• CLEAN
  • OUTLIER
  • HUMAN
  • REPORT

How to clean datasets for machine learning with your AI

Connect the Outlier Detection Engine to your AI client once. From there, your agent handles the statistical heavy lifting to isolate noise in your data.

Ask AI about this page

Short answer

How can I clean my datasets using AI?

Your agent uses mathematical outlier detection to identify and flag data points that deviate from your dataset's distribution. This removes noise and errors before you train your model. You'll see the specific outcomes below.

The results of the scan.

Where your data lands.

The engine processes your raw files and sorts the results into these categories.

CLEAN

Validated data points

The AI identifies points that fit your statistical distribution. You get a clean subset ready for training.

OUTLIER

Detected anomalies

The engine flags points that fall outside expected mathematical bounds. These are isolated for your review.

HUMAN

Manual review required

Edge cases that sit right on the boundary of being an outlier are sent to you. You decide if they stay or go.

REPORT

Statistical summary

You receive a breakdown of the distribution and the specific reasons why certain points were flagged.

The workflow

What your AI does when the dataset arrives.

The AI takes over the heavy lifting of statistical analysis so you don't have to write custom scripts.

  1. Scan the dataset

    Your agent reads the provided file and maps the distribution of all numerical features.

    detect_outliers

  2. Calculate deviations

    The engine runs mathematical checks to find points that deviate significantly from the mean or median.

    detect_outliers

  3. Flag anomalies

    The AI marks specific rows that meet the outlier criteria, separating them from the valid data.

    detect_outliers

  4. Generate clean set

    The agent prepares a version of your dataset with the erroneous points removed.

    detect_outliers

Try it

Copy these to start.

Use these prompts to trigger the cleaning process.

Starting points

Replace the bracketed text with your actual file names or data sources.

Outlier Detection Engine Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_dfkaEjEHL0xoHc706KKIKrxSgf6YRwqXhyMOeE7z/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Outlier Detection Engine capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "outlier-detection-engine-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_dfkaEjEHL0xoHc706KKIKrxSgf6YRwqXhyMOeE7z/mcp"
    }
  }
}
Copy into chat04

  • Run outlier detection on my training_data.csv file.

  • Identify any statistical anomalies in the user_logs.json dataset.

  • Clean the sensor_readings.csv file by removing outliers.

  • Check the customer_metrics.xlsx sheet for erroneous data points.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Outlier Detection Engine once, then ask.

You only need to link the MCP to your client once. Your credentials stay encrypted, and you can immediately start asking your agent to process your files.

Connect Outlier Detection Engine to your AI

FAQ

How this task behaves.

  • 01

    Will the AI delete my original files?

    No. The AI only reads your data and generates new, cleaned versions or reports. Your original files remain untouched.

  • 02

    Can the AI decide which outliers to keep?

    No. The AI identifies outliers based on mathematical rules. You must review the flagged points and decide which ones to keep or discard.

  • 03

    What kind of files can I use?

    The engine works with standard data formats like CSV, JSON, and Excel files.

  • 04

    Does it work on text data?

    No. This specific tool is designed for numerical and statistical outlier detection.

  • 05

    How does it know what an outlier is?

    It uses mathematical models to find data points that are statistically distant from the rest of the set.

  • More questions about Outlier Detection Engine? The Connector page answers them. See everything the Outlier Detection Engine Connector can do

Connect Outlier Detection Engine to Claude, Cursor, ChatGPT & more