Error classification in GraphQL - Step-by-Step Execution

Concept Flow - Error classification

Input Data

↓

Model Prediction

↓

Compare Prediction vs Actual

↓

Calculate Error

↓

Classify Error Type

↓

Bias Error

↓

Variance Error

↓

Irreducible Error

The process starts with input data and model prediction, then compares prediction to actual values to calculate error, which is then classified into types like bias, variance, or irreducible error.

Execution Sample

GraphQL

actual = [3, -0.5, 2, 7]
predicted = [2.5, 0.0, 2, 8]
errors = [a - p for a, p in zip(actual, predicted)]
mean_error = sum(errors) / len(errors)

This code calculates the error between actual and predicted values and finds the average error.

Execution Table

Step	Action	Values	Result
1	Set actual values	[3, -0.5, 2, 7]	actual = [3, -0.5, 2, 7]
2	Set predicted values	[2.5, 0.0, 2, 8]	predicted = [2.5, 0.0, 2, 8]
3	Calculate errors element-wise	errors = [3-2.5, -0.5-0.0, 2-2, 7-8]	errors = [0.5, -0.5, 0, -1]
4	Sum errors	sum = 0.5 + (-0.5) + 0 + (-1)	sum = -0.5
5	Calculate mean error	mean_error = -0.5 / 4	mean_error = -0.125

💡 All errors calculated and mean error computed as -0.125

Variable Tracker

Variable	Start	After Step 3	After Step 5
actual	[ ]	[3, -0.5, 2, 7]	[3, -0.5, 2, 7]
predicted	[ ]	[2.5, 0.0, 2, 8]	[2.5, 0.0, 2, 8]
errors	[ ]	[0.5, -0.5, 0, -1]	[0.5, -0.5, 0, -1]
mean_error	N/A	N/A	-0.125

Key Moments - 2 Insights

Why do some errors have negative values?

Errors are actual minus predicted values; if predicted is larger than actual, the error is negative, as shown in step 3 of the execution_table.

Is mean error always positive?

No, mean error can be negative if predictions tend to be higher than actual values, as seen in step 5 where mean_error is -0.125.

Visual Quiz - 3 Questions

Look at the execution_table at step 3, what is the error for the second element?

A 0.5

B -0.5

C 0

D -1

💡 Hint

Check the 'Values' column at step 3 for the second error calculation.

At which step is the mean error calculated?

A Step 5

B Step 2

C Step 4

D Step 3

💡 Hint

Look for the step where mean_error is assigned a value.

If predicted values were all equal to actual values, what would the errors list look like?

A [-1, -1, -1, -1]

B [1, 1, 1, 1]

C [0, 0, 0, 0]

D Cannot determine

💡 Hint

Errors are actual minus predicted; if they are equal, difference is zero.

Concept Snapshot

Error classification in data science means comparing model predictions to actual values.
Calculate errors by subtracting predicted from actual.
Errors can be positive or negative.
Mean error summarizes average difference.
Classify errors into bias, variance, or irreducible types.

Full Transcript

We start with actual and predicted values. We subtract predicted from actual to get errors for each data point. Some errors are positive, some negative, depending on whether prediction is lower or higher than actual. Then we sum these errors and divide by the number of points to get the mean error. This mean error helps us understand if the model tends to over or under predict. Finally, errors are classified into types like bias error (systematic error), variance error (model sensitivity), and irreducible error (noise).