7 terms
Batch processing runs data jobs over a bounded collection of records at scheduled intervals, favoring throughput and repeatability over low latency.
Data cleaning is the process of detecting and correcting errors, inconsistencies and gaps in a dataset so that analysis and models produce reliable results.
Data transformation converts data from its source structure, format or values into the shape a target system or analysis requires, in batch or streaming.
Deduplication in data processing identifies and removes or merges records that describe the same entity or event, so each appears only once.
Feature engineering is the data processing work of turning raw data into informative input variables that make machine learning models more accurate.
Imputation in data processing replaces missing values with substitutes estimated from the available data so records stay usable for analysis.
Missing data refers to values absent from a dataset; in data processing, its cause and pattern decide whether deletion, flagging or imputation is appropriate.