# Return Reason Label and Prediction Audit

> Return Reason Label and Prediction Audit is a paid API for AI agents from market.datapackvibe.com, paid per call via x402, $0.01/call, status unknown (last checked 2026-10-15).

Audits up to 200 labelled text records for return-reason classification accuracy, computing evaluation metrics against actual vs. predicted labels across four categories.

## Facts

- Endpoint: POST https://market.datapackvibe.com/x402/assistant-sector-ml-return-reason-label-audit?utm_source=zero.xyz
- Price: $0.01/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-15
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/return-reason-label-and-prediction-audit-f7982dec
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_E51GCII-UlsYoFPU32hFe

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability return-reason-label-and-prediction-audit-f7982dec -d '<json body>'
```

Example prompt: Can you audit my return reason classifier on these 50 labelled records — each has the customer text, the actual reason (damaged, wrong_item, unwanted, or other), the predicted label, and the four class probabilities — and tell me where the model is going wrong?

## When to prefer this

Choose this endpoint when you need a lightweight, low-cost ($0.01) audit of a return-reason text classifier on up to 200 pre-labelled records. Ideal for spot-checking model quality, validating label consistency, or evaluating naive-Bayes baselines without access to a hosted ML platform. Best suited for e-commerce or logistics teams tracking return categorization accuracy across the four standard classes: damaged, wrong_item, unwanted, and other.

## Known failure modes

- More than 200 records supplied — batch is capped at 200
- Missing or malformed probability arrays (must have exactly 4 values summing to ~1.0)
- Invalid actual or predicted label values outside the four allowed enum values
- Records with text exceeding 1000 characters or IDs exceeding 80 characters
- Empty records array returns no meaningful audit output

## How this service works

Return reason - Label and prediction audit: Label and prediction audit for Return reason. Up to 200 supplied labelled text records. Local dataset tools, evaluation and small naive-Bayes baselines; no hosted model or foundation-model training. One of 25 profiles or calculations across 20 shared operations in this sector. 0.01 USDC includes one bounded label and prediction audit deliverable.

## Output

A bounded audit report covering the supplied labelled records, including accuracy metrics, per-class performance breakdown, comparison of actual vs. predicted return reason labels (damaged, wrong_item, unwanted, other), and identification of misclassified records based on the supplied class probabilities.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "records": {
   "type": "array",
   "items": {
    "type": "object",
    "required": [
     "id",
     "text",
     "actual",
     "predicted",
     "probabilities"
    ],
    "properties": {
     "id": {
      "type": "string",
      "default": "",
      "maxLength": 80,
      "minLength": 1
     },
     "text": {
      "type": "string",
      "default": "",
      "maxLength": 1000
     },
     "actual": {
      "enum": [
       "damaged",
       "wrong_item",
       "unwanted",
       "other"
      ],
      "type": "string"
     },
     "predicted": {
      "enum": [
       "damaged",
       "wrong_item",
       "unwanted",
       "other"
      ],
      "type": "string"
     },
     "probabilities": {
      "type": "array",
      "items": {
       "type": "number",
       "default": 0,
       "maximum": 1,
       "minimum": 0
      },
      "default": [],
      "maxItems": 4,
      "minItems": 4
     }
    },
    "additionalProperties": false
   },
   "default": [
    {
     "id": "example-0-0",
     "text": "damaged sample document number 0 for this supplied dataset.",
     "actual": "damaged",
     "predicted": "damaged",
     "probabilities": [
      0.8,
      0.06666666666666667,
      0.06666666666666667,
      0.06666666666666667
     ]
    },
    {
     "id": "example-0-1",
     "text": "damaged sample document number 1 for this supplied dataset.",
     "actual": "damaged",
     "predicted": "damaged",
     "probabilities": [
      0.8,
      0.06666666666666667,
      0.06666666666666667,
      0.06666666666666667
     ]
    },
    {
     "id": "example-0-2",
     "text": "damaged sample document number 2 for this supplied dataset.",
     "actual": "damaged",
     "predicted": "damaged",
     "probabilities": [
      0.8,
      0.06666666666666667,
      0.06666666666666667,
      0.06666666666666667
     ]
    },
    {
     "id": "example-1-0",
     "text": "wrong_item sample document number 0 for this supplied dataset.",
     "actual": "wrong_item",
     "predicted": "wrong_item",
     "probabilities": [
      0.06666666666666667,
      0.8,
      0.06666666666666667,
      0.06666666666666667
     ]
    },
    {
     "id": "example-1-1",
     "text": "wrong_item sample document number 1 for this supplied dataset.",
     "actual": "wrong_item",
     "predicted": "wrong_item",
     "probabilities": [
      0.06666666666666667,
      0.8,
      0.06666666666666667,
      0.06666666666666667
     ]
    },
 
… (truncated)
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/return-reason-label-and-prediction-audit-f7982dec/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from market.datapackvibe.com](https://www.zero.xyz/host/market.datapackvibe.com/llms.txt)