---
title: "Get ML Model Summary"
method: GET
path: "/ml/{id}/summary"
tags: ["ML Models"]
---

# Get ML Model Summary

`GET /ml/{id}/summary`

## Request

- Base URL: `https://api.lytics.io/v2`
- URL: `https://api.lytics.io/v2/ml/{id}/summary`
- Auth: API key in header `Authorization`

Get an ML model's summary by ID

## Path parameters

- `id` string, required

## Query parameters

- `account_id` string

## Response `200`

ML Model Summary Response

- object
  - `_meta` object: Response Metadata
  - `data` object: Response Payload
    - `features` ModelsFeature[]: All features of final model build with importances and correlations of each
      - `correlation` number: Correlation coeffient between the feature and the target audience
      - `field_prevalence` ModelsFieldPrevalence
        - `source` number: Field prevalence in the source audience
        - `target` number: Field prevalence in the target audience
      - `impact` ModelsImpact
        - `lift` number: %increase in the target audience when the feature is present
        - `threshold` number: For numeric fields; use for determining lift
        - `value` string: For categorical fields
      - `importance` number: Feature importance as calculated by the model
      - `kind` string: Kind of the feature; segment, lql, score, content, campaign, or unknown
      - `name` string: Name of the feature/data field
      - `type` string: Type of the feature; numeric or categorical
    - `id` string: ID of the model
    - `name` string: Name of the model
    - `state` string: State of the model, "complete", "building", or "invalid"
    - `summary` ModelsMLSummary
      - `accuracy` integer: Measure of 0-10 of how accurate the model is. Interpreted from the R-squared
      - `auc` number: "Area under curve" (0-1) for the ROC (receiver operating characteristic curve). AUC provides an overall measure of performance across all possible thresholds
      - `audience_similarity` number: Jaccard Index/Similarity for source and target segments for a model
      - `error_matrix` ModelsConfusionMatrix
        - `FalseNegative` integer: Predicted negative but actual positive
        - `FalsePositive` integer: Predicted positive but actual negative
        - `TrueNegative` integer: Predicted negative and actual negative
        - `TruePositive` integer: Predicted positive and actual positive
      - `model_health` string: Health of the model; "healthy" or "unhealthy". Unhealthy models are not accurate enough or encountered an error during training
      - `mse` number: Mean squared error is the average squared difference between the estimated values and the actual value
      - `msgs` ModelsModelMessage[]: Messages to help debug and improve the model
        - `severity` string: Severity level of the message; "info", "warn", "error", or "debug"
        - `tags` string[]: Tags for the message; "summary", "workflow", "autotune", "collinearity", or "verdict"
        - `text` string: Model message
      - `reach` integer: Measure of 0-10 of how many users this model can reach. Interpreted from the False Positive Rate
      - `rsq` number: R-squared is a measure (0-1) of "goodness of fit", i.e. how well the model fits the data
      - `source_predictions` object: Test dataset predictions for the source audience
      - `target_predictions` object: Test dataset predictions for the target audience
      - `threshold` number: Computed as the optimal threshold to use when creating predictive audiences. The value that optimizes both reach and accuracy simultaneously
  - `request_id` string
  - `status` integer: HTTP Status Code

## Other responses

- `400`: Bad Request
- `404`: Not Found
- `500`: Internal Server Error

## Changes

- **2026-11-03** `d8e975057aec`: 1 info
  - endpoint added

[Operation history](https://skmtc.dev/lytics/apis/lytics-api/changes/ml/:id/summary/get.md)

---

[API](https://skmtc.dev/lytics/apis/lytics-api.md) - [All operations](https://skmtc.dev/lytics/apis/lytics-api/llms.txt) - [OpenAPI document](https://skmtc.dev/lytics/apis/lytics-api/revisions/d8e975057aec?raw)