In Bayes Theorem, Your Hypothesis is...

A particular outcome you are interested in

Train multiple models on different on subsets of the data, Use ensemble methods

Polynomial Regression

Tends to overfit, Fits the data into a polynomial function

In Bayes Theorem, Evidence refers to...

Feature values that could be related to a specific outcome

Multiple Regression Uses...

Multiple input (feature) variables

In Bayes Theorem, The Set of Outcomes are ...

All of the possible events that could happen

Linear Regression, y=mx+b. M is the...

Multiplicative Coefficient/Slope

Are the inputs to the model, available in both training and runtime

In Supervised Machine Learning

You have outcome data in your training/test sets, Model learn based on correct "ground-truth answers"

Use information theory to split variables, Are fairly powerful, Are the basis of Random Forest models, Create flowchart-like AI models

Logistic Regression is good when...

You have only two classes and they are pretty balanced in the dataset, You want the output to between 0 and 1

Decision trees decide what feature to branch on based on...

Gini Index, Information Gain, Entropy

Train multiple weak models sequentially to reduce errors, Are ensemble methods

Which is true about Linear Regression?

It creates a model of the form y = mx+b

Train multiple decision trees, and use the trees together to get an answer, Are an ensemble technique

Bagging and Boosting Models...

Are typically implemented with decision trees, Are ensemble methods

Linear Regression, y=mx+b. B is the...

Slope Intercept (Additive Graph Offset)

Are what the model generates, Are not available at runtime

Lasso and Ridge Regression are handy when...

Variables are co-linear, You need regularization in your variables

Linear Regression is a good model when

The data relationship is essentially linear, You want an explainable model, You want a simple model or don't have much data

Bayes Theorem - P(H|E) = P(H)*P(E|H) / (P(E). P(E|H) is...

Probability of the Evidence given the hypothesis

Kinda fun, I just want lunch, Highly annoying, I don't care one way or the other

Bayes Theorem - P(H|E) = P(H)*P(E|H) / (P(E). P(E) is...

Usually broken down into hypothesis and non-hypothesis constrained portions, Probability of the Evidence

studbud

906/24 Review Quiz

31 termsStudBud Administration17 plays

Review Quiz for Friday 9/6/24

Flashcards31 cardsFPractice quizPHost liveL

Questions 31

Answers

1
Decision Trees to reduce ____ and increase _____ when splitting
sobel kernels, entropy
entropy, information gain
global warming, carbon sequestration
information gain, entropy
2
In Bayes Theorem, Your Hypothesis is...
Feature values that could be related to a specific outcome
A particular outcome you are interested in
All of the possible events that could happen
The probability of evidence
3
Bagging models....
Train multiple models on different on subsets of the data
Use the kernel trick to fit a hyperplane
Use the gradient boosting algorithm
Use ensemble methods
4
Polynomial Regression
Tends to overfit
Stratifies data into multiple categories.
Fits the data into a polynomial function
Is another name for Kullback Regression
5
In Bayes Theorem, Evidence refers to...
The probability of evidence
Feature values that could be related to a specific outcome
A particular outcome you are interested in
All of the possible events that could happen
6
Multiple Regression Uses...
Multiple output variables
No such thing as multiple regression
Lasso, Ridge, and Sigmoid optimization
Multiple input (feature) variables
7
Bayes Theorem - P(H|E) = P(H)*P(E|H) / (P(E). P(H) is...
Probability of the Evidence given the hypothesis
Usually broken down into hypothesis and non-hypothesis constrained portions
Probability of the hypothesis
Probability of the Evidence
8
In Bayes Theorem, The Set of Outcomes are ...
All of the possible events that could happen
The probability of evidence
Feature values that could be related to a specific outcome
A particular outcome you are interested in
9
You are trying to predict profit based on sales. Profit is the...
Outcome / Y
Root of all evil
Input Feature / X
Confounding Variable
10
Linear Regression, y=mx+b. Y is the...
Independent Variable (Feature)
Multiplicative Coefficient/Slope
Dependent Variable (Outcome)
Slope Intercept (Additive Graph Offset)
11
Linear Regression, y=mx+b. M is the...
Slope Intercept (Additive Graph Offset)
Independent Variable (Feature)
Dependent Variable (Outcome)
Multiplicative Coefficient/Slope
12
Input features
Are available in the training set but not at runtime
Are the predicted output
Are the inputs to the model, available in both training and runtime
Are generated by the model
13
In Supervised Machine Learning
You have outcome data in your training/test sets
Model learn based on correct "ground-truth answers"
Models get outputs at training and ground truth at runtime
You would typically use a clustering algorithm
14
Linear Regression, y=mx+b. X is the...
Multiplicative Coefficient/Slope
Dependent Variable (Outcome)
Slope Intercept (Additive Graph Offset)
Independent Variable (Feature)
15
Decision Trees
Use information theory to split variables
Are fairly powerful
Are the basis of Random Forest models
Create flowchart-like AI models
16
Logistic Regression is good when...
You need a probabilistic Bayesian model
You have only two classes and they are pretty balanced in the dataset
You want the output to between 0 and 1
You need to account for multiple variable co-linear variables
17
Decision trees decide what feature to branch on based on...
Gini Index
Information Gain
Randomly
Entropy
18
Boosting Models
Are part of the SVM model family
Train n strong models on the whole dataset to reduce the loss function
Train multiple weak models sequentially to reduce errors
Are ensemble methods
19
Which is true about Linear Regression?
It creates a model of the form y = mx+b
Invented by Sir Francis Linear
It follows only a Bayesian approach to Statistics
It is short for Linearli Regressionitis
20
Higher Entropy means Higher...
Decision Tree Index
Probability
Uncertainty
Surprise
21
Random Forests
Always train each tree on all the data
Train multiple decision trees, and use the trees together to get an answer
Train multiple decision trees, and pick the best one
Are an ensemble technique
22
Bagging and Boosting Models...
Are linear regression variants
Are typically implemented with decision trees
Use the gradient descent algorithm
Are ensemble methods
23
Linear Regression, y=mx+b. B is the...
Multiplicative Coefficient/Slope
Slope Intercept (Additive Graph Offset)
Independent Variable (Feature)
Dependent Variable (Outcome)
24
Output Outcomes
Are available in the training set and at runtime
Are what the model generates
Are not available at runtime
Are not present in the training set for supervised learning
25
Lasso and Ridge Regression are handy when...
You need models only for categorical variables
Variables are co-linear
You need regularization in your variables
You need to constrain the output between 0 and 1
26
You are trying to predict frost based on the temperature. Temperature is...
Root of all evil
Outcome / Y
Confounding variable
Input Feature / X
27
Linear Regression is a good model when
The data relationship is essentially linear
All of the input feature variables are co-linear
You want an explainable model
You want a simple model or don't have much data
28
Bayesian models are...
Good at handling co-linear variables
Constrained to linear models
Explainable
Decently powerful
29
Bayes Theorem - P(H|E) = P(H)*P(E|H) / (P(E). P(E|H) is...
Probability of the Evidence
Usually broken down into hypothesis and non-hypothesis constrained portions
Probability of the Evidence given the hypothesis
Probability of the hypothesis
30
Quiz Quizzes are
Kinda fun
I just want lunch
Highly annoying
I don't care one way or the other
31
Bayes Theorem - P(H|E) = P(H)*P(E|H) / (P(E). P(E) is...
Usually broken down into hypothesis and non-hypothesis constrained portions
Probability of the hypothesis
Probability of the Evidence given the hypothesis
Probability of the Evidence

906/24 Review Quiz Flashcards | StudBud