What is a model in data science?

A simplified representation of how variables are related

In the equation Price = b0 +b1(Size) + b2(Bedrooms), what does b1 represent?

The change in predicted price for 1 more unit of Size, holding Bedrooms constant

Why is it a problem to include a variable just because it's a number?

Some numbers, like house numbers or jersey numbers, have no meaningful relationship with the outcome

A model finds that people with longer commutes ear higher salaries. What is the best conclusion?

Commute distance is associated with salary, but that doesn't prove one causes the other

You leave an important variable, like house condition, out of a price model. What is most likely to happen?

The model is less accurate, and other coefficients may absorb its effect

Why do we split data into a training set and a testing set?

To check how well the model works on data it hasn't seen before

Which is the correct order of steps?

Split -> Train -> Test -> Predict on new data

A student evaluates their model using the same data they trained it on. What is the problem?

The score will probably look better than the model's real performance

What does an R^2 of 0.70 mean for a regression model

The model explains 70% of the variation in the target

A disease affects 1 in 100 people. A model predicts "no disease" for everyone and gets 99% accuracy. What is the problem?

It never finds anyone who actually has the disease, so accuracy along is misleading

A model predicting loan approval doesn't use race, but it does use ZIP code. Why might it still be unfair?

ZIP code can act as a proxy for race because of segregated neighborhoods

A heart-attack model works well for men but misses more heart attacks for women. What is the best next step?

Check performance separately for each group and investigate why

Why might removing "sex" from a medical diagnosis model make it worse?

Symptoms of some conditions differ by sex, so the variable carries real medical information

Help

Options

focusNode

Didn't know it?
click below

Knew it?
click below

Don't Know

Remaining cards (0)

Know

retry

shuffle

restart

0:00

Embed Code - If you would like this activity on your web page, copy the script below and paste it into your web page.

Normal Size Small Size show me how

Data & Society Final

Question	Answer
What is a model in data science?	A simplified representation of how variables are related
A regression model predicts house prices. After it's fitted, we give it the size and bedrooms of a new house. What is the model doing?	Making a prediction
In the equation Price = b0 +b1(Size) + b2(Bedrooms), what does b1 represent?	The change in predicted price for 1 more unit of Size, holding Bedrooms constant
You want to predict a student's exam score. Which variable is the WORST choice as a predictor?	Student ID number
Why is it a problem to include a variable just because it's a number?	Some numbers, like house numbers or jersey numbers, have no meaningful relationship with the outcome
A model finds that people with longer commutes ear higher salaries. What is the best conclusion?	Commute distance is associated with salary, but that doesn't prove one causes the other
You leave an important variable, like house condition, out of a price model. What is most likely to happen?	The model is less accurate, and other coefficients may absorb its effect
Why do we split data into a training set and a testing set?	To check how well the model works on data it hasn't seen before
A model scores 98% accuracy on training data but only 60% on test data. What is this called?	Overfitting
Which is the correct order of steps?	Split -> Train -> Test -> Predict on new data
A student evaluates their model using the same data they trained it on. What is the problem?	The score will probably look better than the model's real performance
What does an R^2 of 0.70 mean for a regression model	The model explains 70% of the variation in the target
A disease affects 1 in 100 people. A model predicts "no disease" for everyone and gets 99% accuracy. What is the problem?	It never finds anyone who actually has the disease, so accuracy along is misleading
A model predicting loan approval doesn't use race, but it does use ZIP code. Why might it still be unfair?	ZIP code can act as a proxy for race because of segregated neighborhoods
A heart-attack model works well for men but misses more heart attacks for women. What is the best next step?	Check performance separately for each group and investigate why
Why might removing "sex" from a medical diagnosis model make it worse?	Symptoms of some conditions differ by sex, so the variable carries real medical information

Created by: user-2063460

"Know" box contains:
Time elapsed:
Retries: