Fbla data science and ai quizlet

Mean

Sum of all values divided by the number of values (the average)

Median

Middle value of an ordered dataset; if the count is even, the average of the two middle values

Mode

Most frequently occurring value; a dataset can have none, one, or several modes

Range

Maximum value minus minimum value

Variance

Average of the squared differences from the mean; measures how spread out data is

Standard deviation

Square root of variance; the typical distance of values from the mean, in the data's original units

Covariance

Measures how two variables change together; positive means they rise together, negative means one rises as the other falls

Correlation

Standardized measure from -1 to +1 of the linear relationship between two variables

Correlation vs causation

Two variables moving together does not prove one causes the other

Gaussian (normal) distribution

Symmetric, bell-shaped distribution where the mean, median, and mode are equal

Empirical rule (68-95-99.7)

In a normal distribution about 68% of data is within 1 SD of the mean, 95% within 2 SD, and 99.7% within 3 SD

Expected value

Long-run average of a random variable; sum of each outcome multiplied by its probability

Random variable

A variable whose value depends on the outcome of a random event

Probability

A number from 0 to 1 describing how likely an event is

Discrete variable

Takes countable, separate values (e.g., number of students)

Continuous variable

Can take any value in a range, including decimals (e.g., height, temperature)

Outlier

A data point far from the rest of the data; can distort the mean

Skewed distribution

Asymmetric distribution with a longer tail on one side

Histogram

Bar-style chart showing the frequency distribution of one numeric variable grouped into bins

Boxplot

Chart showing the median, quartiles, and outliers of a dataset

Scatterplot

Plots two numeric variables as points to show their relationship

Bar chart

Compares values across separate categories

Line chart

Shows how a value changes over time

Pie chart

Shows parts of a whole; works best with a few categories

Heat map

Uses color intensity to show values in a grid or matrix

Five-number summary

Minimum, first quartile (Q1), median, third quartile (Q3), maximum

Interquartile range (IQR)

Q3 minus Q1; the spread of the middle 50% of the data

Multivariate data

Data with more than two variables recorded for each observation

Dependence methods

Techniques that predict one variable from others (e.g., multiple linear regression, logistic regression)

Interdependence methods

Techniques that find structure among variables without a designated outcome (e.g., clustering, factor analysis)

Linear regression

Fits a straight line to predict a numeric outcome from a predictor

Multiple linear regression

Predicts a numeric outcome using two or more predictor variables

Logistic regression

Predicts the probability of a categorical outcome, usually yes/no

Decision tree

Flowchart-like model that splits data with feature-based questions to reach a prediction

k-means

Unsupervised algorithm that groups data into k clusters around center points called centroids

Data cleaning

Finding and fixing or removing errors, duplicates, missing values, and inconsistent formats before analysis

Duplicate records

Repeated entries that can bias results and inflate counts

Missing or incomplete data

Gaps in a dataset that can reduce accuracy; handled by removing or filling values

Data quality factors

Accuracy, completeness, consistency, timeliness, and source reliability

Garbage in, garbage out

Poor-quality input data produces poor-quality results

SQL

Structured Query Language used to create, query, and manage data in relational databases

SELECT

SQL clause that chooses which columns to return

FROM

SQL clause that names the table to read from

WHERE

SQL clause that filters rows before grouping

GROUP BY

SQL clause that groups rows so aggregate functions can summarize each group

HAVING

SQL clause that filters groups after GROUP BY

ORDER BY

SQL clause that sorts results (ASC ascending, DESC descending)

JOIN

Combines rows from two tables based on a related column

INSERT INTO

SQL command that adds new rows to a table

UPDATE

SQL command that changes existing rows

DELETE

SQL command that removes rows from a table

Aggregate functions

SQL functions such as COUNT, SUM, AVG, MIN, and MAX that summarize many rows into one value

Relational database

Stores data in tables of rows and columns that are linked by keys

Primary key

Column that uniquely identifies each row in a table

Foreign key

Column that references the primary key of another table to link them

Pandas

Python library for working with tabular data in DataFrames

NumPy

Python library for fast numerical computing with arrays

Matplotlib

Python library for creating charts and plots

scikit-learn

Python library of classic machine learning algorithms and tools

PyTorch

Open-source Python library for building and training neural networks

TensorFlow

Google's open-source library for machine learning and deep learning

DataFrame

Two-dimensional table of data used in Pandas

Python data wrangling

Cleaning and reshaping data in Python (e.g., dropna, fillna, drop_duplicates, merge)

R

Programming language built for statistics, data analysis, and visualization

Jupyter Notebook

Interactive document that mixes code, output, charts, and notes

Artificial intelligence (AI)

Computer systems that perform tasks normally requiring human intelligence, such as learning and decision-making

Generative AI

AI that creates new content such as text, images, audio, or code based on patterns learned from training data

Generative AI capabilities

Drafting text, summarizing, generating images and code, translating, and brainstorming

Generative AI limitations

Can hallucinate, reflect bias in training data, lacks true understanding, and depends on its training data

Generative AI uses

Healthcare (drug discovery, imaging), research, digital art, customer service, and coding assistance

Computer vision

AI subfield that lets machines interpret images and video

Natural language processing (NLP)

AI subfield focused on understanding and generating human language

Robotics

AI subfield focused on machines that sense, decide, and act in the physical world

Human-computer interaction

Field focused on how people and AI systems communicate and work together

Large language model (LLM)

Neural network trained on huge amounts of text to understand and generate language

Token

A chunk of text (word or piece of a word) that an LLM reads and produces

Prompt

The input or instruction given to a generative AI model

Transformer

Neural network architecture using attention that powers most modern LLMs

Hallucination

When an AI produces confident but false or made-up information

Narrow AI

AI built for one specific task (e.g., spam filter); all current AI is narrow

Machine learning (ML)

Subfield of AI where systems learn patterns from data instead of following explicit rules

Training set

Data used to teach a model

Validation set

Data used to tune a model and choose settings during development

Test set

Data held back to measure final model performance on unseen data

Overfitting

Model memorizes the training data and performs poorly on new data

Underfitting

Model is too simple to capture patterns in the data

Supervised learning

Learns from labeled data to predict outcomes (classification and regression)

Unsupervised learning

Finds patterns in unlabeled data (clustering, dimensionality reduction)

Reinforcement learning

An agent learns by taking actions and receiving rewards or penalties

Classification

Supervised task that predicts a category (e.g., spam or not spam)

Regression (ML)

Supervised task that predicts a continuous number (e.g., house price)

Clustering

Unsupervised task that groups similar items together

Feature

An input variable used by a model to make predictions

Label

The correct answer or target value in supervised learning

Neural network

Model of connected layers of nodes (neurons) that learns weights from data

Deep learning

Machine learning using neural networks with many hidden layers

Loss function

Measures how far a model's predictions are from the correct answers; training tries to minimize it

Gradient descent

Optimization method that adjusts model weights step by step to reduce loss

Choosing an algorithm

Labeled data with a known target means supervised; unlabeled data to find groups means unsupervised; learning from rewards means reinforcement

Predicate logic

Logic that uses predicates, variables, and quantifiers to make statements about objects and their properties