mis swag

0.0(0)
Studied by 0 people
call kaiCall Kai
learnLearn
examPractice Test
spaced repetitionSpaced Repetition
heart puzzleMatch
flashcardsFlashcards
GameKnowt Play
Card Sorting

1/80

encourage image

There's no tags or description

Looks like no tags are added yet.

81 Terms

1
New cards

What is regression in machine learning primarily used for?

Predicting a continuous numerical value

2
New cards

In regression modeling, the target variable is usually:

Continuous

3
New cards

The main goal of linear regression in ML is to:

Learn the weights of a linear equation relating inputs to outputs

4
New cards

In a simple linear regression model with one predictor, the line is parameterized by:

b0 and b1

5
New cards

What is another common name for prediction errors in regression?

Residuals

6
New cards

To find the best-fitting regression line, we usually try to minimize:

Residual Sum of Squares (RSS)

7
New cards

Which sequence best describes how least squares regression finds the best fitting line?

Visualize the goal, define the error, square and sum errors, find the minimum, solve for coefficients

8
New cards

Why is solving partial derivatives equal to zero often not feasible in many ML problems?

Because of high dimensionality, non-convexity, cost-function complexity, and stochastic training environments

9
New cards

Instead of solving analytically for stationary points in many ML problems, practitioners often use:

Gradient descent or parameter search

10
New cards

A derivative tells you:

The slope of a function at a point

11
New cards

A gradient is best described as:

The direction of fastest ascent represented as a vector of partial derivatives

12
New cards

In gradient descent, the parameters are updated in which direction?

In the direction opposite the gradient

13
New cards

What controls the magnitude of movement in gradient descent?

Learning rate

14
New cards

In the hill-climbing metaphor, the negative gradient tells you:

How to go downhill the fastest

15
New cards

When does gradient descent usually stop?

Once the pre-set convergence criterion is met

16
New cards

If the linearity assumption is violated in linear regression, the model may:

Produce biased predictions

17
New cards

A recommended pre-check for linearity is:

Scatterplots

18
New cards

In ML, independence of errors is generally less critical unless the data has:

A temporal or spatial structure

19
New cards

Normality of errors is less critical in ML unless the goal is to:

Make probabilistic predictions

20
New cards

If residuals deviate significantly from normality, one possible remedy is to:

Apply transformations such as log or square root to the target variable

21
New cards

Equal variance of errors mainly affects:

Coefficient estimation and interpretability

22
New cards

Which methods can help stabilize unequal variance of errors?

Weighted least squares (WLS) and transformations of the target

23
New cards

Multicollinearity in linear regression primarily:

Increases sensitivity to data changes

24
New cards

What are common checks or remedies for multicollinearity?

Check correlations, check VIF (variance inflation factor), use Ridge or Lasso

25
New cards

Which regression metric is intuitive, robust to outliers, and easy to explain?

MAE

26
New cards

Which metric punishes extreme errors more strongly and reflects risk?

RMSE

27
New cards

Which metric is scale-free and gives interpretable percentages?

MAPE

28
New cards

R² mainly indicates:

Explanatory power

29
New cards

How should you choose a regression performance metric?

Based on business objectives, data distribution, and cost of errors

30
New cards

What are good responses to poor regression performance?

Improve data quality, try different algorithms, adjust features, use better metrics and validation

31
New cards

A classic symptom of overfitting in linear regression is:

Low bias and high variance

32
New cards

Very large estimated coefficients in an overfit model often mean:

The model is capturing noise rather than genuine relationships

33
New cards

Ridge regression is also known as:

L2 regularization

34
New cards

What does Ridge regression add to the least squares objective?

A penalty based on squared coefficients

35
New cards

Why can Ridge regression improve generalization?

It adds some bias during training to reduce overfitting

36
New cards

Lasso regression differs from Ridge because it:

Shrinks some coefficients exactly to zero

37
New cards

A major advantage of Lasso is that it can:

Perform feature selection

38
New cards

Which are Lasso solvers or related approaches?

LARS, coordinate descent, PCD, ADMM

39
New cards

A practical issue with Lasso is that shrinking coefficients to zero may:

Introduce extra bias

40
New cards

A debiasing strategy for Lasso is to:

Run Lasso for feature selection, then Least Squares

41
New cards

When features are highly correlated, Lasso tends to:

Select among them arbitrarily

42
New cards

Elastic Net combines:

Ridge and Lasso regularization

43
New cards

Why are regularization methods sensitive to feature scaling?

All features should contribute equally to the penalty

44
New cards

What is regularization in linear regression?

Adding a penalty to constrain coefficients and reduce overfitting

45
New cards

Which learning style describes KNN most accurately?

Lazy or instance-based learning

46
New cards

In eager learning, the model typically:

Finds a global approximation before prediction

47
New cards

Instance-based (lazy) learning usually has:

Less effort in training and more in prediction

48
New cards

Key ingredients of distance-based models include:

Distance metrics, exemplars, and decision rules

49
New cards

The similarity assumption states:

Nearby points in feature space are more similar

50
New cards

Distance is an inverse proxy for similarity because:

Smaller distance means more similarity

51
New cards

Which distance metric is best for continuous data?

Euclidean

52
New cards

Manhattan distance is especially useful for:

Grid-like data

53
New cards

Chebyshev distance focuses on:

Maximum difference along any dimension

54
New cards

Cosine similarity is useful for:

Sparse, high-dimensional data like text

55
New cards

A centroid is:

The arithmetic mean position of all points

56
New cards

A medoid differs because it:

Must be a member of the dataset

57
New cards

In classification with nearest exemplars, a new point is classified based on:

The nearest exemplar or center

58
New cards

In KNN, after calculating distances, the next steps are:

Find neighbors and vote on labels

59
New cards

Why does 1-NN have low bias but high variance?

It perfectly separates training data

60
New cards

As k increases in KNN:

Bias increases and variance decreases

61
New cards

What is the advantage of KD-tree in KNN?

It prunes the search space

62
New cards

The KNIME workflow includes:

Preprocess text, vectorize, compute distance, KNN, optimize k, evaluate

63
New cards

The objective of text preprocessing is to:

Extract key terms and remove noise

64
New cards

Which are noise in text preprocessing?

Punctuation, stopwords, diacritical marks, casing

65
New cards

Tokenization is:

Breaking text into tokens

66
New cards

Stemming vs lemmatization:

Stemming may create non-words; lemmatization uses context

67
New cards

TF-IDF gives high score when:

Term is frequent in document but rare overall

68
New cards

TF measures:

Local frequency

69
New cards

IDF measures:

Global rarity

70
New cards

Why are N-grams useful?

They capture context and word order

71
New cards

N-gram importance measures:

Count and relative frequency

72
New cards

Why is cosine similarity used?

Efficient for sparse vectors

73
New cards

Distance from cosine similarity:

1 - cosine similarity

74
New cards

If vectors are very similar, angle is:

Close to 0 degrees

75
New cards

Which KNN metric statement is correct?

Manhattan is for grid-like data

76
New cards

The Elbow Method is used to choose:

Number of clusters k

77
New cards

What is a challenge of KNN in high dimensions?

Curse of dimensionality

78
New cards

How to mitigate high-dimensional issues?

PCA or feature selection

79
New cards

Why is feature scaling important?

Prevents large-scale features dominating distance

80
New cards

What does regression modeling describe?

Regression modeling describes the relationship between one or multiple features and a continuous target variable.

81
New cards

Which shapes correspond to common distance metrics?

Euclidean: circle, Manhattan: diamond, Chebyshev: square, Cosine: wedge.