# Definition of Overfitting

Get Babylon's Dictionary & Translation Software
Free Download Now!

**Overfitting**Definition from Science & Technology Dictionaries & Glossaries

Electronic Statistics Textbook

When attempting to fit a curve to a set of data points, producing a curve with high curvature which fits the data points well, but does not model the underlying function well, its shape being distorted by the noise inherent in the data.

See also, Neural Networks .

See also, Neural Networks .

**Overfitting**Definition from Computer & Internet Dictionaries & Glossaries

DW and OLAP terms

The characteristic of some data mining algorithms that assigns importance to random variations in data by viewing them as important patterns.

**Overfitting**Definition from Encyclopedia Dictionaries & Glossaries

English Wikipedia - The Free Encyclopedia

In statistics and machine learning,

**overfitting**occurs when a statistical model describes random error or noise instead of the underlying relationship. Overfitting generally occurs when a model is excessively complex, such as having too many parameters relative to the number of observations. A model that has been overfit will generally have poor predictive performance, as it can exaggerate minor fluctuations in the data.See more at Wikipedia.org... |

**Wikipedia**and is licensed under the GNU Free Documentation License and under the Creative Commons Attribution-ShareAlike License

^{®}