Data Mining Concepts and Techniques - Chapter 2 Vocabulary

Flashcards covering vocabulary, definitions, and concepts from Data Mining Chapter 2 notes.

Data Object

An entity represented in a dataset, composed of a collection of attributes (also referred to as a sample, example, instance, data point, or tuple).

Attribute

A data field representing a property, characteristic, or feature of a data object (also known as a variable, field, or characteristic).

Nominal Attribute

An attribute type whose values are categories, states, or names of things without any implicit quantitative order.

Ordinal Attribute

An attribute whose values have a meaningful order or ranking, but where the magnitude between successive values is unknown.

Symmetric Binary Attribute

A nominal attribute with only $2$ states where both outcomes are equally important, such as gender.

Asymmetric Binary Attribute

A nominal attribute with only $2$ states where the outcomes are not equally important, such as a medical test result where positive carries greater significance.

Interval-Scaled Attribute

A numeric attribute measured on a scale with equal intervals and meaningful differences, but lacking a true zero-point and meaningful ratios.

Ratio-Scaled Attribute

A numeric attribute measured on a scale with equal intervals and an inherent true zero-point, allowing meaningful ratios and multiples between values.

Discrete Attribute

An attribute that has only a finite or countably infinite set of values, such as zip codes or document words.

Continuous Attribute

An attribute that takes real numbers as values, typically represented using floating-point variables.

Trimmed Mean

A measure of central tendency calculated by removing extreme values or outliers from the dataset before computing the mean.

Median

The middle value in an ordered set of data if the number of values is odd, or the average of the middle two values if the number of values is even.

Mode

The value that occurs most frequently in a dataset.

Multimodal Data

A dataset that possesses two or more distinct modes.

Quantiles

Data points taken at regular intervals of a data distribution that divide it into equal-sized consecutive sets.

Quartiles

Specific $4$-quantiles that divide a distribution into four equal-sized subsets, comprising $Q_1$ ($25\text{th}$ percentile), $Q_2$ (median), and $Q_3$ ($75\text{th}$ percentile).

Five-Number Summary

A summary of a data distribution consisting of five values in order: Minimum, $Q_1$, Median ($Q_2$), $Q_3$, and Maximum.

Boxplot

A graphical display of the five-number summary where a box spans from $Q_1$ to $Q_3$, a line indicates the median, whiskers extend to the minimum and maximum, and outliers are plotted individually.

Outlier

A data point that deviates significantly from others, typically identified as a value lying more than $1.5 \times \text{IQR}$ beyond the first or third quartile.

Variance

A measure of data dispersion that quantifies the spread of a set of $N$ observations, denoted as $s^2$ for samples and $\text{σ}^2$ for populations.

Standard Deviation

The square root of variance (denoted as $s$ for a sample or $\text{σ}$ for a population), measuring how closely data points cluster around the mean.

Quantile Plot

A plot where each data value $x_i$ is paired with $f_i$, representing that approximately 100 f_i \text{%} of the data is less than or equal to $x_i$.

Quantile-Quantile (Q-Q) Plot

A graphical display that plots the quantiles of one univariate distribution against the corresponding quantiles of another.

Scatter Plot

A graph that plots pairs of values as coordinates in a two-dimensional plane to visually reveal relationships between variables.

Similarity

A numerical measure indicating how alike two data objects are, taking higher values in the range $[0, 1]$ as objects become more alike.

Dissimilarity

A numerical measure indicating how different two data objects are, taking lower values (often with a minimum of $0$) as objects become more alike.

Proximity

A general term referring to a measure of either similarity or dissimilarity between data objects.

Simple Matching Distance

A dissimilarity measure for nominal attributes given by $d(i, j) = \frac{p - m}{p}$, where $m$ is the number of attribute matches and $p$ is the total number of attributes.

Jaccard Coefficient

A similarity measure for asymmetric binary attributes calculated as $\text{sim}(i, j) = \frac{q}{q + r + s}$, where $q$ represents mutual presence ($1\text{-}1$) and $r, s$ represent mismatched attributes.