Browse Definitions :
Definition

overfitting

Overfitting is the incorrect optimizing of an artificial intelligence (AI) model, where the seeking of accuracy goes too far and may result in false positives.

Overfitting contrasts with underfitting, which can also result in inaccuracies. Overfitting is often referred to as overtraining and underfitting as undertraining. Overfitting and underfitting both ruin the accuracy of a model by leading to trend observations and predictions that don’t follow the reality of the data.

False positives from overfitting can cause problems with the predictions and assertions made by AI. Underfitting, on the other hand, can miss data that should be included due to omissions resulting from an over-specific model. In unseen data, an overfit model will make errors reflecting those in its training data. This inaccuracy is often a result of the model beginning to try to memorize results instead of accurately predicting previously unseen data.

For example, an AI hunting is for the number 1 in handwritten data. Depending on the clarity of the handwriting, a false positive might be grouping some 7s as 1s, which would be an overfit. Conversely, an omission that might result could be the failure to recognize 1s in some styles of handwriting, which would be an underfit.

Overfitting can be the result of overtraining, a lack of validation, improper validation or adjustment of weights and attempts at optimization after final testing. Overfitting can also be the result of using training data that is too noisy, containing unsuited information.

This was last updated in April 2018

Continue Reading About overfitting

SearchCompliance
  • compliance risk

    Compliance risk is an organization's potential exposure to legal penalties, financial forfeiture and material loss, resulting ...

  • information governance

    Information governance is a holistic approach to managing corporate information by implementing processes, roles, controls and ...

  • enterprise document management (EDM)

    Enterprise document management (EDM) is a strategy for overseeing an organization's paper and electronic documents so they can be...

SearchSecurity
  • IPsec (Internet Protocol Security)

    IPsec (Internet Protocol Security) is a suite of protocols and algorithms for securing data transmitted over the internet or any ...

  • principle of least privilege (POLP)

    The principle of least privilege (POLP) is a concept in computer security that limits users' access rights to only what are ...

  • biometric authentication

    Biometric authentication is a security process that relies on the unique biological characteristics of individuals to verify they...

SearchHealthIT
SearchDisasterRecovery
  • risk mitigation

    Risk mitigation is a strategy to prepare for and lessen the effects of threats faced by a business.

  • call tree

    A call tree is a layered hierarchical communication model that is used to notify specific individuals of an event and coordinate ...

  • Disaster Recovery as a Service (DRaaS)

    Disaster recovery as a service (DRaaS) is the replication and hosting of physical or virtual servers by a third party to provide ...

SearchStorage
  • cloud storage

    Cloud storage is a service model in which data is transmitted and stored on remote storage systems, where it is maintained, ...

  • cloud testing

    Cloud testing is the process of using the cloud computing resources of a third-party service provider to test software ...

  • storage virtualization

    Storage virtualization is the pooling of physical storage from multiple storage devices into what appears to be a single storage ...

Close