Describe a Hypothesis Space
A dispatcher proposes a rule using route, weather, and departure hour.
Describe a Hypothesis Space explained visually
A dispatcher proposes a rule using route, weather, and departure hour. English captions are optional and start off. The complete transcript follows the written lesson.
What you will learn
A dispatcher proposes a rule using route, weather, and departure hour.
A practical question
Our dispatcher wants a rule for predicting late deliveries. One option asks only whether it is raining. Another combines weather, route, and departure hour. Before fitting anything, we have already made a choice about what kinds of rules the learner is allowed to consider.
The useful idea
That collection of possible rules is the hypothesis space. If the space only contains single-condition rules, it cannot express 'rain on the hill route after five' even if the data strongly supports that interaction. A huge space can represent more patterns but also memorize peculiar training trips.
Work through the example
Write down two candidate rules and ask which future trips each could classify. Then check whether your chosen learning method can actually represent both. It cannot discover a relationship outside its available set of models.
Limits and next step
Keep the task and rule language clear before comparing scores. Next, we'll see how tightening or loosening a rule changes the deliveries it covers.
Worked example
Compare a one-condition rule with one that combines route and weather.
Quick check: test your understanding
Try it yourself: In the situation described above, explain the method and one reason its output might be misleading.
Check the reasoning
A hypothesis space is the set of candidate rules the learning method can represent. The space's design limits what can be learned even from many examples.
What is the main idea behind describe a hypothesis space?
A hypothesis space is the set of candidate rules the learning method can represent.
How is it used in the example?
Compare a one-condition rule with one that combines route and weather.
What limitation should be checked?
The space's design limits what can be learned even from many examples.
Further reading
Machine Learning — Tom M. Mitchell.
Original ToolsInVision teaching examples. Editorial approach · Suggest a correction
Accessible referenceLesson 17 transcript150 words · Video transcript · Matches the published narration and captionsOpen transcript
Our dispatcher wants a rule for predicting late deliveries. One option asks only whether it is raining. Another combines weather, route, and departure hour. Before fitting anything, we have already made a choice about what kinds of rules the learner is allowed to consider.
That collection of possible rules is the hypothesis space. If the space only contains single-condition rules, it cannot express 'rain on the hill route after five' even if the data strongly supports that interaction. A huge space can represent more patterns but also memorize peculiar training trips.
Write down two candidate rules and ask which future trips each could classify. Then check whether your chosen learning method can actually represent both. It cannot discover a relationship outside its available set of models.
Keep the task and rule language clear before comparing scores. Next, we'll see how tightening or loosening a rule changes the deliveries it covers.