Machine Learning Flashcards

1
Q

Supervised learning

A

labeled training data to guide the ML program toward superior forecasting accuracy

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Unsupervised learning

A

the ML program is not given labeled training data

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Deep learning

A

used for complex tasks such as image recognition, natural language processing, and so on

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

reinforced learning

A

Programs that learn from their own prediction errors

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

neural networks

A

a group of ML algorithms applied to problems with significant nonlinearities

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Supervised Learning Types

A

Regression (Continuous)
Classification (Categorical)
Neural Networks
Deep Learning
Reinforcement LEarning

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Unsupervised Learning Types

A

Dimensionality Reduction
Clustering
Neural Networks
Deep Learning
Reinforcement LEarning

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Overfitting

A

when a large number of features are included in the data sample

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

cross validation

A

estimates out-of-sample error rates directly from the validation sample.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

To measure how well a model generalizes, data analysts create three nonoverlapping data sets

A

(1) training sample
(2) validation sample
(3) test sample

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Data scientists then decompose these errors into the following:

A
  • Bias error.
  • Variance error.
  • Base error.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Bias error.

A

This is the in-sample error resulting from models with a poor fit.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Variance error.

A

This is the out-of-sample error resulting from overfitted models that do not generalize well.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Base error.

A

These are residual errors due to random noise.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

k-fold cross validation

A

the sample is randomly divided equally into k parts. The training sample comprises (k − 1) parts, with one part left for validation. Error is then measured for the model in each of the parts.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

common supervised ML algorithms

A

Penalized regressions.

Support vector machine (SVM).

K-nearest neighbor (KNN).

Classification and regression trees (CART).

Ensemble and Random Forest.

17
Q

Penalized regressions

A

reduce the problem of overfitting by imposing a penalty based on the number of features used by the model.

18
Q

Least absolute shrinkage and selection operator (LASSO).

A

popular penalized regression model. In addition to minimizing SSE, LASSO minimizes the sum of the absolute values of the slope coefficients.

19
Q

Support vector machine (SVM)

A

linear classification algorithm that separates the data into one of two possible classifiers (e.g., sell vs. buy).

20
Q

K-nearest neighbor (KNN)

A

used to classify an observation based on nearness to the observations in the training sample

21
Q

Classification and regression trees (CART)

A

Classification trees assign observations to one of two possible classifications at each node.

22
Q

Ensemble and Random Forest.

A

Ensemble learning is the technique of combining predictions from multiple models rather than a single model.

Random forest is a variant of classification trees whereby a large number of classification trees are trained using data bagged from the same data set.

23
Q

common unsupervised ML algorithms

A

Principal component analysis (PCA).

Clustering.

24
Q

Principal component analysis (PCA)

A

summarizes the information in a large number of correlated factors into a much smaller set of uncorrelated factors.

25
Clustering.
clustering is the process of grouping observations into categories based on similarities in their attributes (called cohesion)
26
K-means clustering
partitions observations into k nonoverlapping clusters, where k is a hyperparameter (i.e., set by the researcher).
27
Hierarchical clustering
builds a hierarchy of clusters without any predefined number of clusters
28
agglomerative (or bottom-up) clustering
start with one observation as its own cluster and add other similar observations to that group, or form another nonoverlapping cluster.
29
divisive (or top-down) clustering
starts with one giant cluster, and then it partitions that cluster into smaller and smaller clusters.
30
Neural Networks
constructed as nodes connected by links
31
Deep learning networks (DLNs)
neural networks with many hidden layers
32
Reinforcement learning (RL)
have an agent that seeks to maximize a defined reward given defined constraints
33