Informatieverlies

Informatieverlies refers to the reduction of information content that occurs when data undergoes processing, transmission, or storage, so that the output contains less information than the input. In information theory and related fields, it is the consequence of processes that trade fidelity for efficiency, compression, or privacy. The term is used across disciplines such as data science, computer science, telecommunications, archival science, and journalism.

Causes and manifestations include lossy data compression (for example JPEG or MPEG), data downsampling or quantization, filtering, anonymization or aggregation, and the introduction of noise or truncation. These processes can make reconstruction of the original data impossible or impractical, especially when important details or context are removed.

Consequences center on reduced accuracy, diminished reconstructability, and potential loss of context or meaning. Informatieverlies involves trade-offs between fidelity and goals such as reduced storage requirements, faster transmission, or enhanced privacy. Measuring information loss often relies on information-theoretic metrics such as mutual information, entropy, or divergence between the original and processed data.

Examples include image compression discarding fine textures, audio compression removing subtle timbre, or data anonymization removing identifying attributes. The concept is distinct from information leakage, which concerns unintended exposure of sensitive data. Informatieverlies is related to, but separate from, lossless versus lossy methods, highlighting the balance between information preservation and practical constraints in data handling.