type: object description: >- Evaluation metrics over the training population (present when trained). AUC/log_loss/accuracy/positive_rate for classification; r2/rmse/mae for regression. additionalProperties: true properties: sample_count: type: integer intercept: type: number coefficients: type: array description: Standardised per-feature weights (feature importances). items: type: object properties: feature: type: string weight: type: number reason: type: string description: Present only when `status` is `insufficient_data`. sample_count: type: integer description: Present only when `status` is `insufficient_data`. ResponseMeta: type: object required: - request_id - timestamp additionalProperties: false properties: request_id: type: string description: Unique request identifier (also returned as X-Request-Id header) timestamp: type: string format: date-time docs_url: type: string format: uri nullable: true securitySchemes: Bearer: type: http scheme: bearer bearerFormat: JWT description: Dashboard JWT token from Clerk ApiKey: type: apiKey name: X-API-Key in: header description: Server-to-server API key (dv_live_sk_*) ````