Mean Absolute Percentage Error Calculator

Calculate MAPE, inspect row errors, compare models, manage zero values, visualise predictions, and export reliable machine learning evaluation reports with confidence in one place.

Calculator inputs

Use CSV, tab-separated data, semicolons, pipes, or JSON. Include label, actual, one or more model columns, and optional weight.
Use zero to disable the cap.

Interpretation thresholds

How to use

  1. Enter a header row containing actual values and prediction columns.
  2. Add optional labels and weights for clearer analysis.
  3. Select the preferred zero-value handling method.
  4. Choose rounding, caps, signs, and interpretation thresholds.
  5. Submit the form to calculate every model metric.
  6. Review warnings, row errors, rankings, and charts.
  7. Copy, print, or export the final evaluation.

Example data format

label actual Model A Model B weight
Week 11001101041
Week 22001902051
Week 31501651481
Week 42502302521

Understanding MAPE

MAPE measures average absolute prediction error as a percentage. It is easy to compare across similarly scaled datasets. Small actual values can create misleadingly large percentages.

Zero actual values make standard MAPE mathematically undefined. This calculator offers several transparent handling methods. Always report the selected method with your result.

MAPE should not be judged without supporting metrics. MAE and RMSE reveal errors in original units. sMAPE and WAPE can improve difficult comparisons considerably.

Frequently asked questions

What is a good MAPE value?

A good value depends on data and decisions. Lower MAPE usually indicates better percentage accuracy. Compare against baselines and business requirements before judging.

Why is MAPE undefined for zero actual values?

The formula divides each error by its actual value. Division by zero has no finite result. Select a documented alternative method for those observations.

Can MAPE exceed 100 percent?

Yes, large errors can exceed the actual value. Their percentage errors can exceed one hundred percent. This often occurs near zero actual values.

Does MAPE support negative actual values?

The calculator can include negative actual values optionally. Their interpretation may become confusing in many contexts. Consider MAE, RMSE, or sMAPE as alternatives.

What is the difference between MAPE and MAE?

MAPE expresses average error as a percentage. MAE keeps the original measurement units instead. MAE remains defined when actual values equal zero.

What is sMAPE?

sMAPE uses actual and predicted magnitudes together. It reduces some asymmetry found in standard MAPE. Its denominator can still require careful interpretation.

What is WAPE?

WAPE divides total absolute error by total demand. It handles aggregation differently from averaging row percentages. Large observations therefore receive greater practical influence.

Why compare several metrics?

Every metric emphasises different error behaviour. RMSE penalises large misses more strongly. Multiple metrics provide a more reliable evaluation picture.

Can I compare multiple models?

Yes, add each model as another prediction column. The calculator ranks valid models using MAPE. Detailed charts and metrics support deeper comparison.

Related Calculators

Linear Regression CalculatorPolynomial Regression CalculatorRidge Regression CalculatorLasso Regression CalculatorElastic Net Regression CalculatorLogistic Regression CalculatorRegression Coefficient CalculatorRegression Prediction CalculatorResidual Error CalculatorMean Squared Error Calculator

Important Note: All the Calculators listed in this site are for educational purpose only and we do not guarentee the accuracy of results. Please do consult with other sources as well.