Sign In Get Started
Table of Contents

Pass/Fail Predictor Guide

Learn how to train AI on your historical inspection data to predict whether new parts will pass or fail, and discover which measurements have the most influence.

Start Predictor

What Is This Tool?

The Pass/Fail Predictor uses machine learning (ML.NET) to analyze your historical inspection data and build a model that can predict whether a new part will pass or fail. It also ranks which measurements have the most influence on the outcome, helping you focus quality improvement efforts on the dimensions that matter most.

When Should I Use This Tool?

  • When you have historical inspection data with multiple measurements and a pass/fail outcome per part
  • To understand which dimensions or characteristics drive pass/fail decisions
  • To predict whether a new part will pass based on its measurements before final inspection

Before You Start

Have the following ready:

  • At least 30 rows of inspection data with numeric measurements and a pass/fail outcome column
  • At least 2 measurement columns (features) and at least 8 pass and 8 fail examples
  • Data in a spreadsheet (Excel, Google Sheets) that you can copy and paste

1 Setup

Give your analysis a title and optionally customize the pass/fail labels. Add any notes about the data source or measurement methods for context.

Tip
Use a descriptive title like "Bearing Assembly Line 2 - June 2025" so you can identify the analysis later.
Click to view screenshot

2 Training Data Entry

Paste your inspection data from a spreadsheet. The tool auto-detects column headers, separators, and the pass/fail outcome column. Review the parsed data preview and column detection before proceeding.

Tip
Copy rows directly from Excel or Google Sheets -- the tab-separated format works automatically. Make sure the first row contains column headers.
Click to view screenshot

3 ML Training Results

The tool trains a machine learning model on your data and shows accuracy metrics, a confusion matrix, and a feature importance chart. The feature importance ranking shows which measurements have the most influence on pass/fail outcomes.

Tip
Focus on the feature importance chart -- the top-ranked features are where quality improvement efforts will have the biggest impact.
Click to view screenshot

4 Predict New Parts

Enter measurement values for a new part and get an instant pass/fail prediction with a confidence probability. The gauge shows green for likely pass, yellow for uncertain, and red for likely fail.

Tip
Try adjusting individual measurement values to see how each dimension affects the prediction. This helps identify critical tolerances.
Click to view screenshot

5 Report & Share

Review the complete analysis summary and download a professional PDF report. Share the report with your team via a link, or use QC-Coach AI for additional interpretation of the results.

Tip
Use the QC-Coach AI to get practical recommendations for improving your process based on the feature importance ranking.
Click to view screenshot

Tips for Best Results

  1. More data gives better results. 50+ rows with a balanced mix of pass and fail outcomes produces the most reliable model.
  2. Include all relevant measurement columns. The model automatically determines which ones matter and which ones don't.
  3. If accuracy is low (below 70%), the pass/fail outcome may depend on factors not captured in your measurement data.
  4. After identifying the top features, use SPC Quick Check to monitor those specific dimensions with control charts.

Rejoining the server...

Rejoin failed... trying again in seconds.

Failed to rejoin.
Please retry or reload the page.

The session has been paused by the server.

Failed to resume the session.
Please retry or reload the page.