a good performance on training data and poor performance on unseen data

What are the three major subgroups in Machine Learning?

Supervised learning, unsupervised learning and reinforcement learning

Machine Learning or otherwise called Artificial Intelligence got very popular over the last several years, mainly becaus

machine learning methods have shown unprecedented results in many areas

What is the most important feature of supervised learning methods?

They need labelled data to work

What is the main difference between classification and regression algorithms?

classification predict distinct classes (categories)

What K in K-Nearest Neighbour stands for?

K is the number of neighbours used when predicting a class for a new point

What is the best way to compare machine learning models?

Decide based on results of CV algorithm

No free lunch theorem for machine learning (Wolpert, 1996) states that

no ML algorithm is better than any other across all possible datasets

What is the interpretation of slope, here?

The effect of each additional HW point on final grade

studbud

Supervised learning

11 termsStudBud Administration1,015 plays

Imported from external platform

Flashcards11 cardsFPractice quizPHost liveL

Questions 11

Answers

1
Overfitting is ...
a strategy to make complex models (like the decision tree) more robust
a good performance on training data and poor performance on unseen data
the idea that simpler models should be used more frequently than complex
a problem of applying linear regression algorithm to non-linear data
2
What are the three major subgroups in Machine Learning?
Supervised learning, unsupervised learning and classification methods
Classification, regression, unsupervised and reinforcement learning
Supervised learning, unsupervised learning and reinforcement learning
Regression algorithms, classification methods and clustering
3
Machine Learning or otherwise called Artificial Intelligence got very popular over the last several years, mainly becaus
machine learning methods have shown unprecedented results in many areas
everyone likes science fiction books which feature an omnipotent AI
AI had been good at drawing in Van Gogh style
AI has been selling good to investors at conferences and startup events
4
What is the most important feature of supervised learning methods?
They need labelled data to work
They work better than unsupervised methods
They don't require much data to work
There are very few of them
5
What is the main difference between classification and regression algorithms?
regression help to identify hidden groups in data
classification does not require labeled data to work
classification predict distinct classes (categories)
no difference, essentially the same
6
What K in K-Nearest Neighbour stands for?
K is the number of points needed to ensure high accuracy
K is the number of neighbours used when predicting a class for a new point
K is the number of classes that the Nearest Neighbour can predict
K is the number of times one would need to re-run the Nearest Neighbour
7
What are the most reasonable values for Y and Z in the following example?
Y = 5 and Z = 4
Y = 4 and Z = 5
Y = 3.5 and Z = 4
Y = 4 and Z = 4
8
What is the best way to compare machine learning models?
Decide based on MSE on train
Decide based on test set performance
Decide based on MSE on val
Decide based on results of CV algorithm
9
The decision tree algorithm is similar to the natural tree as they both
use distance travelled to predict ride price
suffer from parasites
have leaves and the root
tend to grow very large
10
No free lunch theorem for machine learning (Wolpert, 1996) states that
no ML algorithm is better than any other across all possible datasets
universally good ML algorithm is hard to design and implement
no ML algorithm can be fully trusted
no ML algorithm should be used on unseen data
11
What is the interpretation of slope, here?
The effect of each additional HW point on final grade
Final grade with 0 HW points ("base grade")
The effect of additional point in final grade on HW points
It shows the type of relationship between sum of HW point and final grade

Supervised learning Flashcards | StudBud