basic definitions

Question 1

what is unsupervised learning

Answer 1

algorithm is trained on unlabeled data

Question 2

Clustering

Answer 2

grouping similar data points together

Question 3

Supervised learning

Answer 3

algorithm is trained on a labeled dataset

Question 4

Reinforcement Learning

Answer 4

A type of machine learning where an agent interacts with an environment by performing actions and receives rewards or penalties. The agent’s goal is to learn a policy that maximizes the cumulative reward over time

Question 5

Overfitting

Answer 5

A situation where a machine learning model performs well on the training data but poorly on unseen test data because it has learned noise and irrelevant patterns in the training data.

Question 6

Underfitting

Answer 6

A situation where a machine learning model is too simple to capture the underlying pattern of the data, resulting in poor performance on both the training and test datasets.

Question 7

Cross-Validation

Answer 7

A technique for evaluating a machine learning model by dividing the dataset into multiple subsets, training the model on some subsets, and validating it on the remaining subset. This process is repeated several times to ensure that the model’s performance is robust

Question 8

Hyperparameters

Answer 8

Parameters whose values are set before the learning process begins and control the behavior of the learning algorithm. Unlike model parameters, hyperparameters are not learned from the data.

Question 9

Neural Network

Answer 9

A machine learning model inspired by the human brain’s structure, consisting of layers of interconnected nodes (neurons) that process input data to produce output

Question 10

Gradient Descent

Answer 10

An optimization algorithm used to minimize the loss function in machine learning models by iteratively adjusting the model parameters in the direction that reduces the error.

Question 11

Feature Engineering

Answer 11

The process of selecting, modifying, or creating new features from raw data to improve the performance of a machine learning model

Question 12

Confusion Matrix

Answer 12

A table used to evaluate the performance of a classification model, showing the actual vs. predicted classifications, including true positives, false positives, true negatives, and false negatives

Question 13

Nearest Neighbors (k-NN)

Answer 13

Nearest Neighbors is an instance-based learning method where the classification of a new instance is determined by the majority vote of its ‘k’ closest neighbors from the training dataset. The distance between instances is typically measured using metrics such as Euclidean distance, Manhattan distance, or others depending on the feature types

Question 14

knn Lazy learning

Answer 14

k-NN is a lazy learner, meaning it doesn’t learn a model explicitly but rather computes results based on the distance from the query point to the stored instances.

Question 15

knn Decision Surface

Answer 15

The decision boundary of k-NN is often very irregular and heavily influenced by the choice of ‘k’.

Question 16

knn- Challenges

Answer 16

k-NN is sensitive to the choice of distance metrics and can be computationally expensive, especially in high-dimensional spaces.

Question 17

Regression Models

Answer 17

Predict numbers

Question 18

Classification models

Answer 18

predicts categories

Question 19

2 types of Supervised learning models

Answer 19

classification and regression

Question 20

Training Set

Answer 20

data used to train the model (features and targets)

Question 21

X

Answer 21

input variable feature

Question 22

Lower case y

Answer 22

output variable or target variable

Question 23

m

Answer 23

number of training examples

Question 24

(X,y)

Answer 24

single training example

Question 25

(X(i), y(i))

Answer 25

Ith training example - ith

Question 26

linear function

Answer 26

fancy name for a line

Question 27

univariate linear regression

Answer 27

1 variable line

Question 28

Cost function

Answer 28

the cost function (also known as the loss function or error function) is a mathematical function that quantifies the “error” between a model’s predictions and the actual, expected outcomes.

Question 29

fitting or training

Answer 29

capturing patterns from data

Question 30

Univariate Feature Selection

Answer 30

his involves analyzing each feature individually and assessing its importance. Methods like the Pearson correlation coefficient or the chi-squared test can help identify features with strong relationships to the target variable

Question 31

Embedded Methods

Answer 31

Some machine learning algorithms, like Lasso regression, have built-in feature selection mechanisms. The algorithm automatically assigns weights to features, and features with low weights can be dropped.

Question 31

Wrapper Methods

Answer 31

These methods use the machine learning algorithm itself to evaluate the importance of features. Techniques like recursive feature elimination (RFE) iteratively remove features and measure how the model’s performance drops.