Definitions

Question 1

Machine learning

Answer 1

Field of study that gives computer the ability to learn without being explicitly programmed

Question 2

Supervised Learning

Answer 2

A type of machine learning where the model learns from labeled data — each input has a known correct output — to make predictions on new data.

Question 3

Unsupervised Learning

Answer 3

A type of machine learning where the model learns from unlabeled data, finding hidden patterns or groupings without knowing the correct output in advance.

Question 4

Semi-Supervised Learning

Answer 4

A type of machine learning that uses a small amount of labeled data and a large amount of unlabeled data to improve learning accuracy.

Question 5

Rainforcement learning

Answer 5

An agent learns to make decisions by interacting with an environment and receiving feedback through rewards or penalties

Question 6

Classification

Answer 6

A machine learning task where the model learns from labeled data to assign new inputs to predefined categories.

Question 7

Regression

Answer 7

A machine learning task where the model learns from labeled data to predict a continuous numerical value for new inputs.

Question 8

Clustering

Answer 8

The model detected many inputs and grouped them where similar inputs are placed together based on patterns in the data

Question 9

Anomaly detection

Answer 9

Finding irregular data that doesn’t match regular patterns.

Question 10

Association Rule Learning

Answer 10

Finding patterns or rules that show how items are related to each other in data.

Question 11

Batch Learning

Answer 11

A learning method where the model is trained on the entire dataset at once, usually offline, and updated only when retrained with new data.

Question 12

Mini-Batch (Online) Learning

Answer 12

A method where the model is trained on small batches of data, allowing it to learn and update continuously as new data arrives.

Question 13

Dataset

Answer 13

A collection of data is treated as a single unit by a computer.
- Oxford dictionary.

A collection of data used for some specific machine learning purpose.
- the encyclopedia of machine learning (kakas, 2010).

Question 14

Labels

Answer 14

The output variables (the target value) the model aims to predict

Question 15

Features

Answer 15

The input variables used by the model to make predictions

Question 16

Training data

Answer 16

The dataset used to train a machine learning model, containing both features and labels.

Question 17

Testing data

Answer 17

The dataset used to evaluate the performance of a trained machine learning model.

Question 18

Validation data

Answer 18

A dataset used to tune the model’s parameters and prevent overfitting during training.

Question 19

CSV Format:

Answer 19

A file format used to store data in a tabular form, where each row represents an instance and each column is a feature.

Question 20

Textual data

Answer 20

Data represented in the form of text, including words, sentences, or documents. It needs to be converted into numbers when working with machine learning models.

Question 21

Numerical data

Answer 21

Data that consists of numbers, used for mathematical calculations and statistical analysis.

Question 22

Bivariate datasets

Answer 22

Datasets with two variables, used to analyze the relationship between them.

Question 23

Multivariate datasets

Answer 23

Datasets with more than two variables, used to analyze relationships among multiple variables.

Question 24

Correlation Datasets

Answer 24

Datasets used to measure the relationship between two or more variables, where the variables must be numerical to calculate correlation.

Question 25

Time series data

Answer 25

A collection of a sequence values that all recorded and collected over specific intervals of time

Question 26

Overfitting

Answer 26

When a machine learning model learns the details and noise in the training data to the extent that it negatively impacts its performance on new, unseen data.

Question 27

Underfitting

Answer 27

When a machine learning model is too simple to capture the underlying patterns in the training data, leading to poor performance on both the training and testing data.

Question 28

Regularization

Answer 28

A technique used to make a model simpler and reduce the risk of overfitting by adding a penalty to the model's complexity.

Question 29

Imbalanced dataset

Answer 29

Datasets where the positive values of the datasat are not approximately the same as the negative values.

Question 30

Resampling

Answer 30

One of the most common techniques that were proposed in order to solve imbalances datasets. Aims to handling the original dataset to make it more balanced

Question 31

Oversampling

Answer 31

A technique used to increase the number of observations in the minority class to achieve balance, either by duplicating instances, or generating the new ones using various techniques

Question 32

Undersampling

Answer 32

A technique used to balance an imbalanced dataset by reducing the number of observations in the majority class, either by removing instances or selecting a subset of the majority class using various techniques.

Question 33

SMOTE

Answer 33

(synthetic minority oversampling technique) a technique focuses on increasing instances in the minority class to balance the dataset, usually by the applicating instances without adding new information

Question 34

Outliers

Answer 34

A natural phenomena in a dataset. The art observations that differ significantly from the rest of the data, often appearing at the extremes

Question 35

Irrelevant features

Answer 35

Features in a dataset that do not have any meaningful relationship with the target variable and do not contribute to improving the model's performance.

Question 36

Feature Engineering

Answer 36

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

Question 37

Feature Selection

Answer 37

The process of choosing the most important features from a dataset, removing irrelevant or redundant ones, to improve model performance and reduce complexity.

Question 38

Pre-training Feature Importance:

Answer 38

The process of assessing the importance of features before training the model, often using techniques like correlation analysis, to determine which features contribute most to predicting the target variable.

Question 39

Post-training Feature Importance

Answer 39

The process of evaluating the importance of features after the model has been trained, using methods like model coefficients or feature importance scores, to understand which features most influence the model's predictions

Question 40

Feature extraction

Answer 40

Combining existing features producing more useful ones (data mining or dimensionality reduction algorithms)

Question 41

Decision tree

Answer 41

A machine learning model used in supervised learning, whether for classification or regression, that splits data into branches based on feature values to make predictions or decisions.

Question 42

Support vector machine (SVM)

Answer 42

Regression approach finds the optimal decision boundary to separate different classes

Question 43

Random Forest

Answer 43

A supervised learning model used for both classification and regression. It builds multiple decision trees during training and combines their predictions to improve accuracy and reduce overfitting.

Question 44

Neural networks approach

Answer 44

A machine learning method that tries to mimic human thinking and cognition

Question 45

K-nearest neighbor (KNN)

Answer 45

A machine learning method that assigns class based on closest neighbors

Question 46

Permutation

Answer 46

A post-training feature importance technique where data is randomly shuffled, and the impact on model performance is measured to determine the importance of each feature.

Question 47

Pandas library

Answer 47

A Python library used for data manipulation, cleaning, and analysis. It allows reading from and writing to various file formats such as CSV, Excel, SQL databases, and JSON.

Question 48

.info() function

Answer 48

A function in Pandas that provides a summary of a DataFrame, including the number of non-null entries, data types of each column, and memory usage, helping to understand the structure of the dataset.

Question 49

.head() Function:

Answer 49

A function in pandas that displays the first few rows of a DataFrame, helping you get a quick overview of the dataset. By default, it shows the first 5 rows.

Question 50

.describe() function

Answer 50

A function in Pandas that generates summary statistics for numerical columns in a DataFrame, including count, mean, standard deviation, min, max, and quartiles, providing a quick statistical overview of the data.

Question 51

.value_counts() function

Answer 51

A function in Pandas that returns the count of unique values in a column, sorted in descending order, providing insights into the distribution of categorical data.

Question 52

.plot() Function

Answer 52

A function in Pandas used to create various types of plots (line, bar, histogram, etc.) directly from a DataFrame or Series, helping visualize the data.

Question 53

Vectorization جات ف امتحان السنة اللي فاتت

Answer 53

Representing text as numerical vectors to make it usable for machine learning models.

Question 54

One hot encoding

Answer 54

A method of converting categorical data into numerical vectors, where each category is represented by a vector with a single active element, called a "1," and the rest are zeros.

Question 55

Bag of words

Answer 55

A method for representing text where each word is treated as a separate feature, and the frequency of each word in the document is counted, ignoring grammar and word order.

Question 56

Term Frequency (TF)

Answer 56

A measure of how often a word appears in a document, usually calculated as the number of times the word appears divided by the total number of words in the document.

Question 57

Inverse document frequency (IDF)

Answer 57

A measure used to evaluate how important a word is within the entire corpus. It is calculated by taking the logarithm of the total number of documents divided by the number of documents containing the word.

Question 58

Decision node

Answer 58

A point in the decision tree where a condition is evaluated. Represents a feature in the dataset that is used to split data

Question 59

Leaf node

Answer 59

The final output node that contains the predicted value or class label. Represents the conclusion of the decision process

Question 60

Decision boundary (hyperplane)

Answer 60

A decision plane which separates between a set of object having different class membership

Question 61

Support vectors

Answer 61

The data points, which are closest to the hyperplane. These points will define the separating glowing bitter by calculating margins.

Question 62

Margin

Answer 62

The distance between the closest data points (support victors) and the decision boundary