Data set

A data set is an organized collection of related data assembled for analysis, reference, or use by a computer system. It may be stored as a table, files, records, or another structured form.

Connect

Data point: A single recorded value or measurement within a collection of data. Individual values form the contents of many data sets.
Data collection: The systematic gathering of information from observations, measurements, records, or other sources. Collection choices shape what a data set contains and represents.
Training set: Data used to fit the parameters of a machine-learning model. It is the portion used to teach a model from examples.
Big data: Data characterized by scale, speed, or diversity that challenges conventional processing methods. A data set need not be large or computationally difficult to qualify.
Variable (statistics): A characteristic recorded for each unit in a statistical study. Variables define the attributes represented in tabular data sets.
Data cleaning: The process of detecting and correcting or handling errors and inconsistencies in data. Cleaning addresses errors that can distort results drawn from a data set.
Test set: Data reserved for evaluating a model after training and model selection. A held-out set estimates performance on previously unseen cases.
Data stream: A continuous or sequential flow of data items generated over time. Unlike a fixed data set, a stream may arrive continuously and change.
Observation (statistics): A recorded case or unit, often represented by one row in a data table. Observations identify the cases whose values a data set records.
Data transformation: The conversion of data from one representation or format to another. Transformations reshape values or structure for analysis and reuse.