Skip to content
Oak National Academy
Oak National Academy
View my library
0
Teachers
Pupils
New
New
Year 9

Improving a model

I can make changes to how my model is trained to improve the accuracy of the output.

Download all resources
Share activities with pupils
New
New
Year 9

Improving a model

I can make changes to how my model is trained to improve the accuracy of the output.

Download all resources
Share activities with pupils

These resources will be removed by end of Summer Term 2025.

Switch to our new teaching resources now - designed by teachers and leading subject experts, and tested in classrooms.

These resources were created for remote use during the pandemic and are not designed for classroom teaching.

View new resources

Lesson details

Key learning points

  1. Data cleaning can remove outliers so the model "learns" real patterns instead of noise.
  2. False positives occur when data is misclassified.
  3. Creating a null class reduces the chances of misclassification.

Keywords

  • Misclassification - when a model predicts the wrong class for a sample

  • Data cleaning - the process of removing incorrect, inconsistent or irrelevant data to improve the quality of a data set before training a model

  • Noise - unwanted or random variations in data that don’t represent meaningful patterns

  • False positive - a sample that has been incorrectly classified

Common misconception

More data is always better as it means the model can be better informed to make a prediction.

Poorly labelled or inconsistent data can make a model's predictions inaccurate. It is better to use a smaller amount of higher quality data to make more accurate predictions.


To help you plan your year 9 computing lesson on: Improving a model, download all teaching resources for free and adapt to suit your pupils' needs...

File needed for this lesson

  • microbit-Oak 1.29 MB (HEX)

Download this file to use in the lesson.

Download lesson file
A null class is a concept that learners may find difficult. Model the importance of recording a null class when training a model.
Teacher tip

Equipment

Pupils will need micro:bits for this lesson and access to a device that can access CreateAI online.

Licence

This content is © Oak National Academy Limited (2025), licensed on Open Government Licence version 3.0 except where otherwise stated. See Oak's terms & conditions (Collection 2).

Lesson video

Skip lesson video

Loading...

Prior knowledge starter quiz

Download quiz pdf

6 Questions

Q1.
A machine learning model can make incorrect predictions if the training data has .
Correct Answer: bias, Bias
Q2.
A model’s score tells us how likely the model’s prediction is correct.
Correct answer: confidence
accuracy
certainty
Q3.
Match the problem with its potential cause in machine learning.
Correct Answer:incorrect predictions,poor-quality training data

poor-quality training data

Correct Answer:unexpected results,presence of outliers

presence of outliers

Correct Answer:model favours certain outcomes,bias in the data set

bias in the data set

Correct Answer:low confidence score,data does not fit the trained model

data does not fit the trained model

Q4.
A classifier learns by analysing between samples and grouping them into classes.
Correct Answer: patterns
Q5.
What is one way to improve a machine learning model’s accuracy?
decreasing the amount of training data
increasing the amount of training data
Correct answer: adding more diverse and well-labelled training data
removing all boundaries between classes
Q6.
Which of the following best describes an outlier in a data set?
a missing value that has no impact on model performance
Correct answer: a data point that does not follow the general pattern of the rest of the data
a data point that is identical to all others

Assessment exit quiz

Download quiz pdf

6 Questions

Q1.
What is the main benefit of training an AI model on diverse data?
it makes the model more complex
it helps the model learn faster
Correct answer: it reduces bias in the model's predictions
it makes the model easier to build
Q2.
Cleaning data enhances the of a machine learning model.
Correct Answer: accuracy, ccuracy
Q3.
Match the following terms with their descriptions.
Correct Answer:data cleaning,removing irrelevant data

removing irrelevant data

Correct Answer:noise,unwanted variations in data

unwanted variations in data

Correct Answer:misclassification,when a model predicts the wrong class

when a model predicts the wrong class

Q4.
What is an outlier in data?
a very important data point
a data point that is similar to others
Correct answer: a data point that is significantly different from others
a data point that is always correct
Q5.
Adding a null class helps prevent in machine learning models
Correct Answer: misclassification, isclassification
Q6.
Why is it important to include data from people with different movement styles when training a model to classify movements?
it makes the model look more impressive
it makes the model run faster
it makes the model smaller
Correct answer: it ensures the model can recognise valid movements from diverse individuals